Home

Using StatView

1. Mac only Show ATI Analyses Peseriptiyp Statistigs Weight Order eS eeu Mean 2957 629 Std Dey 525 664 Ste Frequency Distri Std Error 49 785 Ss ders ea Count te Data Car Data j Paired Comparisons Minimurn 1E2500 Orgel Estase Variable Unpaired Compar Maximum 4265 000 Model Ti browser Analysis browser Correlation Cov a Missing o Country H Regression Type H ANDOY A weight x Contingency Table Nonparametrics Factor Analysis Survival Monpar Survival Regres Univariate Plots Bivariate Plots Cell Plots Box Plot Cannas Parcanti Statview 4 5 Turning Cir Displacement E Horsepower E Gas Tank i Message area Mac only Split pane control Selected analysis result Most of the view window is an empty page area that fills with tables and graphs This is where youll build statistical analysis tables draw graphs and put together presentations Think of this part of the view window as your paper View window controls The view window is a normal window You can move it resize it and scroll it the same way as any other window The Recalculate button at the top of the window lets you control how results recalculate or update to reflect changes in the dataset Recalcu
2. Message area orizontal and vertical scroll bars Mac only Split pane control When your dataset has many columns you might find it helpful to split the window horizon tally into two panes so you can visually compare columns that are far apart in the dataset To split the dataset click and drag or double click the split pane control the black rectangle just to the right of the message area in the lower left corner Double clicking this control splits the dataset into two panes and double clicking again returns it to a single pane When the dataset is split you can use the scroll bars at the bottom to scroll either half of the dataset l Lipid Data S D of x OST Erana Criteria EEE x Cholesterol Trig Dataset preferences Preferences from the Manage menu allows you to set dataset preferences You can choose defaults for decimal formatting fonts and which direction the cursor should move after you enter a cell down to the next row of the same variable or right to the next column of the same case See Dataset preferences p 227 56 2 Datasets Dataset windows Variable browser Variables in the dataset are listed in the variable browser a floating window that appears alongside both dataset and view windows When you are working in a dataset window the variable browser allows you to 1 Choose open da
3. to calculate its own bounds but ensure that zero is included in those bounds Use the lock controls to choose whether these values should be locked so that any recalcu lations of the graph for changing data conditions do not change them or unlocked w so that recalculations do change the values as needed If you specify a value that value is locked by default Click a lock to toggle from locked to unlocked or vice versa Major and minor intervals Specify how wide the interval between major tick marks should be and specify how many minor intervals you want between major tick marks For example the following axis has major intervals of 500 and 2 minor intervals within each major interval 1500 2000 2500 3000 3500 4000 4500 When you first create a graph intervals default to widths appropriate to the bounds If the axis bounds change major and minor interval widths automatically update You change the intervals by typing in the text boxes The Fewer choices dialog box has a pop up suggesting good major interval widths Use the lock controls to choose whether these values should be locked so that any recalcu lations of the graph for changing data conditions do not change them or unlocked w so that recalculations do change the values as needed If you specify a value that value is locked by default Click a lock to toggle from locked to unlocked or vice versa Tick marks
4. Default Text Default Text Font arial a Font Geneva Size E v size r T Initially show hints window EJ Initially show hints window T Beep for each error message Beep for each error message I Show alert messages in Hints window Show alert messages in Hints window Cancel Display hints window when it contains Check Interface hints Windows or Balloon hints Macintosh if you want to see hints for buttons menu commands and other interface items Check Informational hints if you want to see general informational hints Leave both options unchecked if you don t usually want to see any hints at all Font and Size Choose the font and size you want the hints window to use You can type a specific size Initially show hints window Check this option if you want StatView to show the Hints win dow automatically each time it starts Uncheck it if you prefer not to see hints Beep for each error message Check this option if you want StatView to beep when display ing error messages Uncheck it if you want error messages to appear silently Show alert messages in hints window Check this option if you want to see error messages in the hints window rather than in alert boxes Survival Analysis preferences Survival Analysis Preferences Use More choices dialog as default I Nonparametric Methods I Regression Models P Retain covariate matrix after regression calculations Censor var
5. Complex compact variable Expanding a complex compact variable takes several steps e From the Window menu select Complex Compact Variable If you closed the dataset use File Open to reopen it e Select Cholesterol readings click its name in either the dataset or the variable browser e Click the Expand button in either the dataset or the variable browser You ve unpacked the Smoking levels Now you need to expand each gender level e Select Cholesterol readings Male in either the dataset or the variable browser e Click Expand in either the dataset or the variable browser e Select Cholesterol readings Female in either the dataset or the variable browser e Click Expand in either the dataset or the variable browser 2 Datasets Compact vanables 95 Variables Show Criteria Compact Sie Cholesterol readings Fe Data Complex Compact Varia EE s Order Dataset order 194 Male Smoker fp D l P Nunerokar A E r AEG And youre done er Seok Male None Female Sm_ Female Now af ier eee iste e is 224 l 194 s msm l If you expand a compact variable by mistake select Undo from the Edit menu Analyze compact variables Now we learn how to work with compact variables in analyses We assume that you already know the basics of working with analysis objects in a view window If not either review the chapter
6. Create Analysis Show All Analyses Order Default Descriptive Stati h rerFrequency Distrib 2 Score Histog Pie Chart Percentiles One Sample Analy Paired Comparisons Unpaired Compari Correlation Cova Regression ANDOY A Contingency Table Nonparametrics Factor Analysis Survival Nonpar Survival Regres Univariate Plots Bivariate Plots gt eeta p Gompare Percent QC Subgroup Mea QC Individual Mea QC PNP ac Cru Pareto Analysis Histogram Box Plot 4200 4000 M 3500 F 3000 M 2500 M Ht 2000 s 1500 J Box Plot 300 275 7 s 250 7 aL 225 7 A 200 4 H 175 7 i 150 7 125 7 100 7 757 4 50 Horsepower Order Yariables Remove Split By Data Car Data Datase Model Country blah EE i Sei Statview 4 5 First we create a new analysis Open Car Data from the Sample Data folder From the Analyze menu select New View In the analysis browser under Frequency Distribution select Histogram and click Create Analysis Accept the default parameters for the histogram by clicking OK In the variable browser select Weight and click Add Shortcut You
7. Calories Total fat g Saturated fat g Cholesterol g turated fat g gars g Sugars g From different datasets You can assign variables from different datasets in one of two ways The first method is to choose Be applied to different dataset StatView asks you to open the dataset and then presents the Assign Variables dialog box The second method is easier when you want to use variables from several datasets including the original one Choose Open original dataset but check Always show Assign Variables dia log box You can then use the Data pop up menu in the Assign Variables dialog box to select different datasets The Assign Variables dialog box lists all the variables in the dataset and all the variable slots that need to be filled for this view You assign variables from the variables list into slots by dragging or double clicking See the next chapter Templates p 161 Print a view You can print the contents of a view Before you print scroll through the view to see whether any tables graphs or drawn objects fall on a page break Page breaks are shown by red hashed lines You can avoid splitting results across page breaks by fixing them yourself or by having StatView fix them for you From the Layout menu select Clean Up Items e Click Clean Up Another thing to check is the number and configuratio
8. Min Value Selected Value Man Value 107 184 234 The number below Selected value changes as you move the cursor across the bar To define a range click endpoints in the value bar and choose brackets or parentheses at each end of the value bar to specify closed or open intervals Brackets indicate closed intervals which include their endpoints Parentheses indicate open intervals which exclude their endpoints For example if a variable contains integers between 1 and 10 the range 4 6 con tains the values 4 5 6 the range 4 6 contains the values 5 and 6 and 4 6 only contains the value 5 4 Managing data Create criteria 127 To specify a single value click in the value bar To change a value click the vertical line and drag it to its new location To specify an interval click and drag between endpoints For lt gt or Is or ISNOT which take a single value a vertical line indicates that value Select value to include l O O Min Value Selected Value Max Value 35 o 34 For lt gt gt or 2 or lt or lt which take a range a bracket or parenthesis and a gray fill pat tern indicate the included range Select values to include Ol Min Value Selected Value Max Value 44 40 1 42 For ElementOf which takes an interval of a continuous variable brackets or parentheses and a gray fill pattern indicate the included range Select values to
9. Histogram Histogram 35 11 1 1 1 i 40 111111 gp fy fg fT 305 F 35 257 F 307 257 f E209 F E 3 220 7 L 0154 H S 157 N 104 E 107 54 F 57 L 0 T T T TT ry T T L 0 a a TT Ley al Tie Goa aE Saab 30 32 34 36 38 40 42 44 46 48 8 10 12 14 16 18 20 22 24 26 28 Turning Circle Gas Tank Size Histogram 35 1 30 7 F 257 F Z207 A F gt Q 0154 L 10 r 574 L 0 T T T T T T m 50 100 150 200 250 300 Horsepower Any selected results forward their variable assignments to the next results you create Split analyses by groups You can break results down by groups of a nominal variable without reorganizing your dataset Note that StatView also allows you to set criteria to restrict analyses to a subset of your dataset See Create criteria p 124 To try this you can split the correlation matrix into separate matrices for each country e Click to select the correlation matrix previously created e In the variable browser select Country and click Split By Now you have a separate correlation matrix for each country represented in the dataset Correlation Matrix Split By Country Cell Japan Turning Circle Displacement Horsepower Gas Tank Size Weight Turning Circle 1 000 760 649 752 738 Displacement 760 1 000 944 932 920 Horsepower 649 944 1 000 864 822 Gas Tank Size 752 932 864 1 000 908 Weight 738 920 822 908 1 000 30 observa
10. Weight Mean 158 653 Std Dev 28 389 Descriptive Statistics Std Error 2 913 Mean Std Dev Std Error Count Minimum Maximum Missing Count 95 Weight 158 653 28 389 2 913 95 107 000 234 000 0 Minimum 107 000 Maximum 234 000 Missing 0 Change line thicknesses and pen patterns The Draw palette s pen and line tools are handy for calling attention to tables or reducing the visibility of borders in tables You can thicken lines and change their pen patterns EE Line Thickness You can change the thickness for border lines Table borders default to a sin gle point line To print hairlines see View preferences p 232 Choose the top dotted line for invisible lines Pen pattern You can choose from sixteen pen type choices Table borders default to a solid P P pattern Pen patterns have greater visual impact with thicker lines PostScript printers inter pret some patterns as halftones so a single pixel screen line may print as a gray line Change colors Pop up menus at the bottom of the Draw palette let you choose colors for the pen and fill Use the pen color tool to change the color of text inside a table and the fill color to change the color of table borders To change the color of a title or note select it directly and change the pen color Control click Windows or Command click Macintosh to select both pen and fill colors at once
11. G A grouping or factor variable S A Split By variable F An independent variable or covariate forced into a stepwise regression C An item count variable for QC P NP charts A unit count variable for QC C U charts A censor variable for Survival analysis T A time to event variable for Survival analysis A stratification variable for Survival analysis A variable playing multiple roles usually because more than one result is selected If no result is selected in the view any usage markers indicate any preassignments for a pend ing analysis where you select an analysis in the analysis browser and then assign variables in the variable browser before clicking Create Analysis in the analysis browser Number of variables Some analyses can handle any number of variables Some accept only a specific number Still others can handle multiple variables with one usage but only a single variable of another 5 Analyses Analysis windows usage The number of variables an analysis requires is the determining factor in what happens when new variables are assigned to an existing analysis For example an unpaired t test requires a single continuous and a single nominal grouping variable What happens if you assign another continuous variable to the analysis A new sepa rate analysis is generated that performs a t test on the new continuous variable using the groups identified by the original nominal variable The result appear
12. o complete this analysts assign at least one variable using the variable browser Add Button The note on the empty object says what to do next We use the variable browser to add vari ables to the empty analysis e Make sure the object is still selected has black handles if not click it to select it e In the Variable Browser Shift click to select Calories Total fat g and Saturated fat g e Click Add Variables Add N Remove Remove Split By SEE Data Candy Datase Brand i a Name Servings pkg 3 Oz pkg E ries tal fat g aturated f u g Cholesterol g Cholesterol g Sodium mg Sodium mg Carbohydr Carbohydrate m Dietary fib Notice that each of these variables is marked with a icon in the variable browser indicat ing that the variables are continuous Similarly Brand has an x icon for nominal and Name has an icon for informative We made these class settings in the attribute pane of the dataset window and now the browser reminds us StatView also uses these settings to help you assign appropriate variables to each analysis Ta dah You ve completed your first analysis in StatView Black handles indicate the object is still selected 22 I Tutorial Analyze data Pescriptive Statistics Mean Std Dew Std Error Count Minimum Maximum Pissing z Calories z243 027 61 996 75 125 000 450 000 Total fat g 11 573 5 72
13. From the Edit menu select Paste 2 Datasets Save datasets You re done If you keep StatView s selection flexibility in mind you ll find many similar opportunities to save time when entering and editing your datasets Save datasets You can save changes to any active dataset by selecting Save from the File menu If you want to save the changed dataset under a different name and preserve the original dataset unchanged choose Save As from the File menu If the file is untitled when you choose Save a dialog box prompts you to name it Choose the file format you want from the choices listed at the bottom of the dialog box Windows only The normal format that StatView uses is DataSet You can also choose Excel or Text for exporting data to other applications Macintosh only The normal format that StatView uses is StatView Data You can also choose Excel or text for exporting data to other applications StatView 1 x Data for export ing to versions of StatView earlier than version 4 1 or SuperANOVA Data for exporting to SuperANOVA It is best to save files as StatView Data unless you plan to transfer data to another application or to earlier versions of StatView When you save as StatView Data you retain all information about the dataset including formula definitions criteria the current selection which rows are currently included and so on SuperANOVA format is useful for transferring data to StatView vers
14. E StatView moO J F a os I Copyright Copyright 1998 by SAS Institute Inc Second edition First printing March1998 Acrobat edition revised October 1998 All rights reserved Printed in the United States of America No part of this publication may be reproduced stored in a retrieval system or transmitted in any form or by any means elec tronic mechanical photocopying or otherwise without prior written permission of the pub lisher SAS Institute Inc Information in this document is subject to change without notice The software described in this document is furnished under the license agreement packaged with the software media The software may be used or copied only in accordance with the terms of the agreement It is against the law to copy the software on any medium except as specifically allowed in the license agreement StatView and SAS are registered trademarks of SAS Institute Inc in the usa and other countries All trademarks above are registered trademarks or trademarks of SAS Institute Inc The symbol indicates usa registration Other brand and product names are trademarks or registered trademarks of their respective companies Technology License Notices Mac2Win software 1990 94 Altura Software Inc All rights reserved Mac2Win is a regis tered trademark of Altura Software Inc Portions of this software are copyrighted by Apple Computer Inc Apple
15. Show grid lines Show Cancel Axes for nominal variables are inherently different from numeric axes since a cell axis does not define a range of values but instead lists the labels of groups or the names of variables Cell axes use tick marks to identify different groups The tick marks for a cell axis have no relation to a numeric width Small Sporty Compact Medium Large Tick marks Choose the placement for tick marks Stagger Choose the stagger pattern that makes axis labels most legible Show Choose how many labels to show Below are some of the possibilities 7 Customizing results Graphs 193 Cell Axis Cell Axis Tick marks Y Stagger rerit bd Show Every label Show grid lines shear Cancel Tick marks Stagger sept Y Show Every 2nd EJ Show grid lines Show Cancel y Cell Point Chart Grouping Yariable s Type Cell Point Chart Grouping Yariable s Type 42 42 co a co a EH EH F uo a uo o F404 F407 F E E ez ez F ze 4 5 o z j 2a 2a r 536 364 L Oo C3 D 35 35 Small Sporty Compact Medium Large Large Compact Cell Axis Cell Axis Tick marks Y Stagger rerit Y Show Every ard kd Show grid lines Show
16. Drawing and layout StatView includes many drawing features to help you customize the results of your analyses Using StatView as your drawing program has two advantages 1 You can change data and analyses updating results automatically without having to rework presentations 2 You can save your finished presentation as a template to apply to other datasets In this chapter we discuss Draw tools you can use to add lines shapes and text to complete and enhance your results Then we discuss Layout tools that let you assemble your customized tables and graphs in an attractive full color presentation The previous chapter Customizing results p 179 discusses how you can use Edit Display and tools from the Text menu and Draw palette to customize StatView s graphs and tables Draw tools Tools in the Draw palette let you draw shapes and type text in the view Other Draw palette tools let you customize these shapes and text as well as tables and graphs Using drawing tools to customize graphs and tables is discussed in the previous chapter Customizing results p 179 204 8 Drawing and layout Draw tools Windows only To show the Draw palette select Draw Palette from the View menu Drag its title bar to reposition it Macintosh only The Draw menu is a tear off palette Click the menu then quickly and smoothly drag it to a new location If you hesitate or dr
17. US64 USII3 alignment USI85 USI99 US214 SR334 Us217 US218 allow changes see Unlock Numerics AllRows ust08 UsI24 R325 SR326 alpha 0 sR345 control limits 1 R345 c u analyses sR299 individual measurements sR279 258 Index SR StatView Reference US Using Statview p np analysis sr288 range or moving range R475 subgroup measurements sR262 error see type I error significance level alphabetize see Sort alternate mouse button usvit ambiguous data class us51 analyses clone usI46 compact variables us95 us97 example US95 SR9I SR93 SR97 SR99 SRIOI SRIOS SRIO7 create USI3I USI33 by hand or with templates us131 exercises USI50 USI56 tutorial example us20 us27 dialog box hints us222 multiple vs compound results US135 USI37 USI70 US229 objects Us21 uUs22 overview USI3I SRIV SRX parameters US24 USI34 USI35 USI42 tutorial example us31 template exercises USI75 tutorial example us14 us40 variable requirements us145 variable requirements also see data re quirements under specific analysis analysis browser Us141 US143 exercises USI50 USI56 open close us43 analysis generated variables correlation matrix SR46 data source US77 factor scores SRI35 regression SR6I survival regression analysis sRi77 analysis of variance US235 analysis of covariance models sr78 SR80 SR8I SR99 SRIOI analysis of variance models sR77 SR79 SR96 SR97 data requirements sR90 SR95 dialo
18. solve problems see troubleshoot Sort USII6 USII7 SR413 SR418 analyses USI49 USI50 turn off sRr418 tutorial example us14 Undo us117 source see data source space SR333 SPC SR25I Spearman rank correlation coefficient data sR123 Spearman rank correlation coefficient tho SRI21 R123 Special Causes Definitions table sr270 special purpose functions OD 1 0 sx337 lt gt S 2 SR337 sR336 ChooseArg sR359 VariableElement sr429 specification limits SR253 sphericity SR44 spline tool us2z08 us210 Split By us97 US136 USI44 USI46 tutorial example us25 us26 split pane control uss5 USI41 Recode usg Sqrt sR420 square root SR420 squared multiple correlation sR133 SR140 squares US205 SS e i e i 1 srs9 stabilize variance sr386 stack order of objects us215 stagger tick marks us192 standard deviation us6o US184 SR4 bars on interaction plot sR90 lines on univariate plot sR217 standard error us6o bars on interaction plot sR90 descriptive statistics table srs lines on univariate plot sR217 standard error of the mean SR5 SR421 StandardDeviation sR420 StandardError sr421 standardize sR397 SR422 standardized regression coefficients sr56 StandardScores sr422 Static Formula vus116 SR329 SR330 data source US77 reason to use US256 SR344 SR418 also see Formula stationery see templates statistical functions BoxCox sR358 CoeffOfVariation sr360 Correlation sR362 Cou
19. 218 8 Drawing and layout Layout tools Custom Rulers Ruler units Linches Divisions per unit Grid line color Black v Page break color Black v For example below is the upper left corner of a view window in which rulers and grid lines are shown Using Custom Rulers we specified centimeter ruler units and two divisions per centi meter We turned Align to Grid on which forces the cursor to align to a ruler tick mark at all times notice the dotted lines in the rulers showing exactly where the cursor is located Align to Grid Align to Grid snaps selected objects into alignment with the grid You must turn the grid on to use this command Objects align by the rectangle they inscribe in other words what aligns to the grid is the smallest rectangle you could draw around every part of the object including titles legends notes etc with a tiny margin You can see this rectangle when you start to drag the object I i 1 i g eae eae Box Plot Grouping Yariable s Type Split By Country Align Objects Align Objects lets you align selected objects vertically and horizontally e Select the objects you want to align e Select Align Objects from the Draw palette e Choose whether to align or distribute objects according to their tops bottoms rights lefts middles
20. Changing data class from continuous to nominal can be meaningful for some variables this flexibility is discussed in more detail under Data class p 50 Analyses use variables differ ently according to their class When you change a variable from continuous to nominal groups are sorted in increasing numerical or alphabetical order Although you can use any data type as a nominal variable category data offer some special advantages If you plan to use the variable as both nominal and continuous choose a non cat egory type and change the data class See Categories p 80 for more information 2 Datasets Variable attributes 79 Format Data format specifies how to display real currency and date time data The format you choose affects only the display of the data not its contents Values are stored and computa tions are performed to the fullest precision of your platform always Real Free Format Fixed Displays real numbers using Fixed Places format unless the column is not wide enough to display the entire number If that occurs the number is displayed in scientific notation Free Format Ignores trailing zeroes to the right of the decimal point without regard to the number of decimal places specified Fixed Places Displays the data using the number of decimal places specified Data are not displayed in scientific notation Scientific Displays data in scientific notation using the specified number of
21. Page 1 cr Box Plot Plot for Hors Edit Analysis Each analysis presents its own dialog box in which you set analysis parameters choices you make to control how the analysis is performed For example when you create a histogram you can specify how to divide the range into intervals whether intervals show counts or cumulative counts etc Since histograms are only one possible result from a Frequency Dis tribution analysis the dialog box also has options that apply to the other types of output Frequency Distribution Number of intervals E Show normal comparison Do you wish to enter your own interval information no Qyes width initial value Intervals indicate include Lowest value Tables show Counts Percents Relative frequencies Histograms show After the histogram is created you can click Edit Analysis to return to this dialog box and change the parameters e Select one or more analysis results e Click Edit Analysis e Change the parameter settings e Click OK The Edit Analysis dialog box for any analysis is the same one you see when you first create the analysis Some analyses have no parameters so Edit Analysis has no effect If you change the parameters of an analysis any other results associated with that analysis also change For example if you create a freque
22. Cancel Tick marks x Stagger pit Show Every label Show grid lines Show Cancel Cell Point Chart Grouping Yariable s Type Cell Point Chart Grouping Yariable s Type 42 42 2 2 o EH FEIS L 6 o 5 o e404 Edo 5 F E 23 2394 L 38 7 o o 384 o F 5 c 2a 2374 L Tze B36 4 s o o 9 o a5 35 Small Medium Small Mediurn Sporty Large Compact For more information on how StatView orders the groups in a cell axis see How does Stat View use ordering in nominal variables p 238 and How can I reorder category vari ables p 238 Ordinal axes Ordinal axes are used only for univariate plots By default ordinal axes have no tick marks nor scale values By default ordinal axes show the order of points in a dataset only by relative posi tion Check Show ticks and values turn the option on if you want to see a numeric scale for row numbers Note to select an ordinal axis click carefully just below the axis line away from the Observations label 194 7 Customizing results Graphs Ordinal Axis EJ Show ticks and values Show Cancel Ordinal Axis Show ticks and values Show Cancel Univariate Line Chart boii brirtbiritirvitrirtirs Univariate Line Chart Large WT co c
23. Count 116 Num Missing 0 R 707 R Squared 499 Adjusted R Squared 495 RMS Residual 28 299 Selecting an existing result saved you the trouble of reassigning Horsepower as the dependent variable and clicking OK in the Regression dialog box Any selected analysis forwards infor mation to a new analysis Adopt variables for new analyses The variables in a selected result can be used in a completely different analysis just as they can be used to create additional output for the same analysis e Click to select the correlation matrix previously created e In the analysis browser double click Frequency Distribution e Accept the default dialog box settings click OK The default output for frequency distribution appears in the view Your results appear in a single vertical column in the view We rearrange them to conserve space Histogram Histogram 35 1 1 1 1 i 30 1 1 1 30 g 257 F 257 F 204 E 207 F E 3 3157 F O 154 js O 104 L 10 7 F 4 5 5 L 54 m L 0 Tail T r Tsk T 0 dee a T p 50 00 150 200 250 300 350 400 1500 2000 2500 3000 3500 4000 4500 Displacement Weight 5 Analyses Exercise 155
24. H hairlines us232 half open interval sr337 harmonic mean sR3 HarmonicMean sr382 Harrington Fleming test SRISI SRI56 Harris image analysis sr132 hashed red lines see page breaks hatch marks see tick marks hazard function sr178 hazard plot sr150 SRI6I height us229 graphs usi88 tables us199 Us200 help us48 us221 US223 US225 helpful hints see troubleshoot heterogeneous variances sR86 Hide Grid Lines us217 Hide Page Breaks us217 Hide Rulers us217 hierarchical Analyze menu vus169 Hints us48 US221 US222 balloon us222 formulas us13 informational us222 interface Us222 preferences US222 US230 templates Assign Variables us163 window us4 Histogram US1I32 USI36 USI37 US233 capability indices us244 normal curve us244 tutorial example us31 histograms see frequency distribution historical values sr256 homogeneity of slopes sr81 homogeneity of variances US235 SR84 horizontal see casewise Hotelling Lawley trace sr82 Hour sp382 HYP USII2 hyperbolic arccosine sR349 hyperbolic arcsine sR354 hyperbolic arctangent sR356 hyperbolic cosine sR364 hyperbolic sine sr419 hyperbolic tangent sr426 hyperbolic trig functions us112 hypothesis testing SR74 SR76 hypothesized mean sR23 hypothesized variance sR24 I charts SR277 SR281 SR283 I class marker see class marker if then else Us114 R341 tutorial example usi9 ignore characters sR329 illustrations in manual usv1 impact SR201 import categories
25. The third attribute class tells how the data are to function as continuous measurements such as calories fat in grams etc as nominal data such as our brand groups or as infor mative such as name labels or identification numbers Since we ve set type to category Stat View automatically sets class to nominal Having set all the attributes for Brand we are ready to enter data Now we can see the power of categories I Tutorial Manage data Z e Click the first empty gray cell in the Brand column to select it Type the first letter of the first value M8 amp M Mars m StatView supplies the rest of the name If you had several groups beginning with M you would need to type a few more letters All you need to do is accept the value e Press Enter or Return to move to the next cell You can also enter category values by typing its number 1 for the first group 2 for the second group etc in the order that you defined them This alone saves you time Now think of how much time you ll save not having to correct your typing mistakes especially on a hard to type name like M amp M Mars Let s try using both shortcuts to finish entering brand names e Type hand press Enter or Return e Type 2 and press Enter or Return e Type 1 and press Enter or Return e Type cand press Enter or Return User Entered te _ 4 M amp M Mars Now we re ready for the Name variable First we need to change its type
26. You ll notice from this equation that you do not need an estimate of the baseline survival function to evaluate parametric models but only the event times Similar equations can be constructed for the other parametric models as well though the others Weibull log normal and loglogistic require specification of a scale parameter as well The complete equa tions for these other models are in the appendix Algorithms p 433 of StatView Reference How can make comparisons among coefficients for linear hypotheses Sometimes you might need to determine if the contribution of a particular group level or weighted combination of group levels to a survival regression model differs significantly from that of another group level or weighted combination of group levels Suppose for instance you have a covariate in a regression model called Treatment the levels of which are Treat ment 1 Treatment 2 and Control A simple example of a test of a linear hypothesis would be testing whether the contribution to the model of the Treatment 1 group is signifi cantly different from that of the Control group A more complex example might test whether an evenly weighted combination of the groups receiving treatments 1 and 2 is signif icantly different from the Control group If you are familiar with the techniques of analysis of variance ANova you will recognize that these hypotheses are contrasts Using th
27. as an X variable and the for mula column as a Y variable to a bivariate chart If you frequently need to evaluate proportional hazards models at particular covariate values you should probably save this dataset for later use as a sort of template We suggest that you delete all of the rows before saving as a template that way the next time you use it all you will need to do is paste in the new baseline data then type in the new coefficient and covariate val ues You might wish to make the formula column static so that it will not recompute as you are typing in each of the new coefficient and covariate values Once you are ready to compute the survival function reopen the formula dialog box then click Compute As you might have guessed you can also use these same techniques to evaluate any of the parametric models at particular covariate values For example from information provided in 248 9 Tips and shortcuts Common questions the appendix Algorithms p 433 of StatView Reference we know that the formula for evalu ating an exponential regression model at particular covariate values is z z STe S T z e where is the estimate of the intercept parameter The StatView formula for evaluating this function would then be e Event Time e4 MuHat DotProduct Coefficients Covariate values where MuHat is the estimated value of the intercept parameter type the number instead of MuHat
28. nonadjacent rows or columns Control click criteria name in Criteria pop up menu rows that meet a criterion To deselect any rows cells or columns selected in the dataset quickly click once in the empty rectangle above the row numbers and to the left of the first column heading Cut clear and delete data You can remove entire columns or rows using Edit menu commands or your keyboard Windows Macintosh Result Cut Control X Cut Command x Removes selected data to the Clipboard If data do not constitute an entire row or column the row or column remains in the dataset Cells where data have been cut contain missing value symbols If an entire row or column is selected including row number or variable name Cut completely removes the row or column from the dataset Rows below a cut row move up and columns to the right of a cut column move left 66 2 Datasets Edit data Copy data Paste data Clear Clear Command B Removes selected data and replaces them with missing values Data are permanently cleared unless you immediately select Undo from the Edit menu or type Control Z Windows or Command Z Macintosh Clear does not remove rows or columns Delete Backspace key Delete Delete key lf an entire row or column is selected Delete removes it from the dataset Deleted rows or columns are permanently removed unless you choose Undo immediately after
29. over In StatView all you have to do is open a pane in the dataset window If your data change the statistics update automatically e Click and drag the attribute pane control EA downward to expose the twelve rows of summary statistics for each variable e Scroll to the right so you can see the numeric variables Serving pkg Oz pkg Calories Total fat g Saturated fat g Chole Type Rea Real___ integer_ Real Real integ Class Continuous Contin Continu Continuous Continuous Cont Dec Places 1 2 1 1 Std Deviation 9 1 15 61 996 57 3 4 5 42 Sta Errori fa as eos fa J4 26 Coeff of Variation f 7 35 200 A 1 02 Maximum 4 5 6 450 29 15 20 Count f 75 75 75 75 75 75 Sum 164 21 18227 890 5 462 5 336 Sum of Squares 197 5 459 11 4714069 130008 27112 4266 Peel rl These summary statistics can help you spot and fix data entry errors quickly For example let s change the top Oz pkg value from 1 25 to 125 dropping a decimal point is a common data entry error e Click the cell to select it e Type 125 e Press Enter or Return Real Re User Entered User En Continu Dec Places 1 2 3 84 Std Deviation 9 14 23 Std Error Variance 202 38 Coeff of Variation 3 71 Minimum Maximum Notice that the summary statistics change right away probably faster than you can see Now notic
30. Application Color Palette Dataset Formula Graph Hints Survival Analysis Table View When you are finished setting preferences click Done All preferences are stored in the StatView Library If you move or delete the Library all prefer ences return to their original settings While this can be a convenient way to undo numerous changes you should be aware of the other effects of discarding your Library file see p 233 Application preferences Application preferences affect the overall application 226 9 Tips and shortcuts Preferences Application Preferences Windows zoom Application Preferences Leave room for Finder icons x 5 Browsers appearance Browsers appearance Font aia x Font Geneva Size E 7 size Initially hide which browsers IT Analysis l Variable M Results Analysis Variable 4 Results Initially hide which browsers I Use System s Temporary Folder EJ Use System s Temporary Folder cance TEn Windows zoom Macintosh only Occupy full screen tells StatView to make its windows as large as your monitor allows Leave room for Finder icons lets you see your hard disk the Trash etc along the right edge Your choice takes effect the next time a new window is opened Browsers appearance Choose the font and size you want for the variable analysis results and function browsers You can type a s
31. Decimal places Lr Always have leading digit Show Cancel Frame Choose from three different styles in the pop up menu 188 7 Customizing results Graphs di 1 Flip horizontal and vertical axes Check this option turn it on to transpose the horizontal and vertical axes changing the orientation of the graph ge ham Bounds include extra lines Check this option turn it on to ensure that default axis bounds do not exclude calculated lines such as the three standard deviation lines in univariate charts Show legend Check this box to display the graph legend Show title Check this box to include titles showing analysis type variables assigned and cri teria row inclusion in effect Dimensions Specify the horizontal and vertical dimensions in inches centimeters picas or points Numbers Choose a number of decimal places for numbers in axis scales see Decimal places p 80 Check Always have leading digit to include a leading zero in fractional values e g 0 25 Choose box styles You can choose from among four different box styles for box plots e Inside a box plot click the plot the actual box and whisker e Click Edit Display Bou Plot Select a box plot style Ble 2 stow Concer Ga The default style uses regular boxes and includes outliers as points The second style uses reg ular boxes but exc
32. If you ve forgotten what some of these statistics mean you ll be relieved to know that the chapters in StatView Reference include discussions of the theories behind each type of analy sis and they give pointers on which tests to use what you need to check first how to inter pret the numbers and where to turn next 40 I Tutorial Analyze data We also get an interaction bar plot This simply shows us the means and confidence intervals graphically Interaction Bar Plot for Saturated fat g Effect Brand Inclusion criteria Big Three from Candy Bars Data Cell Mean O NOO kU DN OO Hershey M amp M Mars Nestle Cell Notice that these results are no different from results created with the analysis browser You can clone them add variables remove variables reformat them even resave them to be a new template Quiz Are saturated fat values normally distributed We should check that assumption before tak ing our results seriously If we were doing an important analysis we d want to do more tests to be sure For our purposes it s reasonable simply to examine the histogram for Saturated fat g that we created earlier Can we predict calorie content from saturated fat content Use the Regression Simple template specifying Calories as the Dependent Variable and Saturated fat g as the Indepen dent Variable From total fat content Use the Regression Simple template specifying Calories as the Dependent Variab
33. If your formulas contain functions that SuperANOVA does not recognize SuperANOVA opens the Formula dialog box with the dataset and highlights the unknown function You have two choices in this situation 1 Click Cancel in the Formula dialog box This retains the computed values in the formula column If you save this dataset you can read it back into Stat View with all formula infor mation intact 2 Change the formula column from Dynamic Formula to User Entered This removes the formula definition but preserves its current values If you save this dataset and read it back into StatView the formula definition is lost Managing data StatView s Manage menu offers numerous ways to manage your dataset 1 Include Row Exclude Row let you choose which cases rows are used in statistical and graphical analyses 2 Create Criteria and Edit Apply Criteria let you define logically which cases rows are used for analyses e g all cases with Weight values less than 200 and Age values greater than 20 3 Formula Series and Random Numbers create new variables by definitions using Stat View s mathematical expression language For example you could create a variable of Cel sius temperatures from one of Fahrenheit temperatures with a formula like this Temperature 32 5 9 4 Sort reorders cases rows according to the values of one or more key variables 5 Recode creates new variables by grouping the values of continuous variab
34. Mars Nestle Units Saturated fat g This is the candy bar to eat Print a presentation After you ve put the finishing touches on your presentation you ll want to print it Ifyou have a color printer you can print in full color e Make sure the view is the frontmost active window If not click it or select Nutritional analysis from the Window menu e From the File menu select Print e Adjust the printing options and click OK Windows or Print Macintosh Save a presentation You ll want to save the view again so that all your finishing touches are preserved From the File menu select Save If your data change Stat View updates your analyses automatically You don t have to repeat a single step Suppose you discover the morning of your big presentation that someone entered a few data values wrong That never happens No need to panic You just fix the values and StatView does the rest and your layout is preserved Save a template You can also save this whole presentation colors annotations and all as a template Then you could apply all these analyses and artistic efforts to a completely different dataset in one simple step e From the File menu select Save As e Change to your My Projects folder inside the template folder e Type a filename Nutrition presentation e Click Save 48 I Tutorial Notes Notes Where to go from here Now you have a feel for data analysis the
35. Tutorial p 1 or skip ahead to the chapter Analyses p 131 Ifyou closed your Simple Compact Variable and Complex Compact Variable datasets reopen them now e From the File menu select Open e Select a dataset e Click Open We ve already seen that compact variables look like regular variables in the variable browser except that their nominal components are indented underneath their continuous compo nents Small triangle controls and 5 let you show or hide the built in nominal parts j Variables Variables Show Show Compact Compact Expand Expand Data Complex Compact Variable Order Dataset order Data Complex Compact Variable Order Datazet order ip hoeterel readings 1 4 Cholesterol readings 1 4 Gender Fh Smoking status ca You can select any part of a compact variable by clicking it just as you would a regular vari able You can Add or Remove its parts the same way you would Add or Remove regular vari ables You can Split By the groups of a nominal part Let s try a few simple analyses with Simple Compact Variable Add both parts From the Analyze menu select New View 96 2 Datasets Compact variables e From the analysis browser double click Descriptive Statistics e Choose Basic Statistics and click OK e In the variable browser select Simple Compact Variable from the Data pop up menu if it is not already chosen
36. You can delete only categories that are not used by any dataset If you try to delete a category that is being used an error message appears variables Recall our discussion of Data arrangement p 51 in which we introduced the most typical way to arrange data one row in the dataset represents one case For example we could record cholesterol readings and gender data for eight people like this Gender Cholesterol male male male female male female male female The principal advantage of this data arrangement is that one row always represents one case Every cell in any column across one row describes one individual We could record weights ages and names and know that each value in a row corresponded to that person 2 Datasets Compact variables 85 However sometimes it is easier to visualize groups when their members fill separate columns like this Cholesterol readings Male Female 127 193 232 181 224 194 is d 131 You can easily see which readings are male which are female Now you identify a cholesterol readings group by its column not its row The advantages of this other arrangement are even more apparent when you have several nominal variables creating subgroups Here is the usual way to arrange such data Smoking Gender Cholesterol Smoker male 127 HNonsmoker male 232 Monsmoker male 224 Smoker female 193 Smoker male 1397 Monsmoker female 18
37. s a mistake we wont suffer too much Then we came back to the office miraculously without a single candy bar and sat down to enter the data Enter data by hand Here s a small part of the dataset Brand Name Serving Oz pkg Calories Total fat g Saturated pkg fat g M amp M Mars Snickers Peanut Butter 2 310 20 7 Hershey Cookies n Mint 1 55 230 12 6 Hershey Cadbury Dairy Milk 3 5 5 220 12 8 M amp M Mars Snickers 3 3 7 170 8 3 Charms Sugar Daddy 1 7 200 2 5 2 5 StatView s data organization These data are already organized the way StatView wants each case or observation each spe cific candy bar is in a horizontal row and various characteristics or variables appear in verti cal columns Each value occupies a cell in the table This row and column design is important It means that any value in any column belongs to one row one case one observation one candy bar and only one row The very organization of the numbers tells you something For example the 170 in the Calories column is not just any measurement of calories on any subject It corresponds exactly to the Name in the same row Snickers It corresponds exactly to the measurement of total fat in the same row 8 grams StatView holds its data in a dataset a spreadsheet format in which columns represent variables such as gender weight height and rows represent cases such as patients in a medical study or p
38. 2 Using a template is the same as reopening a view except that you can specify different datasets and or different variables with a template I Tutorial Present results 43 Present results So far we ve just explored our data and done some analysis It would probably be pretty hard to get anybody to pay much attention if we printed these analyses and tacked them up to a wall Let s pull it all together into an eye catching presentation Close the analysis browser Were done analyzing these data so lets make more room in the view window by closing its analysis browser pane e Double click the split pane control Ein the lower left corner of the view window ls L You can reopen the browser by double clicking the control now again Clean up results First let s straighten up these results space them evenly and move them off page breaks From the Layout menu select Clean Up Items e Click Clean Up Clean up by Location he Distance between item Vertical 25 inches Horizontal 25 inches T Ignore page breaks I Align to left margin Cancel Add some color ean Up H Distance between items Vertical inches Horizontal 25 inches Ignore page breaks Align to left margin Now let s highlight those analysis objects that concern saturated fat We can automatically select all those objects by working wi
39. Assign variables dialog box us163 variable browser Uss6 USI43 USI46 data source USG US77 R329 SR330 analysis generated Us77 change us77 dynamic formula us77 in examples usv1 Index SR StatView Reference US Using Statview 265 static formula us77 user entered US77 data type USSI US73 US76 SR316 SR318 SR321 SR332 SR359 category US73 US74 SR320 change us75 us76 convert SR33I currency US73 SR320 date time Us73 R321 import USIO3 US254 in examples usv1 integer US73 SR319 long integer us73 real US73 SR318 string US73 SR319 SR320 R338 dataset add columns us54 us62 close us72 common questions Us237 US240 copy US6s5 cut US65 delete us66 edit Us64 Us70 insert columns us62 paste US66 preferences US227 print us72 renamed vust58 save US7O scroll us64 split pane control uss5 summary pane see attribute pane transfer between Windows and Macintosh us7o us99 troubleshoot us252 window vus4 windows US54 US57 Dataset Templates Us233 US237 custom US240 Date sR369 date time data type US73 R321 fix imported values SR369 R427 format Us79 formats USVI US79 R330 SR33I functions R330 SR331 Date sR369 DateDifference sr370 Day sR371 DayOfWeek sr372 DayOfYear sR372 Hour sp382 Minute sr389 Month sr391 Now sR393 Second sr419 Time sR427 Weekday sr430 WeekOfYear sr431 Year SR431 group by month sr370 missing values sR321 valid data range US74 SR3
40. Click Hide definition or double click the triangle to close the split pane Select Print from the File menu to print the definition You cannot edit the formula definition EI Recode of Weight EENE Recode definition How to recode this variable if Weight lt 140 Selected Breakpoint E09 then Low else if Weight lt 200 234 then Medium 140 T 107 Hide definition Cancel Statview 4 5 To change a breakpoint select it and then drag it or enter a new value in the text box to change it or press Backspace Windows or Delete Macintosh to remove it A Recode dialog box is listed in the Window menu where you can select it to bring it to the front You can double click the top area beneath the title bar to bring its dataset to the front Select Print from the File menu to print a formula definition If you prefer you can use Formula to build your own recoding formula Recode is a shortcut Either way you can use the variable s Source pop up menu to view and edit a formula defini tion or to change from static formula to dynamic formula or user entered See Change sources p 77 By default recoded variables are based on dynamic formulas which means that changes or additions to the original variable automatically change the recoded variable New variables are appended at the right side of the dataset To mov
41. Click OK to create a crite rion named n Rows Included and apply it to the dataset A new random sample is 4 Managing data Edit Apply Criteria 129 included each time you select or define a random criterion If you need to do more compli cated random inclusions see RandomInclusion p 409 of StatView Reference A special dialog box lets you create complex random criteria For example you might want to include a random number generator seed so that a random inclusion can be reproduced See RandomInclusion p 409 of StatView Reference When you close a dataset that contains random criteria their definitions are saved but an exact row inclusion is only saved if the random criterion is in effect applied when you save Edit Apply Criteria You can apply an existing criterion two ways 1 From the Manage menu select Edit Apply Criteria choose a criterion and click apply 2 From the Criteria pop up menu in a dataset select the criterion You can edit existing criteria with Edit Apply Criteria only e From the Manage menu select Edit Apply Criteria e Select one or more criteria e Click Edit Selecting any criterion shows a preview of its definition Edit Apply Criteria Select a criterion No Criteria Delete New Random Domestics Imports Cancel Definition Country USA In a criterions dialog box you can eithe
42. Frequency distribu tions percentile plots and comparison percentiles always have two numeric axes Cell plots and box plots have a cell X axis and a numeric Y axis Pie charts have no visible axes Each type of axis has its own dialog box To edit an axis e Click axis numbers to select an axis e Click Edit Display e Choose options e Click OK Axis values can be rotated using commands Rotate Right and Rotate Left from the Text menu Note that long axis values are legible only when they are perpendicular to the axis Numeric axes StatView offers two dialog boxes for numeric axes one with fewer choices and one with more choices 7 Customizing results Graphs 191 Numeric Anis Numeric is S55 Bounds Numbers Upper Format Lower 50 Decimal places Include zero Ensure leading digit Intervals Linear i i Scale Linear Major width 50 Grid Lines at From Major interval width Scale _Linear Grid lines at zero zero Minor divisions Number format Free Decimal places 2 Tick marks Length Width Major 4 1 Always have leading digit More Choices e Cancel Minor 2 1 Fewer Choices sae Cancel Bounds Specify maximum and minimum values Check Include Zero if you want StatView
43. Horizontal 3 inches Numbers Decimal places Lr Always have leading digit ne Cancel Subsequent sections Graphs p 183 and Tables p 197 discuss these dialog boxes in detail Table format Row height gt EJ Transpose rows and columns Show Cancel Preview changes Both formatting dialog boxes have a Show button which lets you preview formatting changes as you make them After choosing options click Show to see the effects of the options on the graph or table in the view The dialog box is still open for you to adjust your choices Click OK to implement the changes or click Cancel to abandon all changes close the dialog box and return the graph or table to its original state Em c Gr Undo changes Any formatting changes you make whether through Edit Display drawing tools or layout tools are reversible If you make a mistake choose Undo immediately to return your graph to its previous state You can only Undo the single most recent action To undo earlier actions you must retrace your steps e g return to Edit Display and restore previous settings Clipboard commands You can use standard Edit menu commands with tables and graphs However please note that any StatView table or graph copied into the clipboard becomes a static picture when pasted back into
44. MANOVA see analysis of variance Mantel Cox test sRIsI SRI56 Mantel Haenszel test sRisi SRI56 marquee select us184 USI98 martingale residuals SRI70 sR178 SRI9O mathematical expression language us113 mathematical functions absolute value sr348 addition SR333 Average SR356 AveragelgnoreMissing SR357 Ceil sr359 Combinations sr361 CumProduct sr367 CumSum sR368 CumSumSquares sRr368 difference SR373 Div sR374 division sR334 DotProduct sr374 SR375 Erf sr376 exponentiation SR334 Factorial SR377 Floor sr380 Lag sr382 Ln sr385 Log sr385 log sr386 LogOdds sr386 Mod sr390 MovingAverage SR391 multiplication sR333 negative SR335 Norm sR392 parentheses sR336 Percentages sR397 Percentile sr398 Permutations SR399 Pi sr400 positive SR335 Remainder sr43 Round sr416 Sqrt sR420 Index SR StatView Reference US Using Statview subtraction SR333 Sum sR423 SumlgnoreMissing sr424 SumOfColumn sr424 SumOfSquares sr425 Trunc sRr429 matrix inversion SR433 Maximum sR387 maximum US60 US255 SR3 Mean sr388 mean USG6O USI84 SRI SR356 SR357 SR388 SR391 SR428 confidence interval around sr217 one sample t test sR23 mean difference confidence interval sr30 mean square SR73 means tables sr78 sR90 tutorial example us39 measure string values sR383 measurement units US217 Median sr388 median sR2 R398 median absolute deviation SR6 R387 memory requirements US232 US25I merge datasets US66 f
45. Regression Models dialog box Add columns to dataset bi Create baseline survival dataset To save the baseline values to a dataset check the Create baseline survival dataset check box then click OK Because formulas cannot be created in datasets created by analyses you must copy the Time and Cum Surv variables from the Baseline Survival Table dataset and paste them into a new dataset For clarity you might want to change the names of the variables in the new dataset to Event Time and Baseline Cum Surv respectively In this new dataset create 2 new columns one called Coefficients and the other called Covariate values This new dataset should then appear as below Baseline Cum Surv Coefficients Real Real Real Real User Entered User Entered User Entered User Entered i Free Format Fixed Free Format Fi Free Format Fi Bs Ee ee Po 9000 99 ef el Pp 5s000 sw e E 423000 9 fo e anc 9 Tips and shortcuts Common questions 247 From the Model Coefficients Table you know that the coefficient associated with cigarette consumption is 0 013809261 and that associated with type A personalities is 0 624020567 Type these two numbers into the first and second rows of the Coefficients variable Because these coefficients are to be used as exponents errors introduced by rounding will be greatly exaggerated For this reason we recommend that you record coefficient value
46. ReturnT sr416 probability value see p value ProbBinomial sr4or ProbChiSquare sr402 ProbF sr4o02 problems see troubleshoot ProbNormal sr403 Probt sr404 process SR251 process capability analysis sr261 product SR333 R367 SR374 product limit method Kaplan Meier sr149 product limit method Kaplan Meier sRi52 progress bar usi1 proportional hazards models sr167 SR175 SRIQI baseline hazard sri69 coefficients SRI69 confidence intervals sRI70 covariate values sRI69 residuals plots sr1z70 significance tests SRI69 stratification SRI7I stratified SRI69 protected least significant difference see Fisher s PLSD Q QC analysis us142 common questions US244 US245 example sR254 introduction sR25I SR256 QC c u analysis c u charts SR300 R303 SR306 control limits sr299 data requirements sR299 R302 SR303 dialog boxes SR301 SR302 discussion sR299 SR301 exercise SR305 SR307 nonconformity variable sr302 results sR303 SR305 standardize inspection criteria sR301 subgroups sR299 templates sr305 QC individual measurements analysis capability analysis sr279 CUSUM SR279 data requirements sR280 dialog boxes sR279 sR280 discussion sR277 SR279 exercise SR283 SR284 results sR280 sR282 templates sRr283 tests for special causes SR278 QC p np analysis sr255 control limits sR288 data requirements sR287 SR290 SR292 Index SR StatView Reference US Using Statview 279 dialog boxes
47. Type Weight Turning Circle Displacement Horsepower Gas Tank Size Cancel A new view opens and a scattergram with the same format appears 6 Templates Build templates 175 400 4 4 Ht 4 4 4 350 7 T 3 300 7 k ee c g 250 7 ot T 8 J DE g 200 7 hate if AO 150 7 E raed a 100 F p383 e F 50 HHHH 30 34 38 42 46 Turning Circle e Close this view and Car Data do not save changes In view preferences you might want to restore the option Save analysis results with view Modify an existing template A template lets you name and save all the parameters and output of a particular analysis You can modify parameters and add results to any of the templates that ship with StatView You can also start from scratch and build your own analysis template with the parameters and results you desire We will modify the Regression Simple template that ships with StatView That template includes a scattergram with a regression line and we want a template that also draws 95 confidence bands for the regression line Also we want to add a plot of residuals vs fitted val ues Again we want to use generic variable names From the File menu select New e Name the first variable Dependent Click the Input Column cell type Dependent and press Enter or Return e Name the second variable Independent From the Analyze menu select Regression Regression
48. USIO4 USI8I USI82 imported pictures and text us2z10 results into datasets Us233 unusual selection shapes us68 us70 Paste Transposed us68 patterns see pen patterns Pearson correlation sRI31 SR362 pen color us202 pen patterns colors US197 US212 graphs us185 us196 shapes vus2I1 tables us199 Us202 Percentages SR397 Percentile sr398 percentile plots SRI9 percentiles convert raw scores to R399 data requirements SRI9 dialog box sr19 exercise SR20 find several at once sR398 results sR20 templates sr20 percents of column totals table sri12 percents of row totals table sri12 period see missing values Permutations SR399 permutations ordered SR377 SR400 unordered sR361 Peto Peto Wilcoxon test sRI51 SRI56 phi coefficient sR113 Pi sr400 Pillai s trace sr82 placeholders us138 UsI5I USI64 R324 formulas usis plateau sR79 platykurtic srz plots us183 colors usI97 plotted lines us183 color usi97 PLSD see Fisher s PLSD plus SR333 R335 SR368 SR423 SR424 point charts see scattergrams point colors us195 us196 point sizes USI95 US229 point types USI95 US229 Poisson distribution sR410 polygon tool us205 US207 US208 polynomial regression sR52 SR64 SR65 polytomous logistic regression sR199 SR204 SR214 population statistics SRI portrait page US213 positive SR335 post hoc tests SR74 SR84 SR9O SRIO3 assumptions SR84 Bonferroni Dunn sr86 cell contributions table sr112 Dunnett s sr87 Fish
49. a Name a Serving pkg Oz pkg my Calories io Saturate Cholesterol g 3 Sodium mg Carbohydra Dietary fib ur analysis is updated in place to show just calories Also the variable browser updates so O ly dated lace to show just cal Also th ble b dat that only Calories has an X marker Pescriptive Statistics a Mean Std Dew Std Error Count Minimum Maximum Missing ae aomes 243 027 61 996 7 159 75 125 000 450 000 a Notice the analysis is really short and wide It might look better if we flipped it sideways Edit a display e Make sure the analysis is still selected e Click the Edit Display button at the top of the view window E_ _ _ _ _hhaal a _ X Untitle E Recalculate Edit Analysis CEN Create Analysis Show aaah Pescriptive Statistics e Check Transpose rows and columns click the checkbox so it has a check mark e Click OK Numbers Format Free Format Fixed x Format Free Format Fired Decimal places 3 x Decimal places La I Always have leading digit Always have leading digit Table format Table format Row height Row height gt I Transpose rows and columns Eg Transpose rows and columns Show Cancel Show Cancel As easy as that we ve transposed the whole table 24 I Tutorial Analyze data
50. and Paste the data into the empty column To create a random variable in another location insert a column and change its source to Static Formula or Dynamic Formula Create criteria When you use a variable in an analysis all the values in that variable are included in the anal ysis unless you exclude some rows cases All rows are included in analyses by default If you prefer to restrict your analyses to a subset of the rows in a dataset you can use Include Row and Exclude Row commands Criteria commands or both Include Row and Exclude Row are discussed separately under Include and exclude rows p 108 Use Criteria to specify algebraically which cases to include in calculations When you apply a criterion any analysis results in any open view window using those data automatically recalcu late to show the new results unless you turn Recalculate off see Control recalculations p 138 You can tell whether a row is included or excluded by its row number Included row numbers appear in regular dark type Excluded row numbers are dimmed or grayed These characteris tics are also visible if you print your dataset Caution Many columnwise functions have a final argument that controls which rows of the column are used for computations AllRows OnlyIncludedRows or OnlyExcludedRows When you use the latter two arguments be aware that any Include Row Exclude Row and Criteria commands you use will cause these fo
51. below and to the right of the original Duplicate is equivalent to using Copy and then Paste Clear removes a selected object or component a table or graph or a part of one from the view but does not place a copy in the clipboard You can select and clear the following compo nents titles entire legends not parts of legends axis labels and notes If you select and clear a plot the representation of a variable or group within a graph axis whole graph or whole table the entire table or graph is cleared If you mistakenly Clear a graph title or legend you can restore it with Edit Display s Show title and Show legend options Paste places the contents of the clipboard in the view at the point of the most recent mouse click If the most recent mouse click location is no longer visible Paste centers the pasted object s in the visible portion of the view If you resize a pasted picture you can restore its original size by double clicking the object 7 Customizing results Graphs 183 Graphs You can edit many aspects of a graph s appearance Some formatting changes apply to an entire graph others to particular components of a graph To make changes to an entire graph you select the entire graph and click Edit Display To change a component of a graph you select just that component and click Edit Display To make still other changes you select graphs or components and work with Draw tools This diagram identifies man
52. decimal places Engineering Displays the numbers in scientific notation using exponents that are multiples of e3 and e 3 Enhanced Free Fixed Displays real numbers like Free Format Fixed except that it switches to scientific notation whenever doing so results in a more accurate representation of the number e g 0 0078 is 008 for three decimal places in free fixed but enhanced free fixed shows 7 8E 3 Currency StatView supports most major international currency formats Examine the pop up menu to see the choices available to you Your choices vary according to platform operating system and international configuration Date time StatView supports most major international date time formats Examine the pop up menu to see the choices available to you Your choices vary according to platform operating system and international configuration Regardless of format chosen date time values are always stored and interpreted as a complete date an exact number of seconds since the earliest possible date which varies by platform Change formats Feel free to choose the format most convenient for the moment Values are always stored and used in computations to the fullest precision of the platform you are using Your format choice affects only the display of values Therefore changing formats is usually harmless However when you exchange data with other applications through the clipboard or text files be sure to
53. e Shift click or click and drag to select both Cholesterol readings and Category for Choles terol readings If you don t see Category for Cholesterol readings click the triangle control e Click Add Descriptive Statistics Cholesterol readings Mean 192 375 Std Dev 31 622 Std Error 11 180 Count 8 Minimum 127 000 Maximum 232 000 Missing 2 Nominal Descriptive Statistics Levels Count Missing Mode Category for Cholesterol re 2 10 0 Because we Added both variables both are treated as X variables We could instead Split By the nominal part after we Remove it from its current X role e Make sure the analysis object is still selected still has black handles In the variable browser select Category for Cholesterol and click Remove e Click Split By Descriptive Statistics Split By Category for Cholesterol readings Cholesterol readings Total Cholesterol readings Male Cholesterol readings Female Mean 192 375 194 200 189 333 Std Dev 31 622 41 385 7 234 Std Error 11 180 18 508 4 177 Count 8 5 3 Minimum 127 000 127 000 181 000 Maximum 232 000 232 000 194 000 Missing 2 0 2 We can work with the Complex Compact Variable similarly Let s look at box plots of Choles terol readings split by both nominal parts e Click in the white space of the view to deselect the tables e From the analysis browser double click
54. exercises USI50 USI56 Force button sR53 Formula dialog box usro keyboard shortcuts see StatView Short cuts card X Variable button sr229 Y Variable button sr229 variable names rules sR317 variable summary pane see attribute pane variable symbols see usage markers class markers variable types see data type VariableElement sr429 variables sr324 SR325 assign USI44 delete us116 names US58 change us58 generic for templates us17o tutorial example us4 requirements also see data requirements under specific analysis slots for templates us163 vs columns Us53 Variables dialog box sr267 variable wise see columnwise Variance SR430 288 Index SR StatView Reference US Using Statview variance USGO SR4 chi square test SR24 comparison sR4I test homogeneity us235 varimax SRI35 vectors R374 also see variables velocity handle usz09 vertical alignment us218 vertical see columnwise View menu Us148 US149 View pop up menu USI47 views background color us212 clean up uS213 US214 document limits us213 documents vs templates Us161 Edit commands us181 file formats UsI56 USI57 fonts Us232 grid lines us217 hairlines us232 Open vusi57 UsI use different datasets us159 use different variables us158 use original variables us157 preferences USI4I USI7I USI79 USI80 US232 US233 print Usts9 Us160 rulers Us217 save results USI7I US232 save templates us157 save views as views USI56 vs templates us161 us162 wind
55. for instance you would like to evaluate the survival function for the proportional hazards model you computed in the example of the individual who smokes 18 cigarettes per day on average and has a type A personality see Exercise p 191 of StatView Reference From the 246 9 Tips and shortcuts Common questions exercise the model coefficients table with results displayed to 9 decimal places for that model is Model Coefficients for Time Censor Variable Censor Model Proportional Hazards DF Coef Std Error Coef SE Chi Square P Value Exp Coef 013809261 007746646 1 782611662 3 177704337 0746 1 013905049 624020567 272504339 2 289947269 5 243858496 0220 1 866417032 a Cigarettes Personality Type A ae Using the saved values of the baseline survival function it is a straightforward task to write a formula in StatView that will compute the survival function for specific covariate values of your own choosing To save the baseline survival function to a dataset e Select one or more of the regression model results in the view then click the Edit Analysis button Note that the following change to the regression models parameters could be made at the time of analysis creation as well The fewer choices version of the Survival Regression Models dialog box now appears e Click the More Choices button This brings up the more detailed version of the Survival
56. g L 2 5 4 E 0 0 T T T T 100 120 140 160 180 200 220 240 100 140 180 220 260 300 Weight Cholesterol Similarly you can assign as many Split By variables to an analysis as you want The way Stat View handles extra variables varies according to the analysis For some analyses Split By changes simple tables or graphs into compound tables or graphs that show results for all the groups at once For example adding a Split By variable to a Descriptive Statistics table breaks the table into a compound table with a column for each 5 Analyses Overview 137 group Similarly adding a Split By variable to a box plot changes it to a compound box plot where a single graph frame contains boxes for each group Descriptive Statistics Split By Gender Weight Total Weight male Weight female Mean 158 653 169 282 127 208 Std Dev 28 389 23 288 16 208 Std Error 2 913 2 764 3 308 Count 95 71 24 Minimum 107 000 107 000 110 000 Maximum 234 000 234 000 190 000 Missing 0 0 0 2405 2207 200 7 180 5 Units 160 5 140 4 1204 Box Plot Split By Gender 8 8 o 9 100 Weight o male o female For other analyses Split By produces multiple tables or graphs for each group For example adding a Split By variable to a frequency distribution analysis produces separate summary tables and histograms for each gr
57. g and another and if we add more text it wraps downward at the chosen Tine width You can also set the width of a text object in advance rather than clicking once and beginning to type click and drag a box with the text tool and then start typing 8 Drawing and layout Draw tools 205 Change text objects Use Text menu commands to format text objects you add with the text tool or to customize text components of graph and analysis objects see Change text items p 187 and p 200 You can change font size style and justification left alignment right alignment or center ing used in text objects Just select characters with the text tool or a whole text object with the selection tool and then choose commands from the Text menu You can rotate entire text objects left or right to make text sideways or upside down Just select the text object s with the selection tool then select Rotate Left or Rotate Right com mands from the Text menu You can edit rotated text just as you do normal text While you are editing rotated text StatView displays the text in normal position When you are finished editing the text again appears rotated You cannot rotate just selected characters Use the pen color tool in the Draw palette to change the color of selected characters or objects see Change colors p 212 Draw objects You can draw the following shapes Shape Tool K
58. gt Percentiles Nna Saranla gran PerOC Subgroup Mea abar Statistics Line Chart Needle Chart Bar Chart Point Chart Results Table e R Statistics S Statistics Summary Table DOC Individual Mea ries These will help us determine whether the variable is normally distributed pan co E AE gt CUSUM Statisti e Click somewhere in the white space of the view window to be sure no results are selected This way we avoid adopting variables from any analyses that are selected e Click the triangle next to Frequency Distribution to show the detailed list e Click and drag to select both Summary Table and Histogram e Click Create Analysis Show All Analyses Order Default Descriptive Stati Frequency Distrib Sur H Score Histog Fie Chart gt Percentiles Nna Saranta Anali e Click OK to accept the default analysis parameters 32 I Tutorial Analyze data Frequency Distribution Number of intervals Show normal comparison Do you wish to enter your own interval information no Qyes width initial value Intervals indicate Count include Lowest value Tables show Counts Percents Relative frequencies Histograms show oa e In the variable browser select Calories and click Add It looks lik
59. slots in the template Template Variables Measurement Data Normality Test LN order Dataset order Actual Ideal measures gt Measurement Mo Actual Ideal measures ki 5 A new view shows a Kolmogorov Smirnov table and two histograms with normal curves If the result of the K S test is significant i e p lt 0 05 then the Actual and Ideal variables are probably not from the same distributions This implies that the Actual variable is not nor mally distributed because the Ideal variable is You should be aware however that the K S test is sensitive to outliers so you should inspect your data for such cases The histograms with fitted normal curves let you compare the distributions visually Compute Bartlett s Test and Compute Welch Test Bartlett s test is a test of homogeneity of variances among groups A significant p value means that group variances are not equal Welch s test is a robust one way anova Use it instead of one way ANOVA when Bartlett s test shows that group variances are not equal Both templates work the same way we ll demonstrate Bartlett s with the Lipid Data From the File menu select Open e Select Lipid Data from the Sample Data folder e Click Open e From the File menu select Open Select Compute Bartlett s Test from the Dataset Templates folder e Click O
60. the target location for the pasted data such as the input row or the input column Size of the target area The selected area can have four basic sizes relative to the data in the Clipboard 1 It can be the exact size of the data in the Clipboard In this case StatView pastes an exact copy of the data 2 It can be smaller than the data in the Clipboard In this case StatView pastes as many val ues as it can starting in the upper left cell and leaving out the additional data For example if the Clipboard contains an array of numbers 3 columns wide and 3 rows deep and the selected area is only 2 by 2 only the first four data points of the source data 2 by 2 2 Datasets Edit data 67 are pasted Paste does not shift cells to the right of the selected area further to the right nor move the cells underneath the selected area down 3 It can be arger than the data in the Clipboard and an exact multiple In this case Stat View duplicates the data as many times as necessary to fill the selected area 4 It can be arger than the data in the Clipboard but not an exact multiple In this case Stat View copies the data only once and fills the remaining cells in the selected area with missing value symbols This is a handy way to fill many cells with one value or several repeating val ues Data type to be pasted When you paste data into a dataset the pasted data are converted to the selected area s data type if the types a
61. 0 191 232 35 674 3 660 95 115 000 285 000 0 Box Plot 300 J 280 7 260 7 240 7 2207 200 4 180 4 160 4 140 4 1204 o 6 0 000 00 es 8 8 100 Weight Cholesterol example assigning two variables to a frequency summary table produces two separate tables and assigning two variables to a histogram produces separate histograms for each group Frequency Distribution for Weight Frequency Distribution for Cholesterol From 2 To lt Count From 2 To lt Count 107 000 119 700 8 115 000 132 000 4 119 700 132 400 13 132 000 149 000 6 132 400 145 100 10 149 000 166 000 11 145 100 157 800 16 166 000 183 000 19 157 800 170 500 12 183 000 200 000 21 170 500 183 200 18 200 000 217 000 15 183 200 195 900 11 217 000 234 000 10 195 900 208 600 3 234 000 251 000 2 208 600 221 300 2 251 000 268 000 4 221 300 234 000 2 268 000 285 000 3 Total 95 Total 95 Histogram Histogram 20 ee eee DES L l AAR AET ADE PE AR ENCENS MERN AEA PAE STE MEA pei ps ee SE DD EN 184 t 20 4 J 164 i 17 5 4 H ta f 154 fay F 5 ne 512 5 7 L 0 107 mi F 8 E L 6 7 54 4 E aa j
62. 000 13 Hershey Cadbury Caramedlo 10 526 14 Hershey Cadbury Dairy Milk 3 091 15 Hershey Cadbury Fruit amp Nut 2524 16 Hershey Cadbury Roast Alnor 3 091 Fi Daweh eres FETTE oa Eel We also made the whole window small for this illustration You can pick a size you like It s easy to see you could have 12 Cup O Golds or 8 Almond Joys or 4 Mr Goodbars or 11 Peppermint Patties Oops We forgot that these data are per serving not per bar and some of the candy bars are so big they have several servings per package We need to fix that formula e Open the attribute pane Double click the attribute pane control From the Source pop up menu for Bars per day select Dynamic formula Click and hold that cell then release the mouse button User Entered Static Formula namic Formula fr piged Now we just edit the formula in the dialog box to have another division term e Click just after the existing formula e Click in the keypad area of the formula dialog box or type a slash e Double click Servings pkg from the list of variables e Click Compute Formula of Bars per day Formula variable definition Order Dataset order Serving pkg Oz pkg Calories Total fat g Saturated fat g Cholesterol g Sodium mg Order by Function Type gt Date Time gt Logical gt Mathematical gt Probabilities gt Random Numbers gt Series gt Special Pu
63. 2 913 Count 95 Minimum Maximum Missing ol 4 e Make sure the rectangle is still selected has black selection handles e Select a fill pattern from the Draw palette F l j bj Descriptive Statistics Eo Weight fill None None Mean 158 653 oo Tine Std Dew 20 409 Std Error 2 913 Count 25 Minimum 107 000 Maximum 234 000 Missing i pp Don t be alarmed Your filled rectangle should now be in front completely hiding the table behind it e Make sure the rectangle is still selected Click the Move to back tool in the Draw palette pee ene e Select the rectangle tool and draw a smaller rectangle around the table e Choose a solid white fill for this rectangle Make sure the smaller rectangle is still selected e From the Layout menu select Move Backward Make sure the smaller rectangle is still selected Now you have a layer of three objects a table in front a white rectangle in the middle and a gray rectangle at the back 8 Drawing and layout Layout tools 217 Descriptive Statistics Weight Mean 158 653 Std Dev 28 389 Std Error 2 913 Count 95 Minimum 107 000 Maximum 234 000 Missing 0 Rulers and grid lines Ruler and grid commands in the Draw palette help you align objects neatly Show Rulers Turn Grid Off Y Show Grid Lines Hide Page Breaks
64. 23 000 0 000 1 0 000 0 000 11 667 1 528 252 3 10 000 3 000 n 11 219 4 301 1 075 16 2 500 20 000 10 000 1 10 000 0 000 14 000 1 14 000 14 000 10 333 1 562 760 5 000 2 000 16 000 1 16 000 16 000 11 500 2 121 1 500 2 10 000 3 000 16 000 1 16 000 16 000 11 000 1 11 000 1 000 5 500 2 121 1 500 2 4 000 7 000 4 000 2 520 2 000 2 2 000 6 000 Cloning an object makes a new copy of the object using the new variables leaving the original object unchanged We could have added Total fat g to the original table instead but two sep arate tables are easier to read Notice that Split By Brand is still in effect Let s clone this analysis for Saturated fat g also e Make sure the analysis is still selected e In the variable browser select Saturated fat g Descriptive Statistics petit By Brand Saturated fat g Total Saturated fat g Adams amp B Saturated fat g Annabelle Saturated fat g Bit 0 Honey Saturated fat g Brown amp H Saturated fat g Charms Saturated fat g Hershey Saturated fat g Just Born Saturated fat g Leaf Saturated fat g M amp M Mars Saturated fat g Myerson Saturated fat g Nabisco Saturated fat g Nestle Saturated fat g Pearson Saturated fat g Sherwood Saturated fat g Standard Saturated fat g Tobler Saturated fat g Tootsie z Saturated fat g Weider B Mean Std Dew Std Error Count Minimum Maximum Control Shift click Wi
65. All rows are included in analyses by default If you prefer to restrict your analyses to a subset of the rows in a dataset you can use Include Row and Exclude Row commands Criteria commands or both Criteria are discussed separately under Create criteria p 124 and Edit Apply Criteria p 129 Use Include Row and Exclude Row to restrict analyses to certain rows When you exclude rows or include rows that were excluded any analysis results in any open view window using those data automatically recalculate to show the new results unless you turn Recalculate off see Control recalculations p 138 e Select the row s by clicking row number s To select multiple adjacent rows Shift click or click and drag their numbers To select non adjacent rows Control click Windows or Command click Macintosh their numbers To select all rows use Select All Rows from the Edit menu e From the Manage menu select Include Row or Exclude Row You can tell whether a row is included or excluded by its row number Included row numbers appear in regular dark type Excluded row numbers are dimmed or grayed These characteris tics are also visible if you print your dataset Shortcut Double click any row number to toggle the row between included and excluded Caution Many columnwise functions have a final argument that controls which rows of the column are used for computations AllRows OnlyIncludedRows or OnlyExcludedRows When
66. Always show Assign Variables dialog bou Open original dataset This choice opens the dataset s the view uses Usually StatView can find the original dataset even if it has been renamed or moved If not StatView asks you to find the dataset Be applied to different dataset s This choice lets you select a different dataset so you can use the view as a template for different data Create new view This choice displays the results in a new view window Add to top view This choice places the results at the end of the topmost current view win dow after all existing tables and graphs If no views are open this choice is dimmed Always show Assign Variables dialog box Check turn on this option when you want to use the view as a template for different data From the original dataset If you want to use the original dataset but make different variable assignments check Always show the Assign Variables dialog box turn the option on The dialog box shows the original assignments Drag variables back and forth between the assignment slots and the scrolling list until you have the assignments you want 5 Analyses Print a view 159 Assign Variables for Nutrition analysis Please double click or drag the desired variables into the proper slots in the template Template Variables Calorie groups A Data Candy Bars 2 Calorie groups Order Dataset order tal fat g
67. Box Plot In the variable browser select Complex Compact Variable from the Data pop up menu e Select Cholesterol readings Gender and Smoking If you don t see the nominal parts click the triangle control Shift click or click and drag to select all three parts e Click Add 2 Datasets Compact vanables 97 Box Plot Grouping Variable s Gender Smoking status 250 5 240 4 230 5 3220 4 z 32104 z 200 4 L 1904 T B 1804 F 21704 9 1604 M 150 7 140 Male Smoker Male Nonsmoker Female Smoker Female Nonsmoker You might want to try removing the nominal parts and then reassigning them to the analysis as Split By variables The results convey the same information in a different way Box Plot Split By Gender Smoking status 250 5 240 f 230 4 F 220 7 Is 210 4 ls Male Smoker 200 C Male Nonsmoker 190 4 El Female Smoker 180 4 F Female Nonsmoker 170 4 160 5 F 150 4 H 140 Units Cholesterol readings For information on using compact variables in analyses see Compact variable p 40 and Repeated measures ANOVA p 82 of StatView Reference 98 2 Datasets Compact vanables Importing and exporting StatView can read and write several file types directly and it can import data saved by other applications in
68. Choose the placement for major and minor tick marks from pop up menus Specify a length in points for major and minor tick marks It is common to make major tick marks longer than minor tick marks Specify the width thickness for major and minor tick Tick ma ee yf 1 mark lines in pixels 192 7 Customizing results Graphs Scale Axis scales are linear by default or you can choose logarithmic scales of base e 10 or 2 Grid lines Choose whether to display grid lines at major or minor ticks at zero or not show any grid lines Horizontal and vertical grid lines can help show where values fall Numbers Choose numeric format and number of decimal places see Format p 79 and Decimal places p 80 Check Ensure leading digit to include a leading zero in fractional values e g 0 25 To change formatting for a date time axis change the display format for the date time variable in the dataset attribute pane Cell axes Your formatting choices for cell axes are not as extensive as for numeric axes You can change the style of tick marks the position of axis values how many of them are shown and deter mine whether or not to show grid lines There is only one dialog box for cell axes To make any of these changes select the axis by clicking on its values and click the Edit Display button The following dialog box appears Cell Axis Tick marks v Stagger rittt 7 Show Every label
69. Click Import mpor meor Please specify how this text file looks Please specify how this text file looks Items may be separated with tabs and Items may be separated with tabs and spaces 7 commas M retums spaces commas returns Number of variable Number of varigibies I Convert small integers to Categories CO Convert small integers to Categories PM Import non numeric data as type string EJ Import non numeric data as type string I Make variables with errors have type string Make variables with errors have type string Cancel Cancel Import The default settings are appropriate for most files most Windows and Mac applications use the tab delimited format for text files Many older pos programs and programs on other plat forms use commas and or spaces When in doubt use a text editor to examine your file Items may be separated StatView recognizes tabs spaces commas returns or any user entered character as separator characters If you use returns as separators you must spec ify the number of variables Convert small integers to Categories If your source application uses integers rather than text values to code the levels of groups you might want to use this option For example a variable containing the values 1 and 2 to represent Male and Female can be converted to a category with two groups Group 1 and Group
70. File menu 162 6 Templates Use templates Templates let you do several things 1 Create any analysis without using the analysis browser or variable browser StatView has numerous pre built templates that perform the analyses in the analysis browser and many other special combinations of analyses suited for particular tasks You select templates from an hierarchical Analyze menu You can add your own templates to the menu and you can reorganize the hierarchy of the Analyze menu to suit your needs 2 Combine different types of analyses into sets you can compute all at once For example you might build your own template combining QC charts histograms cell plots and your own special anova model Then you can use the template to do these analyses with each new dataset 3 Replay complex sequences of analyses graphs annotations and formatting choices consis tently and effortlessly If you have to produce the same sales report each month build the report once save the view and reuse it as a template with each month s new sales numbers 4 Harness the expertise of a consultant You can take a problem to a specialist who can assemble the appropriate analyses in a template that you can use over and over again 5 Build analyses that others can repeat error free without your assistance All they need to learn is how to open the template and assign the right variables Use templates Templates reduce analyses of any complexity to two
71. Format Fixed ld Dec Places e Z 3 Z 3 1 MM Snicker 1 000 2 000 310 20 000 7 000 2 Hershey Cookies 1 000 1 550 220 12 000 6 000 Z Hershey Cadbur 3 500 5 000 220 12 000 8 000 4 M amp M Snickers 3 000 3 700 170 8 000 3 000 S Charms Sugar 1 000 1 700 200 2 500 2 500 E MMA Twix Po 1 000 1 710 260 16 000 5 000 F Hershey Twizzler 1 000 2 200 190 1 500 0 000 8 Tobler Toblero 1 000 1 230 130 11 000 7 000 9 Nestle Crunch 1 000 1 550 220 12 000 7 000 10 Hershey Almond 2 000 3 220 220 13 000 8 000 11 Sherwood Elana M 1 000 1 600 200 10 000 6 000 12 Hershey Krackel 1 000 2 600 390 21 000 13 000 13 M amp Ms M amp Ms 1 000 1 740 250 13 000 5 000 14 Bit 0 H Bit 0 1 000 1 700 200 4 000 2 500 15 Nestle 100 Gr 1 000 1 500 200 8 000 5 000 16 Hershey Skor 1 000 1 400 220 13 000 9 000 17 Hershey Twix C 1 000 2 000 z280 14 000 5 000 18 M amp M Milky 1 000 1 570 160 5 000 3 500 19 M amp M Mars 1 000 1 760 240 13 000 4 000 20 Pearson Peanut 1 000 2 500 340 16 000 3 000 21 Nestle Raisinet 1 000 1 580 200 8 000 4 000 ki a SE StatView converts Excel data types and formats to the nearest StatView equivalents For details on how this works see the chapter Importing and exporting p 99 We only need to change a few of the variable attributes e Change Brand from type string to category Click and hold the Real cell in the Brand column
72. Import non numeric data as type string turn the option off e Check Convert small integers to Categories turn the option on e Click Import Column 1 Column 2 Column 3 Column 4 Column 5 Column amp Column 7 Categor Date Ti User Entered User Ent User Ent User Ent User Entered User Ent User Ent Nominal Continuo Continuo Continuo Nominal Nominal Nominal if if04 C1 234 Free For a 1 Unlabeled group 1 12 1 87 1 1 0 Unlabeled group 1 red Charlie Unlabeled group Z 12 3 87 20 001 3 0 Unlabeled group Z blue Miles Unlabeled group 3 12 5 87 40 001 5 1 Unlabeled group 5 red John Unlabeled group 1 127 7787 60 001 5 3 Unlabeled group F blue Roscoe 4 Unlabeled group 2 127 9 67 70 001 3 4 Unlabeled group amp green Mitchell Unlabeled group 3 127 9787 20 001 5 5 Unlabeled group 9 a 237 Unlabeled group 1 127107 90 001 se red Ra Columns 1 5 are now category variables with simple group labels You can change group labels for categories so they are more informative see Edit category definitions p 83 Col umn 6 is also category notice that in row 3 the value 1 is interpreted as blue the first group in the category and in row 9 the value 10 is missing since the category has only four groups Older StatView products Macintosh only Text Many StatView users also use Super
73. L Continuous Thursday 5 94 10 7 oO Friday Continuous Friday 6 09 15 o Saturday Continuous Saturday 8 97 Continuous Formulas and criteria How can edit formulas From the Source pop up menu select Dynamic or Static Formula again This reopens the Formula window so you can edit the formula definition When you re done making changes click Compute Control click Windows or Command click Macintosh Compute to calcu late the formula without closing the Formula window b sl y User Entered i b Ss Tupe Real Static Formula Dy nar E ggi Free Format F Free Format Fi p _Dec Places How can edit criteria From the Manage menu select Edit Apply Criteria Select the criterion you want to change and click Edit When youre done making changes click Apply to use the criterion Save to save changes without applying the criterion or Select to highlight all rows meeting the crite rion Control click Windows or Command click Macintosh Apply to apply the criterion without closing the Criteria window Copy and Paste your favorite formulas If you find yourself using the same formulas over and over you might want to Copy them from the Formula window and Paste them into a text document or the Scrapbook Macin tosh Then you can quickly reuse them in any dataset Build dataset templates If you regularly need to use the same formulas or criteria on different datasets for instance for a mo
74. M amp M Mars Nestle Start by double clicking Brand Now you have a new set of choices double click ElementOf Your choices change again click Hershey then M amp M Mars and finally Nestle Notice how StatView guides you through each step so you don t have to learn any special rules e In the Criteria name box type a name for the criterion Big Three e Click Apply Criteria 1 of Candy Bars Data E Criteria name Big Three Criteria definition Brand ElernentOt A Select a category group Annabelle Bit 0 Honey Brown amp Haley Charms Just Born Look at the dataset window Notice how the row numbers for candy bars made by other man ufacturers are dimmed indicating that the cases are not included in analyses Also the Crite ria pop up menu shows the criterion in effect I Tutorial Analyze data 29 Candy Bars Data Sate Cup Gold 4 Annabelle Abba 2abba 5 Eig Hunk 40 1 3 Annabelle Look 5 iy Annabelle U No Blue is Annabelle U No Green 13 Bit O Honey Bit O Honey EE Brown amp Haley Almond Roca ti ie Sugar Daddy EE Hershey Sth Avenue 4 Almond Joy ee Hershey Bar Mone 2 iz Harshan PO sdhoen Car artis Now look at the view window click the window or select Untitled View 1 from the Win dow menu Notice how all your analy
75. Only Contingency Tables analysis and certain QC analyses can analyze summary data Two way table In a two way table each column is a column of the contingency table and each row a row of the table The observed frequencies are entered in individual cells There will be as many col umns as groups in one nominal variable and as many rows as groups in the second nominal variable Our example would look like this Column 1 Column 2 The two columns represent the two gender groups male and female The three rows the three eye color groups blue brown and green The values in each cell are the counts for the partic ular combination Note that you cannot record information about individuals you can only count how many individuals fall in a group Only Contingency Tables can analyze two way table data Correlation matrix A correlation matrix is a tabular arrangement of data with a correlation coefficient in each cell of the table You can use Factor Analysis or Correlation Covariance analyses to create a correla tion matrix dataset You can also enter a correlation matrix by hand Only Factor Analysis can analyze correlation matrix data Columns vs variables For most purposes and in the most typical one case per row data arrangement the terms col umn and variable are interchangeable However we do make a distinction Strictly speaking a column is the vertical arrangement of cells in a dataset A column usually contains a var
76. Peseriptize Statistjes Calories Mean 243 027 Std Dew 61 996 Std Error Tisa C i Count 75 Minimum 125 000 Maximum 450 000 Missing a C Just as easily we could have changed the table s number formats borders and row height Edit analysis parameters You may not be surprised about how easy it is to transpose a table s display Would you believe it is just as simple to change the parameters of an analysis e Make sure the analysis is still selected e Click the Edit Analysis button at the top of the view window Now we see the same dialog box of analysis parameters as when we first created the analysis Almost all of StatView s graphs and analyses have a set of options for specifying exactly how to complete the analysis and you can always change your mind by clicking Edit Analysis and making new choices e Click More choices Descriptive Statistics Choose which statistics to compute Basic C Complete Cancel This expanded version of the analysis parameters dialog box lets us choose exactly which sta Descriptive Statistics Choose which statistics to compute Basic Complete More choices Cancel More choicej tistics we want Scroll down to see all the possibilities Since we know our dataset has no missing values let s save space by turning off the Number missing option e Uncheck Number missing Click the box to remove the check mark I Tutorial
77. Saturated fat g qr Descriptive Statistics Plot for Total fat g Br i Flot for Total fat g Sat View Nutrition analysis Order by Location Show Lan v Fage 1 Descriptive Statistics Descriptive Statistics Descriptive Statistics Box Plot Plot for Calor Frequency Distribution Page 2 Frequency Distribution Frequency Distribution Frequency Distribution qr Fage 3 View Nutrition analysis Order by Variable Show Lan sCalorie groups Candy Bar Unpaired Comparisons Unpaired Comparisons Box Plot Plot for Total Calories Candy Bars 2 Frequency Distribution Box Plot Plot for Calor Frequency Distribution Descriptive Statistics vz Carbohydrate g Candy Ba Frequency Distribution Results are grouped under headings appropriate to the Order chosen Each heading has a tri angle control Click a triangle to tip it upward f and hide its results Control click 148 5 Analyses Analysis windows Windows or Command click Macintosh any triangle to tip all triangle controls up or down in one step The Show pop up menu lets you choose whether to list all results in a view or only selected results Since selection plays such an important role in the analysis process knowing what is currently selected helps you understand what will happen when you create a new analysis or assign or remove variables Select results The Select button at
78. Show variables The Show button selects a variable in the dataset and scrolls the dataset so the variable is visi ble If you select more than one variable Show scrolls the dataset to make the leftmost vari able visible Select one or more variable names To select one variable click its name To select several adjacent variables click and drag or Shift click their names To select several nonadjacent variables Control click Windows or Command click Macintosh their names e Click Show A shortcut for Show simply double click a variable name The topmost button of any browser is always the default button double clicking an item has the effect of selecting the item and pushing the top button Compact and expand variables The Compact and Expand buttons let you create and expand compact variables directly from the variable browser rather than by selecting the columns and using the Compact and Expand buttons in the dataset window itself e Select the variables to compact or expand e Click Compact or Expand For a discussion of compact variables see Compact variables p 84 Enter data The chapter Tutorial p 1 presents a step by step exercise for entering a sample dataset If youd like to get a feel for the overall process of entering data and working with a dataset the tutorial is a good place to start This chapter discusses in more detail all the rules variations and shortcuts for entering and ed
79. Simple Drag Dependent to the Dependent Variable slot and Independent to the Independent Variable slot e Click OK The view window fills with a number of empty analyses Notes beneath the placeholders read There were not enough observations to compute this result That s not surprising our dataset has no cases It doesn t matter though we can still make our changes and save a new template e Make sure at least one of the empty results is still selected e Click Edit Analysis e Click More Choices e Check Slope turn the option on for Plot confidence bands e Click OK 6 Templates Build templates Regression Model No intercept in model Felp anier Save to dataset Residuals Fitted Predicted Camnute values for Wir iuie ros O atl rages Confidence level T Plot confidence bands for Mean x Slope Fewer choices Cancel In the analysis browser under Regression double click Residuals vs Fitted We ve changed the regression plot to include confidence bands and we ve added a plot of residuals vs fitted values Now we can save our view as a new template From the File menu select Save Specify a filename Regression Simple conf Choose a location inside the Template folder Windows or StatView Templates folder Macintosh perhaps the Regression folder Click Save StatView war
80. StatView way The easy way There is much more to learn Many more tools are available for helping you manage your data Lots of drawing and layout tools give you complete control over every detail of your pre sentation Scan through the remaining chapters of Using StatView to learn what s available StatView offers a broad range of statistical analyses and graphs not just the sample we ve used A comprehensive StatView Reference devotes an entire chapter to each type of analysis you see in the browser Finally as you get comfortable with StatView you ll find countless shortcuts tricks and pow erful variations for how to get your work done These manuals the Hints window Balloon Help and Apple Guide Macintosh only and Windows Help Windows only are available to help you But no matter where you concentrate your efforts all the basic techniques you ve just learned will serve you well All of StatView s analyses work the same basic way whether you use browsers or templates whether you re using basic tables and graphs or specialized survival and quality control analyses Afterword No candy bars were injured or mistreated in the making of this tutorial All scenes involving candy bars were carefully supervised by humane chocolate loving professionals Datasets The first step of any analysis is to get data into a dataset Stat View s dataset is a column by row spreadsheet style window This chapter discusses how to arran
81. Triglycerides HDL E LDL ideal bod E F ideal body wt ideal body wt Height E Heinht A Heinht A Skinfold o Dataset View Dataset View In the middle part of the browser a Data pop up menu lets you open datasets and select among any datasets that are already open The name of the currently active dataset is shown Click on the menu to choose another open dataset and choose Other to locate and open a previously saved dataset An Order pop up menu lets you choose how to sort variable names in the scrolling list Dataset order The order in which variables appear in the dataset s columns left to right Alphabetical Alphabetical order by variable name with nonalphabetic names first 2 Datasets Fnter data 57 Variable type Grouped in order by continuous nominal and informative Usage Ordered first by variable use in analyses and then in alphabetical order When a dataset window is active ordering is only alphabetical Variables are shown in a scrolling list Icons next to variable names indicate their data class for continuous for nominal and a gt for informative Compact variables see Compact variables p 84 are preceded by a triangle gt and followed by a symbol Click the triangle to tip it downward lt gt and display the category of the vari able These categories are marked nominal n p Effectiveness wrEffectiveness Time E
82. USI67 USI75 with anova procedure sR80 also see logistic regression regroup SR360 SR381 relations sR338 SR340 ElementOf sr345 equal sR340 greater than SR341 greater than or equal to SR341 1S 346 ISNOT SR347 less than sR340 less than or equal to sR340 not equal sR341 relative frequencies SRIS relative risk sR201 Remainder sr43 remainder sR390 remark sR329 Remove vusi44 remove variables us63 templates us164 tutorial example us22 rename datasets Us70 reopen view USI57 USI59 reorder category variable us238 us240 repeat analyses see templates repeated measures analysis of variance see analysis of variance reserved words us255 US256 Reshape us206 us208 us209 spline curves us209 residual mean square sR73 residuals sR57 SR59 SRGOI plots sr58 sR68 proportional hazards models sR171 saving and plotting SRI71 resize columns vus63 graphs us186 imported pictures US210 pasted object us182 shapes Us206 tables us199 uUs200 text US204 restrict computations see Criteria Include Row Exclude Row row inclusion results accuracy USVI align us214 clean up uS213 US214 group US2I5 incorrect US250 layers us215 list in analysis browser us141 lock us214 move US214 selected us133 unexpected uUs251 ungroup US2I5 unlock us214 validation us250 results browser Us147 Us148 selected results us133 UsI tutorial example us43 Results Selected note us133 Us134 USI40 resume wo
83. USIO4 US254 data type US103 US254 date time values SR369 SR370 SR427 SR428 dialog box usror examples UsIo4 USIO5 Excel us254 tutorial example usi1 us12 missing values Us103 US253 US254 non numeric data as type String us254 pictures US210 Index SR StatView Reference US Using Statview 271 previous StatView versions USI06 separator characters US252 US253 SuperANOVA us106 Us256 text USIOO USIO2 tutorial example us12 us13 troubleshoot us252 us254 variable names usIo2 in control SR251 SR253 Include Row usio8 us109 analyses USI49 USI50 compare results us182 subtitles us2z9 vs Criteria Us108 inclusion see Criteria Include Row Exclude Row row inclusion incomplete sr147 incorrect results us250 Independent vus144 independent variables sr77 sR78 index of results see results browser individual measurements analysis see QC individual measurements analysis inequality sR338 sR341 informational hints us222 informative data class US78 USII7 SR332 input column vs4 Us54 USGI input row USGI insert columns us62 insert rows US62 US63 integer data type US73 SR319 interaction effects sR78 SR9O interaction plots SR78 SR90 SR96 SR99 SRIO4 SRIOZ tutorial example Us40 also see cell plots intercept SR55 SR56 SR80 SR81 SR200 Interface hints us222 interior graphs us183 tables usi198 international datasets Us256 R324 international system configurations USVI interquartile range S
84. a scattergram template the Assign Variables dialog box contains slots for the variables assigned to each axis We want these slots to have generic names so we ll use vari ables named X variable and Y variable e Open Tree Data from the Sample Data folder e Rename Trunk Girth to be X Variable Click the variable name cell type the new name and press Tab e Rename Weight to be Y Variable Don t forget to press Enter or Return e From the Analyze menu select New View e In the analysis browser under Bivariate Plots select Scattergram and click Create Analysis e Click OK to accept the default parameters e In the variable browser select X Variable and click the X Variable button e In the variable browser select Y Variable and click the Y Variable button The scattergram appears in the view 172 6 Templates Build templates Scattergram 2750 4 1 i pi 1 1 j 1 1 2500 4 o F 2250 4 o p 2000 4 o F 1750 4 poo F 1500 po t 1250 4 oped g r gt 1000 4 ge g F 7507 89 r 500 4 a 8 2 L 250 7 we F 0 j T m T T T T T ET y T T 150 200 250 300 350 400 450 500 550 X Variable Now we remove the title customize X and Y axis lengths change the font and size of the axis values and labels and change the position of major and minor tick marks e Make sure the graph is selected and click Edit Display e Uncheck Show title turn the option off e Specify a Vertical measurement 2 inches e Specify a Hor
85. an existing category or click New to define a new one Assigning categories in order from least to greatest simplifies later steps in recoding For help with categories see Categories p 80 Choose Category Please choose the column s category Lindt Bata Ei alcohol frequency Gender Heart Attack Gk Smoking After you create or choose a category you must specify how to divide the range of continuous values into groups A value bar represents the range of data its top and bottom edges corre spond to the variable s maximum and minimum values e Either click breakpoints in the value bar or enter values in the text box e Use pop up menus to correct group label assignments StatView automatically assigns group labels in order according to the category definition If your category definition orders labels from least to greatest Recode assigns groups cor rectly e Click Recode 4 Managing data Recode data 119 Recode of eight Saar How to recode this variable Selected Breakpoint 234 Medium W 140 Low kd 107 Show definition Cancel _Recode a Statview 4 5 Click Show definition or double click the triangle in the lower left corner to open a pane showing the formula definition for the recode You can then drag the triangle to resize the split pane
86. as eR EER Pee el o L a 14 124 UCL 74 124 74 3 Boa 74 25 o H B74 08 R g 74 2 o o o E 8o Q 274 15 L T eS L 3 74 06 4 od pf e JaA i PT h 5 74 041 5 7 f Center 74 044 374 05 L om g 74 H 5 74 02 Oo T E ya 2 73 95 i ae ee F g o 73 9 o o E p 73 98 73 85 o a F 73 96 een LOL 73 964 73 8 eereeanRgae 8 eereeangaa amp 8ssggsgsesgege88 sssessssess Date How can add histograms with normal curves to my capability analyses It is often useful to plot a histogram with a normal curve to evaluate whether data are nor mally distributed Many QC analysts also like to plot a histogram with a normal curve together with their tables of capability indices This allows easy visualization of the positions of target values and upper and lower specification limits relative to the actual distribution of the data Simply select a table of capability indices and adopt its variable assignment for a his togram in the analysis browser under Frequency Distribution double click Histogram In the Frequency Distribution dialog box check Show normal comparison turn it on How can get more information about differences among subgroups For subgroup measurement data analysis of variance can often provide a more fine grained comparison of differences among subgroups For instance an ANOVA with post hoc tests can tell you specifically which subgroups have mean values that differ significantly from the oth ers Please consult the chapter
87. asked to Choose Category for the column click New One at a time enter the labels for each group in the order you determined Be sure to type each label exactly as it appears in the variable don change spelling spacing capitalization etc Edit Category Category name Self defined category Group label Saturday Sunday GSTs iuesdas ma S Tieng Cancel Wednesday Thursday Friday ee e Click Done e Change the type for the old variable to String and change its class to Informative e Select the old column by clicking its name e From the edit menu select Copy e Select the new column by clicking its name From the Edit menu select Paste Now you have a new copy of your old column It should look the same however in the new column the levels of the category are ordered according to your definition Any graphs or tables you produce with this variable will show its levels in that order 240 9 Tips and shortcuts Common questions Descriptive Statistics Box Plot R Split By Self defined p Split By Self defined 9 o Sunday Mean 25 4 F o Monday Continuous Total 8 86 20 4 L Continuous Sunday 7 63 r o Tuesday Continuous Monday 10 15 z p O Wednesday Continuous Tuesday 8 08 0 Sl Thursday Continuous Wednesday 15 18 5 4
88. box e Click Compute Browser order pop up menus Definition area text box Formula of Column 27 grada variable definition Variable browser og Cholesteral Ln Function browser Factorial aj Al eal Floor Lag 1 Log LogB 7 SN Windows Calculator keypad The Formula dialog has a variable browser a function browser and a calculator keypad for building expressions interactively You can also type formulas directly into the definition area and as StatView recognizes the variable or function name you are typing it completes the name and supplies all the needed parameters Window controls This dialog box behaves like a regular window you can resize it use Cut Copy and Paste on the text change fonts and move the window behind or in front of other windows IIO 4 Managing data Formula Browser order pop up menus Definition area text box Formula ula variable definition LogiCholestero Loi Order Dataset order Variable browser Triglycerides HDL Order byl Function Type Lagt 1 Ee S Function browser Logl LogB 7 Modf 7 MovingAveragel 7 Normt AlIRows Percentages AllRows Statview 4 5 Macintosh Calculator keypad A Formu
89. box enter the number of columns you want to create e Use the pop up menus to set the attributes for these variables and click OK For more details on variable attributes see Variable attributes p 73 The columns are appended to the right of existing columns The default name for the nth variable to be created in a dataset is Column n You can use the Undo command in the Edit menu to delete the new columns if you make an error You must select Undo immediately after adding columns before taking any other action Insert columns You can insert a column between two existing columns in the same dataset e Control click Windows or Command click Macintosh the border between two variable names Position the cursor between two variable names over the vertical line separating the col umns the cursor changes to a cross arrow shape Hold the Control Windows or Command Macintosh key down the cursor changes to a double arrow p shape Click and release the mouse button Release the key 2 Datasets Enter data 63 Gender j Age Gender Column 26 Age female male d 22 female male 22 The newly inserted column has default variable attributes and is filled with missing values until you enter data Repeat to insert additional columns Remove columns You can remove a column at any time e Click the variable name to select the column e Press the Delete or Backspace key or s
90. but not from the dataset itself You can also combine Include Exclude and Criteria rules Line Chart Line Chart Inclusion criteria Weight lt 200 Row exclusion Lipid Data from Lipid Data 240 1 4 e L 200 J lai li l rtiitititi 220 J E 1907 L 200 l 3 F 1805 E J j oe N C 170 7 p 1807 i i 1 o 1604 F D J Ne F 2160 NA H 53150 4 J C 1407 r 140 5 i T 130 4 L 120 4 po ona H 120 4 i 100 SLE EE ES ES aa 18 20 22 24 26 28 30 32 Oey ae tds eae Age 20 24 28 32 36 Age Exercise The exercise below demonstrates the interaction of the analysis and variable browsers and analysis objects Create an analysis then assign variables Creating an analysis from scratch is a simple two step process involving the variable and anal ysis browsers To create any analysis you must select an item from the analysis browser and assign variables to it from the variable browser You can use the browsers in either order Try both and use whichever feels more comfortable e Open Car Data from the Sample Data folder e From the Analyze menu select New View 5 Analyses Exercise I5I First select an analysis from the analysis browser then assign variables from the variable browser If you are interested in the degree of correlation among several continuous variables in this dataset a good place to begin your analysis is with a correlation matrix From the analysis br
91. calculations are not influenced by order in nominal variables the numerical results are the same no matter how your groups are ordered In other words Stat View s nominal variables are not ordinal variables How can reorder category variables Alphabetical order is not always the best order for informational displays Generally you should sort information in statistical displays according to the meaning of your data For example if you want to demonstrate that groups of patients showed increasing effects from treatments you might want to sort those groups from least affected to most affected in your 9 Tips and shortcuts Common questions 239 plots of the effects Or if you are examining sales patterns for each day of the week you might want to show those days in calendar order You might think you could make changes to the order of labels by directly editing your graph or table However this is not possible Instead you must first create a new column in the Stat View data set and then redo your analysis with that variable e First determine the order you want if the list is long or hard to remember you might want to write it down e Open your dataset and scroll to the variable you want to reorder e Insert a column to the right of this column and name it e Open the attribute pane if it is closed by double clicking the control in the upper right corner e Change the Type of this newly inserted column to Category e When
92. can instead select an analysis in the analysis browser assign variables in the vari able browser and then click Create Analysis For large datasets or complex computations waiting to click the Create Analysis button last can save time because you don t have to wait for calculations after adding each variable Second we adopt the variable assignment Weight from this histogram simply by leaving the analysis selected while we create another analysis this time a box plot e Be sure the histogram is still selected has black selection handles if not click to select it 5 Analyses Overview 133 e In the analysis browser select Box Plot and click Create Analysis Third we clone the box plot to analyze another variable Horsepower e Be sure the box plot is still selected has black selection handles if not click to select it In the variable browser select Horsepower and Control Shift Windows or Command Shift click Macintosh the Add button As you can see result objects help you generate new analyses They retain the information used to create them the analysis you chose the variables you assigned and even the parame ters you specified for example in the histogram above we accepted the default parameters for how to divide the variable into intervals and we chose not to add a normal curve When a result object is selected this information passes directly to your next analysis saving steps you would oth
93. choose formats that display values in sufficient precision We recommend that you choose date time formats that are complete enough to reveal exactly how values you enter or compute are interpreted All date time values are stored as an exact time on an exact date regardless of the format used Many of the date time formats hide this exactness from you For example the format Jan 04 shows the exact time Monday 6 Febru ary 2040 06 28 15 as Feb 40 If you enter an ambiguous date time value such as 8 11 which could mean either August 11 or 8 November StatView warns you a dataset preference lets you suppress this warning see Silently accept ambiguous values p 227 80 2 Datasets Categories If you are unsure how StatView interprets a value choose a date time format that shows more detail Decimal places The Decimal Places attribute specifies how many digits after the decimal point to display for real numbers Values are always stored in the fullest precision of the platform you are using Your choice affects only the display of values Changing the number of decimal places to display is always harmless Categories A category is a special variable type that makes nominal data entry faster and more accurate The category type allows you to define a named set of labels for the groups of one or more variables Other common terms for groups are cells and levels Once you associate a variable with a category
94. clever Copy and Paste steps to finish the data entry Here s one way we could enter these data From the File menu select New e Create three columns by typing the variable names A B and C e Select all three columns by clicking and dragging across the variable names 2 Datasets Edit data 69 In one of the column s variable attribute pane select Integer for the Type this changes the type for all three columns in one step z Integer 7 a Long Integer Input Column Type a Real Category t User Entered String Currency t Free Format Fixed e Enter the value 1 for the first row of A Control click Windows or Command click Macintosh the border under the row num ber label for the first row then specify 7 more rows to insert This technique is detailed under Insert rows p 63 Input Column Insert Rows Source Class Sumber of rows to insert Dec Places Double click the split pane control to close the attribute pane In row 5 for A enter the value 2 Select the cell with the 1 and from the Edit menu select Copy Control click Windows or Command click Macintosh to select all the other cells that should have a 1 From the Edit menu select Paste Now select the cell with the 2 and from the Edit menu select Copy Control click Windows or Command click Macintosh to select all the other cells that should have a 2
95. columns You should specify column attributes for each column before entering data Doing so ensures that each column is set up appropriately for the data it is to contain If you enter values that are not compatible with the attributes specified StatView warns you that the values are incompatible Change attributes You can change the attributes of a variable anytime except 1 You cannot change the class of a variable that is in use in an analysis or formula definition 2 You cannot change the type of a variable that is in use in an analysis or formula definition to category or string 3 You cannot change any attributes for a compact variable except format and decimal places 4 If the source of the variable is analysis generated you cannot change the type or class Additional guidelines for changing each attribute are given under Save datasets p 70 View summary statistics The variable attribute pane also contains a set of summary statistics for each variable Usually these statistics are hidden but you can reveal them either by scrolling the attribute pane downward or by increasing the size of the attribute pane e Click and drag the attribute pane control downward 60 2 Datasets Fnter data User Ent User Entered Z Lipid Data Name Gender Age Weight Cholesterol Trighceriq4 Type Categor Inte Integer Integer Integer User Entere F
96. e E aE AAE T E E E weight Weight weight All components of the graph move to stay in the same relative location to the graph Move graphs or components You can move an entire graph and most components of a graph the title legend axis axis labels notes If you move the axis it is constrained to the respective horizontal or vertical direction If you move the graph as a whole all components move together and stay in the same positions relative to each other e Click and drag the graph or component Do not drag selection handles dragging handles resizes rather than moves the graph or com ponent Shortcut You can use cursor arrow keys to move any selected graph or graph component If the grid is turned on each keystroke moves the object one grid unit in the direction of the arrow If the grid is turned off each keystroke moves the object one screen pixel In this con text grid refers to an underlying grid in the view window itself not the grid lines that can be used to show tick intervals inside a graph s interior for more information see Turn Grid On Off p 217 7 Customizing results Graphs 187 Change text items You can change the font size style and color of any text item in a graph graph titles notes axis labels axis values and legend text You can also change the alignment and orientation of any text item except in legends You can change components individually or all at once To format or
97. edit a text component select only that component To select several text items Shift click to select them at the same time Then use Text menu items to change the font size style justification left alignment right alignment or centering and rotation Use the pen color tool in the Draw palette to change the color of text components Use the text tool in the Draw palette to edit the contents of a text item You cannot edit axis scale values with text tools If you use the text tool to edit any text that text no longer updates when the graph updates Change overall structures While StatView displays many different graphs they share a common overall structure which can be changed with the Graph dialog box You can choose a graph frame transpose two axes hide or show titles and legends and set the height and width for any graph To use the Graph dialog box e Select the entire graph click its frame e Click Edit Display Any changes you make in this dialog box apply only to the selected graph If you want to change several graphs select them all at once Click Edit Display for a series of Graph dialog boxes one for each graph in the order in which you selected the graphs Graph Flip horizontal and vertical axes Bounds include extra lines EJ Show legend EJ Show title Dimensions Vertical inches Horizontal 4 12 inches Numbers
98. else Crime Attributes Cancel Recode Statview 4 0 Fy Use Attributes if you want to specify variable names and attributes before clicking Create You can adjust attributes afterward in the dataset window if you prefer See Variable attributes p 73 A Recode dialog box is listed in the Window menu where you can select it to bring it to the front You can double click the top area beneath the title bar to bring its dataset to the front Select Print from the File menu to print a formula definition If you prefer you can use Formula to build your own recoding formula Recode is a shortcut Either way you can use the variable s Source pop up menu to view and edit a formula defini tion or to change from static formula to dynamic formula or user entered See Change sources p 77 By default recoded variables are based on dynamic formulas which means that changes or additions to the original variable automatically change the recoded variable New variables are appended at the right side of the dataset To move a variable Copy or Cut the data insert an empty column elsewhere in the dataset see Insert columns p 62 and Paste the data into the empty column To create a recode variable in another location insert a column and change its source to Static Formula or Dynamic Formula In this exercise you create a variable that categorizes the risk of hea
99. g 0 25 instead of 25 check Always have leading digit Default frame Select which graph frame which combination of axes you want Default point size Choose a size from the points in the Default Point Size pop up menu Order in which to choose points fills and colors Choose the order in which you want point types fills and colors to be chosen for graphs Click and hold the item you want to modify then select a choice from the pop up menu and release the mouse button Do not drag across the bar itself As you change the position of one color point or fill in the order the others adjust for example if you choose black for your first variable the item that is cur rently black changes to the color that had been first Distinguish cells by For compound graphs see Multiple and compound results p 135 that show more than one group variable StatView automatically assigns a different point type color or fill pattern to each variable so that they can be easily distinguished Choose which parameter you want to vary between variables For example choose color if you want different variables to have the same point type and fill pattern but different colors 9 Tips and shortcuts Preferences Hints preferences Hints Preferentes Display hints window when it contains Display hints window when it contains P Interface hints Balloon hints Informational hints T Informational hints
100. group labels begin with the letter s you type StatView alerts you to your error 2 Type the number of the group label 1 for the first label 2 for the second and so on according to the order in which you defined the labels StatView fills in the label for you For example suppose you defined group levels Low Medium and High You could enter a Low value by typing L or 1 You could enter Medium by typing M m or 2 You could enter High by typing H h or 3 If you typed another letter or number Stat View would alert you to your error Define category levels in the order in which you want results to appear if such an ordering exists for your variable for example Low Medium and High or Monday Tuesday Wednes day Edit category definitions You can edit categories with the Edit Categories command from the Manage menu e From the Manage menu select Edit Categories In the Choose Category dialog box select the category you want to change e Click Edit The Edit Category dialog box appears This dialog box is exactly like the dialog box used to create a new category Edit Category Create a new category Category name Animals Group label Dogs K wie Cancel Replace Delete e Select the label you want to change e Type a new value in the Group label box e Click Replace To delete a group label select the group
101. immediately below it Press Shift Tab to move the selection to the previous slot Remove variables from slots If you assign a variable to the wrong slot by mistake you can remove it three ways 1 Click the variable in the slot and press Delete 2 Drag the variable from the slot back to the variable list 3 Double click the variable in its slot Manipulate results When you finish assigning variables and click OK results appear in a view If the active top most window is a view results appear in that view after existing results If the top window is not a view results appear in a new untitled view If the Recalculate box in the view is unchecked turned off empty graph and table place holders X boxes appear rather than completed results When you are ready to see your results check the Recalculate box turn it on See Control recalculations p 138 6 Templates Use templates 165 Exercise You can manipulate template results the same way as any other results See the previous chap ter Analyses p 131 for instructions on using analysis and variable browsers and the Edit Display and Edit Analysis buttons Also see the next two chapters Customizing results p 179 and Drawing and layout p 203 For instructions on working with each type of analysis and its results see its chapter in StatView Reference In this exercise you apply one of StatView s installed templates to gener
102. include ol Ol Oo Min Value Selected Value Max Value 121 510 i1 To move an existing range click and drag the entire range across the bar To change one of the endpoints of a range click drag the endpoint to the new position You can also change the range values by typing the ranges into the criterion definition Choose a level or levels When the variable is nominal you see a list of its group labels instead of the value bar Select a category group Double click one or more group label to select one or more groups If you need to define a criterion that includes only those cases with or without a missing value for a certain variable use IS or ISNOT and a missing value Complex criteria If you click inside the definition area at the end of a definition the scrolling list presents a list of Boolean operators You can combine as many criteria as you like using AND OR or xor To enter a Boolean operator double click it or type its name You can group expressions by inserting parentheses into the definition 128 4 Managing data Criteria pop up menu Parentheses AND OR and xor are discussed in detail in the chapter Formulas p 315 of StatView Reference Apply Save and Select The buttons at the lower right corner of the Criteria dialog box give you three options when you are finished creating or editing criteria All three buttons save the criterion with the dataset add it t
103. lines us183 USI85 USI92 group US215 height us188 us229 interior USI83 USI85 USI95 layers us215 legends us183 USI86 USI88 USI94 USI95 USI97 line patterns us185 line widths us185 us196 list in analysis browser us141 lock us214 move USI86 US214 move components USI86 notes USI84 USI86 numeric formats US229 overlay us186 pen color us197 pen patterns USI85 USI96 percentile plots sr19 plots us183 US197 plotted lines us183 us197 point colors us195 usI96 point types USI95 point types and sizes us229 preferences USI79 USI8O US228 US229 reference lines us184 resize USI86 select usi84 select components us185 template exercise USI7I text color us185 USI97 tick marks us183 USI85 titles us183 USI86 USI88 ungroup US215 univariate plots sR217 unlock us214 width us188 us229 X axis USI83 Y axis US183 Graphs Only us142 greater than SR341 greater than or equal to SR337 SR341 270 Index SR StatView Reference US Using Statview gremlins us252 grid UsI86 US200 US217 colors us217 spacing US217 grid lines us183 USI85 USI92 Group USs215 unexpected results us251 group labels see category group variable nonparametric analyses sR157 grouped regression SR8I Groups SR381 groups US5O US5I US80 USI9O R327 SR336 SR349 choose see Criteria nonparametric analyses sR147 also see compact variables Split By growth regression SR55 SR67 G statistic SRII3
104. make to your data along the way automatically trickle through all your analyses So you can change any graph or any table any time directly in place without having to redo anything So you can fix that one tiny error you made weeks ago quickly the morning of your presentation All of StatView s advanced analyses work the same way as the simple ones so once you know how to build one analysis you can build any analysis If you need a quick review of some of the statistical techniques we ll give you a hand with that too Why should bother with a tutorial This tutorial is meant to get you started using StatView by stepping through typical activities in each phase of a data analysis project 1 Manage data collect data enter or import data into StatView find and fix errors sort groups get a feel for the numbers 2 Analyze data explore the data look for patterns test your hypotheses turn raw data into information 3 Present information put together persuasive graphs annotate results with your comments call attention to the discoveries that lead to your conclusions It is also meant to give you a taste of chocolate I Tutorial Manage data What No seriously By the time you finish this tutorial you will be craving a candy bar What s more youll know which candy bars are the most nutritionally sound Usually you concentrate on your data not StatView But now we want you to concentrate on how StatVi
105. of the lines e Click Edit Display Cell Line Plot EJ Connect lines between variables stow Concer Gy Cell Line Chart Grouping Variable s Smoking History 220 4 o 210 4 H 200 4 S g 2 1907 M 2 0 1807 F S 170 7 F O oO 160 a 150 22 ow 8 Ox o goe n c 5s5 Lo Ert oga 5g SELT 27g E288 ES DOLE zvao D 0D z Es D Cell Mean Cell Line Plot Connect lines between variables stow Concer Gram 220 210 200 o 180 170 160 150 Cell Line Chart Grouping Variable s Smoking History o oo dl A Ny 2x 8 zee oe 328 Eza Loga 552 S285 z233 g E DS Zg DEST OOPe Sosxd D z0 zs D 190 7 Customizing results Graphs Change axes StatView uses three different types of axes a numeric axis for displaying continuous measure ments a cell axis for displaying information on groups or subgroups and an ordinal axis for displaying the order of a point in a dataset Axis values appear at major tick intervals along both vertical and horizontal axes r T T T T T 1 1500 2000 2500 3000 3500 4000 4500 Numeric axis Small Sporty Compact Medium Large Cell axis T T T T T T 0 20 40 60 80 100 Ordinal axis Bivariate plots can have either numeric or cell axes for both the X and Y axes Univariate plots always have an ordinal X axis and can have either a numeric or cell Y axis
106. pop up menu you can assign variables from several different datasets to a single analysis If you assign variables of different lengths i e if you assign variables from datasets with dif ferent numbers of rows StatView does the following to compensate e Variables with fewer values are padded with missing values at the end so that all variables have the same number of cases The excluded rows are the union of the excluded rows for the datasets which contain the variables For example if variable A comes from dataset A with rows 3 and 4 excluded and variable B comes from dataset B with rows 7 and 8 excluded rows 3 4 7 and 8 will be excluded from each analysis that uses both variables Two analyses are exceptions to these rules descriptive statistics and one sample analyses only if you do not have split by variables assigned For these two analyses no missing values 5 Analyses Analysis windows 147 are padded and the exclusion applies to each variable individually If you do have split by vari ables assigned they calculate the same as other analyses Results browser Results in the view window are listed in the results browser a floating window that appears alongside the view window If a view contains a great deal of output or you are working with several views at once the results browser makes it easy to keep track of what you have done what is currently selected and where particular results are located The results br
107. resize a polygon 208 8 Drawing and layout Draw tools Reshape Select Reshape from the Edit menu to switch to Reshape mode and select it again to exit Reshape mode when you are finished In Reshape mode the cursor changes to a reshaping crosshair and you can 1 Change the shape of a polygon click and drag any vertex to a new location DE AA 2 Open a closed polygon Control Alt click Windows or Command Option click Macin tosh the line segment you want to remove 2S 3 Close an open polygon Control Alt click Windows or Command Option click Macin tosh either the starting point or the finishing point ay 4 Remove a vertex to reduce the number of edges Alt click Windows or Option click Macintosh the vertex 5 Add a vertex to any edge Alt click Windows or Option click Macintosh the edge where you want the vertex ne Shortcut Macintosh only Hold the Control key to switch temporarily to Reshape mode and release the key to exit Reshape mode Splines The spline tool lets you draw open or closed free form curves based on cubic splines similar to Bezier curves An open curve is one that doesn t start and end at the same point Below the left figure is closed and the right is open 8 Drawing and layout Draw tools 209 Select the spline tool e Click and release at the starting point e Click where you want the first vertex the first change of direction e Continue clickin
108. results us182 subtitles us2z9 vs Criteria UsIo8 also see row inclusion exclusion see Criteria Include Row Exclude Row row inclusion exclusive OR sr347 Exercises SR96 Expand us57 us94 expand compact variables us94 us95 expected value sr12 exponential distribution sr407 exponential function sR375 SR385 exponential model sr172 sRI73 SRI75 exponential regression SR54 SR68 ExponentialSeries sR376 exponentiation SR334 export EMF USI57 Excel usu Us99 us100 missing values us1o2 PICT USI57 previous StatView versions USIO6 SuperANOVA us106 text USIOO USIO2 USIO5 WME USI57 expression R325 SR338 SR339 expression language USsI13 extended precision us73 extra Binomial variation sR202 extract text SR423 F F distribution SR402 SR407 factor SR77 factor analysis SRI31 basic output sR138 data requirements SRI37 dialog box sr136 268 Index SR StatView Reference US Using Statview discussion SRI3I exercise SRI38 factor extraction methods sR132 factor loadings SR133 SR139 factor scores SRI33 stepwise regression SR62 survival regression models sr185 force recalculation us138 usI39 foreign versions US256 format sR316 oblique solution sr142 results sR138 save factor scores SRI33 summary table sr138 templates sR138 transformation methods sRI35 unrotated solution SRI4I Factorial SR377 factorial design sR91 factors see category nominal data false SR338 SR345 features USI FibonacciS
109. rs not NOT j Led I x i 0 if if then else Isin ElementOf NOT ElementOf 2 2 To insert an operator or function in a formula definition click its button Click first the Inv button and then another button to insert an inverse functions Click Hyp and a trig button to insert an hyperbolic trig functions and click 1nv then Hyp then a trig button to insert inverse hyperbolic functions 4 Managing data Formula 113 Mathematical expression language The functions found in the browser and keypad constitute a complete mathematical expres sion language which is detailed in the chapter Formulas p 315 of StatView Reference Con sult that chapter for a comprehensive discussion of how expressions are evaluated how each function works and myriad examples To see a quick definition for a function open the Hints window and select the function Build definitions Functions and variables appear in the definition area as you double click them in the brows ers click them in the keypad or type them To insert an element in a particular part of a for mula first click the I beam cursor Y in that location Most functions require that you supply arguments for the functions These are represented by question marks in the formula definition area and the function browser Arguments are the objects of operators and functions When you enter a function into a formula d
110. sr290 discussion sr287 sR28 exercises SR294 SR297 p charts SR256 R266 SR288 SR292 SR295 results sR292 SR294 standardize inspection criteria sR289 subgroup variables sr287 templates sr294 QC subgroup measurements analysis data requirements SR268 SR269 dialog boxes sR262 sR268 discussion sR257 SR262 exercises SR273 SR276 results sR269 SR272 templates SR273 tests for special causes sR259 QuadraticSeries sr404 quantile plots sR174 QuarticSeries sR405 quartiles sR398 quartimax SRI35 question marks us113 R324 questions see common questions quotation marks us255 R325 SR33I quotient sR334 R R partial correlation coefficient sR205 R charts SR259 R266 SR270 R squared sr56 radians R372 SR400 SR405 radius for round corners Us206 RadToDeg sr405 raise to powers R334 random criteria USI24 USI28 SR409 Random Numbers USs123 US124 R317 SR330 hints us222 print definitions us123 unique SR412 RandomBeta sr406 RandomBinomial sr406 RandomChiSquare sr407 RandomExponential sr407 RandomF sr407 RandomGamma sR408 RandomGaussian sR408 RandomInclusion sr409 randomized complete block design srro1 SRIOS RandomNormal sr410 RandomPoisson sr410 RandomT srqit RandomUniform sr4it RandomUniformInteger sr412 Range SR259 SR412 range US6O SR3 ranges SR336 SR338 SR345 Rank sr413 rank tests SRISO SRI5I SRIG2 SRIG4 SRIGG6 raw data contingency tables exercise sRII7 fact
111. sri65 cumulative hazard plot sr1so SRI6O cumulative survival plot sr16o currency data type USZ3 R320 SR321 formats USVI US79 cursor movement see dataset preferences curve tool see spline tool Custom Rulers us217 custom templates us169 UsI77 exercise USI7I USI77 custom tests for special causes sR260 dialog box sr264 save as template sR261 customize graphs Us183 USI97 US203 US212 results us179 Us202 shapes US203 Us212 tables us197 us202 text US203 US212 CUSUM analysis charts sR255 SR261 SR271 SR283 individual measurement analyses sR279 results sR263 SR265 Cut Us65 USI8I USI82 unusual selection shapes us68 us70 cutpoints SR342 also see Recode D D usage marker see usage markers data Copy us66 Cut Clear Delete us65 enter US61 manage USIO7 USI30 select us64 subsets see Criteria Include Row Ex clude Row Sort row inclusion also see dataset data class US6 US50 SR332 change uss1 Us78 continuous US78 discussion us78 example us49 in examples usv1 informative US78 USIIZ nominal us78 us238 us240 data format currency US79 date time us79 engineering US79 enhanced free fixed us79 fixed places us79 free format US79 free format fixed us79 in examples usv1 scientific US79 data loss change type uS75 when pasting us67 data organization US3 US49 US53 arrangement Us50 class Us49 us50 compact variables us84 example us49 structure US49 US53 Data pop up menu
112. steps e Ifyou get an error hint when trying to do something follow the instructions in the Hints window to solve the problem In the Formula dialog box Hints give information about functions that might prove helpful See Hints window p 222 252 9 Tips and shortcuts Troubleshooting Importing e Read this book using the table of contents or index to locate information e See if your question is addressed in this chapter e If your question involves setting up an experiment see if one of the sample datasets pro vided with the program mirrors your situation If the problem disappears you might have a problem in the structure of your dataset Review the data requirements in the StatView Reference chapter for that analysis Make sure each variable s attribute settings are correct see Variable attributes p 73 If you imported your data from another application examine it for possible importing errors e Try again with the simplest possible system configuration On Windows reboot with plain startup files On Macintosh restart and hold the Shift key until you see the message Extensions off If the problem disappears try adding things back one at a time to isolate the problem Try reinstalling StatView A change in your system or network configuration could have caused a problem or an application file might have become corrupted It is important to identify the correct separator characters to make
113. templates for reports and presentations you should also become familiar with the topics in the chapter Drawing and layout p 203 If you need help with the parameters and variable requirements for a particular analysis see its chapter in StatView Reference Template tips When you create a view to reuse as a template there are several things to take into consider ation Some characteristics that are helpful in working views get in the way in templates and vice versa 170 6 Templates Build templates Give variables general meaningful names You use the Assign Variables dialog box to assign variables from any dataset to the slots in the template The templates list contains separate slots for each variable used in the view you save as a template Therefore when you create your template consider giving your variables generic names that could apply to any dataset rather than names that have a meaning only in a specific dataset The picture below shows both H Variable Weight a Variable Cholesterol a Generic variable names Specific variable names Slots take their names from the variables used to create the template so you may want to rename the variables in your dataset while youre creating a template You might create a dummy dataset with generic names you can use for template construction Variables used to build templates needn t contain data W
114. the calorie variation Do a box plot of Total fat g select the Total fat g descriptive statistics table select Box Plot from the analysis browser and click Cre ate Analysis The only surprise is that Hershey has the lowest fat content candy bar while M amp M Mars has the lowest calorie count candy bar Is the saturated fat variation similar Similarly do a box plot of Saturated fat g by starting with the statistics for Saturated fat g Again the results are similar Hershey also has the lowest saturated fat content Since saturated fat is part of the total fat it is not surprising that these two go closely together However the plots show somewhat different distributions which means that the proportions of unsaturated fat content something nutrition labels don t report do vary Are the other brands besides Hershey M amp M Mars and Nestle much different Tempo rarily turn off the Big Three criterion in the dataset window select No Criteria from the Cri teria pop up menu When youre ready to continue select the Big Three criterion again I Tutorial Analyze data 31 Create an analysis with several parts Sideways triangles sit in front of many items in the analysis browser These triangles indi cate that more detailed choices are available Clicking the triangle tips it downward 7 and reveals a list of possible results Some even have subcategories of possible results In all cases the triangles let you sho
115. the correlation analysis created above you can add more variables to the matrix e Select the correlation matrix by clicking on it In the variable browser select Weight and click Add The correlation matrix immediately recalculates Correlation Matrix Turning Circle Displacement Horsepower Gas Tank Size Weight Turning Circle 1 000 747 482 618 752 Displacement 747 1 000 764 719 830 Horsepower 482 764 1 000 666 707 Gas Tank Size 618 719 666 1 000 847 Weight 752 830 707 847 1 000 116 observations were used in this computation Other analyses compute new results for additional variables These new results automatically retain the analysis parameters but produce separate results The simple regression analysis cre ated above is such an analysis Select any of the three regression tables or the regression plot you created above The next variable you assign will use these same analysis parameters 154 5 Analyses Exercise e In the variable browser select Weight and click Independent A completely new regression analysis uses Horsepower as the dependent variable and Weight as the new independent variable The same four regression results are created Regression Summary anova Table Regression Coefficients and Regression Plot and the same parame ter choices are used Here we show just the new summary table Regression Summary Horsepower vs Weight
116. the following three variable dataset 9 Tips and shortcuts Troubleshooting 253 J totum 1 Comm 2 Colurn ilere ape 46 000 Laloma wong 19 000 As you can see StatView split the employee names in two To remedy this you could either separate data points with some other character besides spaces such as Tabs or place double quotation marks around distinct data points having spaces within them The text file should look like this Employee lt tab gt Age Kate Bishop lt tab gt 46 David Wong lt tab gt 19 or this Employee Age Kate Bishop 46 David Wong 19 Commas as separators Another example of an incorrect choice would be to import the following file without specify ing commas as a separator Name Age Yukito 32 Setsuko 35 Armita 30 Choosing Tab for example tells StatView to import the file as a one variable dataset J oni input Cour Thame age 2 Yukito 32 3 Setsuko 35 4 Armita 30 Remedies for common errors 1 If some values in certain rows are shifted to the right see if the source text file contains groups of the separator character you used two commas for instance If it does edit out the extra separator characters from the source text file and import the file into StatView again 2 If some values are shifted to the left in a file imported with space separators you used a space separator to indicate a missing value Two or more spaces are conden
117. the importing process sim ple and trouble free A separator character is a character that occurs between data points and tells Stat View where a data point in one column ends and the next begins Separator charac ters also define the end of rows in the dataset If you are not certain what separator characters are used in a text file you are importing open the file in a word processing application and choose the setting that enables you to see format ting characters within the document Formatting characters or gremlins are non textual char acters that indicate tabs spaces paragraphs and so forth Each word processing application represents gremlins a little differently so check the user manual for your application to see how they appear If a separator character other than a Tab is used such as a comma StatView needs to know Problems can arise if an improper separator is used in the source file or if you tell StatView to use a separator character that does not match the one used in the source application Spaces as separators If you choose spaces and the text file contains strings with spaces within an individual entry StatView expects the spaces to indicate separate data points For example it would be a mis take to import this text file with spaces as a separator Employee Age Kate Bishop 46 David Wong 19 In this dataset spaces appear not only between data points but also inside them Stat View would import this text file as
118. the top of the results browser lets you select and scroll to results in a view e Select one or more result titles e Click Select Shift click or click and drag to select several adjacent results Control click Windows or Command click Macintosh to select nonadjacent results Selecting a heading is equivalent to selecting all the items under the heading Shortcut Double clicking an item is the same as selecting it and clicking the Select button When you click Select in the results browser the view window scrolls to that result or the first result if you selected several and selects the object or objects and you can work with them all at once clone them change their variables delete them or whatever you need to do Thus the results browser can speed up your editing For example if you have several dozen results involving a key variable and you want to make them all green you could use the results browser ordered by variable select the heading for that variable click Select and then select green for the foreground color in the Draw palette to change them all at once Also in a view with many tables and graphs it might not be easy to find the one you want The results browser lets you move to the result you want quickly View Windows only and Window menu The View and Window menus Windows or Window menu Macintosh help you keep track of all the windows you have open and makes it easy to switch between them Both
119. to certain result types or data situations are unique to Edit Display because they are options that can be modified only locally one result at a time 3 Draw tools let you change lines colors fill patterns etc used in graphs and tables 4 Finally a complete palette of drawing and text tools lets you modify annotate and amplify the information shown in graphs and tables Layout tools help you arrange results on the pages of a presentation StatView s drawing and layout tools are discussed in the next chap ter Drawing and layout p 203 This chapter discusses preferences briefly and then shows how to use Edit Display and Draw tools to customize individual tables and graphs For graphs you can change number formats axis bounds tick marks interval widths labels and symbols grid lines fills colors patterns line widths and text formats You can also transpose the horizontal and vertical axes For tables you can change number formats borders row heights fonts sizes and styles lines and thicknesses colors and pen patterns You can also transpose rows and columns Preferences Preferences let you establish global settings for graphs and tables Your preference settings determine the defaults used for every graph and table you create By contrast Edit Display lets you change these and other settings for any single graph or table you select Before producing any results for a presentation you should set preference
120. to tip it downward 7 and reveal the nominal variable within e Widen the browser so you can read the full variable names 2 Datasets Compact variables 89 i Variables E ariables i Variables Show Show Show Compact Compact Compact Expand Expand Expand Data Untitle Data Simple Compact Variable Order Order Order Dataset order gt Cholester T errs E E Cholesterol readings a Category for Cholesterol readi H Notice that Cholesterol readings and Category for Cholesterol readings both look like regular variables Cholesterol readings has a class marker for continuous and Category for Cho lesterol readings has an H class marker for nominal They ook like regular variables in the browser because they work like regular variables in the browser We ll learn in Expand com pact variables p 94 that the parts of a compact variable each act the way regular variables do You select them and apply them to analyses the same way as regular variables First let s save the dataset From the File menu select Save e Specify a filename Simple Compact Variable e Click Save Complex compact variable Entering a complex compact variable a compact variable with more than one nominal vari able follows basically the same two step process 1 First you create a column for each subgroup in this case male smokers male nonsmokers female sm
121. types are converted to the nearest equivalent format in Excel ASCII text Import text Most applications can open and save documents in a plain ascu text file format for exchange with other applications If you are not sure how to save plain text files from an application consult its documentation In most cases a Save As or Export command is available in the File menu If you have troubles importing examine the ascu text file with a text editor or word processor Be sure to display all special characters and formatting symbols and if you need to make corrections remember to save as plain ascii text StatView expects data in text files to be organized in rows and columns with each row case ona single line and with values separated by one or more separator delimiter characters such as a tab comma or space By default StatView expects tab delimited columns Values that contain the delimiter characters should be enclosed in double quotation marks Macintosh StatView expects carriage returns at the end of each line To import a text file cre ated on another platform you must first use a utility such as Apple File Exchange to replace linefeeds LF or carriage returns and linefeeds CR LF with just carriage returns CR From the File menu select Open e Choose Text file type 3 Importing and exporting ASCII text IOI Export text e Select a file e Click Open e Set importing options appropriate for your file e
122. usi191 dialog boxes usr9o tick marks us191 numeric data Us66 numeric formats axes USI92 graphs us229 tables us201 US231 numeric intervals R337 numeric precision US73 nutrition labels us2 0 object oriented technology us133 USI6I objects align us214 clean up uS213 US214 group US2I5 layers us2I5 lock us214 move US214 ungroup US215 unlock us214 oblique factor scores SRI35 SRI37 SRI40 SRI42 odds sR200 off diagonal sR133 old StatView data us106 omnibus tests sR74 one SR345 one sample analysis sR23 data requirements SR25 dialog box sr24 Index SR StatView Reference US Using Statview discussion sR23 SR24 exercise SR26 nonparametric SRII9 results sR26 templates sR26 t test SR23 one case per row data US5I US52 OneGroupChiSquare SR394 OnlyExcludedRows us1o08 Us124 SR325 SR326 OnlyIncludedRows usi108 us124 SR325 SR326 Open us72 US99 USIOO USI57 dataset tutorial example us13 view use different variables us158 views USI5 7 USI59 use different dataset us159 use original variables us157 open interval SR337 open polygon us207 open spline us208 Open View As usi57 usI58 USIGI Operators SR332 SR336 addition sr333 division sR334 exponentiation SR334 multiplication SR333 negative SR335 parentheses sR336 positive SR335 subtraction SR333 OR SR347 order of operations sR326 SR328 SR336 Order pop up menu analysis browser us142 Assign v
123. values remain For help solving common formula errors see Formulas and criteria p 240 8 static formulas Formula variables use dynamic formulas by default If you change data associated with a dynamic formula the formula automatically recalculates You can change the source of a formula variable from Dynamic Formula to Static Formula in the attribute pane Changing formulas to static prevents recalculation Changing to User Entered deletes the formula information completely but saves the values You cannot edit the data in a Dynamic or Static Formula variable unless you first change it to User Entered See Change sources p 77 If you delete a variable that is used in a formula you are alerted to the fact that the variable is used in a formula definition If you continue and delete the variable the formula changes automatically to static and retains its current values You can then change the column to user entered or redefine the formula Sort reorders the rows of a dataset in either ascending or descending order according to the values of one or more key variables in the dataset For example when you sort the rows of Lipid Data on the values of Cholesterol in ascending order the subject with the lowest choles 4 Managing data Recode data 117 terol value appears in the top row and the one with the highest in the bottom row Any sub jects with missing cholesterol values appear at the end e From the Mana
124. 0 2000 1800 Group mean A Lower pe 4 T gt o T A B C D E Group Our plot uses triangle symbols for the error values you can choose others You might also want to delete the legend and use the text tool to change the labels How can generate subgroup and labeling variables Subgroup data often are recorded in sequence with the same number of measurements per subgroup Under these circumstances you could type copy and paste all of the values for the subgroup variable But since the values in the measurement variable follow a defined pattern it is much easier to use a formula to generate the values of the subgroup variable Suppose for instance there are 8 measurements in each of 15 subgroups and all measure ments from a subgroup are consecutive This means that the first 8 measurements are from subgroup 1 the next 8 from subgroup 2 and so on The formula to create these values is 1 Div RowNumber I 8 or more generally Div RowNumber 1 X where X is the number of measurements per subgroup Another pattern in which the measurements could be recorded is that the first measurement from each subgroup is recorded in sequence then the second measurement from each sub group and so on For the example above with 8 measurements from each of 15 subgroups this means that the first 15 measurements are from subgroups 1 15 as are next 15 measure 9 Tips and shortcuts Common questions 243 ments and s
125. 000 f Maximum 234 000 axiraum f Maximum 234 000 a Missing a missing og tMissing Change column widths To change the width of any column including the odeles col umn of row labels position the cursor over a column border When the cursor changes to a cross arrow click and drag the border to a new location 200 7 Customizing results Tables Forrelation Matrix C 7 Weight Cholesterol Triglycerides HEL LBL weight 1 000 022 10g 276 O57 Cholesterol 022 1 000 O1 552 362 ki Triglycerides 108 401 1 000 278 489 HDL 1 000 LDL O57 362 i deg 083 1 000 k 5 observations were used in this computation Forrelation Matrix Weight Cholesterol Trigl HEL LBL weight 1 000 022 105 276 057 k Cholesterol 022 1 000 401 252 962 h Triglycerides 278 489 HEL 1 000 083 LDL O57 SEZ 453 0S3 1 000 ts observations were used in this computation Move tables or components You can move an entire table its title or its note s If you move the table as a whole its title and notes move together and stay in the same positions relative to each other e Click and drag the table or component Do not drag selection handles dragging handles resizes rather than moves the table or compo nent If you want to move only the title or the note select whichever one you want to move and drag it to a new location Moving a title or no
126. 1 Smoker male 131 Monsmoker female 134 Quick How many male nonsmokers do we have How many female smokers It s not easy to tell This arrangement helps Cholesterol readings male 232 ies tat is zza ef 19 e l Now it s easy We have two male nonsmokers and one female smoker This arrangement is also more compact What took 8 rows and 3 columns now takes only 3 rows and 4 columns Smoker StatView lets you record data this way if you prefer Not surprisingly StatView s special struc ture is called a compact variable because it expresses the same information in fewer cells There are several advantages to using compact variables 1 Compact variables help you see which values of a continuous variable fall into which nom inal group or subgroup 2 Sometimes raw data are arranged this way and it is easier to enter data into StatView using the same arrangement 3 Compact variables are visually smaller they take up fewer cells 4 For repeated measures ANOVA StatView requires that your within factor be coded this way Please note though that there are disadvantages 1 Compact variables violate the usual assumption of data arrangement that values on a sin gle row correspond to a single case observation subject patient individual The corre spondence between values of different continuous variables is lost For example we cannot necessarily enter Ages and We
127. 1908 format used date time values are an exact time a specific hour 10 20 13 minute and second of a specific day month and year Currency 142 213 00 Real data displayed in one of many international currency formats Set 113 88 the number of decimal places to display in the dataset with the 872 543 Decimal Places pop up menu Category Female Alphanumeric text data recording group memberships for individual 2 observations or cases You must choose an existing category definition Group or create a new one Values for each type must fall within the following ranges All numerical computations are performed in the fullest precision of the platform you are using so the ranges of acceptable values for real numbers vary among platforms Macintoshes based on 680x0 processors with a numeric coprocessor FPU perform calculations in 96 bit extended precision Macintoshes based on 680x0 processors without numeric coprocessors noFPU perform calculations in 80 bit extended precision Power Macintoshes prc and Windows machines Win perform cal culations in 64 bit double precision 74 2 Datasets Variable attributes Type CPU Minimum Smallest Maximum fraction Real FPU 1 1E4932 1 9E 495 1 1E4932 noFPU PPC 7E308 5 0E 324 1 7E308 Win Long integer 2 147 483 647 2 147 483 647 Integer 32 1617 32 161 Date time Friday January 1 1904 0 00 01 Monday
128. 2 After importing you can edit the category definition to change its labels to Male and Female Integer variables are converted to categories only if the integer variable contains fewer than 255 unique elements and the smallest value falls between 0 and 6 Import non numeric data as type string This option is a safeguard and is checked on by default By default text values are imported with type string However you may turn the option off to import text variables containing repeated group names with type category Make variables with errors have type string If your imported dataset has missing values you don t expect you might want to import again using this option Often missing or incor rect separator characters can cause values to shift to columns where their type is inappropriate which causes error messages or missing values Any variables with errors are imported as string variables this way all data values are displayed and you can investigate and correct the prob lems For more information about data types see Type p 73 and How StatView imports data p 102 StatView can export data to plain ascu text files for use in other applications Before export ing you should investigate the format your target application expects consult its documenta tion so you can make the appropriate choices in StatView s export dialog box 102 3 Importing and exporting How Statliew imports data When you save data a
129. 21 SR330 SR331 DateDifference sr370 Day sR371 DayOfWeek sr372 DayOfYear sr372 decimal characters usv1 decimal places us80 graphs usi88 in examples usv1 see dataset preferences tables us201 US231 defaults see preferences defect variable sr310 degrees sR372 SR405 degrees of freedom sr73 DegToRad sr372 Delete us63 us65 categories US84 Criteria USI29 variables usi16 delimiters USIOI Us252 Us253 denominator df sr41 density plot sR161 Dependent us144 dependent variables sr76 sR77 descending Sort see Sort descriptive statistics US96 USI35 USI36 data requirements SR9 dialog box srz discussion SRI exercise SR9 266 Index SR StatView Reference US Using Statview results sr9 template exercise UsI65 USIG6 templates USI65 SR9 tutorial example us20 design of StatView usi determine whether results are selected Us133 USI deviance sR203 deviance residuals sr170 sR178 SRI9O df see degrees of freedom dichotomous logistic regression sR199 Difference SR373 difference R333 SR370 SR373 unary SR335 differing results usv1 dimensionality reduction sR131 direction of operation sR321 SR323 SR33I SR332 directory of results see results browser disk space us170 US232 distribute space us218 distributions see cDF random numbers func tions Div sR374 divide continuous into groups see Recode division sR334 SR374 SR413 document formulas sr329 document size us213 limit us232 documents vs templates us1
130. 23 e Specify the number of variables 5 e Click Create Five new variables are added to your dataset Random numbers Random Numbers lets you create new variables whose values are generated from a commonly used distribution Random Numbers variables are static formula variables and can be edited accordingly You can also use Formula to generate random numbers with dynamic formulas e From the Manage menu select Random Numbers In the function browser on the left double click a distribution Or type a distribution name directly into the definition area e Replace any arguments and edit any default parameters to suit your needs e Specify the number of rows and columns to create e Click Create Random Numbers of Column 15 RandomBetal 1 is RandomEtinomial co Number of rows to create RandomChiSquare 1 RandomExponential 1 RandomF 1 1 RandomGaramatt RandomGaussian O 1 RandomMormalca 1 RandomPoisson 1 RandomT 1 RandomUniformtd 1 RandomUniforminteger 7 Number of variables to create 1 Random variable definition Attributes Cancel Statview 4 0 This dialog box behaves like a regular window you can resize it use Cut Copy and Paste on the text change fonts and move the window behind or in front of other windows You can double click the top area beneath the title bar to bring its dataset to th
131. 335 Sum sR423 sum USG6O SR333 SR368 SR424 sum of squares US6O R368 SR425 algorithms sr433 SumlgnoreMissing sr424 summary data us52 summary pane see attribute pane summary statistics USI5 US255 tutorial example us8 SumOfColumn sr424 SumOfSquares sR425 SuperANOVA file format uszo formulas us256 import export USIO6 US256 superimpose graphs usi86 supersmoother R221 R225 SR227 R236 suppress recalculation us138 usI39 USI7O survival analysis us142 common questions US245 US250 example sR146 SRI47 functions US245 SRI50 SRIG5 SRI78 introduction sR143 nonparametric methods data requirements sRI57 SRI62 dialog boxes sR152 sRI56 discussion sRI47 SRI5I exercise SRI63 SRI66 results sRI59 templates sR163 plot with confidence intervals us249 preferences US230 US231 regression methods data requirements sRr183 sRI85 dialog boxes SR175 SR183 discussion SRI67 SRI75 exercise SRI9I SRI98 results sRI85 SRI9I stepwise SRI76 templates SRI9I symbols see point types syntax R339 arguments SR323 SR326 SR331 combine functions sR323 constants R324 SR325 expression R325 order of operations sR326 sR328 placeholders sr324 quotation marks SR325 R331 row inclusion R325 SR326 variables sr324 SR325 system configuration US74 troubleshoot us252 system crash us251 system software SR3I7 SR321 R330 SR331 T t distribution sr404 SR4II T usage marker see usage mar
132. 57 As a WME or EMF file 157 Reopen your work 157 Use original variables 157 Assign different variables 158 Print a view 159 Rearrange templates 168 Update the Analyze menu 169 Build templates 169 Template tips 169 Exercise 171 Change overall structures 187 Change axes 190 Change legends 194 Change plotting symbols 195 Change colors 196 Tables 197 Select tables 197 Select components 198 Resize tables 199 Move tables or components 200 Change text items 200 Change overall structure 201 Change line thicknesses and pen patterns 202 Change colors 202 XII Contents 8 Drawing and layout Draw tools 203 Select objects 204 Add text objects 204 Draw objects 205 Import objects 210 Change fill patterns pen patterns and line types 211 Change colors 212 9 Tips and shortcuts Index 257 Tool Bar Windows only 221 Help 221 Hints window 222 Help Windows only 222 Status bar Windows only 223 Tool tips Windows only 223 Apple Guide Macintosh only 223 Balloon help Macintosh only 224 Error messages 225 Alert messages 225 Preferences 225 Application preferences 225 Color Palette preferences Macintosh only 226 Dataset preferences 227 Formula preferences 228 Graph preferences 228 Hints preferences 230 Survival Analysis preferences 230 Layout tools 212 Control page layout 213 Arrange objects 213 Move object
133. 60 1 858 2 161 0328 Weight 005 2 860E 4 847 16 997 lt 0001 Regression Plot 95 Confidence Bands Residuals vs Fitted 28 5 ej biti tit ii 5 267 i S 24 7 g4 L a o o 8 22 E2 f oo Sear io Soc F eo Gi 9 mA 0 04 x e S o Q tee 6187 g e EI Fo E 8 0 Oy a o So 8 08 3167 z i 4 oe ea f O 144 3 2 P rag Po 12 g ENNE L 104 7 8 T T T T T 6 E a e S E E L e a E i 1500 2000 2500 3000 3500 4000 4500 8 10 12 14 16 18 20 22 24 Weight Fitted Gas Tank Size Y 1 858 005 X R 2 717 In view preferences you might want to restore the option Save analysis results with view 178 6 Templates Build templates Customizing results Stat View includes all the features of a graphing program with numerous features enabling you to enhance the clarity and visual impact of your graphs and tables 1 You can set global preferences that establish defaults for every graph and table you create Before producing any results for a presentation you should set preferences to suit your most general needs For example if you are preparing a journal article you should set pref erences to match your results to that journal s style guidelines 2 Edit Display lets you change tables and graphs individually Many options duplicate those that you can set globally with preferences so Edit Display provides a way to override your default settings for specific graphs Other options for controlling features specific
134. 61 us162 dose response sR202 DotProduct sR374 dotted lines us185 dotted red line see page breaks double asterisk SR334 double byte strings manipulating SR379 SR384 SR423 double click row numbers see row inclusion double spacing see line spacing Draw palette us185 UsI95 USI97 USI99 US202 US212 arc tool Us205 Us20 arrow tool us212 tutorial example us45 also see selection tool corner center control us205 curve tool see spline tool ellipse tool us205 fill color us202 tutorial example us43 fill pattern us211 grid us217 line tool usz05 tutorial example us45 line widths us211 pen color us202 tutorial example us43 pen pattern us211 polygon tool us205 US207 US208 rectangle tool us205 rounded rectangle tool usz05 us206 selection tool us185 USI97 US204 spline tool us208 us210 tear off menu Us204 text tool US204 US20 tutorial example us45 drawing and layout Us203 Us212 Drawing Size USI59 US213 dialog box us213 DS Transfer file format us7o us99 Dunnett s sr87 Duplicate us181 us182 Durbin Watson sR59 dynamic formulas USII16 sR329 SR330 SR393 data source US77 also see Formula dynamic links analysis objects us133 Analyze menu vusi68 formulas usio9 USII6 graph text us187 reopen views USI I results and data us138 us139 tables usz2o1 E SR375 SR385 also see hyperbolic functions edit criteria US240 data us64 US70 tutorial example usi5 Index SR StatView R
135. 647 are converted to missing Currency values are rounded up or down to the nearest integer e g 1 234 becomes 1 Values outside the range 2 147 483 647 lt x lt 2 147 483 647 are converted to missing Categories are converted to their underlying codes indices 2 3 Reals are converted to their current text representation as set by Format and Decimal Places this can result in loss of precision Integers are unharmed Long integers are unharmed Date time values are converted to their current text representation as set by Format Currency values are converted to their current text representation as set by Format this can result in loss of precision Categories are converted to their group names as given by the current category definition 76 2 Datasets Variable attributes Date time Reals are rounded to an exact integer number of seconds after the earliest possible date varies by platform Values outside the range 0 lt x lt 4 294 967 295 are converted to missing Integers are interpreted as an exact number of seconds after the earliest possible date varies by platform Negative values are converted to missing Long integers are interpreted as an exact number of seconds after the earliest possible date varies by platform Negative values are converted to missing String values that match valid date time formats see the Formats menu are interpreted accordingly Other string values are converted
136. 8 T3 0 000 23 000 Saturated fat g 393 C While the analysis is selected has black handles notice that the variable browser marks which variables are used each variable has an X marker The X marker means the variable is an X or independent variable in the selected analysis We ll see other markers later Variables ic a Remove Remove Split By Eidi Data Candy Data Candy Bars Data x Order Datase Order Dataset order 7 Erand i l Name ao Servings pkg Brand Name Serving pkg Ozipkg Calories Total fat g S Saturated fat g Cholesterol g 2 Cholesterol g Sodium mg Sodium mg Carbohy dra Carbhnhydrate m Dietary fib Here s the fun part Analysis objects are incredibly flexible You can add variables split the analysis by a nominal grouping variable remove variables 1 2 3 4 5 6 7 replace variables with different variables change the way statistics are displayed choose different statistics etc etc etc We ll try some of these things as we continue our quest for the ideal candy bar Remove variables Let s just look at calories for now Make sure the analysis is still selected has black handles if not click it In the variable browser select Total fat g and Saturated fat g Click the Remove button I Tutorial Analyze data 23 Variables Add Split E Brand fH
137. 9 Enter values 61 Manipulate columns and rows 62 Move and scroll 64 Edit data 64 Select data 64 Cut clear and delete data 65 Copy data 66 Paste data 66 Paste transposed data 68 x Contents Copy and Paste unusual selections 68 Save datasets 70 Exchange datasets between Windows and Mac versions of StatView 70 Close datasets 72 Open datasets 72 Print datasets 72 Variable attributes 73 Type 73 Source 77 Class 78 3 Importing and exporting Microsoft Excel 99 Read Excel files 99 Write Excel files 100 ASCII text 100 Import text 100 Export text 101 How StatView imports data 102 Variable names 102 Data types 103 4 Managing data Manage multiple datasets 107 Include and exclude rows 108 Formula 109 Build definitions 113 Some examples 114 Exercise 114 Shortcuts 115 Errors in formula 116 Dynamic vs static formulas 116 Sort data 116 Recode data 117 5 Analyses Overview 131 Exercise 132 Determine whether results are Format 79 Decimal places 80 Categories 80 Create category definitions 81 Enter category data 83 Edit category definitions 83 Delete unused categories 84 Compact variables 84 Build compact variables 86 Expand compact variables 94 Analyze compact variables 95 Missing values 103 Category definitions 104 Example 104 Older StatView products Macintosh only 105 Text 105 Ol
138. 9 3 672 lt 0001 Group Info for Total fat g Grouping Yariable Calorie groups oelusion criteria Big hree from Candy Bars Qata Count Mean Wariance Std Dew Std Err Blow 45 11 489 12 595 3 545 529 i High 6 24 667 3 067 2 011 1 229 We see that the mean difference is quite large and the p value is well below 0 05 However the groups are vastly different sizes 45 and 6 so we shouldn t take this result too seriously Still it seems apparent that the fat content between groups is significantly different and it does make sense that candy bars with more fat would have more calories What about saturated fat If you took the quiz you probably noticed that calories and total fat were on average pretty similar among the big three brands However the median lines in the box plots for saturated fat looked pretty different In case you re skipping the quizzes here s a plot you would have examined Box Plot Split Ey Brand Inclusion criteria Big Three from Candy Bars Data E Hershey M amp M Mars Nestle Units Saturated fat g From the looks of this plot the M amp M Mars bars are lower in saturated fat per serving than the Hershey and Nestle bars compare the median lines in the middle of the boxes However the Nestle median falls inside the M amp M Mars box height and all three boxes are overlapping with each other We can t be sure just by looking at b
139. 95 o Japan o Japan o Japan o Japan o Japan Other Other Other Other Other a USA USA 2 USA a USA a USA Once you change the layout and frame style of the legend you may want to move it below or above the graph particularly with horizontal legends You can move it by selecting the legend and dragging it to a new location anywhere in the view You can change legend text items with the text tool in the Draw palette and commands from the Text menu If you cut or clear a legend you can bring it back by selecting the entire graph and clicking Edit Display Change plotting symbols For a scattergram you can change point style size and color For a bar chart pie chart or box plot you can change fill patterns colors lines and pen patterns For a line chart you can modify lines and points There are two ways to select the plotting symbols of a graph 1 Select the interior of the graph 2 Select symbols in the legend You can change all the symbols by selecting the entire legend For example you could change every single point to a particular size or color Once a plot is selected use Draw palette tools to make changes Point styles and sizes Points are the plotting symbols in scattergrams line charts percentile plots compare percen tile plots and in box plots To change a point select it and choose a type and or size from the Draw palette Choose the null symbol 9 for invisible points which are ofte
140. 965 61 027 sosse4 53 805 55 197 z 51 709 45 669 56 985 49 259 53 013 55 602 53 950 55 360 4 s 55172 soszz s16o7 51 251 55 095 46 540 55 286 51 664 ef s2797 50313 50 343 seoso 48477 55 612 51 338 51 843 7 sio4o 3337 49 949 48821 53253 s4 077 52 041 57 768 e sss40 50326 49928 51 285 47 170 44825 54445 49 052 f 48 793 51534 48 223 49 940 49615 s271e 55 217 54 638 tof siei 52544 51737 54432 52 335 52217 sot 48 526 When your data are arranged this way you must add both the continuous and nominal por tions of the compacted variable to the chosen analysis as pictured below Add a Split By Data Untitled Dataset 1 Order Dataset order Data arranged this way can be used in any QC analysis that uses subgroup variables i e QC subgroup p np and c u analyses You can also use the continuous portion of any compact vari able in individual measurement analyses If you do use a compact variable for the nonconfor mity and subgroup variables in p np or c u analyses any Item or Unit Count variables you use should match the number of cases in the continuous portion of the compact In the example given above the continuous portion of the compact variable has 10 x 8 80 cases Survival analysis How can estimate the survival function at other covariate values You might wish to evaluate certain regression models for particular covariate values Suppose
141. ANOVA p 73 of StatView Reference Can work with compact variables Sometimes data are recorded so that values from each subgroup are in a different variable column StatView s flexible data handling allows you to use data in this format simply by compacting all of the subgroups into a single compact variable Suppose for instance your measurements or attribute counts fall into 8 subgroups each with 10 measurements in separate variables The data might look something like this 9 Tips and shortcuts Common questions 245 1 sisis s4so6 s6c6s 61 027 sose s3so5 55197 50 103 51709 45669 56985 49 259 53 013 55602 53 950 55 360 4ssoo 49805 49855 sii7o 52535 52275 48 010 53 131 ss 172 50473 sico7 si251 55 095 46540 55 206 51 664 52797 50 313 50 343 s5eoso 4s477 55612 51 338 51 943 51 040 53 337 49 949 48821 53 253 54077 52 041 57 768 53540 50 326 49 928 si1 2e5 47170 44825 54445 49 052 48 793 515z4 48 223 49940 49615 52718 55 217 54 638 Size 52544 Si737 54432 52335 52217 50114 48 526 To compact these variables select all eight of them then click the Compact button in the upper left corner of the dataset When the Compact Variables dialog appears type in a name for the compacted variable then click Compact The compacted variable will look similar to the following 1 2 sisis 54506 s6
142. ANOVA or older versions of StatView You can open and save SuperANOVA and StatView 1 x data files directly with StatView or you can exchange data through intermediate text files When you save data as text only the actual displayed values are saved Be sure to display enough decimal places to meet your needs Formula Recode Series and Criteria definitions are not saved 106 3 Importing and exporting Older StatView products Macintosh only Old StatView data Files saved as StatView Data cannot be read by StatView II StatView SE Graphics or Stat View 512 To export data to any of these packages save the files in StatView 1 x Data for mat All numeric data are saved in full precision Formula Recode Series and Criteria definitions are lost Data class and format attributes are lost Compact variables are saved in expanded format Date time and Currency variables can only be saved if they are first converted to type string SuperANOVA data If you save a dataset in SuperANOVA format nearly all information is kept intact We recom mend using SuperANOVA as an intermediate format to transfer data to Stat View versions 4 00 4 02 Any current selection highlighting of cells rows or columns is lost Row inclusion exclusion settings are lost all rows become included Certain functions used in formulas are not avail able in SuperANOVA although they are interpreted correctly in any version of StatView that supports them
143. Analysis Graphs Nonparametrics QC Analyses Regression Survival Analyses EE blal HvvvvVvV7 CDECPCL You may change the Analyze menu hierarchy simply by rearranging the organization of files and folders inside the Template folder Windows or StatView Templates folder Macintosh e Rearrange how templates are stored within the Template folder Windows or StatView Templates folder Macintosh Use the Explorer or File Manager Windows or Finder Macintosh to create a new folder and file organization You may nest folders insider each other up to five folders deep From the Analyze menu select Rebuild Template List Group related templates into folders according to categories that make sense to you You might want to rearrange templates according to the type of data they use or according to the types of tests they perform or perhaps according to your current projects You can add more folders add more templates put folders inside folders as many as five layers deep and drag files or entire folders to the Recycle bin Windows or trash Macintosh You can even use aliases Macintosh only to point to template files but not folders stored elsewhere 6 Templates Build templates 169 When saving your own views you might want to use the Create New Folder button Win dows or the New Folder button Macintosh in the Save dialog box to create project folders inside the Template folder Windows or StatView Template
144. Analyze data 25 Descriptive Statistics Descriptive Statistics Basic statistics continuous amp Mean E Standard deviation amp Standard error of mean amp Count amp Minimum amp Maximum Basic statistics continuous Mf Mean Mf Standard deviation if Standard error of mean M4 Count Mf Minimum Mt Maximum B statistics nominal B Number of levels amp Count B Number missing x M4 Number of levels M4 Count Denominator for variance etc Cni Cn Denominator for variance etc n 1 On Trimmed mean percentage Trimmed mea Fewer choices e Click OK om Fewer choices Cancel Pescriptize Statistics Calories Mean 244 027 Std Dew m Std Error Count Minimum 125 000 Maximum 450 000 Split by groups How does per serving calorie content of candy bars vary among the different brands We can split this analysis by Brand group to find out e Make sure the analysis is still selected if not click the object to select it e In the variable browser select Brand e Click the Split By button i Variables Pescriptize Statistics Calories Mean 243 027 Std Dew 61 996 E Std Error Count 75 Minimum 125 000 z Maximum 450 000 Data Candy Oz pkg ico Palariar YOA Now we have a table of descriptive statistics broken down by Brand groups Unfortunately it s so wide it runs off the window Ler
145. Calories ANDY A Fisher s PLSD for Brand Saturated fat g Regression Regression Plot for Saturated fat g Calories Regression Coefficients for Total fat g Saturated fat q Bivariate Plot Scattergram for Saturated fat g Calories ANOVA ANDOY Table for Brand Saturated fat g In the view window all the analyses that involve the saturated fat variable are selected so you can change them all at once Next we use StatView s drawing tools to make them red e Select Draw Palette from the View menu Windows or tear off the Draw palette from the Draw menu Macintosh e At the bottom of the tool palette click and hold the pen color tool top rectangle Drag and release to select red from the pop up color palette I Tutorial Present results Apologies for the gray colors e Now the saturated fat results are red and eye catching We won t bother to show the results here a gray color palette looks silly enough Art assignment Call attention to the candy bar lowest in saturated fat per serving e Use the results browser to find the box plot of saturated fat split by Brand If you skipped the quizzes make a box plot of Saturated fat g split by Brand now E Results View Nutrition analysis Order by Yariable Show All Regression Coefficients for Total fat g Saturated fat Bivariate Plot Scattergram for Saturated fat g Calories ANDYA ANOVA Table for Brand Saturated
146. Custom Rulers Align to Grid Align Objects Show Hide Rulers Choose whether to show horizontal and vertical rulers along the top and left edges of the view When rulers are visible hatch marks in the rulers show exactly where your cursor is positioned at that instant which helps you draw and move objects precisely To move the origin zero of a ruler click and drag the zero position tick mark To reset the origin back to the edge of the window click the box in the upper left corner where the rulers intersect Turn Grid On Off Choose whether to have objects snap into alignment with an even grid in the view window The grid affects objects you create and objects you draw Grid lines are spaced at each inch or centimeter depending on your system configuration and Custom Rul ers setting and objects snap into alignment at each tick division along the ruler not just at the drawn grid lines The grid is on by default If you create or position an item when the grid is off you can later snap it to the grid by selecting the object s turning the grid on and selecting Align To Grid Show Hide Grid Lines Choose whether to make grid lines visible in the view window Grid lines are never printed Show Hide Page Breaks Choose whether to make page breaks visible in the view window Custom Rulers Select Custom Rulers if you want to change the ruler measurement units grid spacing or the colors used for page breaks and grid lines
147. F for the latest versions of Excel supported by StatView Read Excel files From the File menu select Open e For List Files of Type Windows or Show Macintosh select Excel e Select the Excel file e Click Open 100 3 Importing and exporting ASCII text Write Excel StatView reads each Excel worksheet into a single StatView dataset so opening workbooks with multiple worksheets results in multiple StatView dataset windows StatView reads only the values in each cell it does not import macros links or functions although it does import the current results of functions Variable names and types are handled the same way for Excel import as they are for text import see How StatView imports data p 102 Stat View can read all built in Excel formats and many custom formats and it assigns imported variables to the nearest equivalent variable types in StatView Any unrecognized formats are converted to type real or string You may change these by using the variable attribute pane see Variable attributes p 73 files From the File menu select Save As e Choose the Excel file type e Specify a filename and folder location e Click Save StatView writes the entire dataset as a single Excel worksheet Stat View writes only the values in each cell it does not export formulas or criteria Currency data are written with zero or two decimal places depending on the current display in StatView All other data
148. F ratio sR74 SR84 free format us79 graphs usi85 tables us199 Us200 views US205 US232 Food Guide Pyramid us2 Force button us144 data requirements sRIG dialog box sr14 discussion SRI3 exercise SRIZ interval settings SRI4 Index SR StatView Reference US Using Statview 269 results sRI6 templates sR16 tutorial example us31 Frequency Summary Table us136 us137 Friedman test sR122 F statistic sR73 F test see unpaired comparisons function browser UsIII USII3 G G usage marker see usage markers Games Howell sr88 gamma distribution sr408 Gaussian distribution sr408 generate data SR338 SR347 also see Series Random Number Formula generator seed see random numbers generic variable names for templates us17o geometric mean SR3 GeometricMean sr380 GeometricSeries SR381 global null hypothesis tests table sr186 golden ratio sR378 graph defaults see preferences Graph dialog box us187 graphs align us214 arrange US213 US214 axes USI83 axis bounds us185 axis frame us185 axis frames usI85 USI87 US229 axis labels us183 axis values usI83 bivariate plots sR221 sR236 box plots sr243 cell plots sr237 chart of types srv colors us229 compare percentile plots sR247 create USI3I USI33 by hand or with templates us131 customize USI83 USI97 decimal places us192 Edit commands us181 edit text us187 fill color us197 fill patterns us196 Us229 fonts usi85 format USI35 frames US183 USI97 grid
149. February 6 2040 6 28 15 Categories Sometimes your data identify particular groups rather than signifying quantities Data like this are nominal You can set any data type to nominal but category data can be used only in a nominal variable When you set a variable s type as category you are prompted to provide information about the groups that make up the variable For example to enter the gender of patients in a study you define a category in which Gender has exactly two group labels Male and Female The only data that can be entered into this column are Male and Female Using a category to enter your nominal data offers you specific advantages over other data types The advantages are explained under Class p 78 See Categories p 80 for more information Date time values It is important to understand the following rules about date and time values 1 Any dates outside the valid time range are invalid However you can type 0 in a date time data cell to get the current date at midnight If you attempt to enter or create by formula any date value outside this range you get either an error message or a missing value 2 If you specify only a date without a time StatView assumes you mean exactly midnight of that day If you specify a time without a date StatView assumes the current day month and year If you specify only a partial date or time Stat View assumes you mean the very beginning of that time for insta
150. Fis Proximity to Charles O Car beta ae Cee Country Names Teaching Effe tiveness Teaching The scrolling list contains all defined categories These are categories used in the active dataset and any other open datasets You can choose an existing category from this scrolling list or create a new category To choose an already existing category select it and click OK For an example of how you might define and use a category see tutorial chapter p 5 Here we give general directions e Choose an open dataset from the Window menu or select New from the File menu e Ifthe attribute pane is hidden double click the FF control above the vertical scroll bar e Select Category from the Type pop up menu in the attribute pane for the column The Choose Category dialog box appears If categories are listed you can choose one and click OK to apply it to the column If no categories are listed it means you have not defined any for this or another open dataset You must click New to create a new category The Edit Category dialog box appears 82 2 Datasets Categories Edit Category Create a new category Category name Category for Input Column Group label a Add Cancel Replace Delete e Type the first group level for the category e Click Add e Repeat these steps until you have named all the group levels you need Gro
151. Format Dec Flaces gt D Input row Attribute pane Input row Attribute pane Each dataset window is a fully independent window You can move it scroll it resize it and close it like any other window in any other program The dimmed cell in the upper left corner of a new dataset represents both an input row and an input column As soon as you enter a value into this cell a new input row appears below it and a new input column appears to the right You create columns variables by entering val ues into the input column and add rows by entering values in the input row Each row repre sents a case and each column contains a variable A new input column appears when you start to fill the current input column or change its name You can use Add Multiple Columns from the Manage menu to add several columns at a time see Add multiple columns p 62 Here s what the window looks like with a dataset entered and the attribute pane closed 2 Datasets Dataset windows 55 Dataset title name Compact and Expand buttons Criteria pop up menu Show hide variable browser Mac onl Click here to EC Mac only deselect any R N Gender 2 Row numbers 3 Variable names 4 A Beal female 6 A Kaufman male T Mubroid male 2 L Fhote male a C Norman male 10 R S Smith Jr male walker male Cholesterol attribute pane
152. Macintosh and Windows The Window menu lists all open datasets views and any open Formula Series Random Numbers Recode and Create Criteria dialog boxes in front to back order The active view or dataset window has a check mark next to it The Window menu also has standard commands for managing windows Windows only The View menu lists 1 Tool Bar 2 Status Bar 3 Show Hide Browsers 4 Variable Browser 5 Analyses Analyze subsets 149 5 Results Browser 6 Draw Palette 7 Hints Window Macintosh only The Window menu also lists 1 The variable browser and results browser 2 The Hints window 3 The Clipboard Analyze subsets Thus far we ve discussed how to analyze all the rows of your dataset at once We ve shown how to use Split By to get separate calculations for each group of a nominal variable or each subgroup from several nominal variables and we ve shown a variety of other analysis controls in StatView s view windows Sometimes you want even more control over the data used in analyses For example rather than analyzing males and females in Lipid Data separately you might want to analyze just males StatView s dataset windows offer still more ways to focus analyses by controlling the data that go into them 1 Sort the dataset according to the values of one or more variables See Sort data p 116 For example if you draw a line plot connecting Weight values graphed against Age va
153. Plot Percentiles Plot Compare Percenti GC Subgroup Mea gt GC Individual Mea Oc PNP Pac cu Pareto Chart Show Order Default gt Survival Nonpar gt Survival Regres An Order pop up menu lets you choose how to sort analysis items in the scrolling list Your Order choice has no effect on how analyses are created Order Default Alphabetigs 5 Analyses Analysis windows 143 Create analysis For multiple results items items with triangle controls you can select any combination of one or more results Click an analysis heading a triangle item to select the default results for that analysis it doesn t matter whether the list is shown or hidden Shift click or click and drag to select several adjacent results items Control click Windows or Command click Macintosh to select several nonadjacent items You can even select items from completely different types of analyses e g you could Control click Windows or Command click Macintosh to select Histogram from Frequency Distri bution and Line Chart from QC Subgroup Measurements Xbar Statistics To perform an analysis select it in the browser and click the Create Analysis button or dou ble click the item in the browser Most selections are followed by a dialog box in which you set parameters for the analysis if you know you want the default parameters you can bypass the dialog box by Right clicking Windows or Control c
154. RS intervals R336 SR338 SR345 frequency distribution sR14 INV USII2 inverse CDF see CDF inverse functions see arc functions inverting matrices SR433 invisible lines us202 IQR see interquartile range IS US255 SR346 IsMissing US255 SR343 ISNOT US255 SR347 IsRowExcluded sr343 IsRowIncluded sr344 item count variable sr290 iterated principal axis factor extraction SRI32 iteration history table sri89 J Japanese characters SR379 R384 SR423 joint significance tests table sRI90 jump point sri49 K Kaiser image analysis sR132 Kaplan Meier sri49 SR163 Kendall rank correlation sr121 data sRI23 data requirements SR123 exercise SRI27 Kendall s tau sri21 key see legends keyboard shortcuts usv1 Us64 draw shapes us205 Edit Analysis Edit Display us233 move graphs usi86 move tables us200 see StatView Shortcuts card Kolmogorov Smirnov test Us86 SRI20 data requirements SRI23 template us234 Kruskal Wallis test sR121 data requirements SR1I23 exercise SRI28 kurtosis Us234 SR6 L Lag sr382 272 Index SR StatView Reference US Using Statview lag R363 lambda srz74 sR76 landscape page us213 Latin square design sR105 sRIO7 layers us215 exercise US2I5 Layout tools us212 us219 tutorial example us43 least significant difference Fisher s protected sr86 Left Justify us205 us218 left to right evaluation sr328 legends us182 us183 USI94 USI95 color us197 frame USI94 layou
155. S248 Cramer s V sRII3 crash us251 create analyses USI3I US133 by hand or with templates us131 templates us161 UsI77 tutorial example Us2o us27 category tutorial example uss us6 compact variables us86 criteria USI24 USI28 graphs USI31 US133 tables us131 Us133 templates us169 us174 exercise USI7I USI77 Create Analysis US132 USI43 exercises USI50 USI56 tutorial example us20 Create Criteria UsI24 USI28 SR317 Criteria USI24 USI29 R325 SR326 SR336 SR339 SR343 SR344 SR409 analyses USI49 USI50 Boolean operators Us127 choose level s us127 compare results with different us182 complex us125 SR409 delete usi29 Edit Apply us129 us240 example us130 hints us222 names USI25 pop up menu US28 US124 USI28 USI29 print definitions us125 random usi28 set values USI26 subtitles usz9 troubleshoot us254 us256 turn off usi29 tutorial example us28 us29 264 Index SR StatView Reference US Using Statview vs Include Exclude Row usi08 windows at Open vus254 US256 also see row inclusion critical values see CDF inverse cross hair cursor see Recode cross platform compatibility us7o us99 crosstabs see contingency tables Csc sR366 CSCC SR261 cube root sR334 cubic spline us208 sR221 R225 SR227 SR228 R233 CubicSeries sR367 CumProduct sr367 CumSum sR368 CumSumSquares sRr368 cumulative distribution function see CDF cumulative hazard function
156. To learn the correct character look at functions listed in the Formula dialog box The separator character varies according to the international number format in use For example a formula with period decimal characters uses commas but a formula with comma decimal characters uses semi colons RandomNormal 1 5 3 0 RandomNormal 1 5 3 0 Try clicking Compute If it doesnt work open the Hints window and check for error mes sages Index SR StatView Reference US Using Statview Symbols A R333 SR335 abbreviations see syntax us64 abort calculations us138 R336 SR337 abscissa see X axis R337 absolute value sR348 SR79 SR333 accelerated failure time SRI72 R487 SR334 actuarial analyses SR148 SR149 R333 SR335 SRI52 SRI53 SRI6I see missing values R334 Add Multiple Columns us62 add results us152 SR329 add variables us132 US144 lt SR337 SR340 multiple vs compound lt gt SR34I results USI35 USI37 USIZO US229 US255 SR340 tutorial example us21 gt SR337 SR341 also see assign variables USII3 add vertex Us208 US210 sR337 addition SR333 R368 SR374 SR423 SR424 sR337 unary SR335 M SR334 adjusted R squared sr56 _sR336 adopt variable assignments US131 U5133 US255 SR34I tutorial example us2z9 uUs30 lt R337 R340 alert messages US222 US225 2 SR337 SR341 algorithms sr433 sR480 T SR400 Align Objects us218 SR345 Align to Grid us218
157. Transposed from the Edit menu The block of cells in the tar get dataset should be as wide as the data to be transposed are tall For example if you have a 3X8 set of data to be transposed you need a block of cells 8x3 for the transposed data To transpose the following data you need to either select the cell at the intersection of the input row and column or an 8X3 pattern of target cells 1 2 4 5 1 z 1 1 4 Ei 10 3 16 19 22 u Caf s etait a e a as EE Cstef st izt ist sat ait 34 13 z0 Z2 23 If the target dataset does not have enough rows or columns to hold the new data either create enough rows or columns to hold the data or highlight the correct number of cells in the input row or input column If you paste into the input row or input column the dataset automati cally enlarges to hold the transposed data Copy and Paste unusual selections As we ve described above StatView lets you Copy and Paste data selections of any shape or size and from one shape to a different one Why would you want to Copy and Paste data from one cell to many or from one nonadjacent selection to another of a different shape Suppose you are entering data for a typical factorial experiment In real life the factor levels would be more tedious to type than simple one digit numbers and you would have more cells to enter Therefore it could be handy to type the values one time each and then use
158. US231 box plots us96 US133 USI35 USI37 SR243 axis labels vs legend text sr245 change style us188 data requirements SR244 dialog box sr243 discussion sR243 exercise SR244 results SR244 subgroups us244 templates SR244 tutorial example us29 BoxCox sR358 braces sR336 brackets sR337 breakpoints see Recode Breslow Gehan Wilcoxon test SRI5I SRI56 browser see analysis browser formula brows er results browser triangle controls variable browser C C class marker see class marker C usage marker see usage markers clu analyses see QC c u analysis calculations background us138 cancel us138 control us138 us139 precision US73 save results with view Us141 calculator keypad us112 cancel calculations us138 Candy Bars Data us2 us48 capability analysis SR252 SR254 SR255 R261 capa dialog box SR263 SR267 capa table sr272 capability indices sr262 Co SR261 Cpm SR261 example sR275 indices US244 SR254 SR476 individual measurements analysis sR279 k centering index sr262 parameters SR267 caret R334 case number sR417 case control studies sR203 case sensitive US255 casewise operation SR32I SR323 SR33I SR332 categories US74 US80 US84 add usg1 advantages us80 compact variables usg1 create US8I example Us92 data type US73 R320 delete us84 Us91 disadvantages us80 edit us83 us84 enter data us83 how StatView uses order us238 us240 import USIO4 US254 multiple us254 nominal dat
159. USI64 compact variables us89 us93 tutorial example us21 class see data class classify results see Split By Clean Up Items us159 us213 tutorial example us43 Clear us65 usi181 us182 Clipboard us66 us181 import pictures text US210 transfer data US66 clone analyses us13I USI33 USI46 tutorial example us26 us27 closed interval sr337 closed polygon us207 closed spline us208 coded raw data USSI US84 SRI14 coded summary data us52 USs53 SRII5 coefficient correlations table sri89 coefficient covariances table sr189 coefficient of determination see R squared coefficient of variation Us6o SR4 CoeffOfVariation sr360 colinearity sR53 SR58 collection of analyses see templates color palette preferences us226 Us227 colors axis USI97 graph frame us185 graph text USI97 Index SR StatView Reference US Using Statview graphs us229 grid Us217 page break us217 plots us197 preferences Us226 Us227 shapes us212 table text us199 tables us199 us202 tutorial example us43 view background us212 column attributes see attributes column charts see cell plots columns add multiple columns us62 insert US62 labels us199 selecting us64 transpose into rows US68 vs variables uss3 widths in tables us199 columnwise operation us108 UsI24 SR32I SR323 SR325 SR326 Combinations sR361 combinations sR358 combine analyses us162 combine datasets see merge combine functions sR323 combin
160. View Guide To use StatView Guide launch StatView and select StatView Guide from the Help menu on the right side of your menu bar 224 9 Tips and shortcuts Help Show Balloons Statview Guide StatView Hints This launches StatView Guide Click Topics for an overview ap StatView Guide Topics Index Look For Select a topic and phrase and click OK to begin StatView Guide leads you step by step 1 Click a topic area 2 Click a phrase then click OK S 7 How do I compute actuarial analysis ANOVA factorial ANOVA repeated measures Bartlett s test of sphericity Bonferroni Dunn test Breslow Gehan Wileoxon test c u charts capability analysis chi square test correlation matrix covariance matrix Techniques through the analysis you want to complete For more information about how to use Apple Guide switch to the Finder and select About D How do I compute a Bonferroni Dunn test Most analyses in StatView are created by 1 selecting an analysis from the analysis browser 2 setting analysis parameters in a dialog box and 3 assigning variables with the variable browser Do This Before you begin make sure a dataset is open and that a view window is open and active To begin click the right arrow Apple Guide from the help menu Balloon help Macintosh only Macintosh Balloon help explains almost all program features the choices in dial
161. View can read files delimited by tabs spaces commas returns or any character you specify e From the File menu select Open e Choose the Text file format Select Candy Bars txt from the Sample Data folder e Click Open e Click Import Our sample file is tab delimited so we don t need to change any settings Import Please specify how this text file looks Items may be separated with tabs spaces commas returns Number of variables Convert small integers to Categories EJ Import non numeric data as type string Make variables with errors have type string Stat View reads the values in each column of the text file and does its best to guess the appro priate attributes I Tutorial Manage data 3 eS Se Real Real User Entered User Class Nominal Continuous Conti Format Free Format Fi Free nena er ee 7 3 Snicker i Harchan Canbiac in Notice that StatView made the same guesses for this file as it did for it s Excel equivalent You would need to make the same adjustments to its attributes You may experiment with it if you like e Close the dataset Open a dataset Often you will begin your StatView data analysis sessions by simply opening a Stat View dataset perhaps one you saved the day before perhaps one you received from a colleague Since all display attributes are saved along with
162. a and criteria definitions You need to update those formulas or criteria to use the current level names from the new category definition s Do your variable names contain reserved words Function names operators and cate gory level names should not be used in variable names Many examples in this reference break this rule so that dataset illustrations are easier to interpret Was the dataset created in SuperANOVA Certain functions unique to SuperANOVA can not be computed in StatView Edit the definition to use StatView functions and click Compute Was the dataset created in a newer version of StatView SuperANOVA and older versions of StatView have fewer functions check to see whether you re using any functions that your program doesn t support StatView warns you about any such formulas before saving to older file formats Was the dataset created on a different type of system Different platforms do their compu tations in different levels of precision Formulas calculated in one version need to be recal culated in another To suppress recalculations between platforms choose the Convert dynamic formulas to static option in Dataset Preferences When opening datasets from other platforms x Recompute dynamic formulas Convert dynamic formulas to staticky Was the dataset created with a foreign system Correct the separator characters used between arguments of a function to match the character used on your system
163. a arrangement Typically data are arranged so that each single row represents a single distinct case or individ ual and each column contains a variable of measurements on each case Most datasets are arranged this way and all StatView analyses can use data arranged this way For example we could record eye color and gender data for eight people like this 2 Datasets Dataset structure Each nominal variable is a separate column in the dataset Each row represents one person and the values in that row identify that person s eye color and gender that is the eye color and gender groups to which that person belongs This is the most typical data arrangement Continuous variables can record other measure ments for each individual for example this dataset could also have columns for weights heights and ages Throughout all columns though one row in the dataset must represent one observation one individual subject one case Within this one case per row arrangement you can use a special compact variable structure in which continuous values are recorded in separate columns representing groups of a nomi nal variable and these columns are linked together compacted into a single structure within the spreadsheet Repeated measures anova requires that the levels of the within factor the repeated measures be stored in a compact variable Elsewhere in StatView compact variables are optional For a complete discussion se
164. a class us74 Us78 problems from editing us256 recode USII8 US238 US255 reorder levels us238 us240 required us81 US9I USII8 Index SR StatView Reference US Using Statview 261 tutorial example us33 CDF Bernoulli sr4o1 binomial sr4or chi square sR402 F sr4o2 inverse chi square sR414 F sr4ts normal sR415 t sR4I6 normal SR403 SR404 Ceil sR359 cell axes usI90 dialog box us192 labels us192 tick marks us192 cell normality sr84 cell plots sr237 axis labels vs legend text sr240 cell bar chart picture sR240 cell point chart picture sR241 data requirements SR238 dialog box sr237 discussion sR237 exercise SR239 results sR239 templates SR239 censor SR488 nonparametric analyses SR147 SRI57 pattern plot sri6r regression methods sR175 SR184 center justify shapes Us205 US218 central limit theorem sR258 central tendency Average SR357 AveragelgnoreMissing SR357 GeometricMean sr380 HarmonicMean sr382 Mean sp388 Median sr389 Mode sr391 TrimmedMean sr428 change analysis parameters US134 USI35 appearance of results us135 criteria US240 data class uss1 Us78 data source US77 data type Us75 us76 formulas us240 templates us171 variable names us58 characteristic roots SRI32 Chinese characters SR379 SR384 SR423 chi square sR24 contingency tables sr12 data requirements SR25 distribution SR402 R407 results sR26 choose group s see Criteria ChooseArg sR359 chords usv1 class markers USIIO USI43 USI63
165. a view This picture is not a dynamic object You can no longer use Edit Display or Edit Analysis change variable assignments or update results with criteria inclusion or Recal culate You should ot use clipboard commands to move objects around 182 7 Customizing results Clipboard commands Recalculate controls all the objects in a view at once so using static pictures is a handy way to maintain a view of data in several states at once If you want to compare results with criteria or inclusion against those without for example Copy and Paste the results as a picture Next change the dynamic object the original results Finally compare the two results side by side An alternative is to extract a subset to a different dataset and create results with variables from that smaller dataset alongside results from the full dataset s variables Note that StatView lets you work in many views and datasets at once Cut and Copy Duplicate Clear Paste Cut removes a selected object or component a table or graph or a part of one from the view and places a copy into the clipboard Copy places a copy of a selected object or component in the clipboard but leaves the original in the view unharmed If you mistakenly Cut a graph title or legend you can restore it with Edit Display s Show title and Show legend options Duplicate places a second copy of a selected object or component a table or graph or a part of one in the view
166. able Click a row or column label or anywhere inside the table Draw palette Font size and style of values inside the table row labels Click a row or column label or Text menu and column labels anywhere inside the table Font size style alignment and angle of title and notes Click the title or notes Text menu Color or text of title Click the title Draw palette Font size style alignment and angle of note Click the note Text menu Color or text of note Click the note Draw palette Resize tables You can resize tables several ways 1 Resize the whole table by dragging selection handles 2 Change column widths one at a time by dragging column borders 3 Choose a different font size for the whole table from the Text menu See Change text items p 200 4 Change row heights with Edit Display See Row height p 201 Resize whole table select the whole table then click and drag one of its selection handles to a new location Drag a corner handle to resize a table s height and width at once drag a side handle to ae Pelee or width alone A peseriwuige Statistjes escript ye Statistjps parietis Statistjes weight weight i i Weight Mean 158 653 ean i i Mean 158 653 Std Dev 28 389 td Dev i i Std Dev Std Error 2 913 td Error 2 913 f Std Error i Count 35 ount p35 i Count f Minimum 107 000 ininum f Minimum 107
167. able assignment slots Variable browser Variable browser Use the Data pop up menu to select among open datasets The name of the currently active dataset is shown Click on the menu to choose another open dataset and choose Other to locate and open a previously saved dataset Use the Order pop up menu to choose how to sort variable names in the scrolling list Dataset order The order in which variables appear in the dataset s columns left to right Alphabetical Alphabetical order by variable name with nonalphabetic names first Variable class Grouped in order by continuous nominal and informative alphabetical order Variables appear in a scrolling list Icons next to variable names indicate their data class E for continuous and E for nominal Informative C variables are not listed because you cannot assign informative variables to analyses or templates Compact variables have triangle controls and symbols Click the triangle to tip it down ward 57 and display the category of the variable marked nominal n Effectiveness E wrEffectiveness IE Time E An M usage marker appears next to any variable that has been mapped to an assignment slot Template list Labeled slots in the Template list show the various roles variables play in the template Hints for each slot tell how the variable s are used 164 6 Templates Use templates Icons to the left of each slot indicate which class the va
168. ables that share the same group level names you should import these variables as type String rather than Category Make sure that the Import non numeric data as type String option is checked in the import dia log box This prevents StatView from creating multiple categories with the same group level names If you are having problems related to printing following these guidelines might remedy the situation Try printing from another application to determine if it is indeed a printing problem or if the problem lies only in the use of StatView Make sure you have your printer and printer drivers correctly installed Some printers require that you install and use special screen fonts otherwise you might get misaligned print or odd spacing Be sure that enough memory is available for printing Formulas and criteria When click compute nothing happens Why Does the Hints window offer any help The Hints window tells you when more or fewer arguments are needed when an argument is the wrong type such as a variable when a number is expected and so on Are any variable names or function names misspelled Are you missing any parentheses You might need to add parentheses to control the order of operations Extra pairs of parentheses are rarely harmful StatView highlights unmatched parentheses so you can fix them Often error messages about the wrong num ber of arguments are due to a problem with parentheses 9 Tips and shor
169. ag too slowly StatView might mistake your movement for an attempt to select a tool Drag its title bar to reposition it If you prefer you can use Draw tools from the menu you do not have to tear off the palette You can place the Draw palette anywhere on the screen so that it is handy but does not obstruct your results Select objects The selection tool activates the usual arrow cursor which lets you select and move tables graphs shapes and text Add text objects You can add text anywhere in the view to label or comment on a result or to modify a text component of a result e From the Draw palette select the text tool e Click where you want to enter text e Type your text press Enter Windows or Return Macintosh to start a new line As you type the box expands to the right and downward to hold your text Here is some text il Here is some text and another line and anather A Drag selection handles Select the selection tool select a text object and drag its selection handles to resize a text box After you resize a text box yourself its width becomes fixed any further text you add wraps to additional lines to stay within the established width You must reselect the text tool to add more text TE E n n Ters is some text Here iF some Here is some Hnd another Tite m text text gnd another mind another m and another Tine Tine gv angther
170. alog box You can change a table s numeric formatting borders and row heights and you can change its orientation To use the Table dialog box e Select the entire table e Click Edit Display Table Numbers Format Free Format Fined Decimal places La Always have leading digit Table format Row height Transpose rows and columns Show Cancel Numbers Choose numeric format and number of decimal places see Format p 79 and Decimal places p 80 Check Always have leading digit to include a leading zero in frac tional values e g 0 25 Table format Choose a set of borders from the pop up menu or select Other to design your own EB EEE ESE OEE Select Other to get a Borders dialog box in which you can design your own table format First choose a line style then click any border that should use the style Borders Click on a row or column border Row height Specify a line spacing for the table e g 1 for single spacing 2 for double spac ing etc The base height for single spacing varies according to font size Transpose rows and columns Check this option turn it on exchange rows and columns thus making a wide table tall or vice versa 202 7 Customizing results Tables Descriptive Statistics
171. alues are interpreted as an exact integer number of seconds since the earliest possible date varies according to platform Currency values are unharmed Categories are converted to their underlying codes indices 2 3 Integer Reals are rounded up or down to the nearest integer e g 1 234 becomes Values outside range 32 767 lt x lt 32 767 are converted to missing Long integers outside the range 32 767 lt x lt 32 767 are converted to missing Strings that are valid integers are unharmed Other strings are converted to missing Date time values are interpreted as an exact number of seconds since the earliest possible date varies according to platform Values greater than 32 767 are converted to missing Currency values are rounded up or down to the nearest integer e g 1 234 becomes 1 Values outside the range 32 767 lt x lt 32 767 are converted to missing Categories are converted to their underlying codes indices 2 3 Long Integer String Reals are rounded up or down to the nearest integer e g 1 234 becomes I Values outside the range 2 147 483 647 lt x lt 2 147 483 647 are converted to missing Integers are unharmed Strings that are valid long integers are unharmed Other strings are converted to missing Date time values are interpreted as an exact number of seconds since the earliest possible date varies according to platform Values greater than 2 147 483
172. an be edited e Select one or more analysis results e Click Edit Display e Change the display option settings e Click OK For more on editing graphs and tables see Customizing results p 179 Shortcut Double clicking an analysis has the same effect as selecting it and clicking Edit Dis play Note that Alt double clicking Windows or Option double clicking Macintosh an analysis is the same as selecting it and clicking Edit Analysis However a view preference lets you switch these shortcuts See View preferences p 232 Multiple and compound results You can assign as many variables to an analysis as you want The way StatView handles extra variables varies according to the analysis For some analyses assigning more variables changes simple graphs or tables into compound tables or graphs that show results for all the variables at once For example assigning two vari ables to a Descriptive Statistics table breaks the table into a compound table with a row for each variable Similarly assigning two variables to a box plot changes it to a compound box plot where a single graph frame contains boxes for each variable 5 Analyses Overview For other analyses assigning more variables causes StatView to produce multiple analyses For Descriptive Statistics Weight Cholesterol Mean Std Dev Std Error Count Minimum Maximum Missing 158 653 28 389 2 913 95 107 000 234 000
173. and Macintosh are registered trademarks of Apple Computer Inc Portions of this software are copyright by Microsoft Corporation Microsoft Windows Windows 95 and Windows NT Microsoft Corporation are either trademarks or registered trademarks of Infinity Windoid wp r is 1991 95 Infinity Systems ISBN 1 58025 162 5 Ongin Credits Stat View began with Jim Gagnon and Daniel S Feldman Jr Subsequent versions of Stat View owe credit also to developers Joe Caldarola Alex Benedict Ania Dilmaghani William F Fin zer Keith A Haycock and Jay Roth StatView has enjoyed its celebrated place in the interna tional marketplace thanks to the ongoing efforts of Hulinks Japanese StatView ALsyD French StatView Cherwell German StatView and many others in the United States and worldwide In 1997 John Sall Senior Vice President and Cofounder of SAS Institute Inc led SAS Insti tute s acquisition of StatView and undertook its development Recent development of StatView was completed by Eric Wasserman Ph D and Charles Soper Clifford Baron led recent product design and development efforts Colleen Jenkins Director of the Statistical Instruments Division guided StatView s entry into the Statistical Instruments offerings of SAS Institute with sales marketing and technical support from Bob McCall Chuck Boiler Bonnie Rigo Nick Zagone and Mendy Clayton Annie L Dudley led the testing efforts of Nicole Hill J
174. ands or normal distribution curves superimposed on a graph Legend Symbols and text identifying how variables or groups are depicted in the graph 7 Customizing results Graphs Notes Text below a graph such as a regression equation or a message explaining why the graph is empty Reference lines Lines indicating means standard deviations or other computed values Select graphs With a whole graph selected you can 1 Change thickness color and pen pattern for all lines in the graph 2 Change the color of the whole graph 3 Add a color or fill to the graph s interior 4 Use Edit Display to change the graph as a whole If the cursor isn t an arrow first choose the selection tool from the Draw palette Ha SNIA To select an entire graph click directly on the frame or interior of a graph Do not click a bar point line text item or any other component To select several graphs Shift click them or click and drag a marquee around them Box Plot Box Plot 425 240 404 a 220 3754 a5 4 a 200 32 5 5 180 z0 J z 275 4 VEN 25 7 140 225 4 rea H 120 175 r 100 r Age Weight You can also use the Results Browser to select objects see Select results p 148 Black handles show that a graph is selected You can also check the Result Selected note in the upper right corner of the view or use Show Selection in the Results browser Box P
175. ane The top five rows of the dataset tell you the type source class format and decimal places for each variable You can change each of these attributes directly just click and hold the cell you want to change then select the correct setting from the pop up menu You can show or hide as many of the rows as you desire by double clicking or clicking and dragging the attribute pane control on the vertical scroll bar Notice that the control changes appearance when the attribute pane is closed E a B open closed attribute pane control open closed The first attribute type indicates whether the data are integers real numbers categories group memberships string currency or date time Since our first variable Brand contains group or category data we need to change the type to category e Scroll back to the first column e Click and hold the mouse button on the Real cell in the attribute pane e Select the correct type Category e Release the mouse button Integer r Integer gt Type Real mie Long Integer gt Tupe Real Real String String Currency Currency Date Time Date Time The category data type is a timesaving feature in StatView When a variable records group memberships the same values are used repeatedly for many cases For example many differ ent candy bars are manufactured by the major brands Hershey Nestle and M amp M Mars Each candy bar is a case belonging to
176. aphs the default font and the thickness of lines in printed output Edit Display dialog boxes The Edit Display button at the top of the view window lets you customize any graph or table in a view window e Click to select the graph or table e Click Edit Display Shortcut Double clicking a result has the same effect as selecting it and clicking Edit Display Note that both Right double clicking and Alt double clicking Windows or Option double clicking Macintosh an analysis is the same as selecting it and clicking Edit Analysis How ever a view preference lets you switch these shortcuts See View preferences p 232 Edit Analysis CRA Box Plot Split By Gender Pescriptize Statistics 240 B Weight 220 4 a L Mean 155 653 q F Std Dev 28389 200 y ati Error 2 913 180 t E mae Count 95 5 160 4 fo female Mini 107 000 a EAE tao 4 L Maximum 234 000 4 i Csi s Missing o 120 i F 100 Weight e Choose the options you want e Click Show to preview your changes e Click OK to apply your changes 7 Customizing results Clipboard commands 181 Graph Flip horizontal and vertical aves Table Numbers Format Free Format Fined Decimal places La Always have leading digit Bounds include extra lines EJ Show legend EJ Show title Dimensions Vertical inches
177. ard deviations 10 4g above or below that if they re normally distributed That s a pretty big spread If you were watching your fat intake which candy bars would be good choices The Mini mum for Total fat g is Og Scroll through and look for the case s with 0 values to see that Super Hot Tamales are your choice Other choices with small numbers are Big Hunk York Peppermint Patty Skittles Sugar Daddy Tiger Sport and Twizzler How many candy bars are in the dataset Look at the Count for Calories The count is 75 which means we have 75 nonmissing cases Since Missing Cells is 0 we also know that our dataset has 75 rows Did any manufacturers refuse to tell us about saturated fat Look at Missing Cells for Satu rated fat g Since it s 0 we know that the manufacturers complied with rules and printed this information for all the candy bars When youre done close the entire attribute pane e Double click the pane control z to shrink the pane e Double click it again to close it completely I Tutorial Analyze data 17 Compute formulas Before we move on to analysis let s generate a formula The guidelines say we can have up to 2000 calories a day How many candy bars would that be if we didn t eat anything else e From the Manage menu select Formula e Use the calculator pad or your keyboard to enter 2000 e Double click Calories from the list of variables in the upper left corner e Click Compute Form
178. are treated as one separator character leaving four variables 3841219 104 3 Importing and exporting How Statview imports data Category definitions Categorical variables with the same levels group names share the same category definition For example if two columns both use only the values Mouse Dog and Cat StatView assigns the same category definition to both This way you can easily Copy and Paste between the col umns Example The following example illustrates data type assignment and the effectiveness of the Convert small integers to categories choice in the Import dialog box Examine the Text File Example in a text editor such as Microsoft Word Notice that it con tains no variable names If you save be sure to save as plain text 12 2 87 10 001 2 12 3 87 20 001 3 3 1244 87 20 001 4 4 12 6 87 50 001 5 2 AN SCN H Oo a 12 10 87 90 001 5 6 12 1 87 1 1 1 red Charlie blue Parker 1 Miles yellow Davis 12 5 87 40 001 5 1 5 red John red Coltrane 12 7 87 60 001 5 3 7 blue Roscoe 12 8 87 70 001 54 8 green Mitchell 12 9 87 80 001 55 9 10 23 7 red Ra In StatView import Text File Example using default settings Column 1 Column 2 Column 3 Column 4 DateTime Currence Real User Ent User Ent User Ent Integer User Ent Column 5 Column Column 7 Continuous Continuous Continuous Continuous User Ent User Ent User Ent Con
179. ariables dialog box us163 function browser uUsiII results browser us147 variable browser us56 USIIO USI43 ordinal axes UsI90 USI93 ordinate see Y axis orientation page US213 also see transpose orthogonal factor solution SRI35 SRI37 out of control sr253 out of memory US232 US25I outliers sR57 SR391 and variance sR4 box plots sr243 in descriptive statistics SRI output list in analysis browser us141 ovals us206 overlay graphs us186 graphs tables us215 P p value analysis of variance sR74 contingency tables sR112 correlation SR44 one sample t test sR23 paired comparisons SR29 regression SR57 regression intercept SR55 unpaired comparisons SR37 z test SR3I p np analyses see QC p np analysis page breaks us27 us159 color us217 print us214 show hide us217 page orientation US213 paired comparisons SR29 data sR33 data requirements SR33 dialog box sr32 discussion sR29 SR32 exercise SR34 SR36 nonparametric SRI2O SRI2I paired t test SR29 results SR33 templates SR34 z test SR3I z test SR3I pairwise deletion SR45 parametric models SRI7I SRI72 parentheses SR327 SR336 SR349 intervals SR337 Pareto analysis US238 SR256 SR3II Index SR StatView Reference US Using Statview 277 data requirements SR3IO SR3II dialog box sR309 sR310 discussion SR309 exercise SR3II SR3I3 results sR311 partial correlation SR45 SR205 partial F ratio SR53 Pascal s triangle sR358 Paste US66
180. asks Win 222 9 Tips and shortcuts Help dows Help is a hypertext system you can click any highlighted words or phrases to get brief definitions or to jump to related topics Status bar messages Tool tips and Balloon Help teach you the functions of items you see on the screen menu items buttons and so on Apple Guide answers questions about how to do things by leading you step by step through the process if necessary asking you questions along the way pointing out where you should look and even doing steps for you Hints window The Hints window is a floating window that contains helpful information about every item on the screen You can make it visible any time by selecting Hints Window from the View menu Windows or Hints from the Window menu Macintosh The Hints window appears automatically in some situations but does not close automatically You can hide it by clicking its close box You can resize it and move it anywhere on your screen E Hints Welcome to StatYiew This is the Hints window It Welcome to Statview This is the Hints window tt a provides information about every aspect of this provides information about every aspect of this program To display a hint click on any item To close program To display a hint click on any item To this window click the close box To show this close this window click the close box To show window choose Hints from the window or Z menu this window choose Hint
181. ate a basic set of descriptive statistics from these data then continue by adding results from a regression tem plate Please close any datasets or views before starting this exercise e From the Analyze menu select Descriptive Statistics and then Descriptive Statistics Point to Descriptive Statistics and then select Descriptive Statistics from the submenu Analyze New View M Rebuild Template List For Beginners b ANOVA t tests b Correlations ld Descriptive Statistics Descriptive Statistics s Factor Analysis b Descriptive Stats Complete Graphs b Frequency Dist Continuous Nonparametrics b Frequency Dist Nominal OC Analyses b Nominal Descriptive Stats Regression b Percentiles ld Survival Analyses e Open Car Data from the Sample Data folder e Double click to assign Weight Turning Circle Displacement and Horsepower Assign Variables for Descriptive Statistics Please double click or drag the desired variables into the proper slots in the template Template Variables Variable s Data Car Data Weight Order Dataset order Turning Circle Model Displacement Country Type Weight Turning Circle Displacement Horsepower G Gas Tank Size kki Cancel CN H The slot has a g marker indicating that you can only assign continuous variables to it Try double clicking a nominal variable H An alert tells you that the class of the
182. atforms StatView stores real numbers and does its computations to the fullest precision of the machine you are using This precision varies between platforms see Type p 73 When you import a dataset created on another plat form StatView usually recomputes dynamic formulas so that its values are computed the way they would be computed on your current platform If you prefer they can be converted to static formulas that are not recomputed See the online documentation for details about file formats that can be opened by your version of StatView 228 9 Tips and shortcuts Preferences Formula preferences gt gt Formula Preferences 5S Formula Preferences Default appearance Default appearance Font vial x Font Geneva Size E 7 size P m Cancel Font and Size Choose the font and size you want the formula window to use You can type a specific size Your choice takes effect for the next new formula window You can change any individual Formula window with the Text menu Enlarging the font size beyond 12 or 14 point might render the buttons in the keypad of the Formula dialog box illegible If you want to use formulas to work with double byte text strings such as with Kanji charac ters you must set the Formulas font to an appropriate font such as Osaka Graph preferences Graph preferences let you change the default display for all graphs while Edit Display lets you fine tune each gra
183. ays from the date of entry into the study until the date of that patient s next seizure or the date of censoring which will be the last day of the study Your data might look like this Emy Date End Date Date Time Date Time User Entered User Entered User Entered Continuous Continuous Nominal Nominal J Jan 1 1904 Jan 1 1904 e Je 2 dnni 1985 Sen 1929 Ineancared Cantral You can use the DateDifference to compute event times or elapsed times from these data DateDifference End Date Entry Date 4 The third argument specifies which time units to use 1 for years 2 for months etc In this case we specify 4 for time in days 244 9 Tips and shortcuts Common questions QC analysis How can draw box plots of subgroup measurements Many QC analysts find that box plots give a useful summary of the distribution of sub grouped data To generate a box plot quickly simply select a completed subgroup measure ment result such as an Xbar chart from the view then double click Box Plot in the analysis browser This adopts variable assignments to produce a single chart with box plots for each subgroup in the same order as those in the QC result Xbar Line Chart Box Plot Control Limits 3 Sigma Grouping Variable s Date 74 145 74 35 Set epee ME
184. by other applications f I Tutorial Manage data Il Read an Excel file Check the WhatsNew PDF document finstalled in the StatView folder for the latest informa tion on versions of Excel files that StatView can read e From the File menu select Open e For Files of type Windows or Show Macintosh select Excel e Select the file Candy Bars xls from the Sample Data folder e Click Open While StatView is importing the dataset the cursor changes to a yin yang and a message win dow shows its progress Reading Data A StatView reads the entire Excel worksheet into a single StatView dataset StatView reads only the values in each cell it does not import functions macros or links This is the complete Candy Bars dataset Take a moment to scroll right and left to see all the variables then scroll up and down to view all 75 rows g Candy Bars xls imported Brand Name Serving pkg Dz pkg Calories Total fat g Saturated fat g Type String String Real Real Integer Real Real b Source f User En User En User Entered User Ent User Ente User Entered User Entered b Class Nominal Nominal Continuous Continuous Continuous Continuous Continuous b Format Free Format Fi Free For Free Format Free
185. cell Click one cell and drag vertically or horizontally to highlight all of them Or Shift click the several adjacent cells cells 2 Datasets Edit data 65 Control click the cells Command click the cells several nonadjacent cells Click the corner cell and drag diagonally to the opposite corner a small block of cells Shift click that corner cell Click a cell in one corner of the block Scroll to the diagonally opposite corner of the block a large block of cells Select one block with either of the above techniques Control click and drag to select another block Select one block with either of the above techniques Command click and drag to select another block nonadjacent blocks of cells From the Edit menu choose Select All Rows or Select All Columns all rows or all columns Click the row number Be careful to single click not double click double click is used for Include and Exclude see Include and exclude rows p 108 an entire row Click the variable name an entire column the end of the group and Shift click Click a row number or variable name and drag over adjoining row numbers or variable names To select a large block of data click the first row or column scroll to the row or column at adjacent rows or columns Control click the row numbers or variable names Command click the row numbers or variable names
186. ck OK and repeat these choices for the second axis e Click in the empty space of the view to deselect the graph 6 Templates Build templates ee 2500 F eal oan I 1500 eee t Z 1000 d e t st 500 T Poe n T 0 Se 150 250 350 450 550 X Variable You have finished customizing the graph and are ready to save this view as a template From the File menu select Save e Specify a filename My Scattergram e Choose some location in the Template folder Windows or StatView Templates folder Macintosh e Close both the view and the dataset do not save changes You are now ready to use this template to format a new scattergram You can use this template with any dataset e Open Car Data from the Sample Data folder e From the Analyze menu select Rebuild Template List e From the Analyze menu select My Scattergram If you chose a subfolder in the Template folder Windows or StatView Templates folder Macintosh the template appears under that heading in the hierarchical Analyze menu Drag Turning Circle to the X Variable slot Drag Displacement to the Y Variable slot e Click OK Assign Variables for My Scattergram Please double click or drag the desired variables into the proper slots in the template Template Variables Variable Data _Car Data order Dataset order H Variable Model Turning Circle Country
187. cs Survival Analyses Descriptive Statistics Descriptive Stats Complete Frequency Dist Continuous Frequency Dist Nominal This menu is dynamic you can rearrange it simply by rearranging the contents of the Tem plate folder Windows or StatView Templates folder Macintosh This flexibility lets you organize your StatView templates for the way you like to work The default organization of the menu reflects the way templates are installed in folders within the main folder amp X Exploring Descriptive Statistics File Edit View Tools Help All Folders 3 Statview a d Template C _For Beginners 9 ANOVA and t tests C Correlations 43 Descriptive Statistics C Factor Analysis CI Graphs C Nonparametrics C GC Analyses C Regression Survival Analyses 8 Bookshel94 E People on Bad dream F a Control Panel m li Contents of Descriptive Statistics a Descriptive Statistics a Descriptive Stats Complete a Frequency Dist Continuous Frequency Dist Nominal Nominal Desc a Percentiles 1 object s selected 9 63KB Rearrange templates b C For Beginners D Statview Templates H Name b QI ANOVAA tests b 2 Correlations 7 2 Descriptive statistics Descriptive Statistics Descriptive Stats Comp Frequency Dist Contin Frequency Dist Nominal Percentiles Factor
188. d StatView data 106 SuperANOVA data 106 Continuous data to nominal groups 118 Missing values to a specified value 120 Exercise 120 Series 121 Exercise 122 Random numbers 123 Create criteria 124 Define criteria 125 Criteria pop up menu 128 Edit Apply Criteria 129 Exercise 130 selected 133 Edit Analysis 134 Edit Display 135 Contents Multiple and compound results 135 Control recalculations 138 Analysis windows 139 View window 139 Analysis browser 141 Variable browser 143 Results browser 147 View Windows only and Window menu 148 Analyze subsets 149 Exercise 150 Create an analysis then assign variables 150 Assign variables then create an analysis 151 6 Templates lt Use templates 162 Assign variables to templates 162 Manipulate results 164 Exercise 165 Manage templates 167 7 Customizing results Preferences 179 Edit Display dialog boxes 180 Preview changes 181 Undo changes 181 Clipboard commands 181 Cut and Copy 182 Duplicate 182 Clear 182 Paste 182 Graphs 183 Select graphs 184 Select components 185 Overlay graphs 186 Resize graphs 186 Move graphs or components 186 Change text items 187 Add more results 152 Add variables to existing analyses 153 Adopt variables for new analyses 154 Split analyses by groups 155 Save a view 156 As a view 156 As a template 157 As a text file 157 Asa PICT file 1
189. d ellipses appear in place of text data Attributes and variable names in the attribute pane are similarly abbreviated Move and scroll Edit data Select data You can move the cursor from one cell to another by clicking or by using keyboard shortcuts Windows Mac Movement Tab Right with wrap around Shift Tab Left with wrap around Enter Return Down except in input row with wrap around Shift Enter Shift Return Up with wrap around Enter on numeric Enter Right or down as set in Dataset Preferences keypad Shift Enter on numeric Shift Enter Left or up as set in Dataset Preferences keypad Page Up Scroll up one page Page Down Scroll down one page Home Upper left corner of dataset End Lower right corner of dataset You can edit an individual value directly by clicking its cell and typing a new value You can also use the standard Cut Copy Paste Clear and Delete commands from the Edit menu on selected cells in a dataset If you need to see what you have selected choose Show Selection from the Edit menu to scroll the dataset to the highlighted section Datasets and views are dynamically linked so if you make any changes to your data StatView automatically recalculates graph and analysis results in the view The following tables tell you how to select cells rows and columns for editing Windows Macintosh Type of selection Click the cell a single
190. d event times and any nonzero values indicate uncensored event times Table preferences Table preferences govern the way table results are displayed by default In this context tables are results from any statistical analysis that take the form of words and numbers arranged in rows and columns as opposed to graphs which are results from any analysis that take pic ture form such as scattergrams and box plots Tables include but are not limited to contin gency tables You can change the appearance of any individual table by clicking Edit Display SS Table Preferences SS Default numbers Default number Format Free Format Fixed x Decimal places 3 pA T Always have leading digit Table format Table format Row height Format Free Format Fined Decimal places La Always have leading digit Row height X Cancel E cua CO Default numbers Choose the numeric format and number of decimal places to display in tables To include a zero before the decimal point e g 0 25 instead of 25 check Always have leading digit Table format This option lets you choose how borders are drawn between rows and columns Choose an appearance type from the pop up menu Row height Specify how many lines of text to use for each row e g choose 1 for single spac ing 2 for double spacing etc 232 9 Tips and shortcuts Preferenc
191. deleting With nonadjacent rows or columns the dataset shrinks by the number of deleted rows or columns You can Copy selected data to the Clipboard by selecting Copy from the Edit menu You can Paste data from the Clipboard back into a selected area of any active StatView dataset When you cut or copy numeric data to the Clipboard they are converted to text when you switch to a different application When data are converted from numeric to text format only the num ber of decimal places currently displayed are saved If you paste data outside of StatView be sure to display enough decimal places before copying to preserve your data values A yin yang cursor spins while data are converted to text When you return to StatView the information is in the Clipboard assuming you have not placed anything there from another application You can merge data from different datasets using Copy and Paste commands Then you can copy data from the source dataset and paste them into the empty cells of the target dataset It is important to understand StatView s rules for pasting data described next Pasting data into a dataset is much easier if you first familiarize yourself with the data you want to paste You should also know the row by column size You should consider the follow ing things 1 the size of the data relative to the size of the selected target area 2 the type of data to be pasted and the data type in the target area 3
192. e Compact variables p 84 This arrangement with or without compact variables is accepted by all analyses in StatView Other arrangements A small number of analyses in Stat View can also handle other data arrangements in addition to the usual one case per row arrangement 1 Contingency Tables also accept summary data and two way table data 2 Certain QC analyses accept summary data 3 Factor Analysis can analyze correlation matrix data Summary data Summary data show how many individuals fall in each combination of group memberships The hair color and gender example would contain two nominal grouping variables in col umns and an additional column with the count in each combination of groups cell The dataset contains six rows one for each possible combination of eye color and gender blue eyes female blue eyes male brown eyes female and so on Each combination is made up of entries in the nominal Eye Color and Gender columns The count for each combination 2 Datasets Dataset structure 53 appears in the count column Note that you cannot record information about individuals you can only count how many individuals fall in a group You are not required to have as many rows as there are combinations If duplicate combina tions appear in your data StatView sums the counts for that combination Also if a fractional value appears in a count column the value is rounded to the nearest integer
193. e a comparison operator and a value You choose these three parts from the scrolling list below the definition area The scrolling list coaches you through each step of the definition process by showing you first a list of variables then a list of comparison operators and finally either a list of the levels of a nominal variable or a value bar for a continuous variable To create a complex criterion one using AND OR or XOR click in the definition area after the expression and then select a logical conjunction from the list You may use the logical Not by editing the definition You can edit definitions by clicking in the part of the definition you want to change Again the scroll list coaches you it changes to the type of list you need to edit that part of the defini tion You can type a definition directly into the definition box if you prefer You can print a crite rion definition by choosing Print from the File menu Define criteria Defining a criterion is a four step process name the criterion select a variable choose a com parison operator and set a value or range of values Name the criterion The default name is Criteria n You should supply a more meaningful name because you select and apply criteria by name from the Criteria pop up menu and the Edit Apply Criteria Also the name of any criterion in effect appears automatically in the titles of statistical analy ses and graphs as a reminder that results are ba
194. e application preferences and category definitions It also stores the locations of the Templates and Tools folders If the Library file is discarded or misplaced StatView creates a new one with default preference settings You will have to specify preference settings again Example Views and Datasets In your StatView folder is a folder of example views and datasets that demonstrate special ways of working with StatView To try these examples open the view files in StatView Dataset Templates Also in your StatView folder is a folder of dataset templates These are datasets that use spe cial formula variables and criteria to compute specialized statistics Normality Test StatView gives you a number of useful techniques for evaluating how well continuous data conform to a normal distribution Some common techniques are 1 Plot a histogram of the variable with a normal curve see Frequency Distribution p 13 of StatView Reference 2 Plot a frequency distribution of the variable s Z scores see Frequency Distribution p 13 234 9 Tips and shortcuts Dataset Templates of StatView Reference 3 Calculate the variable s skewness and kurtosis see Descriptive Statistics p 1 of StatView Reference Unfortunately none of the preceding techniques gives an objective criterion upon which to decide whether a distribution is normal Therefore StatView provides a pair of dataset and v
195. e printer settings for page orientation You cannot set the drawing size smaller than the number of pages needed by any results cur rently present A view preference lets you limit drawing size for printing or export to other applications see View preferences p 232 Drawing Size Drawing Size Height 20 22 in Height 131 44 in Width 7 67 in Width 99 67 in Pages 2 Pages 169 Arrange objects Clean Up Items from the Layout menu lets you neatly arrange graphs tables and drawn objects in the view This command always works with all results in a view it is not necessary to select objects before using Clean Up Items e From the Layout menu select Clean Up Items Clean Up Clean up by Distance between items Vertical inches Horizontal 25 inches Ignore page breaks Align to left margin 214 8 Drawing and layout Layout tools Clean up by Choose whether to keep objects in their current order or to sort them by analy sis type Distance between items Specify how far apart to space each item The distance you specify is rounded to the nearest grid unit if the grid is on see Exercise p 215 The default distance is 0 25 inches Ignore page breaks By default Clean Up Items forces extra space between objects as needed to keep objects from being split over page breaks avoiding page breaks is preferable f
196. e Calories is bimodal the histogram has two humps This is likely to be a problem in further analyses Dotted red lines indicate page breaks we won t worry about them for now Frequency Distribution for Calories goclusion erjeria Big Three from Candy Bars Data From 23 Tott Count 160 000 189 000 E 218 000 247 000 18 247 000 305 000 334 000 1 362 000 392 000 1 392 000 421 000 421 000 450 000 Total Histogram Inclusion criteria Big Three from Candy Bars Data 20 18 te aa i4 12 5 310 o on bom 150 200 250 300 350 400 450 500 Calories Quiz If Calories are bimodal perhaps it s because some candy bars have more fat or carbohy drates than most Are these variables bimodal as well Clone the frequency distribution analysis with the other variables that are likely to be bimodal Total fat g Saturated fat g and Carbohydrate g I Tutorial Analyze data 33 A glance at the histograms doesn t show bimodality in the fat and carbohydrate variables although both Saturated fat g and Carbohydrate g seem to have small jumps at the high end More information is needed What about when we include the smaller brands Does increasing the sample size make any relationships more apparent Turn off the Big Three criteria Bring the dataset window for ward by clicking it or selecting Candy Bars from the Window menu From the criteria pop up menu at the top of th
197. e Criteria pop up menu select Men Only those rows in the dataset with values for men are included The rest have dimmed row numbers From the Criteria pop up menu select Low Lipid Males Notice how the included rows change If you were using this dataset in some analyses the tables and graphs in the view would change as you selected different criteria from the pop up menu Analyses Overview StatView provides two ways of creating statistical analyses tables and graphs 1 You can choose analyses from the analysis browser and use the variable browser to assign variables to roles in the analyses 2 You can select templates from the Analyze menu and assign variables to them Templates are preassembled sets of analysis results This chapter discusses the first method using analysis and variable browsers to construct anal yses directly We discuss the second method in the next chapter Templates p 161 That chapter also shows how to use browsers to build your own templates for the statistics and graphs you use most frequently Once you have created an analysis either with browsers or with templates you can change its parameters or its variable assignments adopt its variable assignments for another type of analysis and clone it into an analysis of different variables You can split an analysis into sepa rate analyses for each group in a nominal variable or each subgroup formed by crossing sev eral nominal
198. e In the analysis browser double click Regression Plot A scattergram of the two variables and the calculated regression line appears 5 Analyses Exercise 153 Regression Plot 300 4 1 1 1 1 1 1 1 275 4 2 F 250 _ 225 200 9175 6 150 7 125 100 75 50 T T T T T T 50 100 150 200 250 300 350 400 Displacement Y 50 444 504 X R 2 584 Because you had regression results selected StatView already knew the appropriate analysis parameters and variable assignments you wanted If you double click Regression Plot with no results selected you have to set analysis parameters again and assign variables again Some analysts like to see every possible result for an analysis all at once for example regres sion and anova tables plots interaction plots residuals and confidence intervals Others pre fer to look at just a few key results and then decide whether the analysis deserves further examination they might try several different variable combinations before finding a satisfac tory model and only then will they start looking at residual statistics StatView accommodates either style of work Start by choosing as many or as few results you want and then add more results when you want them By selecting one of the results you save yourself the trouble of rebuilding the model Add variables to existing analyses Many analyses can incorporate multiple variables in a single analysis For example if you select
199. e Select the application StatView e Select the document type svpx e Click OK e Close the control panel Close datasets Close a dataset by clicking the close box in the upper left corner of the dataset or by selecting Close from the File menu StatView prompts you to save any changes made since last saving the dataset If you attempt to close a dataset whose variables are used in an open view you will be alerted to this fact If you continue and close the dataset without closing the view first all of the vari ables from the dataset will be removed from the view We advise you to close views that use the variables of a dataset before closing the dataset In addition if you close an Untitled dataset one not yet saved or a text file that has not been saved as a StatView document and do not save it StatView will not able to match the dataset with any view s that use the variables of that dataset Open datasets To open an existing dataset select Open from the File menu You can choose which types of files are shown in the dialog box from the pop up menu at the bottom of the dialog box Supported Files Windows or All Available Macintosh displays files of any format that StatView can read or import Other choices are any types of files that can be imported on the current platform The various dataset types are discussed under Save datasets p 70 See the chapter Importing and exporting p 99 for more information o
200. e a variable Copy or Cut the data insert an empty column elsewhere in the dataset see Insert columns p 62 and Paste the data into the empty column To create a recode variable in another location insert a column and change its source to Static Formula or Dynamic Formula 120 4 Managing data Recode data Missing values to a specified value Exercise Recoding missing values to a specific value works by building an if then else formula You must supply a value in place of its question mark placeholder The value you specify can be a number a string or some other function such as the mean of the variable You can choose any function from the browser at the left For more information about each function refer to the chapter Formulas p 315 of StatView Reference or examine the Hints window when you click a function the Hints window describes the function and its parame ters briefly O Recode of Crime Recoded variable definition Order by Function Ty pe Pe Mathematical Normf AlIRows par Statistical Coeff0fVariation ATR Count AlRows GeometricMean AliRo HarmonicMeant ARo MADL AllRows Maximum AlIRows Mean AlRows Median AllRows Minimumt AllRows Mode AlIRows NurnberMissing ATiRo Number OfRiows Percentilel 7 Allows Rangel Allows StandardDevistiont Al if IsMissing Crime then
201. e cumulative survival plot sr187 baseline estimates Kaplan Meier sr175 baseline hazard sri69 baseline In cumulative hazard plot sr189 baseline survival table sr188 Basic Statistics Us142 batch mode see templates beep for error messages US225 bell shaped curve sR1z Bernoulli distribution sr4or beta distribution sr406 beta see type II error between subjects sr83 Bezier curves see spline tool bimodal distribution us32 binary logistic regression SRI99 binary operators see operators binomial distribution SR287 SR401 SR406 BinomialCoeffs sr358 bivariate plots sR221 sR236 axis types USI9O confidence intervals us242 sR221 SR224 SR228 SR232 correlation SR3I SR48 cubic spline SR221 SR225 SR227 SR228 SR233 data requirements SR229 dialog box sr228 discussion SR221 SR228 error bars us242 Index SR StatView Reference US Using Statview exercises SR35 SR230 SR236 fitted lines SR221 SR228 interaction plots SR99 lowess SR221 SR225 SR226 SR235 multiple variables sr222 nominal data sR232 results sR229 split by variables sr222 strategy SR222 SR223 supersmoother R221 R225 SR227 R236 templates sRr230 exercise USI7I black selection handles us21 us184 usI86 USI98 USI99 US204 US207 US209 US2I0 US214 blocking factor SRIOI SRIO5 Bonferroni Dunn sr86 sR87 SRIO4 Boolean operators Us127 Boolean variables sr338 borders dialog box us201 tables us199 US201
202. e disadvantages 1 You might not have an exhaustive list of all the group levels you ll need in advance in this case a string variable might be more convenient 2 Category definitions are limited to 255 groups If your variable has more groups than that youll need to use another type 3 Variables with type category can only have class nominal or informative If you might also want to use the variable as a continuous variable you should use another type 2 Datasets Categories 81 You must use categories in two cases 1 When you recode a continuous variable to nominal using cutpoints to group ranges of values together you must define a category or choose an existing category See Continu ous data to nominal groups p 118 2 When you create compact variables you must use a category to define its groups See Compact variables p 84 Finally whenever possible you should re use the same category definition with variables that share the same group names this reduces storage space and eliminates naming conflicts Using the same labels in two separate category definitions can lead to problems with formu las criteria and other data manipulations Create category definitions Whenever a category is needed for a category variable compact variable or when recoding continuous data to nominal groups you see a dialog box Choose Category Please choose the column s category Hosion Housing Baia
203. e front Select Print from the File menu to print a formula definition Windows only A Random Numbers dialog box is listed in the Window menu where you can select it to bring it to the front You can create any number of rows and columns of random numbers from the same distribu tion By default Random Numbers creates one variable with the number of rows currently present in the dataset If you specify a larger number of rows rows are added to all columns and filled with missing value symbols in the other columns unless they are formula vari ables whose definition specifies otherwise 124 4 Managing data Create criteria Use Attributes if you want to specify variable names and attributes before clicking Create You can adjust attributes afterward in the dataset window if you prefer See Variable attributes p73 For more information about each distribution refer to the chapter Formulas p 315 of Stat View Reference or examine the Hints window when you click a distribution the Hints win dow describes the function and its parameters briefly Use the variable s Source pop up menu to view and edit a formula definition or to change from static formula to dynamic formula or user entered See Change sources p 77 New variables are appended at the right side of the dataset To move a variable Copy or Cut the data insert an empty column elsewhere in the dataset see Insert columns p 62
204. e infor mation provided by survival regression models in the Coefficient Covariances Table you can evaluate such hypotheses In their most general form such contrasts can be computed as follows If the linear hypothe ses for the regression coefficients B a vector are expressed in the form Hy HB c where H is a matrix of rank r of weights for the hypotheses and c is a vector of constants then the Wald chi square statistic with r degrees of freedom is given by x HB cl THV B H ILHB c For the less general case of testing for the equality of two coefficients B and B H is a vector with all elements equal to 0 except H 1 and H 1 and 0 The chi square statis tic in this case has 1 degree of freedom and is given by g2 BB Vii V2 2V2 This equation is applied as follows Suppose that you compute a regression model that gives results such as these in StatView 9 Tips and shortcuts Common questions 249 Model Coefficients for Time days Censor Variable Censor var Model Proportional Hazards DF Coef Std Error Coef SE Chi Square P Value Exp Coef Treatment 3 X 4 549512 2079 Treatment 1 522397 409513 1 275654 1 627294 2021 593097 019237 347218 055402 003069 9558 1 019423 298397 339070 880045 774479 3788 1 347697 Treatment 2 Treatment 3 Coefficient Covariances for Time days Censor Variable Censor var M
205. e levels sr360 combine strings sR361 command syntax see syntax commas R324 comment R329 common intercepts test SRBI common problems see troubleshoot common questions Us237 US250 dataset US237 US240 formulas us240 uUs243 QC analysis us244 us245 survival analysis Us245 Us250 common slopes test sR81 communality summary sRI40 Compact us57 Us88 Us90 compact variables uss7 us81 US84 US97 USHI USII7 USI43 USI63 USI164 advantages us85 analyses usS95 US97 example US95 SR9I SR93 SR97 SR99 SRIOI SRIOS SRIO7 categories US9I compact US57 create US86 US245 complex example us89 us93 simple example us87 us89 disadvantages us85s us86 expand US57 US94 US95 QC analyses us244 Us245 repeated measures analysis of variance US85 SR9I triangle controls us88 Us95 compare distributions box plots sr243 compare percentile plots sr247 data requirements SR247 dialog box sr247 discussion sR247 exercise SR248 results sR248 templates sr248 comparison operators USI26 complete sr147 complex criteria USI25 SR409 compound vs multiple results USI35 USI37 USI70 US229 Concat sR361 conditional transformation SR341 confidence intervals us241 us242 bivariate plots SR221 R224 SR228 SR232 chi square test SR24 interaction plots sr9o logistic regression SR205 SR209 SR214 SR215 mean difference sR30 SR37 SR38 SR44 one sample t test SR23 proportional hazards models sr1zo survi
206. e same Let s just look at Total fat g for now e Click somewhere in the white space of the view window to be sure no results are selected e In the analysis browser double click Unpaired Comparisons e Click OK to accept the default parameters The default options produce an unpaired t test with a null hypothesis difference of 0 n n Unpaired Comparisons Unpaired Comparisons Mean difference EJ Unpaired t test Hypothesized difference 95 confidence interval Mean differenc V Unpaired t test Hypothesized difference E o5 confidence interval Tail Both Tail Variance ratio Variance ratio I F test F test Hypothesized ratio fi Hypothesized ratio 1 B js S conidenceinteval 95 confidence interval Tail fe otk 7 The note below the empty analysis objects says we need to add both a nominal and a continu ous variable o complete this analysis assign a continuous variable and a nominal variable using the variable browser Add Button I Tutorial Analyze data 37 In the variable browser Control click Windows or Command click Macintosh Total fat g and Calorie groups e Click Add Unpaired t test for Total fat g Grouping Yariable Calorie groups Hypothesized Difference 0 jnclusion criteria Bjg Three from Candy Pars Data Mean Diff DF t Yalue P value E howe High 13 178 4
207. e second mouse button If you are right handed the first mouse button is the left button and the second mouse button is the right button We describe clicking with the second button as Right click VII Overview Keyboard and mouse chords Contents Tutorial Data analysis the StatView way 1 Why should I bother with a tutorial 1 Manage data 2 Our sample data 2 Enter data by hand 3 Import data 10 Open a dataset 13 Analyze data 14 Sort data 14 Examine summary statistics 14 Edit data 15 Compute formulas 17 Build an analysis 20 Remove variables 22 Edit a display 23 Edit analysis parameters 24 Split by groups 25 Clone an analysis with different 2 Datasets Dataset structure 49 Example 49 Data class 50 Data arrangement 51 Other arrangements 52 Columns vs variables 53 Dataset windows 54 Dataset preferences 55 Variable browser 56 Enter data 57 Name variables 58 variables 26 Use Criteria to examine a subset 28 Adopt variable assignments for a new analysis 29 Create an analysis with several parts 31 Create a new grouping variable 33 Grouped box plots 35 Create an unpaired test 36 Create an ANOVA using a template 38 Save your work 40 Present results 43 Clean up results 43 Add some color 43 Print a presentation 47 Save a presentation 47 Save atemplate 47 Notes 48 Set attributes 58 View summary statistics 5
208. e that the Maximum is 125 That would be a big candy bar Also notice that the Mean candy bar is 3 84 oz but the standard deviation is 14 We know that if candy bar sizes are normally distributed most candy bars should fall within two stan 16 I Tutorial Analyze data dard deviations of the mean That would mean some candy bars have negative weight Even if sizes arent normally distributed these statistics would vot seem likely Either discovery would tell us to look for an error e Click the 125 cell e Change it back to 1 25 e Press Enter or Return Servings pkg Oz pkg b Type Real Real a Source User Entered User En Class b Format Free Format d Dec Places Mean Std Deviation Std Error f Variance These statistics make more sense Quiz Try to answer these questions by examining your summary statistics pane You may need to scroll through the dataset to answer some questions You also might want to resize the data window to be taller and wider What s the smallest number of calories per serving you can find in a candy bar Look at Minimum for Calories The least value is 125 How much does the per serving total fat vary from candy bar to candy bar What s the aver age See Range or Minimum and Maximum for Total fat g The candy bars range from 0g to 29g per serving They average just below the halfway mark at 11 9g and most should fall within two stand
209. e tools This diagram identifies major table components Title Column labels Borders Correlation Matrix Weight Cholesterol Triglycerides HDL LDL Weight 1 000 022 108 276 057 Cholesterol 022 1 000 401 352 962 Row labels 0 Sehe 108 401 1 000 278 489 HDL 276 352 278 1 000 083 Note LDL 057 962 489 083 1 000 Select tables With a whole table selected you can 95 observations were used in this computation 1 Change thickness pen pattern and color of all lines in the table 2 Change the color of text for the whole table 3 Change the font size style alignment and angle for all text components 4 Use Edit Display to change the graph as a whole If the cursor isn t an arrow first choose the selection tool from the Draw palette 198 7 Customizing results Tables To select an entire table click directly on the borders or interior of a table Do not click a title or note To select several tables Shift click them or click and drag a marquee around them Descriptive Statistics Descriptive Statistics E Weight Cholesterol Hean 158 653 Mean 191 232 Std Dev 28 389 Std Dey 35 674 Std Error 2 913 Std Error 3 660 Count 35 Count 95 tinimu 107 000 tinimu 115 000 Maximum 234 000 Maximum 285 000 Missing o Miszing D You can also use the Results Browser to select objects
210. e window select No Criteria Still no apparent bimodality except in Calories Reselect the Big Three criterion to continue the analysis No Critenja Criteria Create a new grouping variable One way to cope with this problem in further study is to divide the candy bars into two groups high calorie and low calorie StatView s Recode feature lets you do this quickly e From the Manage menu select Recode From the scroll list on the left select Calories e Click Continuous values to nominal groups Recoding Data Select a variable Choose desired recoding Brand Serving pkg Continuous values to nominal groups Oz pkg K Calories Total fat g Missing values to a specified value Saturated fat g Cholesterol g Since we re recoding a continuous variable Calories into a grouping variable we need to choose a category for the groups The only category defined thus far in the dataset is for Brand which wouldn t work Therefore we need to define a new category with values low and high e Click New Choose Category Please choose the column s category Landy Bare Pati Gta Brand names ants lt r OK e Specify a name for the category Category for Calorie Groups 34 I Tutorial Analyze data e Specify the first Group label Low e Click Add e Specify the second Group label High e Click Add e Cl
211. each value you enter in the column must be one of that category s defined group labels This prevents data entry mistakes Also you can use shortcuts to enter group labels quickly For example you might create a category definition called Color that has the group labels Red Yellow Green and Gray in that order to use for any variables that record color groups You cannot enter any other values such as Redd Blue or 4 3 in those variables You can enter Red by typing R r or 1 You can enter Yellow by typing Y y or 2 You can enter Green by typing GRE gre or 3 You can enter Gray by typing GRA gra or 4 Since Green and Gray both begin with GR it is necessary to type a third letter You are not required to use the category type for nominal data Remember variables of any type real integer long integer string date time currency or category can have class nomi nal However categories offer several advantages 1 You can enter data faster Type the first letter or two of a group name and StatView fin ishes the name for you Or type the number of the group 1 for the first group label in the definition 2 for the second group etc and StatView fills in the label 2 You can prevent data entry errors StatView wont let you enter a value that isn t defined in the category 3 You can save memory and disk space You might not want to use the category type for all your nominal class variables Categories have thre
212. eate USI69 USI74 exercise USI7I USI77 dataset Us240 exercises USI65 USI67 USI7I USI75 formats USIZO US233 generic variable names us170 graph formats usI71 manage USI67 USI69 modify Us165 USI7I exercise USI75 open USI57 pre assigned variables usrz7o rearrange USI68 USI69 repeat analyses us162 save USI57 save views USI57 tips UsI69 variable slots us163 vs views USIGI USI62 temporary files us233 tension SR226 SR228 test differences among covariate levels us248 test normality Us86 US233 SRI3 SR416 tests for special causes sR259 SR260 c u analyses SR260 R300 false signal sr259 I analyses sr260 individual measurements sR278 individual measurements analyses sR260 p np analyses SR260 SR289 SR290 subgroup measurement analyses sr262 subgroup measurements analysis sR260 text attributes us205 colors USI85 USI97 US202 US212 edit graph text us187 edit table text us200 import export USIOO USIO2 tutorial example us12 us13 resize US204 rotate US205 286 Index SR StatView Reference US Using Statview Save As USIO5 views and templates us233 see text tool text editor UsIoo text functions sR33I SR332 ChooseArg sR359 Concat sR361 Find sp379 Len sR383 Substring sr422 Text menu us185 USI99 text tool Us200 US204 US205 tutorial example us45 thickness see line widths tick marks us183 USI85 USI9I USI93 stagger USI92 ties SRI24 Time sri84 sR427 time func
213. ee decimal places which matches the current attribute setting for decimal places This setting only affects the way numbers are dis played StatView stores values exactly as you specify them and carries the fullest precision sup ported by the hardware platform through calculations and analyses e Enter the values for Oz pkg 2 1 55 5 3 7 1 7 Enter the values for Calories 310 230 220 170 200 e Enter the values for Total fat g 20 12 12 8 2 5 e Enter the values for Saturated fat g 7 6 8 3 2 5 Now take a moment to check your work If any values are wrong just click the cell type a new value and press Enter or Return Serving pkg Oz pkg Total fat g Saturated fat g Real Rea Real Real User Entered User Entered User Entered User Entered User Entered Continuous Free Format Fi Free Format Fi Free Format Fi Free Format Fi Free Format Fi B ds le J 1 000 2 000 310 000 20 000 7 000 1550 230 000 12 000 6 000 i 170 000 8 000 3 000 1 000 1 700 200 000 2 500 2 500 An easy way to check for data entry mistakes is to view all rows of the attribute pane and look at the summary statistics e Click and drag the attribute pane control downward to expose the twelve rows of summary statistics for each variable I Tutorial Manage data 9 Name Serving pkg Oz pkg Calories Total fat g a Type String Real Real Real b Source User Entered User Entered User Ent
214. eference US Using Statview 267 formulas us240 table text us200 Edit Analysis us134 US135 USI40 USI42 US233 shortcut USI4I tutorial example us24 us2 Edit Categories us83 us84 Edit Display us135 US140 US141 Us180 US182 US185 USI87 USI95 USI97 USI99 US201 US233 dialog boxes us180 us181 Edit Apply Criteria us124 USI29 SR317 effects sR73 eigenvalues sR132 table sr139 ElementOf vus28 us126 sR345 ellipse tool us205 ellipses us64 USII3 empty cell see missing values empty graphs us138 Us164 empty tables us138 UsI51 USI64 engineering format us79 enhanced free fixed format us79 enter data US57 US64 tutorial example us3 US6 USIO values us6I US62 equal sR340 equamax SRI35 Erf sr376 error bars Us24I US242 cell plots sr237 interaction plots SR90 error function R376 error messages US222 US225 US251 beep us225 Formula us16 error of intercept sR56 error free analyses us162 Euclidean norm sR392 evaluation sR326 sR328 event time variable us243 SRI47 SRI49 discrete vs continuous sR148 nonparametric analyses sRI57 pattern plot sri62 regression model survival plots sR175 regression models sRi82 SR184 survival regression models sr184 Example Views and Datasets us233 examples USVI SR316 SR317 Excel import export Us99 USI00 US254 tutorial example usi1 us12 excess risk SR201 Exclude Row us1o8 usio analyses USI49 USI50 compare
215. efined in any open datasets The Chosen scroll list on the right shows the categories that are currently selected to identify this compact variable none yet Variables selected below the Chosen shows the number of variables being compacted Cells in compact below the Categories list shows how many group labels are defined for any category you select This is important Recall that each column represents one group Therefore we need to choose a category with as many group labels Or we need to choose several categories whose group labels combine to make as many subgroups The product of the numbers of cells must equal the number of variables selected Our example has four columns and they represent four subgroups pro duced by combining gender and smoking 2 groups X 2 groups 4 subgroups The buttons in the middle create categories and move them from one list to another as fol lows Control Description Select Adds the selected category from the Categories list to the Chosen list Remove Removes selected categories from the Chosen scrolling list If you select a category by mistake click Remove to remove it from the definition of the compact variable New Lets you create a new category which is added to both scrolling lists Use New to build a compact variable when categories describing its structure do not already exist For more details see Create category definitions p 81 Control cl
216. efinition you need to replace the with the desired argument For example if you click the Log button or double click Log in the function browser Log appears and is selected highlighted in the definition text box Replace a selected with the desired argument by typing it or selecting it from the keypad or a browser To select the next for replacement press Tab To select the previous for replacement press Shift Tab Arguments to a function can be constants variables or expressions For example the function Sum Weight Ln Age 10 adds for each row the value of the Weight variable the natural log of the Age value and the number 10 Some functions contain default arguments which you can change such as LinearSeries 1 1 which accepts two arguments When you enter it into a formula it appears with a 1 for each of its arguments You can change these default argu ments to any values you want Arguments are discussed in detail in Arguments p 323 of StatView Reference Many functions can take a varying number of arguments Functions of this type contain ellipses to indicate that they allow any number of arguments After supplying as many arguments as you need you must remove the trailing ellipsis To learn what is expected for each argument of a function open the Hints window and select the function For a complete discussion of each function and its arguments and some general rul
217. elect Delete from the Edit menu Insert rows Inserting a row is similar to inserting a column e Control click Windows or Command click Macintosh the border between two row numbers Position the cursor between two row numbers over the horizontal line separating the rows Hold the Control Windows or Command Macintosh key down the cursor changes to a double arrow f shape Click and release the mouse button Release the key e In the dialog box specify how many rows to add by typing a number Number of rows to insert _ Number of rows to insert e Click Insert Rows are added between the two rows where you clicked Cells of the input rows contain missing values until you enter data Repeat to insert additional columns Resize column widths To increase or decrease the size of columns e Click and drag the border between variable names Position the cursor to the right of a variable name on the vertical line separating columns the cursor changes to a cross arrow shape Click and drag to the left or right to make the column narrower or wider To resize several columns at once select the columns and resize one of them Shift click or click and drag to select several adjacent columns Control click Windows or Command click Macintosh to select several nonadjacent columns 64 2 Datasets Edit data If a column is too narrow to display its data pound signs appear in place of numbers an
218. em within each brand e From the Manage menu select Sort Select Brand and click Make Key Double click Name The up arrows 4 next to each sort key indicate ascending sort least to greatest numerical sorting alphabetical text sorting If you preferred descending sort you could click the arrow to change it to a down arrow 4 Sort Select the keys for this sort Variables Sort Keys Serving pkg Brand Oz pkg t Name Calories Total fat g Saturated fat g Bempt Key Cholesterol g Sodium mg Carbohydrate g Cancel e Click OK Quiz Quizzes are optional If youre in a hurry skip ahead to the next section Now scroll through the dataset and get a feel for the data See if you can answer some questions just by looking at the data Which candy bar manufacturers make the most candy bars Notice that Hershey Nestle and M amp M Mars appear in big clumps This wasn t obvious before we sorted Which candy bars have been popular enough to spawn sequels Snickers Reese s Peanut Butter Cups Milky Way and others have several varieties Before sorting the names we couldn t see these easily Examine summary statistics In most data analysis packages if you want basic descriptive statistics means standard devia tions and so forth you need to type some commands If your data change you need to start I Tutorial Analyze data 15 Edit data
219. ement and Horse power will play e In the analysis browser click Create Analysis e Accept the default parameter settings click OK The default regression output appears in the view Regression Summary Horsepower vs Displacement Count 116 Num Missing 0 R 764 R Squared 584 Adjusted R Squared 580 RMS Residual 25 796 ANOVA Table Horsepower vs Displacement DF Sum of Squares Mean Square F Value P Value Regression 1 106510 675 106510 675 160 061 lt 0001 Residual 114 75859 765 665 437 Total 115 182370 440 Regression Coefficients Horsepower vs Displacement Coefficient Std Error Std Coeff t Value P Value Intercept 50 444 6 744 50 444 7 480 lt 0001 Displacement 504 040 764 12 652 lt 0001 Add more results You generated only the default results for a simple regression Regression has a triangle next to it in the analysis browser indicating that several tables and graphs are available as output In this exercise you add additional output from the regression analysis to the default results e Click the triangle next to Regression Now you can see all the output available for a regression analysis There are nine results in the indented list We also want to see a regression plot for the analysis e Make sure at least one of the regression results is still selected has black handles if not click one to select it
220. er Cholesterol readings Gender K Smoking status fH Data Complex Compact Variable Order Dataset order pitolestero readings i The nominal variables have the names we gave to their category definitions and they appear in the same order as we chose them Again Cholesterol readings and its nominal components all look like regular variables Cholesterol readings has a E class marker for continuous and Gender and Smoking status have r class markers for nominal Again let s save the dataset From the File menu select Save e Specify a filename Complex Compact Variable e Click Save 94 2 Datasets Compact vanables Expand compact variables Sometimes you might want to unpack compact variables back into simple regular columns Removing the compact structure is easy Simple compact variable For a simple compact variable all you do is select the compact variable and click the Expand button You can do this either in the dataset window or the variable browser From the Window menu select Simple Compact Variable If you closed the dataset use File Open to reopen it e Select Cholesterol readings click its name in either the dataset or the variable browser e Click the Expand button in either the dataset or the variable browser Criteria Lu Variable Show Compact Data Simple Compact Yariable Order Dataset order
221. er s PLSD sR86 Games Howell sr88 interaction effects sR89 purposes sR84 repeated measures sR88 Scheff s F sr86 sRI04 Student Newman Keuls sr88 Tukey Kramer sr87 type I errors SR85 pound signs us64 power SR74 SR76 SR9O SR334 power regression SR55 precision US73 predicted values sRGO SRGI SR66 preferences US225 US233 application us225 US226 color palette us226 us227 dataset Us227 formula us228 graph us228 us229 graphs us179 us180 hints us222 us230 Survival Analysis us230 Us231 table us179 usI80 US231 US250 278 Index SR StatView Reference US Using Statview view USI4I USI7I USI79 USI80 US232 US233 presentation USI6I1 USI69 tutorial example us43 us47 prevent changes see Lock prevent errors USI62 prevent recalculation us138 us139 USI7O preview format changes us181 previous versions data UsIo6 primary pattern solution sri40 principal components analysis sR132 principal values see arc functions print Criteria definitions usi25 dataset us72 Formula definitions ust10 line widths usts9 us232 presentation tutorial example us47 Random Numbers definitions us123 Recode definitions ust19 us120 Series definitions us122 troubleshoot us254 views USI59 USI60 prior probability sr211 probabilities functions ProbBinomial sr4or ProbChiSquare sr402 ProbF sr4o02 ProbNormal sr403 Probt sr404 ReturnChiSquare sR414 ReturnF srqis ReturnNormal sr415
222. ered a Class Informative Continuous Continuous Continuous Continuous a Format e Free Format Fi Free Format Fi b Dec Places Mean 1 900 2 790 236 000 10 900 Std Deviation 1 245 1 505 Std Error 55 E73 23 367 2565 i Yariance 1550 2 266 27zo 000 4ioso sd 5d Coeff of Variation EF 539 231 588 i Minimum 1 000 1 550 170 000 500 Ts un LEESE SLEE in a Maximum fe 3 500 5 000 Range e 2 500 3 450 140 000 17 500 Countife SS Sum 13 950 1130 000 54 500 Sum of Squares 2 758 250 1 Snickers Pean 1 000 2 000 210 000 20 000 Fi Cankias n Mint 1 000 1550 oan ann 17 0NA If any of those statistics seem wrong look for an error in the column With large datasets this trick can be a big time saver Results can differ slightly among different platforms due to dif ferences in numerics handling For example on some systems the Sum of Squares for Oz pkg is 47 982 on others it is 47 983 We ll examine this summary pane later in the tutorial For now let s close the summary pane e Double click the pane control to hide the summary statistics Now let s make a few aesthetic adjustments First notice that all the Calories values are whole numbers We can save memory by storing this variable with type integer From the Type pop up menu select Integer Also notice that the other variables have only one significant decimal place It will be easier to view these numbers
223. eries sR378 file formats US70 US72 US99 USI56 USI57 file size USI70 US232 filename us99 fill patterns color us197 colors Us202 US212 graph interior us185 graphs us196 us229 shapes vUs2I1 final communality estimate sR140 find and replace us241 Fisher s exact test SRII3 Fisher s PLSD sR86 SRIO3 tutorial example us39 Fisher s 7 to z transformation R32 SR44 fitted lines see bivariate plots fitted values sRGO SR6I date time data sR330 SR331 graphs Us183 Us197 multiple columns tutorial example Usg numeric data R332 tables ust97 Us202 US231 templates USI70 US233 also see data format Formula USIO9 USII6 SR318 R330 SR338 SR339 analysis generated variables sr61 build definition us113 common questions Us240 US243 compute USII3 date time data us74 dialog box us1o9 dynamic links ustog us116 edit usi13 Us240 errors USII6 examples usi14 USII5 hints us222 import from SuperANOVA vus256 missing values us255 preferences us228 print definitions ust1o shortcuts USII5 troubleshoot us116 Us254 US256 tutorial example us17 us20 variable attributes us110 windows at Open us254 US256 fix page breaks tutorial example us43 fix page breaks see Clean Up Items fixed places us79 flip see transpose free format fixed us79 Floor sr380 free form curves see spline tool fonts frequency distribution us233 SRI3 fractional values sR359 R374 SR380 SR390 SR413 SR416 SR429
224. ervation for one reason or another These 62 2 Datasets Fnter data cases are missing values A missing value is exactly that it s a data point that has no value Cases with missing values on a particular variable are excluded from most analyses of the vari able Do not enter 0 zero or any other number for missing values You can enter a missing value by simply leaving the cell blank Missing values are displayed as periods If you want you can enter missing values in numeric variables by typing a period For string data simply leave an empty cell If you need to recode missing values to some specific value perhaps for exporting to another application use Recode from the Manage menu see the chapter Formulas p 315 of Stat View Reference Manipulate columns and rows You create one column at a time by entering a value in the input cell pressing Tab and repeat ing the process You create a new row by entering a value in the input row pressing Return and repeating the process Additional commands let you create several columns at one time as well as insert columns and rows between existing rows and columns Add multiple columns You can add several columns at once using the Add Multiple Columns command in the Man age menu This is especially convenient for adding several variables with the same attributes To add several columns e Choose Add Multiple Columns from the Manage menu In the dialog
225. erwise have to repeat Important All results know which kind of analysis produced them which parameters were set and which variables were involved If any result is selected when you create another result it passes that knowledge to the new result To prevent new results from learning such things from old results i e to create completely different analyses you must be sure that existing results are not selected do not have black selection handles Click in the empty space of a view to deselect objects Determine whether results are selected Because information from a selected result passes to new results it is always important to know whether any results are selected You can find out three ways 1 Look at the objects If they have black selection handles they are selected If they do not they are not selected Box Plot Box Plot 4500 300 7 7 275 4 j r 4000 4 250 4 L J 225 4 a 5500 200 4 E 000 4 E175 4504 H 2500 5 125 4 L H 100 4 H 2000 4 a 759 dl F 1500 50 weight Horsepower Not selected Selected 2 Look at the Results Selected note in the upper right corner of the view 1 Result Selected If no results are selected this area is blank 3 Look at the Results browser and choose Show Selection To open the Results browser select Results Browser from the View menu Windows or select Results from the Window menu Macintosh 134 5 Analyses Overview stow
226. es View preferences View preferences govern the way view windows behave print and save Your choices take effect with the next new view window E SS View Preferences Default text Font vial hd Font Geneva Size fp x Size El I Print and copy lines at 1 4 width E Print and copy lines at 1 4 width Default text Limit document size to MacDraw II T Limit document size for printing EJ Save analysis results with view IV Save analysis results with view Templates appear using view text defaults Copy tables graphs as both TEXT and PICT I Templates appear using view text defaults I Copy tables graphs as both text and Metafiles Double click on table graph same as Double click an table graph aa as Edit Display C Edit Analysis Edit Display Edit Analysis Cancel Font and Size Choose the font and size you want the view window to use You can type a specific size These settings affect any results you generate with the analysis browser including table and graph titles numbers axis labels and legends You can override these settings for any individ ual table or graph element with the Text menu Your choice also controls the defaults used for the Draw palette s text tool Print and copy lines at 1 4 width Check this option on to reduce all single line widths to approximately one quarter of their specified size when they are printed with mo
227. es look up the function in the chapter Formulas p 315 of StatView Reference Compute Save Cancel Click the Compute button to create a new variable defined by the formula Click Cancel press Escape or press Control Period Windows or Command Period Macintosh to cancel a formula and close the dialog box If you are editing an existing formula variable Cancel reverts to the original definition and Compute recomputes the variable 114 4 Managing data Formula New variables are appended at the right side of the dataset To move a variable Copy or Cut the data insert an empty column elsewhere in the dataset see Insert columns p 62 and Paste the data into the empty column To create a formula variable in another location insert a column and change its source to Static Formula or Dynamic Formula To edit an existing formula select Dynamic Formula from the variable s Source pop up menu Some examples Exercise The following are all valid formula expressions 3 1 2 2 10 3 4 6 442 log 100 The expression 3 4 6 means 3 times the quantity 4 plus 6 and it returns 30 The expres sion log 100 means evaluate the base 10 logarithm of the value 100 and it returns 2 00 A formula can also operate on variables in the same dataset Age 2 Log Weight A formula of this type is evaluated on a row by row basis For example the expression Age 2 is first evaluated for ro
228. es represent various groups in the dataset If you set a variable to class continuous Stat View understands that its values represents measurements that could fall anywhere along an interval Because StatView understands these things in a properly constructed dataset it can make sure that you use variables appropriately in analyses 2 Datasets Dataset structure SI In other words your setting the correct class for each variable enables StatView to help you analyze your data For each analysis StatView coaches you to assign variables of the appropri ate class and prevents inappropriate assignments Also StatView can break down any analysis by groups Suppose you are studying heart dis ease but you want to study men and women separately If you have a nominal gender vari able you can split all your analyses you ll get separate results for men and for women Data types One of StatView s most powerful features is that it allows you to set variables of any type to be nominal For example gender can be text with values such as Male and Female or it could be numeric with values such as 0 and 1 Similarly informative data can be any type numeric or text However continuous data can only be numeric real currency integer long integer or date time To look at this the relationship of type and class the other way around any numeric data can be continuous nominal or informative However text data type string can be either
229. ew works so we re going to use a simple fun dataset something you won t have to think about too much Dont take it too seriously were not trying to get a grant cure cancer or influence public opinion That s what you do We just try to help by providing simple powerful software Manage data Our sample In any data analysis project the first thing you have to do is collect data We ll introduce the sample dataset you ll be using in this tutorial Then you need to get it into StatView Some common ways to do this 1 Enter the data by hand into StatView s data window 2 Import the data from a text file or another application such as Excel 3 Open a StatView dataset that somebody else created We ll step you through each possibility data Background Since 1994 the United States Food and Drug Administration FDA has required uniform easy to read nutrition labeling for nearly all foods The purpose of the new label is to reduce confusion and help consumers choose more healthful diets Unlike prior labeling laws the reform requires that even such items as candy bars carry full nutrition labels and also requires that nutritional facts per realistic serving must be reported including total calories total fat saturated fat cholesterol sodium total carbohydrate dietary fiber sugars protein vitamins A and C calcium and iron The United States Department of Agriculture uspa and the Department of Health and Huma
230. eyboard modifiers text line Hold the Shift key while drawing to constrain to a 45 or 90 degree angle arc Hold the Shift key while drawing to constrain to an arc from a circle rectangle Hold the Shift key while drawing to constrain to a square rounded rectangle Hold the Shift key while drawing to constrain to a rounded square ellipse Hold the Shift key while drawing to constrain to a circle polygon Hold the Alt key Windows or Option key Macintosh while drawing to preview a closed polygon spline Hold the Alt key Windows or Option key Macintosh while drawing to preview e R lo lo ay LA LM e a closed spline To draw a shape other than a polygon or a free form curve e Select the tool for that shape e Click and drag until the shape is the desired size The cursor changes back to the selection tool as soon as you release the mouse To return to the previous tool hold the Control key Windows or Command key Macintosh A corner center control lets you draw some shapes beginning either in the shape s center or at its corner To choose the starting point of a rectangle rounded rectangle ellipse arc or line 206 8 Drawing and layout Draw tools click the control to toggle it between the two different states draws and reshapes objects from center to corner and 2 draws and reshapes objects from corner to corner You can resize and reshape most shapes by selecti
231. f E g E k 2 Iron BROIL fe Fiber rule 5 Boo Bs Vee A gE Bars per day wea Composite my E rE 5 E a 5 Total fat rule t H Calorie gro H D 2 a 5 E One thing we notice right away is that the large range of Sodium mg values makes the vertical scale too large for the other variables which are squashed together in the lower half of the graph Since sodium content probably doesn t contribute significantly to calorie content sodium may be bad for people with high blood pressure but it s not fattening Let s remove it from the analysis e Make sure the analysis is still selected has black handles if not click it e In the variable browser select Sodium mg and click Remove 36 I Tutorial Analyze data Box Plot Split By Calorie groups Inclusion criteria Big Three from Candy Bars Data 60 34 3 S E 40 7 M 20 Units High p E Low Total fat g Saturated fat g Cholesterol g Carbohydrate g Dietary fiber g Sugars g Protein g Now the most likely culprits are easier to examine And sure enough the fat content both total and saturated carbohydrates sugars and protein all seem to be greater for the high than for the low calorie candy bars Create an unpaired test A statistical test for this conclusion is an unpaired test A test tests the null hypothesis that the means of two groups are the same and a significant p value say less than 0 05 means they are not th
232. fat g Regression ANOYA Table for Total fat g Saturated fat Regression ANDYA Table for Saturated fat g Calories Descriptive Statistics Results for Saturated fat g Bran tox Plot Plot for Saturated fat g Brand Correlation Covariance Correlation Matrix for Total f Regression Summary Table for Total fat g Saturated f ANDY A Means Table for Brand Saturated fat g Regression Coefficients for Saturated fat g Calories e Choose the line tool from the Draw palette e Change the arrow to have an arrow head on the left end 46 I Tutorial Present results e Draw an arrow pointing to the lowest point Click at the lowest point and drag to below the plot zjn e Use the text tool to add a caption This is the candy bar to eat Select the A tool click where the arrow ended and type a caption Saturated fat g eene Use the regular arrow cursor and Shift click to select both the arrow and the caption Saturated fat g Friz is the cang bar to eat H y E e Use the pen color tool to make the arrow and caption blue Click the upper rectangle and drag to select blue from the palette Again apologies for the gray depiction You could continue to refine this plot and the others to your heart s content We ll stop here I Tutorial Present results 47 Box Plot Split By Brand Inclusion criteria Big Three from Candy Bars 7 E Hershey M amp M
233. formats US70 USI56 USI57 template tutorial example us41 us47 text USIO2 views USI56 USI57 tutorial example us41 Us47 Save As see save scattergrams cell plots sr237 compare percentiles plot sr247 confidence intervals us242 error bars us242 factor plots sR141 format USI95 regression plots sR57 residual plots sR57 scree plot SR134 templates usr71 example us174 univariate plots sR217 also see bivariate plots univariate plots Scheff s F sr86 sR104 scientific format Us79 score residuals sRI7I SRI78 score test SRI7O SRI79 Scrapbook us240 scree plot SR134 search sR379 Sec sR418 Second sr419 second mouse button usv seed see random numbers Select us148 select graph components us185 Index SR StatView Reference US Using Statview graphs usi84 rows and columns us64 shapes us204 table components us198 tables us197 variables uss7 Select a Dataset dialog box us107 selected results us133 UsI selection handles us21 us184 us186 USI98 USI99 US204 US207 US209 US2I0 US214 selection tool us185 UsI97 SEM SR42I semi colons SR324 separator characters USIOO USIOI US252 US253 importing US252 serial autocorrelation sR59 Series USI2I USI23 SR3I7 R330 example vus122 hints us222 print definition us122 series functions BinomialCoeffs sr358 CubicSeries sR367 ExponentialSeries sr376 FibonacciSeries sR378 GeometricSeries sR381 LinearSeries sR384 QuadraticSeries sR404 QuarticSeries
234. from real to string because the candy bar names are text not numbers From the Type pop up menu select String In the Name column click and hold Real select String from the pop up menu and release the mouse button Again we don t need to do anything with source format or decimal places We do need to change Name s class from nominal to informative Name is not a grouping variable because each value is unique Rather the names identify each case e From the Class pop up menu select Informative Now we can enter values e Click the empty cell in the first row for Name e Type the first name Snickers Peanut Butter e Press Enter or Return to store the value and move to the next cell Enter the rest of the names the same way Cookies n Mint Cadbury Dairy Milk Snick ers Sugar Daddy I Tutorial Manage data C ea e E User Entered User Entered The rest of the variables are all numeric so we don t need to change their attributes Notice that the cells all contain missing values periods for numeric variables blank cells for charac ter variables right now indicating that no values have yet been specified Let s just enter the values e Click in the first cell for Serving pkg and type the first value 1 e Press Enter or Return to store the value and move to the next cell Enter the rest of the values the same way 1 3 5 3 1 Notice that StatView reformats the numbers to have thr
235. g all the way left to close the analysis browser pane When closed the split pane control is a right arrow double click or click and drag to reopen the browser pane A preference lets you choose whether the analysis browser pane is open or closed by default see Application preferences p 225 To create an analysis just select the analysis from the analysis browser then click the Create Analysis button Analyses with triangle controls f can produce more than one type of results or output Click a triangle to tip it downward and reveal a list of possible results Some even have subcate gories of possible results The triangle controls let you show or hide levels of detail as seen in the picture below For example Frequency Distribution analysis can produce summary tables histograms Z score standardized histograms and pie charts QC Subgroup Measurements is a more complex example It has four categories of measure ments Xbar R S and cusum Statistics and a Summary Table Each category produces sev eral types of results line charts needle charts bar charts point charts and results tables Descriptive Stati gt Frequency Distrib D Percentiles One Sample Analy Paired Comparisons Unpaired Compari Correlation Cova gt Regression D ANOVA gt gt Contingency Table Nonparametrics D Factor Analysis gt Survival Nonpar gt Survival Regres gt Univaria
236. g box sr89 sr90 discussion sR73 SR89 exercises SR96 SRIIO hypothesis testing SR74 SR76 interaction plots sR90 SR96 SR99 SRIO4 SRIOZ Latin square SRIO5 SRIOZ means tables sr78 model building sr76 sr78 multivariate models sr81 sr82 SRIO8 SRIIO nonparametric SRI2I SRI22 post hoc tests sR103 randomized complete block srIOI sRIOS repeated measures models us85 us86 SR82 SR83 SR88 SR89 SR9I SR93 SR97 SR99 results SR95 templates sR96 tutorial example Us38 Us40 analysis windows US139 US148 View Window menus us148 us149 also see analysis browser results browser variable browser views Analyze menu New View USI32 USI39 USI50 rearrange USI168 USI169 Rebuild Template List us168 templates us162 US174 analyze subsets us149 uUsI50 ANCOVA see analysis of variance AND SR345 angle usi85 USI99 ANOVA see analysis of variance Apple Guide us48 us221 US223 application preferences Us225 US226 Arabic characters SR379 SR384 SR423 arc functions USII2 arc tool _us205 US20 ArcCos SR349 ArcCosh sR349 ArcCot SR350 ArcCsc SR35I ArcSec sR352 ArcSin R353 ArcSinh sr354 ArcTan sp355 ArcTanh sr356 arguments USII3 R323 SR326 SR331 arithmetic operators USI12 Index SR StatView Reference US Using Statview 259 arrange results Us43 US213 US214 arrow tool us212 tutorial example us45 also see selection tool ascending
237. g each new vertex e To finish either click the starting point for a closed curve or double click a final point for an open curve Preview While youre clicking vertices hold the Alt key Windows or Option key Macin tosh to preview what a closed curve would look like in other words to see what the final shape would look like if you were to finish by clicking the starting point To force closure hold the key while double clicking the final vertex Drag selection handles Click and drag any of the eight black selection handles to resize a curve Reshape Select Reshape from the Edit menu to switch to Reshape mode and select it again to exit Reshape mode when you are finished In Reshape mode the cursor changes to a reshaping crosshair and you can 1 Change the shape of a curve click and drag any single vertex to a new location Or select any vertex to see its velocity handles tangents and then drag those handles to change the angles of the arcs connecting that vertex to adjacent vertices Velocity handles _Vertices As you drag the end of a velocity handle you change both the angle of the line and the veloc ity of the curve at that point the curve redraws so that the line is still tangent Usually the opposite velocity handle remains aligned to the one you are dragging but you can hold the Alt key Windows or Option key Macintosh while dragging to leave the opposite handle fixed thus creating a corner Draw
238. g statistical and data analysis software that is easy to use and is therefore accessible to those who practice teach or are learning data analysis Finally we thank the Technical Support and Professional Services Divisions of SAS Institute whose contributions enable us to move forward and truly deliver on that mission Overview The StatView manual comes in two volumes Using StatView and StatView Reference This vol ume Using StatView shows how to work with StatView 1 The first chapter Tutorial steps you through every phase of a data analysis product with StatView from collecting and entering data to analyzing data to presenting your results If you read nothing else read the tutorial Subsequent chapters expand on the concepts introduced in the tutorial 2 Datasets discusses the structure of StatView datasets how to arrange your data for Stat View how to enter and edit data in the dataset window how to work with StatView s data attributes and how to use category and compact variables 3 Importing and exporting shows how to get datasets into StatView from other programs through Microsoft Excel files and plain text ascin files 4 Managing data shows StatView s special tools for generating and transforming variables sorting data and studying subsets of your data 5 Analyses shows how to build and modify analyses statistical tables and graphs and how to work with tables a
239. ge datasets and how to enter and edit data in StatView This chapter also discusses in detail how to work with variable attributes how to tell Stat View what kind of data each column contains The time you take in organizing your dataset pays big rewards when you analyze your data StatView can guide you in building analyses with the appropriate kinds of data and it can prevent you from using variables incorrectly Subsequent chapters discuss how to exchange data with other applications how to manipulate data with formulas and how to restrict analyses to data subsets by excluding cases manually or with logical expressions Dataset structure Example Understanding the structure of your data is a very important first step in creating your dataset To enter data properly you must determine two things 1 Data class whether data are continuous or nominal Continuous data are measurements that can assume any value within a given range nominal data identify group memberships Some analyses expect continuous data some expect nominal data and some expect both A third data class informative identifies cases in a dataset but is not used in analyses 2 Data arrangement how to organize data into rows and columns properly for analysis in StatView We ll discuss each of these in detail after looking at an example Suppose youre studying heart disease Your data might include each patient s name gender age weight cholesterol leve
240. ge menu select Sort e Select a variable and click Make Key Or select and click Make Key for several variables in order e For each key click the arrow icon to choose ascending 4 or descending 4 order e Click OK Sort Select the keys for this sort Variables Sort Keys Year S amp P 500 NASDAQ Make Key London Index Tokyo Index Remoue Kou MS Europe Index MS EAFE Index Shearson Corp LT Cancel Ascending order is from lesser to greater numbers or from A to Z and numbers come before letters Case and accent marks are ignored when sorting text Missing values are sorted as if they were the largest values in a variable Nested sorts sorts with more than one key vari able order cases with matching values on the first variable according to values of a second variable etc Only complete datasets are sorted if you want to sort only a few rows or a few columns you must first extract them to another dataset Caution You can unsort data only by immediately selecting Undo from the Edit menu or typing Control Z Windows or Command Z Macintosh Otherwise you cannot unsort data unless you took precautions to preserve the order before sorting For tips on how to pre serve a sorting order permanently see RowNumber p 417 of StatView Reference Recode data Recode creates new variables two different ways by grouping the values of conti
241. good source of fiber Most would take more than 20 packages and some don t have any fiber at all see the missing values If you open the summary statis tics pane you see that 4 2 is the Minimum scrolling down you find that s the Almond Joy Is there any candy bar that would give you enough fiber without putting you over the calo rie and fat limits This is a complex one and we ll need to rely on StatView s logical func tions to do it efficiently Create another formula variable called Composite with this for mula if Fiber rule lt Bars per day AND Fiber rule lt Sat fat rule AND Fiber rule lt Total fat rule then else 0 Click the if button in the keypad area of the formula dialog box Or type if or use the function browser in the lower left corner of the formula dialog box We ll show you that in the Managing data chapter under Function browser p 111 Dont worry about formatting the formula exactly like you see here we just broke the lines like this to make it easier to read Now scroll through the results and see whether any candy bars have a 1 meaning they meet all the requirements Tiger Sport is the only one 12 5 Tiger Sports give you all the fiber you need but you could have 16 before you broke the calorie rule according to Bars per day 20 I Tutorial Analyze data FJ Brand J P01 Bors per Total fot oat fatr_ Fiber rule Compost Charlesto
242. gression results appear underneath the table we already had 6 Templates Manage templates 167 Regression Summary Gas Tank Size vs Weight Count 116 Num Missing 0 R 847 R Squared TAT Adjusted R Squared 715 RMS Residual 1 643 ANOVA Table Gas Tank Size vs Weight DF Sum of Squares Mean Square F Value P Value Regression 1 780 014 780 014 288 914 lt 0001 Residual 114 307 779 2 700 Total 115 1087 793 Regression Coefficients Gas Tank Size vs Weight Coefficient Std Error Std Coeff t Value P Value Intercept 1 858 860 1 858 2 161 0328 Weight 005 2 860E 4 847 16 997 lt 0001 Regression Plot 28 1 1 ft 1 267 244 g2 207 x 6187 2167 144 124 10 8 7 7 7 7 7 1500 2000 2500 3000 3500 4000 4500 Weight Y 1 858 005 X R 2 717 Now we have a Regression Summary an anova Table a Regression Coefficients table and a Regression Plot Manage templates The Analyze menu contains an hierarchical menu of templates These are prebuilt templates for producing most of StatView s analyses and graphs organized according to type of analysis 168 6 Templates Manage templates Analyze New View Correlations Graphs QC Analyses Regression Rebuild Template List For Beginners ANOVA t tests Descriptive Statistics Factor Analysis Nonparametri
243. h and position of tick marks Click the axis numbers Edit Display Whether and where grid lines appear Click the axis numbers Edit Display Line width pen type fill pattern point type point size color Click the plot box bar point or line Draw palette of a plot box bar point or line Thickness pen type and color of plotted lines Click the line Draw palette 186 7 Customizing results Graphs Orientation and frame of legend Click anywhere in the legend Edit Display Point type point size fill pattern and color of legend symbols Click the symbol Draw palette Overlay graphs You can superimpose a graph on another to create a double layer graph Select one graph and drag it into position over the other graph If one plot hides another adjust the plot s fill attributes and consider using Move to Front and Move to Back commands from the Layout menu Also consider deleting or making invisible redundant scales axes etc Resize graphs To resize any graph select the whole graph then click and drag one of its selection handles to a new location Drag a corner handle to resize a graph s height and width at once drag a side handle to change height or width alone Box Plot A Box Plot fr Box Plot 240 4 E ee A 8 f a B B i 220 4 a 220 4 a 2204 i 200 I ai 2004 180 180 b 180 H a i a 3 gt i 160 4 e0 Z 107 k i404 ido omi H 120 4 ia o 1204 E
244. habetical Alphanumeric order by variable name Variable class Grouped in order by continuous nominal informative Usage Ordered first by variable use in analyses and then in alphabetical order When a dataset window is active ordering is only alphabetical Variables are shown in a scrolling list Icons next to variable names indicate their data class for continuous for nominal and gt for informative 144 5 Analyses Analysis windows Compact variables are preceded by a triangle gt and followed by a symbol Click the trian gle to tip it downward and display the category of the variable These categories are marked nominal n e Effectiveness E sr Effectiveness IE Time E Assignment buttons Each analysis requires certain variables before it can calculate Notes below empty place holder results tell you what is expected For example a box plot coaches you To complete this analysis assign at least one continuous variable using the variable browser Add Button StatView Reference chapters detail the variable requirements of each analysis Add assigns variables to analyses and Remove removes them The buttons at the top of the variable browser change according to the analysis you choose For most analyses you use these three buttons LE Variables Remove Split By For most analyses StatView determines each variable s role from its class continuous or nominal F
245. han SR341 greater than or equal to SR341 if then else sR341 1s sr346 IsMissing sR343 ISNOT SR347 IsRowExcluded sr343 IsRowIncluded sr344 less than sR340 less than or equal to sR340 NOT SR345 not equal sr341 OR SR347 true SR345 Index SR StatView Reference US Using Statview 273 logistic regression SRI99 SR215 assumptions sR202 binary SRI99 case control studies SR203 confidence intervals SR205 SR209 SR214 SR215 data requirements sR205 dialog box sr204 dichotomous sRI99 discussion sRI99 SR204 estimating coefficients sR203 exercises SR207 SR21 iterations SR205 multiple sR201 sR202 SR212 nominal data coding sr206 polytomous sRI99 R204 SR214 random samples sR203 reference level sr206 results sR207 simple sR200 SR201 R207 templates sr207 loglogistic model sr172 lognormal model sR172 sR175 LogOdds sr386 logrank Mantel Cox test SRI5I SRI56 long integer data type us73 lowess SR221 SR225 SR226 SR235 tension SR226 SR228 LsD see Fisher s PLSD M M usage marker us163 MacDraw II size limit Us232 macros see templates MAD see median absolute deviation magnitude sr348 main effects sR78 manage data USIO7 USI30 tutorial example us2 us13 Manage menu commands USIO7 R315 SR3I7 SR318 Preferences us225 manage templates Us167 US169 MANCOVA see analysis of variance manipulate columns and rows us62 us64 Mann Whitney U test sriz0 data sR123 exercise SRI25
246. he button on the keypad The entire expression is enclosed in parentheses so the formula now reads Triglycerides Trig 3yrs e Click in the blank area of the text box to deselect the expression e Click the operator and then 2 the keypad The formula is now complete It reads Triglycerides Trig 3yrs 2 e Click Compute The new variable the average of Triglycerides and Trig 3yrs appears at the right of your dataset to the left of the input column As you observed in the exercise StatView automatically creates placeholders for arguments and anticipates the rest of a formula or function when you start typing In a similar way Stat View anticipates arguments for the mathematical operators and For example if you double click a variable and then click the operator the definition area highlights a to indi cate that it is ready to take the second argument gt Gender Age Hoe eight H Cholesterol Triglycerides ei HDL a LDL ei Fe a ee In addition if any part of a formula is already selected when you insert a function name the selection is interpreted as the first argument of any function that can take arguments For example if you select Weight in the formula definition and click the Sin button you see sin Weight 16 4 Managing data Sort data When you type variable names or functions in the formula definition you have t
247. hen the Assign Variables dialog box first appears it pre assigns any variables to slots with matching names So if you are repeating an analysis on datasets with the same variable names it is worthwhile to use specific variable names in the template This saves you the trouble of dragging variables into slots and removes still another opportunity for error Do not save results You can save disk space by not saving results with views you plan to use as templates In View Preferences uncheck Save analysis results with view turn the option off The option takes effect for the next view you create it does not change a view already created Remember to turn this option back on after creating templates so that your regular views retain their results See View preferences p 232 Turn Recalculate off You may want to disable calculation when adding results from several templates to one view so you do not have to wait for each set of results to calculate before you can add results from another template See Control recalculations p 138 Format information In a few cases formatting information saved with a template does not apply to all output 1 For analyses that generate multiple results rather than compound results for additional variables formatting information applies only to the number of tables or graphs present when the template was created See Multiple and compound results p 135 2 For analyses tha
248. iable 0 means observation is uncensored z corea Survival Analysis Preferences Use More choices dialog as default Nonparametric Methods Regression Models Retain covariate matris after regression calculations Censor variable 0 means observation is uncensored Cancel 9 Tips and shortcuts Preferences 231 Use More choices dialog as default By default you see the Fewer choices versions of the nonparametric and regression model dialog boxes If you do not routinely need to edit the More choices parameters leave the boxes unchecked If you prefer to see all the parameters each time check the boxes Each dialog box lets you switch freely between more and fewer choices your choices determine what you see first Retain covariate matrix after regression calculations StatView can either save or discard the intermediate covariate matrices it computes Retaining the matrices consumes RAM discarding them saves RAM but can slow down the subsequent calculations of additional model results Censor variable 0 means observation is This option specifies how your censor variable is coded If you choose Uncensored 0 values in the censor variable are taken to mean that corre sponding event times are uncensored and any other nonzero values in the censor variable indicate censored event times If you choose Censored the opposite is true 0 values indicate censore
249. iable A variable is the data contained within a column Therefore when we refer to a column in this chapter and throughout the manual we are referring to the structure itself not the data in it When we refer to a variable we mean the variable itself not the cells it fills Columns and variables are not the same thing in the context of compact variables see Com pact variables p 84 54 2 Datasets Dataset windows Dataset windows In this section we discuss the windows you will use when working with StatView datasets the dataset window and the variable browser The variable browser is useful with both dataset and view windows we discuss how to use the variable browser with view windows in the chapter Analyses p 131 StatView stores data in a special dataset window Each dataset has its own window and you can have thousands of datasets open at a time if enough memory is available You can even analyze variables from several datasets at once The first step in creating a dataset is to open a new data window From the File menu select New Input column Input cell Input column Input cell Untitled Dataset Me Criteria Input Colu fn ea y Source User Entered f Continuous Free Format Fixed a Untitled Dataset 1 Compact Type Source Class Format Dec Places Criteria
250. ick Windows Builds a category definitions for a simple compact variable from the names of the selected or Command click columns The Fewer choices dialog box does this transparently for simple compact variables Macintosh New Edit Lets you edit a selected category in the Categories list For more details see Edit category definitions p 83 If you make a mistake in creating a category use Edit to correct it 92 2 Datasets Compact variables Since we don t have any categories defined yet we must create two one for gender and one for smoking First we define a gender category e Click New e Type a Category name Gender e Type a Group label Male e Click Add Type another Group label Female e Click Add e Click Done Edit Category Category name Gender Group label MEITA tt iy Male E sca cancel Le e Repeat the process for smoking e Click New e Specify a Category name Smoking status e Specify a Group label Smoker e Click Add e Specify another Group label Nonsmoker e Click Add e Click Done Edit Category Category name Smoking status Group label WEAGE ts Smoker Hc faa _ pone Now Cells in compact and Variables selected both say 4 indicating that we have s
251. ick Done Edit Category Category name Category for Calorie Groups Group label MEIA LE Low ah sca Cancel EOR Now we need to specify the cutpoint for the groups which Calories values go in the Low group which in the High The rectangular area represents the total range of Calories from 125 to 450 we need to click some value below which values should be grouped Low and above which they are High e Move the cross hair cursor up and down until you find an appropriate cutpoint some value in the gap between high and low values like 360 e Click at that cutpoint Recode of Calories How to recode this variable Selected Breakpoint 450 High ka Low 125 Show definition Cancel _Recode al StatView shows the group assignments in pop up menus to the right Because we defined cat egory groups in order from small to big StatView s initial guesses were correct If we had not defined them in order we d have to use the pop up menus to fix the group assignments e Click Recode The dataset has a new variable showing Low and High group memberships e Click the dataset window to bring it forward or select it from the Window menu e Click the variable s name to select it I Tuto
252. iew templates for checking normality First paste into the template a copy of the continuous variable you want to check In the original dataset select the variable double click its name e From the Edit menu select Copy e Open the Normality Test template from the Dataset Templates folder Z0 Normality Test Criteria Heasurement Actual Input Column Ideal Normal Real Real Real User Entered Dynamic Form User Entered Continuous Continuous Continuous Free Format Fi Free Format Fi Free Format Fixed a a Format Click the first cell in the Actual column of the Measurement compact variable From the Edit menu select Paste A formula automatically computes Ideal Normal values from a normal distribution with the same mean and standard deviation as the variable you pasted Next we use the K S Normality Test template to test the hypothesis that your Actual values and the computed Ideal Normal values come from the same distribution e From the Analyze menu select QC Analyses K S Normality Test The Assign Variables dialog box appears and automatically assigns the variables to the proper slots since the variables and the slots share the same names e Click OK 9 Tips and shortcuts Dataset Templates 235 Assign Variables for K S Normality Test Please double click or drag the desired variables into the proper
253. ights for male and female smokers and nonsmokers and then match up one male smoker s cholesterol reading with his weight and his 86 2 Datasets Compact variables age unless we were careful to enter the subjects numbers in the same order for each compact variable However this is not always a disadvantage Sometimes you need to break the one case per row connection between two variables Suppose that you want to perform a Kolmogorov Smirnov test to compare a variable against a variable that is known to be normally distributed i e you want to test whether your variable is normally distributed You don t want to compare the variables case for case Putting the actual and ideal values in two columns of a compact vari able lets you get around that problem In fact the Normality Test in the Dataset Templates folder uses exactly this trick see Normality Test p 233 Note that when compact variables store within factors for repeated measures ANOva the com pact variable must conform to the one case per row rule each value must belong to the indi vidual case represented in that row of the dataset A other analyses interpret the data in compact variables as though they were stored in the usual long and skinny continuous and nominal variable columns 2 You cannot use the nominal portion of a compact variable in an analysis unless you also use its continuous portion 3 The nominal portion of a compact variable mu
254. iles us99 graphs usi86 mesokurtic SR7 message area USI40 method default sr135 Microsoft Excel see Excel Minimum sR389 minimum US60 US255 SR3 minus R333 R335 SR370 SR373 Minute sr389 Missing Cells us255 also see missing values missing values UsGo Us62 USII8 SR326 SR335 SR338 SR339 SR343 SR346 SR347 SR357 SR365 SR393 SR394 SR424 date time data SR321 R330 formulas us255 import USIO3 US253 US254 in criteria USI27 multiple datasets us146 Recode usi20 US255 recode sR34I Mod sr390 Mode sr390 mode sR2 model building sr76 sr78 model coefficients table sr186 modify templates us171 modulus sR390 Month sr391 Most SR324 mouse shortcuts USVI also see StatView Shortcuts card move graphs us186 objects us214 table components us200 tables us200 Move Backward us215s us216 Move Forward us215 us216 Move to Back us186 us215 Move to Front us186 us215 moving range SR477 MovingAverage SR391 MR charts SR278 SR281 R283 multiple sr52 multiple categories us254 multiple comparisons see post hoc tests multiple logistic regression sR20I sR202 SR212 multiple regression SR52 SR69 SR77 multiple vs compound results USI35 USI37 USIZO US229 multiplication SR333 SR367 SR374 N class marker see class marker name variables see variables natural logarithm sR385 negation SR345 negative SR335 also see absolute value nest functions R323 nested groups sR381 new datase
255. ill pattern to add a pattern to its interior Choose None for a transparent empty fill None and white fill are not the same Closed shapes fill as you would expect open curves and polygons fill as though a line connected the first and last vertices and open arcs fill as though lines connected their end points to their center points E gt Pen pattern Select any drawn shape and choose a pen pattern to change the appearance of its edges All lines default to a solid pattern Pen patterns have greater visual impact with thicker lines PostScript printers interpret some patterns as halftones so a single pixel screen line might print as a gray line with some patterns Ta IZN 3 zat gt 3 ae 3 Line thickness Select any drawn shape and choose a line thickness for its edges All line widths default to a single point line To print hairlines see View preferences p 232 Choose the top dotted line for invisible lines A 8 Drawing and layout Layout tools Ay NO Arrow Select any drawn line or arc and choose one or two arrowheads You cannot add arrowheads to the other kinds of shapes AND Change colors Pop up menus at the bottom of the Draw palette let you choose colors for the pen fill and background the empty view area outside graphs You can change the pen color for any object including text and you can change the shape color for any object you can fill You can also change the backg
256. in it 5 Analyses Analysis windows 139 ANOVA Table for lt None gt Assign a continuous and one or more nominal variables using the variable browser Add button Analysis windows In this section we discuss the windows you will use when analyzing data in StatView the view window the variable browser and the results browser Each of these windows has controls for creating and changing analyses View window StatView builds and displays analyses in a special view window A view window is a living document window for both graphs and text tables with an analysis browser for creating anal yses on the left and Edit Analysis and Edit Display buttons for modifying results along the top The document area scrolls both vertically and horizontally to provide as much room as you need for a complete presentation For a self guided tour of the view window From the Analyze menu select For Beginners View Window Intro Or to open a new empty view window From the Analyze menu select New View Or in any open dataset double click the empty rectangle before the first variable name 140 5 Analyses Analysis windows Click to change Click to change how results are How many results analysis parameters formatted are di Untitled Vie 1 a 4 T Click to turn E Recalculate Edit Analysis Edit Display 1 Result Selected recalculation Create Analysis on and off Show hide variable browser
257. ing and layout Draw tools Import objects Qo IVY Open a closed curve Control Alt click Windows or Command Option click Macin tosh the line segment you want to remove a Close an open curve Control Alt click Windows or Command Option click Macin tosh either the starting point or the finishing point ear eer Remove a vertex Alt click Windows or Option click Macintosh the vertex IMD Add a vertex to any curve Alt click Windows or Option click Macintosh the edge where you want the vertex IIS Shortcut Macintosh only Hold the Control 2 to switch temporarily to Reshape mode and release the key to exit Reshape mode You can paste objects from other applications into the view In the other application copy text or a picture In a Stat View view window paste the object Pasted text can be manipulated the same was as text created by StatView However you can not manipulate a picture except to resize it by dragging its black selection handles To return a resized picture to its original size double click it 8 Drawing and layout Draw tools 211 Change fill patterns pen patterns and line types The Draw palette s fill pen and line tools let you change the appearance of drawn objects You can fill shapes with patterns change lines to arrows and change the thickness and pen pattern of lines Fill pattern Select any drawn shape except a line and choose a f
258. ing the foursome produces two grouped pairs that could in turn be ungrouped When you change variables for grouped results it affects all the results Overlap and overlay objects Objects in the view are layered in front of previously created objects The last object you cre ate or draw is on top or at the front Each previous object is one layer behind its succes sor If objects do not overlap this layering is irrelevant However if objects overlap the objects in front obscure objects behind You can re layer objects with Draw tools and Layout menu commands To move an object select it and choose the appropriate command or tool If you select several objects they move as a group Draw palette Layout menu Action Move to back Moves the selected object all the way back behind all other objects Move to front Moves the selected object all the way forward in front of all other objects Move Backward Sends the object one layer back Move Forward Brings the object one layer forward Exercise This exercise shows how to accentuate a table by layering it above a drawn filled rectangle to create a frame e Create any graph or table e g a Descriptive Statistics analysis yg 8 y e In the Draw palette select the rectangle tool 216 8 Drawing and layout Layout tools e Click and drag to draw a rectangle around the table Descr iptire Statistics Weight Means Std Dev Std Error N
259. ing under lined words To start Help press F1 or select a command from the Help menu 9 Tips and shortcuts Help 223 Status bar Windows only The Windows status bar at the bottom of the main StatView window describes menu items and buttons in the dataset view and browser windows The status bar also informs you of the progress of StatView s calculations EJ exit Text Analyze Layout Draw M New Ctri N Open Ctri 0 Close Untitled View 1 Ctrlew Save Ctrl S Print Setup Print CtP Ctr status bar aves the active document with a new name To show or hide the status bar select Status Bar from the View menu Tool tips Windows only To learn the function of a tool on the tool bar pause with the cursor positioned over a tool Apple Guide Macintosh only StatView harnesses the power of Apple Guide active assistance with its own StatView Guide StatView Guide goes beyond the passive information found in typical online help systems enabling you to learn while you work Rather than searching through lengthy help text you can simply follow StatView Guide s step by step instructions and actually complete the task while you learn StatView Guide gives you the same sort of help you get when an expert sits down with you and not only explains things but shows you what to do it guides you from start to finish as you create any statistical test table or graph in StatView Use Stat
260. ions 4 0 4 01 and 4 02 because it retains formula and cri teria definitions See update documentation for other formats that might be available in your version of Stat View For more information about the Excel file format see Importing and exporting p 99 Exchange datasets between Windows and Mac versions of StatView The DataSet Transfer format is a cross platform dataset format that can be read by both Win dows and Macintosh versions of StatView The DataSet Transfer format preserves all data for mula definitions and values and criteria definitions It does not preserve the current selection the current inclusion and exclusion of rows or custom column widths DataSet Transfer files have the filename extension ssp Windows Note that DataSet Transfer files save data values in double precision which may cause slightly different results for Feu and norpu Macintosh versions of StatView which operate in extended precision Datasets that are opened or saved in DataSet Transfer format remain in that format and you might want to use DataSet Trans 2 Datasets Save datasets 71 fer as your usual format if you routinely exchange data between Mac and Windows versions of Stat View Moving a dataset from Mac to Windows In StatView for Macintosh e Open the dataset From the File menu select Save As From the File Format list select DataSet Transfer e Type a filename ending with the extension ssp Including the
261. is integer 2 If more than half the values are currency or date time that type is assigned If fewer than half the values are currency or date time the type with the next highest count is used 3 If strings have the highest count type is string unless there are repeated values If there are repeated values and fewer than 256 distinct strings type is category with group labels in alphabetical order Sometimes data entry errors or missing separator characters shift different types of data into a single column When a data point is incompatible with its column s type it is replaced with a missing value If a variable has many unexpected missing values use the import option Make variables with errors have type string String variables can contain any data point so you can examine the variable to investigate the problem Missing values StatView imports two adjacent separator characters as a missing value Consider the row Tab 3 Tab 5 Tab 7 Tab Tab 11 StatView would import these values as 1357 11 However multiple space separators are not read as multiple missing values because many applications use as many spaces as necessary to align columns Rather multiple spaces are read as a single space and a single separator For example 38 space 4 space 12 space space space 19 You might expect the three consecutive spaces to translate as two missing values in six vari ables 38412 19 Instead the spaces
262. iting data and it goes into more detail about what every thing means 58 2 Datasets Enter data Name variables The default variable name is Column 7 where 7 is the number of the current column It is better to give variables meaningful names so it is easy to remember which variable is which when you are building analyses You can enter or edit variable names at any time e Click the cell containing the current name to select it e Type a new name Do this while the old name is selected and you have an I beam cursor To change only part of the name select the part you want to change and correct it e Press Enter or Return to store the new name Or press Tab to select the next variable name Analyses and graphs are labeled by the names of the variables in use so you should use names that will be meaningful You should capitalize and spell names in the dataset the way you d like them to appear in results Variable names can be up to 80 characters long and can use any characters except colons and quotation marks Each variable name must be unique within its dataset you can have several datasets open at a time in StatView and it is not a problem for variable names to be repeated among datasets Variable names should not begin with numerals and they should not be the same as any func tion name for example Log is a function that computes the logarithm of a variable Log should not be used as a
263. izontal measurement 2 inches e Click OK Graph Flip horizontal and vertical axes Bounds include extra lines Show legend Show title Dimensions Vertical inches Horizontal inches Numbers Decimal places Lr Always have leading digit Shift click to select the axis labels and scales In the Text menu change Font to Helvetica or your favorite and change Size to 12 6 Templates Build templates 173 2500 2000 4 Y Variable a f oO t 1000 7 500 4 o 0 08 o o o o go H o o o 2 iai E sE o 0 TP ab ahah af tat oh el 150 250 350 450 550 X Variable e Shift click to select the X and Y scales e Click Edit Display e Click More Choices and use pop up menus to select major and minor tick marks that straddle axes ed variable n Variable Numeric Axis Bounds upper E Lower 150 Include zero Numbers Format Free w Decimal places Ensure leading digit Intervals Scale Linear Grid Lines at zero lick m Major Minor Major width 50 Minor divisions Length Width 4 1 2 1 Fewer Cho me Cancel e Cli
264. kers t value one sample t test sR23 paired t test SR30 regression SR57 unpaired test SR37 table defaults see preferences Table dialog box uszo1 tables align us214 arrange US213 US214 borders USI99 US201 US231 colors USI99 US202 column widths usi99 Index SR StatView Reference US Using Statview 285 components USI97 create USI3I USI33 by hand or with templates us131 customize USI97 US202 decimal places uszo1 us231 Edit commands us181 edit text US200 fonts UsI99 Us200 format USI35 US2OI US231 group US215 height and width ust99 us200 interior USI98 layers us2I5 line spacing UsI99 US201 US23I line widths us199 us202 list in analysis browser us141 lock us214 move US200 US214 move components US200 notes USI99 number format vus201 numeric formats US231 pen patterns usI99 US202 preferences UsI79 USI80 US231 US250 resize USI99 US200 row and column labels us199 row heights us199 US231 select UsI97 select components us198 structure US20I text alignment us199 Us200 text angles us199 Us200 text colors us199 titles ust99 transpose USI99 US2OI ungroup US215 unlock us214 tails F test SR38 one sample analysis sR24 paired comparisons SR30 unpaired test SR38 Tan sR426 Tanh sr426 target value SR254 Tarone Ware test SRI51 SRIS6 Template folder usi61 us174 templates us161 UsI77 assign variables dialog box us162 combine analyses us162 cr
265. la dialog box is listed in the Window menu where you can select it to bring it to the front You can double click the top area beneath the title bar to bring its dataset to the front Select Print from the File menu to print a formula definition Macintosh only The triangle gt at the bottom of the dialog box is a split pane control click and drag it to resize or close the browser pane If you drag it all the way to the left the variable and function areas disappear you can still type and use the keypad to create formulas Use Attributes if you want to specify variable names and attributes before clicking Compute You can adjust attributes afterward in the dataset window if you prefer See Variable attributes p 73 Variable browser The variable browser lets you choose variables from the current dataset Use the Order pop up menu to choose how to sort variable names in the scrolling list Dataset order The order in which variables appear in the dataset s columns left to right Alphabetical Alphabetical order by variable name with nonalphabetic names first Variable type Grouped in order by nominal continuous and compact Usage Ordered first by variable use in analyses and then in alphabetical order When a dataset window is active ordering is only alphabetical As in the main variable browser variables appear in a scrolling list Icons next to variable names indicate their data class f for conti
266. label in the scrolling list and click Delete 84 2 Datasets Compact variables You should only delete group labels that have no data associated with them If you delete a group label from a category that is in use the group does not disappear Instead it is replaced with the next group label and all groups adopt the label of the following group all the labels slide up one group The last group label is replaced by a generic Unlabeled Group For example a category called Color contains four group labels Green Blue Red and Black If all the group labels are used in a variable and you delete the group label Red using the Edit Cat egories dialog box every occurrence of Red is replaced with Black and every occurrence of Black is replaced with Unlabeled group 4 See also How can I reorder category variables p 238 For more information on how Stat View uses group ordering in graphs and analyses see How does StatView use ordering in nominal variables p 238 Delete unused categories Compact You can create a category and later delete the variable that used the category Doing so does not delete the category definition itself it is still available for use with other variables To delete an unused category definition e Open the dataset that contains the category e From the Manage menu select Edit Categories e Select the category you wish to delete e Click Delete e Click Done
267. late is checked on by default If you want to suspend calculations temporarily perhaps you need to edit several variables and don t want to wait for dozens of tables and graphs to recalculate just uncheck Recalculate click it to turn it off see Control recalculations p 138 The Edit Analysis button lets you change the parameters of an analysis see Edit Analysis p 134 The Edit Display button lets you change how results are formatted and drawn see Edit Display p 135 Near the upper right corner a note tells you whether any results are selected and if so how many If no results are selected the area is blank Knowing whether results are selected is important when you are creating new analyses because new analyses adopt variable assign ments from any selected results see Determine whether results are selected p 133 The status bar Windows or the area underneath the analysis browser tells what StatView is doing at the moment 5 Analyses Analysis windows 141 View window preferences You can set preferences to customize view windows to suit your style of working See View preferences p 232 Analysis browser StatView s analyses are listed in the analysis browser The analysis browser lets you create anal yses The sigma in the lower left corner of the view window is a split pane control click and drag it to resize the browser pane Double click it or click and dra
268. le and Total fat g as the Independent Variable From both total and saturated fat Use the Regression Multiple templates specifying Calo ries as the Dependent Variable and both fat variables as Independent Variables Notice that the Independent Variables slot grows to accommodate more than one variable How are these variables correlated Use templates or browsers to create a correlation matrix and some bivariate scattergrams Since we never quite resolved the question of Calories bimodality among other reasons it s probably best to refrain from drawing any major conclusions We ll leave interpretation of these results up to you Save your work Normally at this point in a data analysis project you would want to save your work if you haven t already I Tutorial Analyze data 41 Save a dataset Since we ve made some changes to the dataset new variables and criteria we should save e Make sure the dataset is the frontmost active window If not click it or select Candy Bars Data from the Window menu e Select Save As from the File menu e Type a filename Candy Bars Data 2 e Click Save StatView saves everything about the dataset the values the variable names the attributes and summary statistics the criteria and whether one is in effect and more This way you can resume working exactly where you left off Save a view We also want to save our view full of graphs and tables e Make sure the view i
269. le in use you add delete or clear data rows you apply a criteria to the dataset you add or remove variables NN BR WO NHN you change the sort order of the dataset only affects certain analyses If you plan to make several changes to analyses or variables it is a good idea to turn off recal culation This way you can make your changes without waiting for recalculation after each change Force recalculation Uncheck turn off and then Control click Windows or Command click Macintosh the Recalculate checkbox to force recalculation of all analyses Ordinarily turning on Recalculate just recalculates those results whose data have changed Cancel calculations Anytime the cursor is a spinning yin yang jg you can cancel the operation by typing Esc Windows or Command Period Macintosh Recalculate in the background You can use other applications while your analysis calculates StatView can perform calcula tion in the background so you can switch to another application Placeholder results StatView ensures that all results appearing on the screen are consistent with the current analy ses variables and state of the dataset If recalculate is on and the analysis or data are changed all tables and graphs are updated However if recalculate is not on and the analysis or data change tables and graphs are replaced with placeholders until you turn recalculate on A placeholder is an empty graph or a box with an X
270. les into nominal values or by replacing missing values with a specific value The other Manage menu commands are discussed elsewhere 1 Edit Categories lets you change category definitions for group labels and is discussed under Edit category definitions p 83 2 Add Multiple Column enables you to add any number of empty columns to the dataset in a single step and is discussed under Manipulate columns and rows p 62 3 Preferences let you customize StatView s behavior to suit the way you work and are dis cussed in Tips and shortcuts p 221 Manage multiple datasets If a dataset window is active frontmost when you select a Manage menu command the command takes action on that dataset If a view window is active and several datasets are open you must choose which dataset to change e Select the dataset you want to manage e Click Use 108 4 Managing data Include and exclude rows Select a Dataset Select a Dataset Choose a dataset Choose a dataset Candy Bars Data Lipid Data Candy Bars Data SVD Use Cancel To make changes to more than one dataset repeat your actions on the other datasets in turn You can also Copy and Paste formula definitions between datasets Include and exclude rows When you use a variable in an analysis all the values in that variable are included in the anal ysis unless you exclude some rows cases
271. licking Macintosh Create analysis Another way to bypass the dialog box is to choose Basic Statistics or Graphs Only from the Show pop up menu but doing so also limits you to those analyses Variable browser Variables in the dataset are listed in the variable browser a floating window that appears alongside both dataset and view windows When you are working in a view window the vari able browser allows you to 1 Choose variables from open datasets and open new datasets 2 Assign variables to analyses and specify what role they should play in analyses 3 Remove variables from analyses 4 Split analyses into separate results for each group in a nominal variable To show or hide the variable browser select Variable Browser from the View menu Windows or select Variables from the Window menu or click the Ej button in the upper right corner of the dataset and view windows Macintosh A preference lets you choose whether the variable browser is shown or hidden by default see Application preferences p 225 In the middle part of the browser a Data pop up menu lets you select among open datasets or open other datasets Click on the menu to choose another open dataset and choose Other to locate and open a previously saved dataset The Order pop up menu lets you choose how to sort variable names in the scrolling list Dataset order The order in which variables appear in the dataset s columns left to right Alp
272. lot Box Plot 4500 300 F 1 Result Selected F 275 4000 250 z500 aa 2 v 200 5 2000 c 175 150 cae A 125 Show soog a pe fae Fage 1 oy 1500 s50 Box Plot Plot for Hors Weight Horsepower 7 Customizing results Graphs 185 Select components If the cursor isn t an arrow first choose the selection tool from the Draw palette To select a component click that component directly Shift click to select several components at once Black selection handles show when certain types of components are selected while dotted lines show selection for other types of components at 140 azo azo mE 10 a m00 cf gt wie 30 Hgg 120 140 160 180 200 220 24 eygh Bbystolic BP 117 445 038 Weight R 2 02 Dpiastolic BF 70 272 D47 Weight RZ O14 This table shows how to select each graph component and which tool to use to change an aspect of the component Component How to select it Tool legend text axis labels axis values notes Color of text items Clickthetet Draw palette Axis frame type Click the frame Edit Display Line width pen type color of graph frame Click the frame Draw palette Color and fill of graph interior Click blank space inside the graph Draw palette Line width pen type color Click the axis numbers Draw palette Bounds intervals scales numeric format Click the axis numbers Edit Display Length widt
273. lots in a field study I Tutorial Manage data Start StatView e Double click the StatView icon Statview The first thing you see after a splash screen is a welcome message in a Hints window if Hints Welcome to StatView This is the Hints window It Welcome to Statview This is the Hints window t 4 provides information about every aspect of this provides information about every aspect of this program To display a hint click on any item To close program To display a hint click on any item To this window click the close box To show this close this window click the close box To show window choose Hints from the Window Cor E menu this window choose Hints from the Window menu Choose Hints in the Preferences dialog box Manage Choose Hints in the Preferences dialog box menu to hide this window at program start up Note Manage menu to hide this window at program this is a scrolling resizeable window teed sun hades Hain j Late Keep an eye on the Hints window as you begin to work with StatView As its welcome says it gives helpful information about what youre doing how to handle errors and what to do next You can close the window if you prefer To reopen it at any time select Hints from the View menu Windows or Window menu Macintosh Start a new dataset From the File menu select New e You see an empty untitled dataset You can resize or move the window if you want Take a moment to look at
274. ls and alcohol use Class continuous or nominal Names are informative they merely identify each case You can t do any statistical or graphical analyses on an informative variable Gender is nominal since each patient belongs either to a male group or a female group 50 2 Datasets Dataset structure Data class Age is continuous since each patient has an exact measurement that falls somewhere along the continuum between birth and death Similarly weight and cholesterol are continuous variables of exact measurements Alcohol use could be either continuous or nominal You could record each patient s average daily number of drinks exactly in a continuous variable or you could group their daily aver ages into a nominal variable with several intervals say fewer than two between two and six and more than six For a heart disease study a nominal variable might make the most sense Arrangement You would most likely arrange this dataset to have a row for each patient in the study and a column for each quantity measured like this Name Gender Age Weight Cholesterol Alcohol use J Suds male 22 197 2 6 T Wilson D S Quintent male 2 fermale 22 181 if 190 42 1 2 a Zz S FR James male 25 172 6 F 5 q 2 6 t Mubroid male 23 154 194 a2 L Phote male 24 185 155 i2 C Narman male TA 175 oad A This one case per row arrangement is the most typical and it is the arrangeme
275. ludes outliers The third style uses boxes with notches showing the 95 confidence interval for the median and includes outliers The fourth choice uses notched boxes but excludes outliers Connect or separate lines for univariate line plots You can choose whether all points in univariate line plot are connected with a single line or with a separate line for each group Examples of both are shown below the default is separate line segments for each group e Inside a univariate or cell line chart click the plot one of the lines e Click Edit Display Turn the option on or off 7 Customizing results Graphs 189 e Click OK Line Plot EJ Separate line for each group T Line Plot Separate line for each group T Univariate Line Chart Split By Gender 240 2207 200 7 5180 J v Z160 4 140 4 120 4 100 o male female Observations Connect or separate lines for cell line charts 240 220 200 180 oO 160 140 120 100 Univariate Line Chart Split By Gender o a o Bs o Oj J 1G lo fi L o male Koy Hi Jol i female K o ro Observations You can choose whether all points in cell line charts are connected with a single line or with a separate line for each variable Examples of both are shown below the default is to connect the variables e Inside a cell line chart click the plot one
276. lues youre likely to get a chaotic mess of criss crossed lines however if you sort the data according to increasing Age values the lines fall into place and become meaningful 240 220 200 180 160 140 120 100 Line Chart 17 5 22 5 27 5 32 5 37 5 42 5 Age 240 220 200 4 180 v 160 140 120 100 17 5 22 5 27 5 32 5 37 5 42 5 Line Chart lal ltt Age 2 Use Include Row and or Exclude Row to control on a row by row basis which rows are used in calculations See Include and exclude rows p 108 For example the line chart above has a few outliers with extreme Age values these commands let you remove those rows from analyses without removing them from your dataset 150 5 Analyses Exercise i Line Chart mne Chare Row exclusion Lipid Data 240 1 240 H 220 220 4 B 200 4 L Z 00 e e 200 4 F E 180 4 L la Pla L 3 ih 180 NoE z160 H D p iig E 2160 4 lj e3 140 4 H 140 T 120 120 4 i L 100 4777 ido 17 5 22 5 27 5 32 5 37 5 42 5 Pe ee eed Age 18 20 22 24 26 28 30 32 Age 3 Use Criteria to choose by logical expression which rows are used in calculations See Cre ate criteria p 124 For example you might instead restrict your line chart to cases whose Weight values are under 200 pounds by applying a criterion Weight lt 200 Again you remove rows from analyses
277. moves to the next cell either down to the next row or right to the next column Continue entering values in input cells until you complete one row or column Windows Macintosh Direction cursor moves Enter on numeric Enter Either down or to the right depending on your Dataset Preferences keypad Enter on main part of Return Down From an input cell at the bottom of a column the cursor moves keyboard to the top of the next row Tab Tab Right From an input cell at the end of a row the cursor moves down to the beginning of the next row Enter across If you prefer to enter values across filling in all the variables for one observation or case you should name and set attributes for all your variables in advance To begin a new row either click the first cell for the row or press Tab from the empty input cell at the end of the first row Enter down If you prefer to enter values down filling in all the cases for a variable you might prefer nam ing and setting attributes for each variable individually To begin a new column either click the first input cell in the column or press Return from the empty input cell at the bottom of the previous column Missing values Sometimes you don t have a value for a given case of a variable Perhaps a respondent didn t answer a question or a doctor forgot to take a blood pressure reading for one patient or per haps a variable is simply not applicable to an obs
278. n Cha 2 8 696 9 286 3 333 25 000 0 000 Tootsie Jr Mints 4 10526 16 250 8 000 e l 74 Weider Tiger Mik_ 15 12500 10 833 20 000 ie Lis 16 000 32 500 40 000 12 500 BENA Build an analysis Weve already analyzed these data quite a bit without ever having left the dataset window Now it s time to see some real analysis power e From the Analyze menu select New View ntitled View 2 M Recalculate Edit Analysis Edit Display Create Analysis Show fa Analyses x Order Defaut x Descriptive Statistics a gt Frequency Distribut gt Percentiles One Sample Analysis Paired Comparisons Unpaired Comparis Correlation Covaria gt Regression p ANOVA z A Contingency Tahle zl jz KI 4 Now you see StatView s view window This is where you ll build statistical analysis tables draw graphs and put together presentations Think of the view window as your paper SS titled View 1 Sa EJ Recalculate Edit Analysis Edit Display l Create Analysis Show All Analyses Order Default Descriptive Stati h gt Frequency Distrib in Percentiles One Sample Analy Paired Comparisa Unpaired Compari Correlation Cova gt Regression pb ANDYA gt Contingency Table Along the left side of view window is the analy
279. n Service Hus have teamed up to produce the Food Guide Pyramid which recom mends eating a variety of foods an appropriate number of calories and a modest amount of fat specifically 30 or fewer of your total number of calories per day should be calories from fat and only a third of those should be calories from saturated fat Fat is the densest source of calories at 9 calories per gram Alcohol is a close second at 7 calories per gram For adults consuming 2000 calories per day which is about right for moderately active women and somewhat sedentary men that works out to no more than 65 grams of fat no more than 20 grams of which are saturated fat I Tutorial Manage data 3 Chocolate We want to know how many candy bars can fit into this daily diet The first thing we need to do is gather nutritional data We went clipboard in hand to a few stores near our Berkeley California offices and stood in the candy aisle copying down nutritional facts about every candy bar we could find We also included some non bar candies like M amp Ms and Reese s Pieces because they re very similar to many candy bars Once we d made that decision it seemed only fair to include other non chocolate candies such as Skittles and Super Hot Tamales Was this a good deci sion theoretically Maybe maybe not We ll have to study that in our analysis Fortunately StatView has all sorts of tools for excluding weird cases from analyses so if it
280. n desirable for line plots JO oo Fill patterns pen patterns and line thicknesses The Draw palette s fill pen and line tools are handy for calling attention to items in graphs You can thicken lines and fill them with patterns or add patterns to the plotted shapes within a graph You can also add a fill to the interior of a graph making the plot itself stand out more 196 7 Customizing results Graphs Fill You can add a pattern to the interior of an entire graph or to bars in a bar chart and his togram slices of a pie chart and boxes of a box plot To add a fill to a graph s interior select the entire graph and choose a pattern from the fill pop up menu Pen pattern For any line you can choose from sixteen pen type choices All lines default to a solid pattern Pen patterns have greater visual impact with thicker lines PostScript printers interpret some patterns as halftones so a single pixel screen line may print as a gray line with some patterns Line thickness You can change the thickness for the following lines a graph frame tick marks boxes in a plot pie chart divisions bars in a chart the simple and polynomial regres sion lines and the normal curve of a histogram All line widths default to a single point line To print hairlines see View preferences p 232 Choose the top dotted line for invisible lines Change colors Pop up menus at the bottom of the Draw palet
281. n importing Print datasets To print a dataset make it the active frontmost window and select Print from the File menu If the attribute pane is open attribute settings and summary statistics print along with the data The entire dataset is printed even if it does not all fit in the window Row numbers and variable names always appear on printouts 2 Datasets Variable attributes 73 Variable attributes Type Here we discuss in detail the five variable attributes and how to change them Please pay par ticular attention to the consequences of changing variable attributes Data type specifies whether a variable contains real numbers integers long integers strings date time values currency values or categories The default type for a new column is real Type Examples Description Real 9 8765 Numbers with fractional parts Set the number of decimal places to 1 2E34 display in the dataset with the Decimal Places pop up menu Full 3 141592 precision is stored and used in all calculations Integer Whole numbers between 32 767 and 32 767 inclusive 156 7 142 Long Integer 98 165 Whole numbers between 2 147 483 647 and 2 147 483 647 inclusive 3 348 920 String John Doe Alphanumeric text data Entries in string columns can be as long as 436 39 9976 255 characters Patient 28A Date time T 8 08 Points in time Many display formats are available Regardless of the July 8
282. n of pages in your document Use Drawing Size from the Layout menu to set the size and shape of the overall document This determines how many pages will be printed Check Print and copy lines at 1 4 width in the View Preferences dialog box to reduce all line widths to approximately 1 4 of their specified size See View preferences p 232 This option affects only printed and copied results there is no discernible difference on the screen 160 5 Analyses Print a view To print a view e From the File menu select Print Templates StatView view documents collections of completed analyses and annotations are two doc uments in one 1 You can reopen a view and its original dataset s to resume analysis where you left off You can continue working with analysis objects tables or graphs which are still linked to their dataset You can add and remove variables clone and adopt to build new results and use drawing and formatting tools to put results together in a complete presentation 2 You can open a view and apply it as a template to a different dataset A template is a collec tion of one or more analyses and all their data links parameter settings layout and all other visible characteristics except variable assignments You simply assign new variables to use throughout the template and StatView computes all the results at once Views saved in the Template folder Windows or StatView Templates folder Macinto
283. n which levels are defined If you define Categories manually using Edit Categories before entering data that means the order in which they appear in the scrolling list If you import data or enter data in a variable and then change its type to Category StatView automatically defines the levels of the Category from the values in the variable sorted in alphabetical order Box Plot Descriptive Statistics i Split By Category Split By Category Friday M 30 F ean 25 4 F o Monday Continuous Total 8 86 20 4 L Continuous Friday 6 09 15 74 F o Saturday 2 Continuous Monday 10 15 7 E F o Sunday Continuous Saturday 8 97 0 _ Bl Thursday Continuous Sunday 7 63 5 4 L Continuous Thursday 5 94 10 7 o Tuesday Continuous Tuesday 8 08 15 cni o Wednesday Continuous Wednesday 15 18 eee There is one exception to this rule Pareto Analysis charts and tables show cells in decreasing order of frequency i e groups with the most cases first least cases last Pareto Table for Category Pareto Chart k Counts Continuous Counts Continuous K A Count 160 4 L Wednesday 46 7 M Monday 30 120 4 H E 0 E Saturday 27 8 80 4 z Tuesday 24 dl Sunday 23 40 F Thursday 18 F F Friday 17 gt gt gt gt gt Bae Sgssgss Oe EDEL GE 338334 gr ore z Category Note StatView s statistical
284. nce midnight of the first day if you specify a month mid night of the first day of the first month if you specify only a year etc 3 StatView works with the current date and time set for your system Be sure you have set the correct time date time zone etc 4 If you enter an ambiguous date time value such as 8 11 which could mean either August 11 or 8 November StatView warns you A dataset preference lets you suppress this warn ing see Silently accept ambiguous values p 227 If you are unsure how StatView interprets a value choose a date time format that shows more detail 2 Datasets Variable attributes 75 Change types Exercise caution when changing variable types if you have already entered data in the column Some changes in data type cause data loss For example changing real variables to integers rounds values to the nearest integer e g 1 456 becomes 1 If you then change type back to real you get integers with zero fractional values e g 1 000 If you ever change a data type by mistake immediately select Undo from the Edit menu to avoid permanent data loss Following are the possible consequences of changing data types New type Real Result of changing from other types Integers are unharmed Long integers are unharmed Strings that are real numbers are unharmed except that numbers exceeding 19 significant digits lose precision Other strings are converted to missing Date time v
285. ncy distribution analysis with both a table and a histogram changing the number of intervals of the frequency distribution updates both the summary table and the histogram If you select more than one type of analysis and click Edit Analysis you see a series of dialog boxes one for each type of analysis Not all analyses have parameters you can change for these Edit Analysis has no effect Each analysis has different parameters so each has its own unique dialog box For details on any analysis parameters see its chapter in StatView Reference Also take advantage of online help see Help Windows only p 222 and Hints window p 222 Shortcut Alt double clicking Windows or Option double clicking Macintosh an analysis has the same effect as selecting it and clicking Edit Analysis Note that double clicking an 5 Analyses Overview 135 analysis is the same as selecting it and clicking Edit Display However a view preference lets you switch these shortcuts See View preferences p 232 Edit Display You can also edit how each result is displayed While Edit Analysis changes the parameters and therefore the results of an analysis Edit Display changes how the results are presented For example you can edit any table to have double borders and format numbers with only one decimal place Or you can transpose a graph flip it sideways and give it a different axis frame Not all results displays c
286. nd graphs in the view window 6 Templates shows how to recycle a complete set of analyses by applying it as a template to other datasets a quick set of steps that can save you countless hours 7 Customizing results shows how to redesign the appearance of StatView s tables and graphs to suit your needs and preferences 8 Drawing and layout shows how you can assemble analysis results into a full color presen tation complete with annotations and enhancements you draw yourself 9 Tips and shortcuts shows how to use various online help facilities and how to set prefer ences to suit your way of working Then it demonstrates example documents and tem plates answers common questions and gives troubleshooting suggestions The second volume StatView Reference presents a reference chapter for each analysis review ing statistical ideas then detailing data requirements dialog box options results and related templates Finally each chapter demonstrates the analysis with a step by step exercise The last chapter in StatView Reference is Formulas a comprehensive reference for the mathematical expression language used throughout StatView Finally StatView Reference is where youll find algorithms and formulas a glossary of statistical terms and references StatView s keyboard and mouse shortcuts are summarized on the StatView Shortcuts quick ref erence card VI Overview Examples Examples If yo
287. ndows or Command Shift click Macintosh the Add button 6 167 4 500 3 407 393 75 0 000 15 000 6 100 2 500 4 656 11 000 2 500 2 500 2 500 11 000 11 000 2 103 0 000 3 424 639 2 500 2 500 0 000 15 000 5 667 5 000 3 215 1 556 123 4 500 4 500 2 100 S00 11 000 00 0 000 0 000 2 000 8 000 5 000 1 500 1 F 10 000 5 000 5 000 5 823 1 500 1 500 4 000 7 000 3 000 6 500 TOF 3 000 3 000 6 000 7 000 5 000 7 000 5 000 5 000 7 000 7 000 2 500 6 000 4 250 2 475 1 750 T50 554 250 Dotted red lines in the view window indicate page breaks 00 1 000 28 I Tutorial Analyze data Use Criteria to examine a subset The numbers in the Count columns of these tables reveal that only Hershey Nestle and M amp M Mars have six or more different candy bars Annabelle has five so you might want to include that brand as well it s your choice Perhaps we should narrow our study to these three major brands StatView makes it easy to do that e Bring the dataset window to the front by clicking it or by selecting Candy Bars from the Window menu e From the Criteria pop up menu select New Candy Bars Criteria al Horchow Dankiac n Mint e In the Criteria dialog box double click items in the scrolling selection list to build the cri teria definition Brand ElementOf Hershey
288. ng Statview 287 turn off Criteria us129 tutorial usi us48 two way table uss3 SRIIS type I error sR75 SR76 SR85 type II error SR75 SR76 U u charts SR256 SR266 SR300 SR304 R306 UE in variable name sr47 uncensored SRI47 Undo vus95 graph table formats us181 Sort USIIZ unexpected results us251 Ungroup vusz215 uniform distribution sR41I sR412 unique random integers SR412 univariate plots sR217 axis types USI9O confidence intervals sR217 connect lines us188 data requirements sR218 dialog box sr217 discussion sR217 exercise SR219 Line Plot dialog box sr217 ordinal axes us193 results sR218 templates sR219 Unlock vus214 unpaired comparisons sR37 data sR39 data requirements SR39 dialog box sr39 discussion SR37 exercise SR4I nonparametric SRI2O nonparametric test SRI2O results SR41 templates SR41 unrotated factor solution SRI39 SRI4I unsort data sR418 update Analyze menu us169 update see dynamic formulas usage markers Us57 USI45 USIG3 USI65 tutorial example us22 user entered data source US77 y valid range date time data US74 R321 SR330 SR33I each data type Us73 integer data sR319 validation of StatView results us250 variability CoeffOfVariation sR361 MAD SR387 StandardDeviation sR421 StandardError sr421 Variance SR430 variable attributes see attribute pane variable attributes in examples usv1 variable browser us21 Us56 uUs57 USI43 USI47 buttons us144
289. ng the shape and dragging one of its black selecting handles Hold the Shift key while you drag to constrain movement vertically hori zontally or at a 45 diagonal Check the state of the corner center control before resizing or reshaping Rounded rectangles After drawing a rounded rectangle you can use Edit Display to change how corners are rounded e Select the drawn shape e Click Edit Display You can either round the ends of the rectangle for an oval or you can round just the corners with an arc of the radius you choose The unit of measure for the radius is either inches or centimeters depending on your Custom Rulers settings see Custom Rulers p 217 Round Corners Round Corners Round Ends Round Ends Round Corners EFA m inches Round Corner imbes Show Cancel saea Cancel C Arcs After drawing an arc you can reshape it by dragging its selection handles or using Reshape from the Edit menu Drag selection handles Click and drag any of the eight black selection handles to change the shape of an arc a E E Lad E LI a i E La E Lad a E 7 7 a Reshape mode Select Reshape from the Edit menu to switch to Reshape mode and select it again to exit Reshape mode when you are finished In Reshape mode the cursor changes to a reshaping crosshair 4 Reshape mode lets you change the way an arc is defined Usually arcs are de
290. nomi nal or informative and category data type category can only be nominal For more detail on data types see Type p 73 Change classes Another powerful feature is that you can easily change variables from continuous to nominal and vice versa Sometimes the class of a variable is flexible or ambiguous Many variables could be either continuous or nominal depending on how they are analyzed For example alcohol usage might be recorded as a nominal variable for a heart study but if youre studying liver and stomach enzymes it might be more appropriate to record exact con tinuous measurements of alcohol consumption As another example three levels of a drug 10 25 and 50 i u could be seen as three drug administration groups nominal or as three values along a range of possible values continuous Class can be ambiguous Many variables although they are continuous in the strictest sense are effectively nominal For example body weights are strictly continuous because their val ues can fall anywhere within an interval if you measure them precisely enough However most people tend to report their weights rounded to the nearest five or ten pounds or kilo grams You should assign the class that makes the most sense in the context of your research Occasionally researchers will study variables both ways for example they ll use nominal weights in ANOvA tables but continuous weights in descriptive statistics tables Dat
291. not directly supported in StatView Using a few simple steps however you can create your own step function plots using the bivariate line chart These steps can be summarized as follows e Copy the variables including the censor variable for which you want to create step func tions then paste them into a new dataset If you were creating a step plot of the cumulative survival function you would copy the Time Cum survival and Status variables from the dataset created by the analysis If you want to plot confidence limits copy the upper and lower confidence limit variables as well Survival results can be saved to datasets for both nonparametric analyses and regression 250 9 Tips and shortcuts Troubleshooting models for more information see the chapters Survival Nonparametric p 143 and Sur vival Regression p 167 of StatView Reference Delete the censored cases from the new dataset You could do this in a variety of ways Probably the easiest way is to define a criterion for the censored cases e g Censor 1s 1 then use that criterion to select only the censored cases then delete these cases Hold the Control key while selecting a criterion from the Criteria pop up menu In the new dataset select and copy all rows of all variables Paste the rows in the input row at the bottom of the new dataset Sort the new dataset by Time ascending sort and Cum survi
292. ns that you haven t saved your dataset Since you don t intend to use the empty dataset again that doesn t matter or more datasets that have not been saved When you re open the view the original dataset s may not be found A Results in this view use variables from one Click OK Close both the view and the dataset You are now ready to use this template with any dataset to perform a simple regression From the Analyze menu select Rebuild Template List From the Analyze menu select Regression Regression Simple Open Car Data from the Sample Data folder Drag Gas Tank Size to the Dependent slot Drag Weight to the Independent slot Click OK 6 Templates Build templates 177 A new view with simple regression results appears Note that the regression plot includes con fidence bands and we also have a plot of residuals against fitted values We ve rearranged the results to fit the page ANOVA Table Gas Tank Size vs Weight Regression Summary DF Sum of Squares Mean Square F Value P Value Gas Tank Size vs Weight Regression 1 780 014 780 014 288 914 lt 0001 Count 116 Residual 114 307 779 2 700 Num Missing 0 Total 115 1087 793 R 847 R Squared T17 Regression Coefficients Adjusted R Squared 715 Gas Tank Size vs Weight RMS Residual 1 643 Coefficient Std Error Std Coeff t Value P Value Intercept 1 858 8
293. nt SR365 Covariance SR365 GeometricMean sr380 Groups SR381 HarmonicMean sp382 MAD sR387 Maximum sR387 Mean sp388 Median sr388 Minimum sR389 Mode sr390 NumberMissing sr393 NumberOfRows sR394 OneGroupChiSquare SR394 Range sr412 Rank sr413 RowNumber sr417 StandardDeviation sr420 StandardError sr421 StandardScores sr422 TrimmedMean sr428 Variance SR430 statistics texts recommended sr481 status bars USI40 US221 US223 StatView 4 x data us7o StatView Guide us223 StatView II SE Graphics file format ustos StatView Library us225 US233 US25I categories US9I StatView Templates folder us161 us174 step function plots us249 stepwise regression USI44 SR52 SR54 SR69 F to enter sR53 F to remove SR53 survival analysis sR176 sR186 Strata button survival regression models sr185 284 Index SR StatView Reference US Using Statview stratification variable nonparametric analyses sRI53 SRIS7 regression models sR183 strike through text sR323 string data type US73 SR319 SR320 R338 string functions see text functions Student Newman Keuls sr88 Style us205 style sheets see templates preferences subgroup measurements analysis see QC subgroup measurements analysis subgroup variables us242 sR290 differences us244 formulas sR291 subject group sRr83 subsets of data see Criteria Include Row Exclude Row row inclusion Substring sr422 subtitles for inclusion us29 subtraction SR333 SR370 SR373 unary SR
294. nt in the dataset If you specify a larger number of rows rows are added to all columns and filled with missing value symbols in the other columns unless they are formula variables whose definition specifies otherwise Use Attributes if you want to specify variable names and attributes before clicking Create You can adjust attributes afterward in the dataset window if you prefer See Variable attributes p73 For more information about each series refer to the chapter Formulas p 315 of StatView Reference or examine the Hints window when you click a series the Hints window describes the function and its parameters briefly Use the variable s Source pop up menu to view and edit a formula definition or to change from static formula to dynamic formula or user entered See Change sources p 77 New variables are appended at the right side of the dataset To move a variable Copy or Cut the data insert an empty column elsewhere in the dataset see Insert columns p 62 and Paste the data into the empty column To create a series variable in another location insert a column and change its source to Static Formula or Dynamic Formula This exercise creates a new dataset with five variables and one hundred observations From the File menu select New e From the Manage menu select Series e Double click CubicSeries 1 0 0 1 e Specify the number of rows 100 4 Managing data Random numbers 1
295. nt usually expected by StatView Some analyses also accept other arrangements The class of a variable is crucial to how it is used in an analysis StatView interprets a variable differently depending on whether it is continuous or nominal Continuous data are measurements that can assume any numerical value over a given interval such as weights of cars or batting averages of baseball players Nominal data identify group memberships such as the countries in which cars are built or whether baseball players com pete in the National League or American League Informative data give case identifications such as model names player names or jersey numbers Why do need groups Many statistical and graphical analyses try to answer in various ways one basic question do measurements differ among groups of individuals For example do husbands and wives vote for the same or different political candidates How do American and National baseball league batting averages compare Do men and women perform equally well on math tests Do patients treated with a new drug have more or fewer heart attacks than patients not treated Such analyses include unpaired tests ANOVAs unpaired nonparametric tests contingency tables cell plots grouped box plots and percentile comparison plots Therefore these analyses require that data identify the group to which each observation belongs If you set a variable to class nominal StatView understands that its valu
296. ntered User Entered i Nominal Mominal Nominal Nominal Mominal Now delete the extra cells and change the columns of numbers to have type real e Click the variable name cell for Column 5 to select the standard error column e From the Edit menu select Delete e Click and drag over the first three row numbers to select the rows with title text From the Edit menu select Delete e Select Columns 2 3 and 4 by dragging over their names From the Type pop up menu for one of the columns select Real to change all three To the warning about loss of data click Yes Column 2 Column 3 Column 4 Dec Places male female 2 Select the data and paste it into the dataset template e Click and drag to select all data cells e From the Edit menu select Copy Tips and shortcuts Common questions 237 From the Window menu select the Compute Bartlett s Welch s test window Select the empty data cells in the first four columns From the Edit menu select Paste p value Finally read the results from the variables on the right pane of the window Make your own dataset templates Examine the formulas and criteria used in these dataset templates to get ideas for your own dataset templates Instructions for building dataset templates are given under Build dataset templates p 240 Common questions Dataset How can record dataset comments You can ea
297. nthly report you can build a dataset template Create a dataset with all the variable names attributes and formulas you use every week 9 Tips and shortcuts Common questions 241 1 2 3 4 5 6 7 Delete all the rows in the dataset and save it You might want to make this file read only Windows or lock this file or make it Stationery using Get Info Macintosh so that you don t accidentally save over the template with some month s data __tates Gross Over short Tax _Net__ Net weet Real User Ent Continuat Free Fori b Dec Places Je J2 d2 d2 J2 2 k v vivy Now each week import the data into a new dataset and Copy all the cells Open the tem plate Paste in the data and save under a new filename When is it faster not to use formulas Sometimes it s faster not to use formulas For example if you want to recode all the missing values in a dataset with many variables it might be faster to export or Copy and Paste the data into a word processor do a global find and replace to change all missing values to your new code say 999 then import or Paste the data back into StatView How can add confidence intervals or error bars to a plot Use a formula to calculate your favorite type of confidence interval or error bar for your Y variable Then add and subtract those values to the values of your Y variable Plot all three variables against the X variable and
298. nuous and for nominal informative variables cannot be used Nominal variables have triangle controls Click a triangle to tip it down ward and reveal group labels This mechanism provides an easy way to enter a group label from a nominal variable into a formula which is particularly useful in formulas with if then else statements 4 Managing data Formula I Compact variables are shown as several continuous variables for example a compact variable of Male and Female Weight values is listed as two continuous variables Male Weight and Female Weight To insert a variable name or a nominal value in a formula definition double click it in the variable browser Function browser The function browser offers an array of date time logical mathematical probabilities ran dom numbers series special purpose statistical text and trigonometric functions you can use for creating variables You can use the Order pop up menu to sort functions alphabetically or by type When you order functions by type each type has a triangle control Click the triangle to tip it downward and reveal its functions Function type Functions discussed Date Time Date DateDifference Day DayOfYear Hour Minute Month Now Second Time Weekday WeekOffear Year Logical lt lt or S gt or gt gt lt gt or AND ElementOf IS ISNOT OR XOR false if then else IsMissing IsRowExcluded IsRo
299. nuous vari ables into nominal values or by replacing missing values of a nominal or continuous variable with a specific value The original variable is unchanged You cannot recode informative vari ables or compact variables To recode a nominal variable into a different nominal variable use a formula see if then else p 341 and ChooseArg p 359 of StatView Refer ence Grouping values of continuous variables lets you derive nominal variables from continuous data For example ranges of temperatures in degrees recode into hot temperate or cold response times in seconds convert to slow medium and fast Date time data with day month year values can be grouped into seasons 118 4 Managing data Recode data Recoding missing values is useful for preparing a dataset for export to a data analysis package that expects missing values to be coded a certain way for example some packages expect miss ing values to be coded 99 e From the Manage menu select Recode e Select a variable e Choose which way to recode Recoding Data Select a variable Choose desired recoding Median Value Zone Industry Charles Missing values to a specified value NOH Rooms Continuous data to nominal groups When you choose Continuous values to nominal groups you must specify a category to define the group labels to be used for the recoding You can either select
300. o if Op Large iai w b Medium 7 PEIA K FOE GO PRR P PT Medium 7 OCG Q K al O i 0 pl Kayl 2 Megat Compact 00 ccd ca ii O F compact bs P P Hdl i Wy Sporty CO jj i I l i Ik F Sporty 4 i iP i 4 ii Small DEO A LS One th Small 4 O 000o CCO CO 0 OOGO rrr 0 20 40 60 80 100 Observations Observations Change legends Compound graphs graphs that depict several variables or groups within a single axis frame have legends that identify how each variable or group is distinguished See Multiple and compound results p 135 A legend serves as a key to the graph and you can use it to change how lines points bars etc are drawn in the interior of a graph L span D Weight A o Triglycerides o male Ei Other O Horsepower so sTrig Pyrs Es female To change the overall format of a legend e Click the legend to select it e Click Edit Display Legend Layout Show frame Round corners Shadow Show Cancel Layout Choose vertical horizontal or two column orientation for the legend Each are shown below o Small Sporty o Small Sporty a Compact Small Sporty Compact Medium Large a Compact 4 Medium A Medium Large Large Frame Check options on and off to control whether and how the legend is framed Possibili ties are shown below 7 Customizing results Graphs 1
301. o on for a total of 8 sequences of measurements from subgroups 1 15 for a total of 120 rows The formula to create these values of the subgroup variable is mod RowNumber I 8 or more generally mod RowNumber I X where X is the number of measurements per subgroup No matter which of these patterns fits your data once the subgroup numbers are generated you should make the variable s class nominal the variable is then ready to be used in any anal ysis that takes a subgroup variable If you wish to give the subgroups alphanumeric names e g Sample 1 Sample 2 etc simply create a category in which the ordering of the group labels corresponds with the numeric values of the subgroups For instance if the groups are weeks then make the variable s Type Category and create a category whose group labels are Week 1 Week 2 Week 14 Week 15 See Categories p 80 for more information on creating and using categories How can create event times elapsed time from start and end times You can easily compute event time variables from survival data containing start times and end times Suppose you have data on patients suffering from epilepsy You are interested in whether a particular treatment significantly affects the time that elapses until a patient s next seizure In this example each patient has entered the study on a different date The event time for each patient will be the elapsed time in d
302. o the Criteria pop up menu and bring the dataset to the front Save Save the criterion and include it in the Criteria pop up menu Apply Evaluate the criterion save it and apply it include in analyses only those rows that evaluate to true Select Evaluate the criterion save it and select highlight those rows that evaluate to true Do not apply the criterion If the criterion definition is incorrect you cannot apply or save it until you correct the error Criteria pop up menu The Criteria pop up menu at the top of the dataset lists all criteria defined for the dataset A criterion is dimmed if it has not been saved or applied if its dialog box is open for editing or if it has become invalid because the variable it is based upon has been deleted or renamed e Select No Criteria to turn off criteria to include all rows Select New to define a new criterion Select Random to define a random criterion e Select a criterion to apply it e Control select a criterion to select in the dataset rows that would be included by the crite rion Mew Criteria Fandom 10 Rows Included Overweight No Criteria is the default indicating that all dataset rows are included in any analysis Random criteria use the following dialog box Random Criteria En Percent of rows to include Percent of rows to include 50 Ear Enter a percentage between 0 and 100 for each row to be included
303. o type only enough characters to identify the word uniquely For example if you type we and you have a variable named Weight the remaining characters are automatically entered However if you have a variable named Weight and a variable named Weight1 you have to type the whole word since the letters we are not unique When you begin to type the name of a function the function is inserted along with any arguments and the first argument of the function is selected If StatView anticipates a term incorrectly simply delete any incorrect character Errors in formula Dynamic vs Sort data If you make an error in a formula definition you get a warning after you click Compute The Hints window advises you about your error and the formula window highlights the problem atic area Formulas do not compute until their expressions can be interpreted unambiguously Most errors involve invalid arguments e g constants where variables are expected a wrong number of arguments missing parentheses or misspelled function or variable names Many common errors can be prevented simply by using the browsers and the keypad to build for mulas rather than trying to type formulas by hand Any formula with an error will not compute For a new formula missing values fill the col umn until the formula can be computed If you edit the formula for an existing valid formula variable and cause an error the original definition and its
304. odel Proportional Hazards Treatment Treatment 1 Treatment Treatment 2 Treatment Treatment 3 Treatment Treatment 1 167701 Treatment Treatment 2 067560 120561 Treatment Treatment 3 067107 066885 114969 The variable Treatment has 4 levels treatments 1 3 and a control group The data were entered so that the coefficients displayed are all relative to that for the control group whose coefficient is 0 See Coefficient Initial Values dialog box p 182 of StatView Reference for details Notice that these results are displayed to 6 decimal places this is recommended to avoid loss of precision in the following calculations Suppose that you wish to test the hypothesis that the coefficient for treatment 1 is the same as that for treatment 3 The Wald chi square statistic would therefore be 0 522397 0 298397 0 167701 0 114969 2 0 067107 98 Then to compute the p value associated with this chi square value use the following formula in StatView ProbChiSquare 4 538 1 In this example the p value is 0 033 You would therefore conclude that the coefficients from treatments 1 and 3 are significantly different at the 0 05 level How can create step function plots In some cases such as when you would like to plot confidence limits about a Kaplan Meier survival estimate you might wish to create your own step function plots Unfortunately this plot type is
305. of decimal places To change a single table select the table and click Edit Display 2 All statistical routines in StatView have been validated using standard datasets against stan dard references as well as other commercially available software packages If you feel that there is an error please send us information about what you feel is incorrect and why If you have compared the result to that of another application please send us the input and 9 Tips and shortcuts Troubleshooting 251 output from that package as well Out of memory If Stat View does not have enough memory to perform an operation it alerts you with a dialog box If this happens you need to increase the amount of memory available to StatView Unexpected results StatView s behavior in the view is largely dependent on what results are selected If StatView creates several tables and graphs when you expect only one it is probably using variables or analyses from results that are selected in a part of the view not visible in the window Open the Results browser from the View menu Windows or Window menu Macintosh to see which results are selected and look at the Results Selected note in the upper right corner of the win dow to see the number of results currently selected See Determine whether results are selected p 133 Grouped objects can account for unexpected numbers of results appearing You might intend to add a variable to a single res
306. og boxes but tons in the view and dataset windows and all browsers all the parts of the dataset and views To use balloons e Select Show Balloons from the Help menu on the right end of the menu bar Menu commands are only described by balloons and status bar messages in Windows All other balloon help is duplicated in the Hints window 9 Tips and shortcuts Preferences 225 Error messages When an error occurs StatView generates an error message The Hints window automatically appears to display the error message often an explanation of why the requested action cannot be completed In the Hints Preferences dialog box you can set an option to have StatView beep when it displays an error message Alert messages An alert message warns you of potentially dangerous situation or advises you of the conse quences of an action you requested Alert messages sometimes appear in a box with a single OK button You can set a preference to have these appear in the Hints window instead which automatically opens when there is an alert message to display The advantage to this setting is that your work is not interrupted by an alert box yet you still see the alert message Preferences You can set preferences that govern the behavior of many different parts of StatView e From the Manage menu select Preferences e Select the type of preferences you want to change e Click Modify Choose Preferences Choose a preference to modify
307. okers and female nonsmokers 2 Second you compact the columns into a single variable However you must also assign category definitions for each grouping distinction a gender category with labels male and female and a smoking category with labels smoker and nonsmoker Again we ll start with a new empty dataset From the File menu select New Name four variables one for each group e Click the cell that reads Input Column e Type a new variable name Male Smokers e Press Tab to move to the next variable name e Type a name Male Nonsmokers e Press Tab e Type a name Female Smokers e Press Tab e Type a name Female Nonsmokers All our cholesterol readings are whole numbers so we give all columns type integer e Shift click or click and drag all variable names to select all columns 90 2 Datasets Compact variables From the Type pop up menu for one column select Integer Male Smoker Male Nonsm Female Smoker Integer Integer Integer Female Monsm Inp Source User Entered User Entered User Entered User Entered User En Class Continuous Continuous Continuous Continuous Continu Format Free Fo Dec Places Now we can enter the values e Click the first input cell e Type the first male value 127 e Press Enter Continue entering all the values Your dataset should look like this Male Smoker Male N
308. ommon analyses by reducing clutter They also speed the process by automatically using the default analysis parameters rather than presenting a dialog box for you to specify parameters You can still change the parameters with Edit Analysis button see Edit Analysis p 134 Survival Analysis and Quality Control restricts your choices to those analyses since specialists using those analyses may not need StatView s more general routines Show Basic Stati Order Default Order Show All Analyses Order Default Descriptive Stati gt Frequency Distrib gt Percentiles One Sample Analy Paired Comparisons Unpaired Compari Correlation Cova gt Regression e ANDYA gt Contingency Table Nonparametrics gt Factor Analysis gt Survival Nonpar gt Survival Regres e Univariate Plots gt Bivariate Plots gt Cell Plots Box Plot Compare Percenti GC Subgroup Mea gt GC Individual Mea PG PNP pac cfu gt Pareto Analysis Descriptive Statisti Histogram Univ Soattergram Biv Scattergram One Sample t test Paired t test Unpaired t test Cell Line Chart Show Quality Con Order Default gt GC Subgroup Mea gt GC Individual Mea pO PNP pac cU e Pareto Analysis Show Graphs Only Order Default p Univariate Plots gt Frequency Plots gt Bivariate Plots gt Cell Plots Box
309. one of the brand groups Creating a category defini tion makes it faster to enter the data prevents data entry errors and saves RAM and disk space Here s how it works first we create a category definition containing the group or level names we plan to use Then we use this category definition to enter the data e Click New to create a new category definition I Tutorial Manage data Choose Category Please choose the column s category iin New OK Type the first Brand name in the Group label box M amp M Mars e Click Add e Type the next name Hershey and click Add e Type the last name Charms and click Add e Click Done Edit Category Category name Category for Brand Group label MUELA L MoiM Mars had a ea canta f Done StatView automatically sets Format and Decimal Places to missing since those attributes dont make sense for category variables We ll discuss missing values some more later not applicable is the idea here es User Entered 2 Nominal The second attribute source answers the question Where did these data come from Most data are user entered raw values but others are computed from static or dynamic formulas such as the sum of several variables and others are generated by analyses such as residuals from a regression
310. ones and Brenda Sun early testing was done by Sid Butts and Bruce Gilbert Ann Lehman Ph D Kristin Rinne Terri March and Lynn Scott made production possible The manuals were written by Erin Vang Authorship credit for significant portions of the Stat View Reference is due to Nicholas P Jewell Ph D Division of Biostatistics and Department of Statistics University of California Berkeley logistic regression and survival analysis Clifford Baron QC analysis and survival analysis Additional contributions were made by Daniel S Feldman Jr Samantha Sager Pete Schorer and Phil Spector Ph D Department of Statis tics University of California Berkeley Erin Vang also coordinated localization and revised the help systems originally created by Clifford Baron Statistical consulting was provided by Phil Spector Ph D Alan Hopkins Ph D and Nicho las P Jewell Ph D Acknowledgements We are grateful to StatView users around the world for their continued support and particu larly for their comments criticisms suggestions and praise A special note of recognition is due the users who have volunteered their time to beta test StatView IV Acknowledgements It is through the efforts of all the people named above and many more whom we could not list that we are able to accomplish our mission to enable anyone not just an expert to perform data analysis and present results We do so by creating marketing and supportin
311. onsm Female Smoker Female Monsm Integer Integer Integer Integer User Entered User Entered User Entered User Entered Continuous Continuous Continuous Continuous Format Dec Places Since we have fewer female values than male the female column has missing values at the bot tom Now we can compact the two columns into one variable e Shift click or click and drag all the variable names e Click the Compact button in the top left corner of the window Or select all the names in the variable browser and then click the Compact button in the variable browser Criteria e Type a variable name Cholesterol readings e Click More choices Now we see the More choices version of the Compact Variables dialog box 2 Datasets Compact variables 9I Compact Variables Choose one or more categories to identify this compact variable Compact variable name UIEEEDOUEE CIO Categories Chosen a Select Remoue H Cells in compact 0 Variables selected 4 Cancel femenct You must assign categories to structure your complex compact variable For simple compact variables StatView takes care of this for you For a complete discussion of category variables see Categories p 80 The Categories scroll list on the left shows the names of categories that have been d
312. or analysis SRI31 real data type US73 SR318 rearrange templates us168 us169 Rebuild Template List us168 us169 exercise USI76 Recalculate us138 uUsI40 USI64 USI8I background vus138 templates us17o recalculate see dynamic formulas reciprocal powers SR334 Recode USII7 USI21 SR7I SR317 R330 SR338 SR359 categories USII8 US238 dialog boxes us117 UsI20 example us120 examples sR342 SR343 hints us222 missing to specified value usi20 missing values us255 print definitions USI19 US120 troubleshoot us255 tutorial example us33 us35 record macros see templates rectangle tool us205 recycle formulas us240 recycle results see template reference level sr206 reference lines us184 reference structure solution sRI40 regression SR5I SR7I Index SR StatView Reference US Using Statview data requirements sR61 dialog boxes SR59 R204 discussion sR51 error distribution sR5r error of intercept sR56 exercises USISI SR64 SR69 SR7I exponential sr54 SR68 growth SRS5 SR67 line equation Us184 lines see bivariate plots logarithmic SR54 models sr77 multiple SR52 SR69 SR77 nonlinear SR54 SR55 SR67 SR68 also see logistic regression plots SR231 polynomial SR52 sR64 SR65 power sR55 residual plots sr57 residuals sr57 results sR62 SR207 simple SR52 SR64 SR77 stepwise USI44 SR52 SR54 SR69 tvalue sR57 templates US166 USI75 SR63 SR207 exercise USI66
313. or example you add both continuous and nominal variables to an aNova and StatView knows that the continuous variable should be the dependent variable and the nomi nal variables are independent variables or factors StatView s ability to assign roles auto matically is one of the many benefits of setting attributes for each variable in the dataset see Set attributes p 58 However some analyses have buttons that assign variables to specific roles in an analysis For example for bivariate plots you need to assign an X Variable for the horizontal axis and a Y Variable for the vertical axis For regressions you need to specify a Dependent variable and one or more Independent variables Otherwise StatView would have no way of knowing what you intend These and other specialized buttons are discussed for each type of analysis in StatView Reference Split By shows results separately for each group of a nominal variable or each subgroup defined by two or more nominal variables You can split any analysis For example if you have a Gender variable that divides your dataset into male and female groups assigning Gender as a Split By variable in a box plot produces separate boxes for males and females in many cases it also produces results for the total sample You can Remove a Split By assignment to stop analyzing groups separately Assign variables All variable browser buttons work the same way e Select one or more variable
314. or print ing unless you need to tile especially large objects across several pages e g for a poster How ever some people prefer to check this option turn it on while working on analyses to save screen space On the screen page breaks are depicted with dotted red lines Align to left margin By default Clean Up Items doesn t reposition items horizontally You can check this option turn it on to align all items along a left margin determined by the Horizontal distance setting Additional alignment commands appear in the Draw palette see Align to Grid p 218 and Align Objects p 218 Move objects To move objects individually select the object and drag it to a new location Be careful not to drag black selection handles dragging selection handles resizes objects If you accidentally resize an object immediately select Undo from the Edit menu Here as in other chapters objects are tables graphs table or graph components drawn shapes or text items Most objects can be arranged moved and edited with the same basic techniques For drawing and layout purposes all objects work the same way unless a particular type of change would destroy the nature of the object For example you cannot edit numbers in a table wrong numbers would do you no good and you cannot move a plot outside its graph frame a plot would be meaningless outside its axis context Lock and unlock objects You can lock any object to p
315. or sizes e Click Align 8 Drawing and layout Layout tools 219 Align Objects ie ie Vertically C Distribute _Heights No Alignment Distribute As you make choices the sample objects in the corner move to demonstrate how your choices will affect the objects The choices above centered all the objects and distributed them verti cally for even graphs You can quickly bring together two objects that are far apart in the view by selecting them in the Results Browser and then using Align Objects 8 Drawing and layout Layout tools Tips and shortcuts Tool Bar Windows only The StatView tool bar provides one step access to many of the most frequently used StatView commands To hide or show the tool bar select Tool Bar from the View menu Help StatView offers a number of ways to get help online while youre working 1 Hints both Windows and Macintosh 2 Help Windows only 3 Status bar Windows only 4 Tool tips Windows only 5 Apple Guide Macintosh only 6 Balloon Help Macintosh only Each of these systems meets a different need Hints give suggestions about what you are doing in StatView how to use a new window how to complete an analysis what to do next how to resolve an error etc Help has brief discussions of the Stat View environment its windows commands and general organization and gives general advice and step by step instructions for common t
316. ormat Dec Places a a Class Informative Nominal Cont e e J b Continuo Continuous Continuous Mean Std Deviation 243 158 653 191 232 97 263 Std Error Variance 1272 648 3714 387 Coeff of Variation Minin AST 107 115 627 43 Hasimum 234 295 450 Range Count D D 20 0 5 127 000 170 000 25 35 437 000 35 Missing Cells Sure z2 4i 5 E 0 0 in a Q 2310 15072 19167 000 9240 000 Sum of Squares 571 2466970 3593733 000 1247864 ot Le T Wilson female 22 D S Quintent male 2 R Beal female R James male 2 22 25 M Mubroid male 23 L 4 2 For continuous variables you see summary statistics mean minimum maximum standard deviation standard error variance coefficient of variation count range sum sum of squares and number of missing cells For nominal variables only count and missing cells are calcu lated and the other cells contain missing values No statistics are shown for informative vari ables When you open the attribute pane you always have complete current summary information the statistics update automatically whenever you make any change to your dataset You can copy cells from the attribute pane and paste them into other windows or applications as needed You cannot edit summary statistics or
317. ou can replace StatView s missing value symbols in one of two ways Either recode missing values in StatView with Recode from the Manage menu or use a text editor to perform a global search and replace on the exported text file How StatView imports data The following sections provide more detail about how StatView interprets data it imports These rules apply to all file formats alike Variable names StatView can import variable names directly only if every variable name contains nonnumeric characters No variable names can be strictly numbers For example Year1975 is okay but 1975 would cause problems 3 Importing and exporting How Statview imports data 103 Data types If variable names are not unique StatView numbers occurrences after the first to ensure that all names are unique StatView assigns each variable the most appropriate data type based on the number of values of each type found in the column Types are discussed under Type p 73 Most variables contain values of a single type and StatView assigns that type When variables contain values of more than one type StatView counts how many values have each type and then uses the following rules 1 If any numeric type has the highest count and there is at least one real value the type is real If there is no real value but there is at least one long integer value the type is long inte ger If there are no real or long integer values the type
318. oup ANOVA Table for Weight Split By Gender Cell male DF Sum of Squares Mean Square F Value P Value Alcohol use 3 155 258 51 753 092 9644 Residual 67 37809 108 564 315 Model II estimate of between component variance ANOVA Table for Weight Split By Gender Cell female DF Sum of Squares Mean Square F Value P Value Alcohol use 3 785 596 261 865 996 4148 Residual 20 5256 362 262 818 Model II estimate of between component variance Histogram Split By Gender Cell male 20 jo 18 4 m 167 144 127 rl gio J 8 4 6 pa 44 24 D E R H ER 100 140 180 220 Weight Histogram Split By Gender Cell female 14 Peis i 127 T04 87 3 fo O 64 4 24 0o Het 100 40 180 220 Weight For information on how each analysis uses split by variables see the StatView Reference chap ter for that analysis 138 5 Analyses Overview Control recalculations Analysis results and their data are dynamically linked and results update immediately to reflect any changes you make to the dataset You can turn this link off temporarily by turning off unchecking the Recalculate control in the upper left corner of the view window When recalculate is on checked all tables and graphs are recalculated when you change the value of a variable in use you delete a variab
319. ow US20 USI39 USI4I Results Selected note us133 violations of control limits sR251 W Wald test sRI70 SRI79 SR209 Wald Wolfowitz runs test sR120 SRI23 Weekday sr430 WeekOfYear sr431 Weibull model sr172 sr173 SRI75 SRI96 Welch s test template us235 Us237 Western Electric rules sr259 Westgard rules sr278 width us229 graphs usi88 tables us199 Us200 Wilcoxon signed rank test sR120 data sRi23 exercise SRI26 Wilks Lambda sr82 Window menu vus148 us149 within subjects sr83 X X axis USI83 X boxes us138 usI64 X usage marker see usage markers X Variable us144 Xbar charts sR255 SR257 SR266 SR269 SR274 XOR SR347 Y Y axis USI83 Y usage marker see usage markers Y Variable us144 Year SR431 yin yang cursor USII US66 USI38 L Zero SR345 z scores SRI3 SR422 z test see paired comparisons
320. owser select Correlation Covariance and click Create Analysis e Accept the default parameters click OK An empty table placeholder X box appears and says that you still need to specify variables e In the variable browser click and drag to select Turning Circle Displacement Horse power and Gas Tank Size e Click Add The analysis calculates and the table updates to show its results Correlation Matrix Turning Circle Displacement Horsepower Gas Tank Size Turning Circle 1 000 747 482 618 Displacement 747 1 000 764 719 Horsepower 482 764 1 000 666 Gas Tank Size 618 719 666 1 000 116 observations were used in this computation Assign variables then create an analysis This time you assign variables first and then create the analysis We add another result to the same view e Make sure that no results are selected have black selection handles by clicking in empty space of the view e In the analysis browser select Regression Notice that the variable browser automatically replaces the Add button with Independent and Dependent buttons You use these buttons to assign variables and specify variable usage simul taneously 152 5 Analyses Exercise Independent Dependent Split By e In the variable browser select Displacement and click Independent e Select Horsepower and click Dependent X and Y usage markers in the variable browser now show the roles Displac
321. owser allows you to 1 Choose open views and open saved views 2 Locate graphs and tables by location analysis or variable 3 Select results one or many at a time The results browser is closed by default To open it if it is closed or bring it to the front if it is hidden select Results Browser from the View menu Windows or select Results from the Window menu Macintosh A preference lets you choose whether the results browser is shown or hidden by default see Application preferences p 225 The View pop up menu lets you select among open views or open other views The name of the currently active view is shown Click on the menu to choose another open view and choose Other to locate and open a previously saved view The Order pop up menu lets you choose how to sort result names in the scrolling list By Analysis sorts results according to the type of analysis By location sorts results by page in order of appearance in the view By variable sorts results according to the variables assigned to them analyses involving more than one variable are listed multiple times in the browser once for each variable Changing browser order has no effect on the order of objects in the view itself Results View Nutrition analysis Order by Analysis Show Lan ANOVA Table for Brand Fisher s PLSD for Bran Interaction Bar Plot for Means Table for Brand vr Box Plot Plot for Calories Brand Plot for
322. ox plots The nutrition guidelines tell us to keep tabs on saturated fat so let s look into this some more 38 I Tutorial Analyze data Create an ANOVA using a template Let s do an analysis of variance anova to see whether the numbers back up our visual inter pretation Our null hypothesis is that there is no significant difference in saturated fat content between brand families and the box plot suggests we ll be able to reject that hypothesis We could create an anova by using the analysis and variable browsers as we ve been doing In fact you can try that right now if you like you can probably figure it out quite easily Instead though we re going to look at one of Stat View s most powerful features templates A template is simply a way of saving a view of analysis and graph results in any combination so that it can be recycled in a future analysis with a twist you can apply the template to differ ent datasets and different variables So where other data analysis packages only allow you to save batches of commands for repeating analyses exactly StatView lets you repeat an analysis strategy with completely different variables no editing no typing no mistakes What s more you can save all your annotations layout and color settings so that the template puts together not just the analysis but your complete presentation StatView ships with dozens and dozens of pre built templates that produce complete analy se
323. paste into their cells Control window panes To hide the summary statistics double click the control To close the attribute pane com pletely double click it again When the pane is closed the pane control looks different FF Double click the control to reopen the attribute pane Opening closing and resizing the attribute pane does not affect the lower data pane of the window Row 1 or the input row in a new dataset always starts right under the attribute pane You can scroll the attribute and data panes independently Summary statistics are calculated only when the attribute pane is open If you need to make many changes to your dataset close the attribute pane to avoid delays from frequent recalcula tions 2 Datasets Enter data 61 Enter values When you have set the attributes for a variable with the controls in the attribute pane you are ready to enter the data for that variable in the dataset The tutorial gives step by step instruc tions for entering a sample dataset here we give general rules e Click a cell to select it e Type a value Press Enter Return or Tab Column 1 Input Column 3 i 3 Source b Class Continuous Continuous a Format Free Format Fi Free Format Fixed a Dec Places 3 a When you enter a value a new input column appears to its right and a new input row appears below it so you can enter another value Also the cursor
324. pecific size not in the pop up menu Your choice takes effect the next time you open StatView Initially hide which browsers Usually the analysis and variable browsers are automatically shown the results browser is only shown when you ask for it Check any browser you dont want to see uncheck any browser you do want to see Use Systems Temporary Folder Check Use Systems Temporary folder if you want tempo rary files stored inside the temporary folder Windows or the System folder s temporary folder Macintosh Uncheck the option if you want to store them in the StatView folder Temporary files are written to store information from memory intensive operations Color Palette preferences Macintosh only StatView can display any color in each slot of the color palette the number of colors available depends on your monitor If you have a monochrome monitor you can still use colors with a color printer Color Palette S Click on a color to edit it Default Cancel On a color system you can edit the color palette 9 Tips and shortcuts Preferences 227 e Click the color slot you want to change e In the standard color picker specify a new color e Click OK Your changes take effect immediately Dataset preferences dataset Preferences H Dataset Preferences Enter key moves Enter key moves Right Right Down Down Does not move Doe
325. pen In the left pane you must provide the counts means and standard deviations on separate rows for each group in your model The easiest way to do this is to copy and paste results from ANOVA means table analysis From the Analyze menu select New View From the analysis browser under anova select Means Table e Click Create Analysis e Click OK to accept the default analysis parameters e Select a continuous variable Weight and click Add Select a nominal variable Gender and click Add 236 9 Tips and shortcuts Dataset Templates Means Table for Weight Effect Gender Count Mean Std Dev Std Err male 71 169 282 23 288 2 764 female 24 127 208 16 208 3 308 Then copy the resulting table as text and paste it into an empty dataset First be sure that you have view preferences set to allow copying the table as text e From the Manage menu select Preferences e Select View and click Modify e If it is unchecked turned off check Copy tables graphs as both text and metafile Win dows or Copy tables graphs as both Text and pict Macintosh and click OK e Click Done e From the Edit menu select Copy be sure the means table is still selected e From the File menu select New e Click in the first empty data cell e From the Edit menu select Paste Column 1 Column 2 Column 3 Column 4 Column 5 String String String String User Entered User Entered User Entered User E
326. ph individually Graph Preferences Default number Axis format Free Format x Axis decimal places Other decimal places I Always have leading digit Default siz Vertical ns inches hes Horizontal 225 inches Order in which to choose points fills and color Fist DECRE ERE jest Fist EOS COOOSS CORSO Default frame Default point Distinguish cells by Point and color x Cancel 9 Tips and shortcuts Preferences 229 Graph Preferences Default size Default numbers Anis format Free Format 7 Vertical inches Free Format T Axis decimal places Horizontal 3 inches Other decimal places Lay Always have leading digit Order in which to choose points fills and colors Default frame First hOooOA xXe ma 0007 FS Last e First RT Last Default point First size Distinguish cells by Default size Specify length of the vertical Y and horizontal X axes by typing values and choosing scale units inches centimeters picas or points Graphs can be resized individually Default numbers Choose the numeric format and number of decimal places to display along axes titles notes and labels To include a zero before the decimal point e
327. plain ascu text files StatView automatically lists any files it can read in the Open dialog box If you prefer to see only a certain file type select that type for List Files of Type Windows or Show Macintosh StatView imports each file into a new dataset If you need to combine several imported files in one dataset or add imported data to an existing dataset you can do so by importing the data copying it into the clipboard and pasting it into the other dataset To prevent accidental file loss the default name for an imported file is the original filename plus imported You can change the name yourself with Save As be sure to supply a unique name to avoid replacing the original file by mistake You can also exchange data with other applications through the clipboard Copy data from one application and Paste it into the other StatView interprets data in the clipboard using the same algorithms it uses for importing files For transferring data among Windows and Macintosh versions of StatView and StatView for SAGENT use the DS Transfer format DataSet Transfer preserves all data formula definitions and values and criteria definitions It does not preserve the current selection the current inclusion and exclusion of rows or custom column widths Microsoft Excel StatView can read and write Excel data files directly Importing and exporting are as simple as opening and saving StatView files See the online document WhatsNew PD
328. plate file is smaller 9 Tips and shortcuts StatView Library 233 Templates appear using view text defaults Usually when you use a template the text of its output is formatted the way it was when you saved the template This is what you would expect a template to do You might however wish to override these settings and use the default text settings from view preferences This is useful when you use templates provided with StatView and want their results to match analyses already in the view Copy tables and graphs Usually when you copy a table or graph a wmr Windows or PICT Macintosh graphic version of that result is copied so that you can paste the result into any application and have it look the same Check this option turn it on to copy the result both as text and as a graphic object one advantage to turning this option on is that you can paste analysis results tables into datasets for further study Double click on table graph Usually double clicking a table a graph in the view is a shortcut to clicking the Edit Display button Right double clicking or Alt double clicking Windows or Option double clicking Macintosh is a shortcut to Edit Analysis You can switch the shortcuts by choosing Edit Analysis for double click StatView Library StatView stores information in a Library in the StatView folder The Library contains all pref erence settings and information the program needs to speed up certain operations for exam pl
329. r by hand or import this is the default Dynamic Formula Data created with Formula or Recode commands A dynamic formula variable updates if the variables it uses change or if there are other changes e g sorting row insertion in the dataset Static Formula Data created by choosing the Static Formula pop up or Series or Random Numbers commands Such formulas do not recalculate automatically although they can be updated manually Changing a dynamic formula to static stops dynamic recalculation of the formula but you can change it back to dynamic at any time Analysis Generated Data generated by an analysis residuals fitted and predicted values from a regression factor scores from a factor analysis Factor Analysis Correlation Covariance Survival Nonparametric and Survival Regression can create a new dataset containing a correlation matrix See the appropriate analysis chapter for a discussion of analysis generated variables Change sources Be careful when changing the source of variables that already have data values Some changes in source can cause loss of information For example changing Formula variables to User Entered retains the current data values but discards permanently the formula definition New source Result of changing from other sources User Entered Values from Dynamic Formulas are saved but the formula definition is lost Values from Static Formulas are saved but the formula definition is lo
330. r edit a definition directly in the text box or you can select part of the definition and make another choice from the scrolling list or value bar If you edit a criterion currently applied to the dataset the dataset immediately reflects the change You can also delete criteria so that they are no longer saved with a dataset From the Manage menu select Edit Apply Criteria e Select one or more criteria e Click Delete You can turn criteria off two ways 1 From the Manage menu select Edit Apply Criteria select No Criteria and click Apply 2 From the Criteria pop up menu in a dataset select No Criteria 130 4 Managing data Edit Apply Criteria Exercise In this exercise you create two criteria one to include only men in analyses and one to include only men with low lipid counts Open Lipid Data from the Sample Data folder From the Manage menu select Create Criteria Double click Gender Double click Double click Male Name the criterion Men Click Apply You have just created a criterion to include only those data for male subjects Now we create one that includes only men with low lipid counts From the Criteria pop up menu select New Double click Gender Double click Double click Male Click after the end of the definition Double click AND Double click Cholesterol Double click lt Click in the value bar when the Selected Value is 200 Name this criterion Low Lipid Males Click Apply From th
331. re compatible Recall that you can set data types in the attribute pane When you paste data of one type into a selected area of a different type the conversion in data type can cause a loss of data See the data loss precautions in the section Save datasets p 70 Location of pasted data You can paste data into four areas of a dataset the input row the input column the intersection of the input row and input column or the body of the dataset When you paste data into the input row the dataset grows to accommodate the new rows However if your source data have more columns than you have highlighted in the input row the extra data are not pasted To paste data into the input row of a dataset highlight the entire row or as many cells across the input row as your source data occupy and choose Paste from the Edit menu When you paste data into the input column the dataset grows to accommodate the new col umn s The type of the new column s is determined by the same procedures used in import ing data To paste data into the input column of a dataset select the entire column or as many cells as your source data occupy and use Paste from the Edit menu If you are pasting into a new dataset simply highlight the input cell The dataset adds the appropriate number of rows and columns to accommodate the data When you paste data into the intersection of the input row and the input column the dataset grows to accommodate the new row
332. recoded variable wrong e Does your original variable have any missing values Look at Missing Cells in the attribute pane Use the Recode command or use a formula with if IsMissing var if you need to recode missing values to a special group such as unknown Consider using 1s and IsNoT instead of and e Does your original variable have more values than you thought Look at Minimum and Maximum in the variable attribute pane For nominal variables which don t have sum mary statistics sort on the variable and look for any unexpected values e Did you use else to set your last group If so any values in the original variable that you forgot end up in that group e Did you forget that strings are case sensitive A and a are different values e Is your original variable a category variable If so be sure to use if then else statements rather than arithmetic functions See if then else p 341 of StatView Reference for some ways to handle these problems 256 9 Tips and shortcuts Troubleshooting Why do get Formula and Criteria windows when open a dataset These open windows indicate definition errors that you need to fix before StatView can com pute the variables Did you rename any variables used in formula or criteria definitions You need to update those formulas or criteria to use the current names Did you edit any category definitions for variables used in formul
333. revent it from being moved or edited accidentally A locked object cannot be moved edited resized or reshaped The black selection handles for a locked object are grayed or dimmed However you can still select and copy objects change variable assign ments and change analysis parameters StatView still recalculates results if their data condi tion changes e Select the object From the Layout menu select Lock To unlock a locked object so you can again move or edit it e Select the object From the Layout menu select unlock 8 Drawing and layout Layout tools 215 Group objects You can select two or more objects and group them together so they act as one object You can group any combination of results drawn shapes and text items If you use Draw tools to make modifications to results you might want to group them to those results so that they move together and stay intact when you Clean Up or rearrange results Unless text additions and shapes are grouped with their tables and graphs they do not move together e Select the objects e From the Layout menu select Group To ungroup grouped objects e Select the group or several groups e From the Layout menu select Ungroup You can also group grouped objects together Ungrouping compound groups works one group level at a time For example if you group a pair of objects then group another pair of objects and finally group the two grouped pairs later ungroup
334. riable s must have for continuous H for nominal and for compact Since markers in the variable browser show which class each variable has it is easy to match variables to slots correctly You cannot put a nominal vari able in a continuous slot or a continuous variable in a nominal slot You can drag entire com pact variables into compact variable slots and you can drag the nominal and continuous portions of compact variables into any slots calling for nominal and continuous variables Template slots are flexible 1 You do not need to use every slot some can remain empty 2 You can assign as many variables to each slot as you want Each slot grows to accommodate as many variables as you assign StatView produces multiple or compound results as needed to analyze all the variables you assign for details see Multiple and compound results p 135 Assign variables to slots You can assign variables to slots in any of several ways 1 Click and drag a variable from the list into a slot 2 Double click a variable to assign it to the highlighted slot Then the next slot is selected and you can double click to assign variables to it 3 Control double click Windows or Command double click Macintosh to assign several variables to the current selected slot 4 Click and drag to move a variable from one slot to another Press the Tab key to move the selection from the currently highlighted slot to the next slot
335. rial Analyze data 35 e Type a new name Calorie groups e Press Enter or Return Grouped box plots Now let s try to learn the reason for Calories bimodality In the quiz above we determined that fat and carbohydrates were not clearly bimodal However their values could still differ between groups or other nutrients could be relevant A grouped box plot is a quick way to examine several possibilities all at once Previously we grouped a box plot of Calories by brand name This time let s examine several variables at once and split it by calorie grouping e Click somewhere in the white space of the view window to be sure no results are selected e In the analysis browser double click Box Plot This is a shortcut for selecting Box Plot and then clicking Create Analysis e In the variable browser click and drag from Total fat g down to Protein g and click Add e In the variable browser click Calorie groups and click Split By Box Plot Variables Split By 5 Calorie groups Inclusion criteria Big Three from Candy Bars Data Add 160 F 1 A A Remove Remove a Split By 140 120 Data Candy 100 7 Order 80 Low Froteing Ola 5 g0 HE High Vitamin A Ey 40 Fl es H Vitamin E eh E on Calcium R 2 n E kon ZRD D af EER r a anaiei inai E FE Bars per day 20 r r r x witamin Total fat rule ETTEN ARE GEA CR ae ae TE Calcium R Sat fatrule S
336. ries variables are based on static formulas and you can edit them the same way you edit formulas You can also use Formula to generate series with dynamic formulas From the Manage menu select Series Double click a series in the scroll list Replace any arguments and edit any default parameters to suit your needs Specify the number of rows and columns to create Click Create 122 4 Managing data Series Exercise DEEE Series of Column 16 BinomialCoetfs CubieSeries 1 0 0 1 ExponentialSeriest 1 FibonacciSeries GeometricSeries 1 2 LinearSeries 1 1 QuadraticSeries 1 0 1 QuartieSeries 1 0 0 0 1 RowMumber Number of rows to create 506 Number of variables to create 1 Series variable definition Attributes Cancel Statview 4 0 This dialog box behaves like a regular window you can resize it use Cut Copy and Paste on the text and change font You can double click the top area beneath the title bar to bring its dataset to the front Select Print from the File menu to print a formula definition Windows only You can move a Series window behind or in front of other windows It is listed in the Window menu where you can select it to bring it to the front You can create any number of rows and columns of the same series By default Series creates one variable with the number of rows currently prese
337. rk Us157 USI59 USI I ReturnChiSquare sR414 ReturnF srqis Index SR StatView Reference US Using Statview 281 ReturnNormal sr4i5 ReturnT sr416 reuse formulas us240 reuse results see template reverse code Likert scale sR348 tho see Harrington Fleming Spearman Right Justify us205 right mouse button usvn right to left evaluation sr328 right justify shapes us218 root nth sR334 square SR420 root curve SRI34 roots greater than one sRI34 Rotate Left Right us190 rotate text US205 axis values UsI90 rotation methods sr133 Round sr416 Round Corners dialog box us206 rounded rectangle tool usz05 us206 rounded squares us205 row and column organization see data orga nization row exclusion see Criteria Include Row Ex clude Row row inclusion row heights us199 US231 row inclusion USIO8 USI24 R325 SR326 R339 SR343 SR344 SR409 multiple datasets us146 subtitles usz9 also see Criteria Include Row Exclude Row row labels usr99 row numbers dimmed Us28 US108 UsI24 SR326 R339 RowNumber sr417 rows selecting us64 transpose into columns vus68 row wise see casewise Roy s Greatest Root sr82 rs range span SR477 rulers us217 S charts SR255 SR259 SR266 SR271 compared to R charts sR259 S usage marker see usage markers sample size SR365 sample statistics SRI save analysis results with view usy USI70 USI7I US232 datasets Us70 tutorial example us41 Excel ustoo file
338. rmula variables to recalculate Criteria can be as simple or complex as you like Criteria are saved when you save the associ ated dataset A list of all defined criteria appears in the dataset s Criteria pop up menu There are four ways to select and define new criteria 1 From the Manage menu select Create Criteria 2 From the Manage menu select Edit Apply Criteria and click New 3 From the dataset Criteria pop up menu select New 4 From the dataset Criteria pop up menu select Random Any method leads you to a Criteria dialog box except the fourth which leads you to a ran dom criterion dialog box 4 Managing data Create criteria 125 Criteria 1 of Investment returns Hae Criteria name rrp sis Criteria definition Select a variable fear S amp P 500 NASDAQ London Index Tokyo Index MS Europe Index MS EAFE Index Shearson Corp LT HoOoeogoao see seist This dialog box behaves like a regular window you can resize it use Cut Copy and Paste on the text change fonts and move the window behind or in front of other windows A Criteria dialog box is listed in the Window menu where you can select it to bring it to the front You can double click the top area beneath the title bar to bring its dataset to the front Select Print from the File menu to print a formula definition A criteria definition consists of three parts a variabl
339. round of the entire view Control click Windows or Command click Macintosh to select both pen and fill colors at once Color preferences You can use Color Preferences to change StatView s color palette StatView can display any color in each slot of the color palette the number of colors available depends on your moni tor If you have a monochrome monitor you can still use colors with a color printer See Color Palette preferences Macintosh only p 226 Layout tools Commands in the Layout menu let you manage page layout and arrange move lock group overlay and clean up objects Layout tools apply also to drawn objects grouped to analysis results 8 Drawing and layout Layout tools 213 Control page layout Drawing Size from the Layout menu lets you control how many pages wide and tall your view is As you produce results StatView automatically adds new pages to the bottom of the view and then starts a second column of pages etc If you want a view to be wide rather than tall you must create the additional pages using the Drawing Size dialog box You can also use Drawing Size to remove blank pages after deleting results e From the Layout menu select Drawing Size e Click and drag to indicate the page layout you want e Click OK Each white rectangle represents one page in your document Pages are added to or removed from the active view when you click OK StatView can handle both portrait and landscap
340. rpose gt Statistical I Tutorial Analyze data 19 The sad truth is we can only have 4 Almond Joys that one had 2 servings per package Accu rate data analysis disappoints coconut and almond lovers everywhere Quiz If youre in a hurry you can skip past these quizzes If you have the time though they ll give you some practice and help you learn more about the functionality you ve learned You re supposed to limit total fat intake to 65 grams per day How many candy bars could you eat if you were only worried about total fat Create a new variable called Total fat rule with this formula 65 Total fat g Serving pkg Now you are reduced to 8 Cup O Golds or 2 Almond Joys or 2 Mr Goodbars But now you could have 16 York Peppermint Patties Only 20 grams of that total fat should be saturated fat How many candy bars could you eat if you were only worried about saturated fat Create a new variable called Sat fat rule with this formula 20 Saturated fat g Serving pkg Yikes More bad news You re down to 4 Cup O Golds 1 Almond Joy or 1 Mr Goodbar And York Peppermint Patties are back down to 8 Most people should try to get 25 grams of fiber per day How many candy bars would that take What s the best choice for fiber Create a new variable called Fiber rule with this for mula 25 Dietary fiber g Serving pkg Clearly candy bars are not a
341. rt attack in patients as high medium or low depending on their measured Hpt Cholesterol values The lower the 4 Managing data Series 121 Series HDL cholesterol values the higher the risk of heart attack Values below 35 denote high risk between 35 and 60 medium risk and 60 and above low risk Open Lipid Data from the Sample Data folder From the Manage menu select Recode Select HDL and click Continuous values to nominal groups Click New Define a category called upt risk with three group labels High risk Medium risk Low risk in order Click Done Type the value 35 and press Enter Windows or Return Mac Type the number 60 and press Enter Windows or Return Mac Click Show Definition to see the corresponding formula 2 SSS Recode of HDL Z Recode definition How to recode this variable TRADE G23 KH Selected Breakpoint then High risk else if HDL 60 W then Medium risk else Low risk Low risk W 60 Medium risk W z5 High risk wW Ee 26 Hide definition Cancel Recode Statview 4 0 gt ay Click Recode The recoded variable is appended to the right side of the dataset Recoded variables are based on dynamic formulas which means that changes or additions to HDL automatically change the recoded variable Series generates new variables with values based on common types of series Se
342. s e Click the appropriate button to assign the variable s 5 Analyses Analysis windows 145 Shift click or click and drag to select several adjacent variables Control click Windows or Command click Macintosh to select several nonadjacent variables Shortcut Double click a variable to assign it to the default role this is equivalent to clicking the topmost button whatever that may be See the Shortcuts card for tips on making other types of variable assignments Usage markers When a variable is assigned to an analysis and that analysis is selected in the view a usage marker in the variable browser indicates which role the variable is playing For example this browser shows that in the selected analysis Age is the Y variable Weight is the X variable and the analysis is split by Gender groups Other variables have no usage markers because they are not involved in the analysis Name Gender 5 Age Y Weight a Cholesterol 3 Triglycerides 23 mm unt SME If you select a different result in the view the usage markers change accordingly Following are some common usage markers you might see X A continuous variable in any analysis An independent variable or covariate in a regression An X horizontal axis variable in a bivariate plot A nominal variable in a frequency distribution or descriptive statistics Y A dependent variable in a regression or ANOVA AY vertical axis variable in a bivariate plot
343. s The differences between nominal and continuous variables are crucial and are discussed under Data class p 50 Class defaults to continuous unless type is category or string in which case class defaults to nominal Continuous Continuous data are measurements that have magnitude and rank and can assume any numerical value over a given interval such as weights of cars or batting averages of baseball players The default class unless type is category or string Real integer long integer currency and date time type variables can be continuous Nominal Nominal data identify group memberships such as the countries in which cars are built or whether baseball players compete in the National League or American League The default when type is category or string Any type variable can be nominal Informative Informative data give case identifications such as model names player names or jersey numbers Any type variable can be informative Informative variables are useful only when viewing the dataset They cannot be used in analyses or formulas change class to nominal or continuous if you need to use the data in formulas or analyses Change classes You can change data class for a variable at any time unless the variable is currently used in an analysis or formula Only data class choices compatible with the current data type selection can be made Other choices are dimmed inactive in the Class pop up menu
344. s for example the anova template we ll use assembles an anova table a means table and a bar chart of the effects and even a Fisher s ptsp Protected Least Significant Difference post hoc test e From the Analyze menu select the aNova t tests submenu and select ANOVA or ANCOVA Analyze New View 38M Rebuild Template List For Beginners ANOVA and t tests ANOVA or ANCOVA d b Correlations gt ANOVA Post Hoc Tests Descriptive Statistics gt Equality of Variances F Test Factor Analysis gt Interaction Bar Chart Graphs gt Interaction Line Chart Nonparametrics gt MANOVA or MANCOVA QC Analyses gt One Group Variance Test Regression gt Repeated Measures ANOVA d Survival Analyses t Test One Group t Test Paired t Test Unpaired e Drag Saturated fat g to the Dependent Variable slot Drag Brand to the Factor s slot e Leave the Covariate slot empty e Click OK I Tutorial Analyze data 39 Assign Variables for ANOVA or ANCOVA Please double click or drag the desired variables into the proper slots in the template Template Yariables Data 50 Candy Bars Order Dataset order Name Serving package 0z package Calories Total fat g Saturated fat g Dietary fiber g Dependent variable Saturated fat g ctor s Brand variate s StatView does all the work putting
345. s as well as some word processing applications Such files contain a PIcT representation of all the information in the view tables graphs all text and drawn objects As a WMF or EMF file Windows only You can also save views as wmr or EMF files for use in drawing or painting applications as well as some word processing applications Such files contain a metafile repre sentation of all the information in the view tables graphs all text and drawn objects Reopen your work You can reopen any view you save From the File menu select either Open or Open View As see below e Select a view file e Click Open Use original variables To open your work exactly as you left it select Open from the File menu Open reopens the view in its own window with its original variable assignments 158 5 Analyses Reopen your work Assign different variables To open a view with different variables either from the same dataset or from a new dataset select Open View As from the File menu When you open a view using Open View As Stat View asks whether to use the original dataset or a different dataset whether to add the view s results to a new view window or the current topmost view window and whether to offer you a chance to Assign Variables differently Opening a View This view should Open original dataset s Be applied to different dataset s Where should work go Create new view Add to top view O
346. s 214 Lock and unlock objects 214 Group objects 215 Overlap and overlay objects 215 Rulers and grid lines 217 Table preferences 231 View preferences 232 StatView Library 233 Example Views and Datasets 233 Dataset Templates 233 Normality Test 233 Compute Bartlett s Test and Compute Welch s Test 235 Make your own dataset templates 237 Common questions 237 Dataset 237 Formulas and criteria 240 QC analysis 244 Survival analysis 245 Troubleshooting 250 General problems 250 Importing 252 Printing 254 Formulas and criteria 254 Tutorial Data analysis the StatView way Data analysis is hard enough without software getting in your way That s why StatView is designed to be easy to use We re not saying your research will be easy If research were easy wed have a cure for the common cold by now We re just saying that you should be able to concentrate on your research instead of your software So we designed StatView to be simple consistent flexible and powerful We think that if you spend just an hour with this tutorial you ll learn everything you need to know to get around in StatView And you ll also know where to look in StatView for the trickier techniques you want to know We designed StatView to do everything you need to do starting with data spreadsheets and going all the way through your project to full color presentations And we made it dynamic so that any changes you
347. s and columns This is the one place where the dataset can grow to accommodate any number of pasted cells To paste data into the intersection high light the single cell where the input row and input column meet and use Paste from the Edit menu You can paste columns of data rows of data or a block of data into existing dataset cells In this case data are only pasted into the highlighted area of the dataset See the earlier section Size of the selected area If you paste nonadjacent data into a dataset be sure to paste rows into rows columns into col umns and blocks into blocks the same size StatView joins nonadjacent data in the Clipboard and cannot preserve the column row structure of copied data unless they are pasted into areas the same shape as their origin This capability is mainly useful for extracting subsets to a dif ferent dataset or for replacing rows or columns in one dataset with nonadjacent discontigu ous rows or columns from another Pasting blocks of data into rows columns or blocks of different shapes should be done with caution 68 2 Datasets Edit data Paste transposed data Data in the Clipboard can be transposed while you paste Use the Paste Transposed command to change entire rows into columns and entire columns into rows For example the following integers in a 3X3 section of a dataset transpose like this 1 2 3 1 4 7 al s el 2 5 3 Lj ela alel s To transpose data choose Paste
348. s as an additional table in the view On the other hand a descriptive statistics table and a correlation matrix analyze multiple variables As you assign new variables the tables simply expand to include the new variables For examples of how this works see Multiple and compound results p 135 To understand what will happen when you assign new variables to an analysis you need to know the number and class of variables it requires See the StatView Reference chapter for the data requirements of each analysis You can force additional variables to appear in their own tables or graphs by cloning an analy sis see below rather than simply adding variables to an existing analysis Clone analyses One shortcut for creating subsequent analyses is to clone an analysis to make another set of results using the same analysis parameters but different variables To clone an analysis e Select a result in the view e Select one or more variables in the browser e Control Shift click Windows or Command Shift click Macintosh the appropriate browser button such as Add An example of this appears earlier in the chapter see Exercise p 132 Shortcut Control Shift double click Windows or Command Shift double click Macin tosh variables to clone an analysis with default topmost button roles Analyze variables in several datasets The variable browser s Data pop up menu lets you choose among several open datasets Using this
349. s folder Macintosh You should not move the template folder itself This main folder should remain at the top level of the StatView folder If you move or rename the template folder StatView asks you to locate it the next time you use a template StatView then remembers the new location and the new name If you change the name or location of the template folder the only way to restore its default name and location is to discard the StatView Library file StatView builds a new Library con taining the default folder name and location Note that you lose all preferences you have set as these are saved in the Library file Update the Analyze menu Any time you add new templates or rearrange templates within the main template folder you must update the Analyze menu e From the Analyze menu select Rebuild Template List StatView examines the folder structure inside the template folder and then builds an hierar chical menu Folders and files are alphabetized Build templates You are not limited to templates provided with StatView You can also make custom templates of your own The previous chapter Analyses p 131 discusses how you can create your own analyses from scratch and save them as templates The next chapter Customizing results p 179 shows how to use Edit Display to format table and graph results Once you are skilled in these areas of StatView you are ready to design basic templates If you want to design
350. s from the Window menu Choose Hints in the Preferences dialog box Manage Choose Hints in the Preferences dialog box menu to hide this window at program start up Note Manage menu to hide this window at program this is a scrolling resizeable window teed un hades Hai late You can also use Hints Preferences to control when the window automatically appears When you are first learning StatView it might be helpful to keep the Hints window visible so you get instant feedback on how to use the program The Hints window displays two different levels of helpful information Interface hints Win dows or Balloon hints Macintosh appear in the Hints window when you click on an item of interest in the dataset a view or a dialog box Similar information is available through Balloon help see below Informational hints offer more detailed information They explain the fol lowing features 1 definitions of the functions in the Formula Recode Series Random Numbers and Crite ria dialog boxes 2 variable usage for template slots in the Assign Variables dialog box 3 dimmed options in analysis dialog boxes when you edit an analysis 4 error messages Help Windows only Help offers brief discussions of StatView s windows menus dialog boxes and commands It also gives step by step instructions on how to complete common tasks in StatView StatView s Help is a hypertext system in which you can navigate freely between topics by click
351. s not move Font facial x Font Geneva Size E 7 size I Decimal places g 7 Decimal places 3 m I Silently accept ambiguous values Silently accept ambiguous values When opening datasets from other platforms When opening datasets from other platforms Recompute dynamic formulas Recompute dynamic formulas Corcel Cancer Enter key moves When you type a value and press Enter on the numeric keypad the cursor either moves down to the cell in the next row of the same column moves right to the same row in the next column or stays in place according to your choice Your choice takes effect immediately While editing you can use cursor arrow keys up down left right to enter a value and move a different direction Font and Size Choose the font and size you want the dataset to use You can type a specific size Your choice takes effect for the next new dataset You can change any individual dataset with the Text menu Decimal places Choose how many decimal places you usually want to see for variables with type real Your choice takes effect for the next new column You can change individual col umns with the Decimal Places pop up menu in the attribute pane Silently accept ambiguous values If you enter an ambiguous date time value such as 8 11 which could mean either August 11 or 8 November StatView can either warn you or silently trust that you know what youre doing When opening datasets from other pl
352. s text only the actual displayed values are saved Be sure to adjust Deci mal Places settings before exporting so that real values are saved with sufficient precision e Open your dataset in StatView e From the File menu select Save As e Choose Text file type e Specify a filename e Click Save e Choose an appropriate separator character e Specify whether to include variable names whether to enclose text values in quotation marks and whether to save category values as integer codes Please specify how to save this text file Separate items with Tabs C Commas Retums C IV Save column names T Enclose text items with quotes I Save Category columns as small integers End line as appropriate for C Macintosh UNIX DOSAWindows Cancel e Click Export Export Please specify how to save this text file Separate items with Tabs Commas Returns EJ Save column names Enclose text items with quotes Save Category columns as small integers Cancel Export End lines Windows only Most Macintosh and unrx applications expect lines in a text file to have just carriage returns CR while most pos and Windows applications expect both car riage returns and line feeds CR LF This option lets you choose which line ending characters to use for the exported file Missing values Some applications require a particular code for missing values Y
353. s the frontmost active window If not click it or select Untitled View 1 from the Window menu Select Save from the File menu Type a filename Nutrition analysis e Click Save StatView also saves everything about a view the analyses and graphs variable assignments the dataset s in use etc We can later reopen the view and continue our analysis resuming right where we left off All objects are still dynamic you can still add and remove variables change analysis parameters and so forth For best results always save datasets first then views Otherwise when you reopen the view StatView might have trouble locating its dataset s Save a template Now suppose you have some candy bar data of your own perhaps you ve collected data on your own favorites Perhaps you live in Japan and would prefer to study Japanese candy bars Perhaps you prefer salty snacks and want to do a similar analysis of potato chips corn chips pretzels and crackers If you were to save your view in the Template Windows or StatView Templates Macintosh folder you could use it as a template to redo this entire analysis on another dataset e Make sure the view is still the frontmost active window If not click it or select Nutrition analysis from the Window menu e Select Save As from the File menu e Navigate to the template folder inside the main StatView folder e Create a new folder named My Projects Windows 3 1 or Windows NT before
354. s to at least 6 decimal places Now because you want to compute the survival estimate for a person with a type A personal ity who consumes 18 cigarettes per day enter 18 and 1 on the first and second rows of the Covariate values variable Note that if the person had a type B personality you would enter 0 in the second row The dataset should then appear as below Baseline Cum Surv Coefficients User Entered User Entered User Entered User Entered t Free Format Fixed Free Format Fi Free Format Fi Pot 18 000 ggo00 399 624 1000 356 000 O f z 4ooo 99 ef e o ef ee 2 423 000 996 ANN AAN anc The final step is to create a formula that uses all of this information to compute the survival estimate From the appendix Algorithms p 433 of StatView Reference you can see that the equation for evaluating a proportional hazards model at particular covariate values is A A b a T z SoC TPT where So T is the baseline survival function So the StatView formula for evaluating this function is the following Baseline Cum Surv e DotProduct Coefficients Covariate values Note that DotProduct is a Mathematical function Now simply click Compute and the results are displayed in Column 5 You can of course rename this formula column whatever you like and then save the dataset If you would also like to plot this computed survival function assign Event Time
355. s to suit your most general needs You should establish through preferences the usual measurements fonts and colors you expect to see You should decide in advance whether you want square or rectangu lar graphs big points or little points differing colors or fill patterns You should decide 180 7 Customizing results Edit Display dialog boxes whether you like tables with borders or without single or double spacing few or many deci mal places If you are planning to publish your results you should set preferences to match any style rules your publication may have established Preferences are discussed in detail in a later chapter Tips and shortcuts p 221 Here we merely call your attention to the choices available See Graph preferences p 228 to learn how to set default height and width numeric for matting for scales axis frames how many edges are drawn around a graph point size and types e g circles squares diamonds stars etc fill patterns blank solid hatched check ered etc and colors Also choose how compound graphs see Multiple and compound results p 135 should distinguish between variables or groups by different point types by different colors or by different fill patterns See Table preferences p 231 to learn how to set numeric formatting borders and line spacing See View preferences p 232 to learn how to control for both tables and gr
356. s untranspose it e Make sure the analysis is still selected e Click the Edit Display button at the top of the view window e Uncheck Transpose rows and columns click the checkbox to remove the check mark e Click OK 26 I Tutorial Analyze data Descriptive Statistics ppelit By Brand Calories Total Calories Adams amp Brooks Calories Annabelle Calories Bit O Honey Calories Brown amp Haley Calories Charms Calories Hershey Calories Just Born Calories Leaf Calories MAM Mars Calories Myerson Calories Nabisco Calories Nestle Calories Pearson Calories Sherwood Calories Standard Calories Tobler Calories Tootsie Calories Weider Hean Std bev Std Error Count Minimum Maximum 243 027 75 125 000 450 000 160 000 1 160 000 160 000 244 000 a 190 000 250 000 200 000 1 200 000 200 000 270 000 1 270 000 270 000 200 000 1 200 000 200 000 267 931 z9 170 000 450 000 220 000 1 220 000 220 000 226 667 3 210 000 240 000 236 250 16 160 000 310 000 230 000 1 230 000 230 000 230 000 1 230 000 230 000 241 667 200 000 280 000 240 000 1 340 000 340 000 215 000 2 200 000 230 000 262 000 1 262 000 262 000 130 000 1 190 000 190 000 210 000 z 190 000 230 000 142 500 24 749 17 500 2 125 000 160 000 This table shows descriptive statistics for the candy bars made under each brand name for e
357. sR405 sets SR336 SR338 SR345 braces sR336 75 variance rule sR134 shapes arcs US205 US207 colors us212 corner center control us205 curves see spline tool ellipses us205 fill patterns us211 line widths us211 lines Us205 ovals us206 pen patterns us211 polygons us205 Us207 US208 rectangles us205 reshape Us206 resize US206 rounded rectangles us2z05 us206 rounded squares us205 select Us204 spline curves us208 us210 squares US205 starting point US205 text US204 US20 shortcuts see StatView Shortcuts card Show us57 USI8I Show Balloons us224 Show Definition us121 Show Grid Lines us217 Show Page Breaks us217 Show pop up menu analysis browser Us142 results browser us148 Show Rulers us217 Show Selection us64 US133 US134 side by side column charts sRr237 sigma limits sR258 sign of coefficients SR54 sign test SRII9 SRI2I exercise SRI25 significance level sr90 discussion sR75 SR76 post hoc tests SR84 also see p value simple logistic regression sR200 SR201 SR207 simple regression sR52 SR64 SR77 Sin SR419 single byte strings manipulating SR379 SR384 SR423 single spacing see line spacing singular matrix SR43 Sinh sr4i9 Size US205 skewness US234 SRG SRIZ slash sR329 SR334 slope sr80 sR200 slots for variables us163 sMC see squared multiple correlation smooth sR391 smoothing see bivariate plots snap to grid us217 us218 Index SR StatView Reference US Using Statview 283
358. saving use File Manager to create a new folder I Tutorial Analyze data Save View as Save in a Template lee E ANOVA and t tests LJOC Analyses Scene E Tempe Correlations L Regression Descriptive Statistics LI Survival Analyses _For Beginners LJ Nonparametrics Factor Analysis ANOVA andttests QC Analyses Graphs Correlations LI Regression Descriptive Statistics C Survival Analyses Factor Analysis N File name Nutrition analysis Graphs Save as type ViewSet ESY z Cancel S Statliew Templates Y Hard disk For Beginners Elect ANOUA t tests Correlations Desktop Descriptive Statistics Factor Analysis My Projects Save View as Cancel Nutrition analysis File Format Statview 4 1 View e Save in the new folder Name of new folder Next we rebuild the Analyze menu so that it offers your new folder and its template e From the Analyze menu select Rebuild Template List And here s the new customized Analyze menu Analyze New View 3M Rebuild Template List For Beginners ANOVA t tests Correlations Descriptive Statistics Factor Analysis Graphs My Projects Nonparametrics QC Analyses Regression Survival Analyses Nutrition analysis rvvwyEsavvvvyVvW Remember two things 1 Saving a template is the same as saving a view except that you put it inside the templates folder
359. sed into one If two or more spaces are grouped together edit those out of the source text file Type a period with a space on either side into the source file where the missing value belongs and import the file again 3 If you have an empty data cell without a missing value symbol you probably chose the 254 Tips and shortcuts Troubleshooting Printing return character to separate values on a line Since duplicate returns are not compressed into a single missing value your data will be imported improperly Check the text file to see if there are any duplicate return characters and remove them Ifyou want to import a variable of mixed data type real integer and string for instance or one with errors choose Make columns with errors have type string in the Import dialog box This setting turns a variable of mixed data type or a variable with field errors into a variable with a String data type You can import the dataset and examine the variable to see what caused the errors Correct the errors and change the variable to the appropriate data type through the attribute pane If you have an inordinately large number of missing values in your dataset check to see if the source text file has formatting characters in it such as dollar signs percent signs etc This is more likely to occur if you import a text file from a spreadsheet application such as Excel If you are importing a dataset with several nominal vari
360. sed on a subset of the data 126 4 Managing data Create criteria Select a variable Double click a variable from the list or type its name to select it A criterion can be based on any continuous or nominal variable in the dataset Select the variable whose values determine whether cases are included or excluded from analyses For example if you want to analyze only those cases whose Weight value is greater than 160 select Weight Choose a comparison operator Next choose a comparison operators For example if you want to analyze only those cases whose Weight value is greater than 160 select gt for greater than lt less than equal to greater than lt gt ore not equal to lt ors less than or equal to gt ore greater than or equal to ElementOf is an element of IS equal to or both are missing values ISNOT not equal to or one value missing and the other is not For a detailed discussion of StatView s logical operators see the chapter Formulas p 315 of StatView Reference What you see for the final step varies according to the variable you select in the first step For a continuous variable you see a value bar for nominal a list of its levels group labels Set a value or range with a value bar A value bar is a linear representation of the range of the variable from its minimum to its max imum Both extrema are labeled Select values to include Oi
361. see Select results p 148 Black handles show that a table is selected You can also check the Result Selected note in the upper right corner of the view or use Show Selection in the Results browser Descriptive Statistics Peseriptive Statistics 1 Result Selected Weight Cholesterol Hean 155 653 Hean 191 232 Std Dew 25 359 Std Dew 35 674 Std Error 2 913 Std Error 3 660 Count Count ss Minimum 107 000 Minimum m5000 Shew setection Maximum 234 000 Maximum 225 000 aw Lee A Statistics Missi Missi ol Results for Cholesterol issing ol issing T Select components If the cursor isn t an arrow first choose the selection tool from the Draw palette To select a component click that component directly Shift click to select several components at once Black selection handles show when components are selected Torrelatipn Matriog ne Cholesterol Triglycerides a Cholesterol 022 1 000 Triglycerides EOR es LDL MEPA _ 489 083 1 000 eS observations were Pred in this computations 7 Customizing results Tables This table shows how to select each table component and which tool to use to change an aspect of the component Component How to select it Tool Borders line spacing and orientation of entire table Click a row or column label or anywhere inside the table Edit Display Line thickness pen pattern and color of entire t
362. select Category and release the mouse button StatView automatically figures out how to define the category by scanning the values present in the column You may examine the definition if you want by selecting Edit categories from the Manage menu I2 I Tutorial Manage data e Change Name from class nominal to informative If you were going to use this dataset for a real analysis you might also want to make some aes thetic adjustments You aren t going to use this dataset so you may prefer to close the file and skip ahead to the next section Open a dataset p 13 e Control click Windows or Command click Macintosh the names Serving pkg Oz pkg Total fat g and Saturated fat g to select all four columns e From the Decimal Places pop up menu select 1 e Click and drag over the variable names for Brand and Name to select both columns e Click and drag the border between their names to widen both columns e Shift click or click and drag over all the numeric variables names to select those columns e Click and drag the border between any two variable names to make all the columns nar rower at once e Double click the pane control to close the attribute pane From the File menu select Save e Change the filename if you want then click Save e Close the file Read a text file In this exercise we will import a plain text asc file Most applications have an option to save in a plain text format Stat
363. ses have already updated themselves to show just the results for the big three brands And so that you don t forget that youre looking at just a subset of your data each object s title now includes Inclusion criteria information Descriptive Statistics Split By Brand Inclusion criteria Big Three from Candy Bars Data Mean Std Dew Std Error Count Minimum Maximum Calories Total 160 000 450 000 Calories Hershey 267 931 50 462 14 341 29 170 000 450 000 160 000 310 000 Calories M MMars 236 250 37 394 9 349 Calories Nestle 231 667 39 200 16 003 200 000 280 000 Descriptive Statistics Split By Brand Inclusion criteria Big Three from Candy Bars Data Mean Std Dew Std Error Count Minimum Maximum Total fat g Total 13 039 5 511 Te 51 1 500 23 000 Total fat g Hershey 14 603 6 140 1 140 23 1 500 23 000 Total fat g M amp M Mars 11 219 4 301 1 075 2 500 20 000 Total fat g Nestle 10 333 1 562 760 8 000 12 000 Descriptive Statistics Split By Brand fnclusion i i Big Three from Candy Bars Data Mean Std Dev Std Error Count Minimum Maximum Saturated fat g Total 6 865 3 242 454 51 ooo 15 000 B Saturated fat g Hershey 8103 2 434 0 000 peo Saturated fat g M amp M Mars 5 000 2 339 i 500 10 000 Saturated fat g Nestle 5 833 1 169 ATT 6 4000 7 000 Adopt variable assignments for a new analysis If a picture is worth a thousand words a graph m
364. sets Compact variables 87 Simple compact variable Entering a simple compact variable a compact variable with only one nominal variable is a two step process 1 First you create a column for each group in this case male and female 2 Second you compact the two columns into a single variable We ll start this exercise with a new empty dataset This isn t necessary You can also add com pact variables to an existing dataset If you do though please remember that the cells in your compact variable cannot necessarily be read the way other cells are they might or might not belong to the cases that fill each row Only repeated measures ANOVA interprets the values as belonging to their rows for other analyses case memberships are assigned as though the sepa rate columns were stacked on top of each other in a single column From the File menu select New Now we name two variables one for each group e Click the cell that reads Input Column e Type a new variable name Male e Press Tab to move to the next variable name e Type a new variable name Female All our cholesterol readings are whole numbers so we give both columns type integer e Shift click both variable names to select both columns From the Type pop up menu for one column select Integer Male Female Input Column b Type Integer Integer Real a Source User Entered User Entered User Entered k Class Continuous Continuous Contin
365. sh are listed in the Analyze menu for fast easy use as templates How does opening a view differ from applying a template Opening a view is the same as opening a document created in any program In Excel you reopen a document and resume working on your spreadsheet In Word you reopen your doc ument and resume writing and editing In StatView you reopen a view and continue working with your analysis results Applying a template is more like opening a stationery document in Word or Excel where the page is blank but certain things such as your company name and address margins headers and font choices are set up But StatView goes even further a template sets up the analyses formatting annotations and everything for any number of analyses and variables For exam ple you could do an entire drug study save the view and then reopen it with data for a differ ent drug You specify which new variable is like each old variable and then StatView redoes the whole analysis with those substitutions How can use templates Because a template is simply a view document that is opened a certain way building a tem plate is simple you just build a view the usual way make any formatting changes and annota tions you want and save it If you want to use it from the Analyze menu save it in the Template folder Windows or Stat View Templates folder Macintosh To use a view saved in other location as a template select Open View As from the
366. sily annotate datasets by adding one or two string variables to the dataset and typ ing your comments in those columns You might record the source of the data and any notes on methodology that you will need to know later For example we could document the Candy Bars Data this way NOTES Brand b Source User Entered User Entered Class a Deo Places 2 2 records brand name and Hershe E 4 information for 75 popular M amp M Mars S candy bars The data were Charms 6 gathered by hand at several M amp M Mars T7 Berkeley California convenience 8 stores in July of 1995 Tobler q Masta To prevent yourself from accidentally trying to use the variable s as nominal data be sure to change the class to informative Also note that you can copy the cells and paste them into your view document to make the same information accessible in a presentation 238 9 Tips and shortcuts Common questions How does StatView use ordering in nominal variables StatView automatically sorts the levels of a nominal variable for its graphs and tables from left to right or bottom to top in graph scales or from top to bottom in tables For numeric nominal variables those with type Real Integer Long Integer Date Time or Currency that means that levels appear from smallest to greatest For string nominals that means levels appear in alphabetic order Category variables are sorted according to the order i
367. simple steps you select a template then you assign variables Stat View does the rest If a view window is the active topmost window template results appear in the window below after any previous results If some other type of window is topmost or if no view is open template results automatically appear in a new view window e Ifyou want to add results to an existing view make that view the active window From the Analyze menu select a template Drag variables from the browser on the right into the assignment slots on the left If necessary use the Data pop up menu to open a dataset or choose among open datasets e Click OK Assign variables to templates After you select a template you must assign variables Fill each assignment slot in the left side of the Assign Variables dialog box by dragging a variable from the browser on the right 6 Templates Use templates 163 Click and drag to assign variables to slots Assign Variables for t Test unpaired Please double click or drag thg desired variables into the proper slots in the template Template Variables Grouping Variable Data Lipid Data _ _ Choose or open 4 dataset Gender f Order Dataset order Sort variables in Measured Variable J Name Ei i H ponade mot the scrolling list Cholesterol Triglycerides Vari
368. sis browser which is a scrolling list of the sta tistical and graphical analyses you can create in StatView To create an analysis just select the analysis from the analysis browser then click the Create Analysis button Some analyses like Frequency Distribution have little triangles in front of their names Click any triangle to reveal a more specific set of choices For example we can get Descriptive Statistics about the candy bars e Click Descriptive Statistics e Click Create Analysis Create Analys Create Analysis N Show ATN Anal Show Al Analyses nayses Order Default he Descriptive tics gt Frequency Distribut gt Percentiles One Samole Analvsis A dialog box asks which descriptive statistics we want Take a look at More choices if you want to see all the descriptives Stat View can do For now we will just compute the basic set of statistics gt Frequency Distrib gt Percentiles One Sample Analy Peden i Benen eet I Tutorial Analyze data 21 Descriptive Statistics Descriptive Statistics Choose which statistics to compute Basic Complete More choices Cancel e Choose Basic e Click OK Choose which statistics to compute Basic Complete More choices Cancel ca Now you should see an empty analysis object with black selection handles indicating that the object is selected escriptive Statistics
369. sort see Sort assign variables us144 dialog box us158 us159 US222 exercises USI50 USI56 from other analyses see adopt templates us162 US164 variables assign from other analyses see adopt assignable causes SR252 SR254 association see correlation covariance asterisk sR79 SR329 SR333 double sr334 attribute pane us58 us60 US73 US80 US255 change attributes uss9 control uss Us60 set attributes uss8 uss9 show uss tutorial example uss autocorrelation sR363 automate analyses see templates Average sR356 average SRI SR388 SR391 AveragelgnoreMissing SR357 avoid errors US162 axis bounds us185 us188 USI9I cell dialog box us192 colors us197 decimal places us188 us192 frames UsI83 USI87 USI97 US229 grid lines us185 USI92 labels us183 Us186 logarithmic and linear scales us192 move USI86 numeric dialog box us190 numeric formats USI92 US229 ordinal dialog box us193 rotate text USI9O select usi85 three types UsI90 tick marks us185 USI9I USI93 transpose USI88 values us183 B background calculation us138 background colors us212 backward stepwise regression see regression Balloon help us48 us221 us222 Us224 bar charts fill patterns us195 frequency distribution sri3 univariate plots sR217 also see cell plots Bartlett s chi square sr138 Bartlett s test of sphericity sr44 Bartlett s test template us235 Us237 baseline cumulative hazard plot sri89 baselin
370. special functions A chord is any combination of simultaneous keystrokes and or mouse actions For example if you hold the Shift key while mouse clicking several variables in the variable browser and then release the key and the mouse button you can select multiple adjacent variables at once We write Shift click to describe this action Or you can Copy the current selection into the clipboard by holding the Control key Windows or the Command key Macintosh and typing the letter c then releasing both keys We write Type Control C Windows or Type Command C Macintosh to describe this action Another example you can select several nonadjacent variables in the variable browser by holding the Control key Windows or the Command key Macintosh and clicking the variable names We write Control click Windows or Command click Macintosh When Windows and Macin tosh chords differ we describe each separately Windows first then Macintosh StatView s keyboard and mouse shortcuts are summarized on the StatView Shortcuts quick reference card Some chords require holding several keys as well as clicking with the mouse For example you can Control Shift Alt double click Windows or Command Shift Option double click Macintosh a variable name in the browser to clone an analysis with the Split By button Overview Keyboard and mouse chords VII Windows only You can perform special functions by using th
371. ssp extension identifies the file to Windows as a StatView dataset in the DataSet Transfer format e Click Save Use PC Exchange or another utility to copy the file onto a floppy disk formatted for Windows Dos or copy it to a Windows volume by using a network connection Then in StatView for Windows e From the File menu select Open Locate the file If you didn t include the ssp extension you must choose All Files from the Files of Type list e Click Open Moving a dataset from Windows to Mac In StatView for Windows e Open the dataset e From the File menu select Save As From the Files of Type list select DataSet Transfer e Type a filename ending in the ssp extension Including the ssp extension identifies the file to PC Exchange as a Stat View dataset in the DataSet Transfer format e Click Save Copy the file onto a floppy disk and use PC Exchange or another utility to read the disk with your Macintosh or copy it to a Macintosh volume by using a network connection Then in StatView for Macintosh From the File menu select Open e Locate the file e Click Open If you would like to be able to open DataSet Transfer ssp files automatically on Macintosh by double clicking them open the PC Exchange control panel and add an assignment match ing ssD files to StatView 72 2 Datasets Close datasets e Open the PC Exchange control panel e Click Add e Type the extension ssp
372. st Changing from Analysis Generated to User Entered breaks the variable s link with the analysis that created it and allows you to save the variable with the dataset An immediate Undo can restore the link Dynamic Formula Changing User Entered to Dynamic Formula replaces all data with values generated by the formula Changing a static formula to dynamic forces values to recalculate when changes to the dataset occur but you can change it back to static at any time Changing from Analysis Generated to Dynamic Formula breaks the variable s link with the analysis that created it and new values are computed by the formula An immediate Undo can restore the link Static Formula Changing User Entered to Static Formula replaces all data with values generated by the formula Changing a Dynamic Formula to Static stops dynamic recalculation of the formula but you can change it back to dynamic at any time Changing from Analysis Generated to Static Formula breaks the variable s link with the analysis that created it and new values are computed by the formula An immediate Undo can restore the link Analysis Generated You cannot change any other source to Analysis Generated Only StatView analyses can create Analysis Generated variables 78 2 Datasets Variable attributes Class Data class describes the information contained in a variable You can assign variables a nomi nal continuous or informative data clas
373. st have type category Usually nominal vari ables can have any type 4 You cannot change the type source or class of a column that is part of a compact variable you must expand the variable change the attributes and then recompact the columns 5 You cannot compact columns of differing types sources or classes You must set matching attributes before compacting If you have not already done so please read the section Categories p 80 The following dis cussions assume you understand categories Build compact variables You already know how to enter data in the usual row and column arrangement you create a nominal variable in one column and you enter group names down the column for each case Then you create a continuous variable in a second column and enter measurements down the column for each case If you have several nominal variables you simply create more columns and fill them with grouping variables However if you want to enter data in a compact variable you need to follow special steps so that Stat View understands that the columns are actually different groups of a single variable In this section we present two exercises in which you learn step by step how to enter the compact variables we examined above First we ll enter the simple example the one with just one grouping variable Gender Then we ll enter the complex example the one where values are broken down both by Gender and by Smoking 2 Data
374. st printers Otherwise lines are single pixel lines Your choice also affects tables graphs or drawn objects that are copied and pasted into another application Your choice takes effect immediately and applies to existing as well as subsequently drawn objects The option does not affect screen display Limit document size Windows Check this option turn it on to limit the number of pages in view documents this prevents out of memory problems when printing by establishing a maximum document size of approximately 54 x54 that should work trouble free for any 300dpi printer Uncheck the option turn it off if you need to print larger documents Macintosh If you save your views as pict files and read them into MacDraw II you need to constrain the document size so it does not exceed the capabilities of MacDraw II Clicking Limit document size to MacDraw II limits the total number of pages in a view to the number of pages that MacDraw II can contain MacDraw drawing size is 9 pages high by 13 pages across on a LaserWriter a bit less than 100 X100 StatView s document size is 23 pages high by 30 pages across about 227 X227 Save analysis results with view By default calculations are saved when you save a view so the results do not need to be recalculated when the file is reopened with its original dataset You can turn this option off to save disk space If you create a view to use as a template it makes sense not to save results so the tem
375. t tutorial example us4 Index SR StatView Reference US Using Statview 275 New View USI32 USI39 USI50 nominal data sR204 bivariate plots sr232 coding sr206 nominal data class us74 US78 US238 US240 SR332 also see category Split By nonconformity variable c u analyses sR302 p np analyses sr290 nonlinear regression SR54 SR55 SR6O7 SR68 also see logistic regression nonparametric tests SRII9 data sR123 data requirements SRI23 dialog box sr122 discussion SRII9 exercises SRI25 Friedman test sRI22 Kendall s tau sri21 Kolmogorov Smirnov test sRI20 Kruskal Wallis test sR121 Mann Whitney U test sriz0 one sample sign test SRII9 paired sign test SRI2I results sRI24 Spearman rank correlation coefficient SRI21 templates SR124 ties SRI24 Wald Wolfowitz runs test sR120 Wilcoxon signed rank test sR120 Norm sr392 normal count sRI5 normal curve us244 normal distribution sR403 SR410 curve on histogram SRIS definition sr4 normality SRI SR416 Normality Test us86 us233 NOT SR345 not enough memory US232 US25I not equal sR341 notation keyboard mouse shortcuts usv1 notation see syntax notched box plots us188 notes USI84 move USI86 tables us199 Now SR393 np charts SR266 SR288 SR293 nth root sR334 null hypothesis sr74 sR75 number of cases sR365 number of seconds see date time functions NumberMissing sr393 NumberOfRows sr394 numerator df sr4r numeric axes USI90 bounds
376. t generate multiple results rather than compound results for the groups of Split By variables formatting information is retained only for the first group of the Split By variable See Multiple and compound results p 135 6 Templates Build templates I7I Exercise Note that many default formatting instructions can be globally set using View Table and Graph preferences Change templates The easiest way to change a template is to Open it from the File menu make your changes and save it again To avoid renaming the variable assignment slots either use the original dataset or create a new dataset with the same generic names This exercise shows you how to build two simple templates One generates a customized graph and the other generates a custom analysis with special parameter settings Customize a graph If you are writing an article for publication you may want all your scattergrams to have exactly the same format You can customize a single scattergram save it in the Template folder Windows or StatView Templates folder Macintosh and use the template to create all the scattergrams There is no need to repeat all the formatting steps for each new scattergram First change one of the preference settings so the template you create uses less hard disk space e From the Manage menu select Preferences e Select View and click Modify e Uncheck Save analysis results with view e Click OK e Click Done When you use
377. t usI95 move USI86 orientation USsI86 show us188 symbols us186 us194 text USI95 Len sR383 length of string sR383 length of vector sR392 leptokurtic srz less than sR337 SR340 less than or equal to sR337 R340 levels see categories Library usg1 Us225 US233 US25I life table method sri52 likelihood ratio test sRI70 SRI79 SR209 Likert scale reverse sr348 limit document size Us232 line charts cell plots sr237 connect lines us188 Line Plot dialog box sr217 univariate plots sR217 also see bivariate plots and univariate plots line patterns graphs usI85 tables us202 line spacing us199 tables us201 US231 line tool usz05 tutorial example us45 line widths us232 graphs usi85 UsI96 shapes vus211 tables us199 Us202 linear algebra DotProduct sr374 Norm sR392 linear axis scale usr92 linear predictor SRI7O sRI78 standard error SRI78 LinearSeries sR384 listwise deletion sr45 Ln sr385 In cumulative hazard function SRI5O SRIGI SRI66 localized versions us256 locally weighted scatterplot smoother see lowess locate results see results browser Lock us214 Log sr385 log odds sr200 logarithmic axis scale us192 logarithmic regression sR54 logarithms SR375 SR385 SR386 also see hyperbolic functions LogB sr386 logical expressions see Criteria Formula logical operators sR328 SR338 SR340 AND SR345 ElementOf sr345 equal sR340 exclusive OR sr347 false sR345 greater t
378. tasets and open new datasets 2 Select variables 3 Create and expand compacted variables To show or hide the variable browser select Variable Browser from the View menu Windows or Variables from the Window menu Macintosh You can also show or hide the variable browser with the WI button in the toolbar Windows or the Ej button in the upper right corner of the dataset and view windows Macintosh An Application Preference lets you choose whether the variable browser is shown or hidden by default see Application prefer ences p 225 You can move the variable browser resize it or close it by using its window controls Buttons at the top of the browser change according to whether a view or dataset window is active When a dataset window is active its buttons let you Show Compact and Expand vari ables When a view window a window where you create analyses is open its buttons let you assign variables to play various roles in analyses Variables Variables Variables Show Compact Remove Rm Expand Split By Expand S em A joo soy Data Cimiano Data Lipid Daa svD m Order Dataset order x Data Lipid DataSVD m Order Dataze Order Dataset order 7 Mame lth Gender ta Gender bade Age icy E Age Anea Weight Weight Age Weight Cholesterol E Cholesterol Cholesterol Cholesterol Triglycerides Triglycerides
379. tcuts Troubleshooting 255 e Do you have quotation marks around variable names containing spaces Around text val y q 8 ues such as e and 7 which are also function names Why do get a column of missing values e Are any variable or function names misspelled e Does the formula variable have the correct type New variables have type real by default you might need to change this to date time category string or currency before you ll see the results you want e Are you missing any parentheses e Are any text values misspelled e Are all arguments listed in the correct order e Do you have quotation marks around variable names containing spaces Around text val y q 8 ues such as e and 7 which are also function names e Do your variable names contain reserved words Function names operators and cate gory level names should not be used in variable names Many examples in this reference break this rule so that dataset illustrations are easier to interpret For a list of function names see the table of contents e Ifyou cant figure out what might be wrong Copy just the columns the formula needs Paste them into a new dataset and try your formula there If it works then the problem has to do with naming conflicts reserved words variable name changes etc If you don t have time to figure out where that naming conflict is Copy the results and Paste them into the full dataset as a user entered variable Why is my
380. te Plots p gt Bivariate Plots gt Cell Plots Box Plot Compare Percenti D QC Subgroup Mea gt QC Individual Mea pac P NP pac ceu D Pareto Analysis Control click Windows or Command click Macintosh any triangle to tip all triangle con trols up or down in one step Descriptive Stati 4 frequency Distrib Percentiles fine GSaranla nat Lompare Fercent QC Subgroup Mea PRC Individual Mpa pO PAN bac C U gt Pareto Analdsis PGC Subgroup Mea Drar Statistics AR Statistics S Statistics gt CUSUM Statisti Summary Table Descriptive Stati p PorFrequency Distrib Summary Table Histogram Score Histog Fie Chart gt Percentiles Nna Sarina grade LPA cn E oE FAC Subgroup Mea a abar Statistics Line Chart Needle Chart Bar Chart Point Chart Results Table e R Statistics gt S Statistics gt CUSUM Statisti Summary Table QC Individual Mea 142 5 Analyses Analysis windows If you select the heading of an item with a triangle control you get the default results for that analysis To determine what those results are just tip the triangle downward and see what is highlighted when you select the heading Show In the middle part of the browser a Show pop up menu lets you select which types of analyses to show in the browser All analyses are shown by default Basic Statistics and Graphs Only simplify the most c
381. te has no effect on the body of the table Shortcut You can use cursor arrow keys to move any selected table or table component If the grid is turned on each keystroke moves the object one grid unit in the direction of the arrow If the grid is turned off each keystroke moves the object one screen pixel For more informa tion about the grid see Turn Grid On Off p 217 Change text items You can change the font size style and color of text inside tables row and column labels titles and notes You can also change the alignment and orientation of table titles and notes You can change text components individually or all at once To format or edit a text component select only that component To select several text items Shift click to select them at the same time To select all text items at once select the entire table Then use Text menu items to change font size style alignment or angle Use the pen color tool in the Draw palette to change the color of text components Use the text tool in the Draw palette to edit the contents of a text item You cannot edit table values or labels with text tools 7 Customizing results Tables 201 If you use the text tool to edit any text that text no longer updates when the analysis updates Change overall structure While StatView produces many different types of analyses in tables they share a common overall structure which can be changed with the Table di
382. te let you choose colors for the pen fill and background the empty view area outside graphs For each component you select the pen and fill color menus change different aspects of the graph Control click Windows or Command click Macintosh to select both pen and fill colors at once 7 Customizing results Tables 197 The table below describes the components that can be selected and what each pop up menu controls Component How to select Pen color changes Fill color changes Frame Click the frame All lines connected to the frame The interior of the graph including axis tick marks Axis Click an axis value Axis values axis line tick marks grid lines Plot Click the plot Points or lines in the plot The fill of bars boxes or pies Regression line or Click the line The line normal curve Legend symbol Click the symbol Points or lines in the plot The fill of bars boxes or pies Legend as a whole Click the interior of the legend Points or lines in the plot legend text The fill of bars boxes or pies Any text item Click the text The text Tables You can edit many aspects of a table s appearance Some formatting changes apply to the entire table others to a particular component of a table To make changes to an entire table you select the entire table and click Edit Display To make other changes you select tables or components and work with Draw palett
383. termined as the arc quarter section that inscribes the rectangle so no matter where you drag the corners of the rectangle the arc you get is just 90 degrees of a circle see the left 8 Drawing and layout Draw tools 207 figure By comparison in Reshape mode the arc is constrained to the same circle but you can choose a bigger or smaller angle of the circle as in the right figures Shortcut Macintosh only Hold the Control key to switch temporarily to Reshape mode and release the key to exit Reshape mode Polygons The polygon tool lets you draw open or closed polygons A closed polygon is one that has no gaps it ends where it starts An open polygon is one with an empty edge Below the left figure is closed and the right is open To draw a polygon e Select the polygon tool e Click and release at the starting point e Click where you want the next vertex e Continue clicking each new vertex e To finish either click the starting point for a closed polygon or double click a final point for an open polygon Preview While youre clicking vertices hold the Alt key Windows or Option key Macin tosh to preview what a closed polygon would look like in other words to see what the final shape would look like if you were to finish by clicking the starting point To force closure hold the key while double clicking the final vertex Drag selection handles Click and drag any of the eight black selection handles to
384. th the results browser This browser is just like the analy sis and variable browsers we ve already seen and it lets you work with analysis results From the View menu Windows or Window menu Macintosh select Results browser e In the results browser select By Variable for the Order pop up menu 44 I Tutorial Present results View Nutrition analysis Order by Analysis Show by Location bynariable 4NOV A Table for Br Fisher s PLSD for B Interaction Bar Plot Means Table for Bra v Box Plot Plot for Calories B Plot for Saturated f Plot for Total fat g Plot for Total fat g J lt 7 Descriptive Statistics Your list of results may be somewhat different since Quiz sections are optional e Resize the browser to make it wide enough for its entries Scroll down to the heading Saturated fat g Candy Bars Data 2 e Click this heading to select all its entries You may instead Control click Windows or Command click Macintosh individual entries underneath it if you only want to highlight certain results e Click the Select button Results View Nutrition analysis Order by Yariable Show Al Box Plot Plot for Total fat g Saturated fat g Choleste Protein g Candy Bars Data 2 Box Plot Plot for Total fat g Saturated fat g Choleste at dy B J Frequency Distribution Histogram for Saturated fat g Regression Summary Table for Saturated fat g
385. the parts of the window Untitled Dataset 1 Jo x untitled Dataset 1 Compact Expand Criteria No Criteria Compact Expand Criteria No Criteria Input Col Input Column ee a d Type Real gt Type Real b Source User Entered gt Source User Entered b Class Continuous b Class Continuous Ld Format Free Format Fixed ni gt Format Free Format Fixed Dec Places 3 z gt Dec Places 3 z k a am First notice the top row of variable names Right now we have only one name Input Col umn Lers change that to be our first variable name Brand e Click the Input Column cell to select it e Type a new name Brand e Press Enter or Return DOO OE Brand Input Column Re Real Continuous Continuous F Fi ree Format Fixed p Dec Places 3 TS User Entered User Entered Now you have another empty Input Column I Tutorial Manage data 5 e Click the next column and type Name e Press Tab to move to the next column and type Serving pkg e Name the rest of the columns Oz pkg Calories Total fat g Saturated fat g es o Tun EEA e mE e Real Rea Real Real User Entered User Entered User Entered User Entered b Class Continuous Continuous Continuous Continuous Dec Places 3 3 3 3 3 D Next notice the attribute p
386. the values you just open the file and begin your analysis e From the File menu select Open e Select Candy Bars Data from the Sample Data folder e Click Open A complete dataset with attributes all set and ready to go appears Candy Bars Data Brand Name Serving pkg Oz pkg Calories 1 M M Mars Snickers Peanut Butter 1 2 z10 2 Hershey Cookies n Mint 1 155 230 3 Hershey Cadbury Dairy Milk 3 5 5 220 4 M amp M Mars Snickers Z a7 170 5 Charms Sugar Daddy 1 LF 200 6 M amp M Mars Twix Peanut Butter 1 171 260 T Hershey Twizzler 1 22 130 3 Tobler Toblerone 1 1 23 190 3 Nestle Crunch 1 155 230 10 Hershey Almond Joy 2 3 22 230 11 Sherwood Elana Mint 1 200 12 Hershey Krackel 1 330 13 M amp M Mars M amp Ms Peanut 1 250 14 Bit O Honey Bit 0 Honey 1 200 15 Nestle 100 Grand 1 200 16 Hershey Skor 1 220 Tt Hershey Twix Caramel 1 280 18 M amp M Mars Milky way Lite 1 160 ra M M Mars Mars 1 240 20 Pearson Peanut Nut Roll 1 340 21 Nestle Raisinet 1 200 22 Sherwood Elana M 1 230 Fe am 14 I Tutorial Analyze data Analyze data Sort data One of the most important and overlooked data analysis tools is sorting It would be easier to get a feeling for these candy bars if they were grouped by brand Let s also alphabetize th
387. then edit the display of the plot to use appropriate sym bols for the upper and lower confidence or error values Suppose you have summary data like these below entered from a published table and you want to do a cell plot of group means with standard deviation error bars 1 2250 227 221 453 2 B 2812600 360 266 ajo 3304 833 318 581 sje 3es7 353 293 487 Ordinarily you could use one of the cell plots but they plot means and error bars computed from raw data Since you have only the statistics you must instead compute error values with formulas and then draw a bivariate scattergram e Use formulas to compute the upper and lower error bar values Group mean Group sd Group mean Group sd 242 9 Tips and shortcuts Common questions 2250 227 221 453 2471 680 2028 774 1 2 2812 600 360 266 3 2815 682 170 847 2986 529 2644 835 o 3304833 319 581 3623 414 2986 252 sje 3657353 293487 3950 840 3363 866 From the Analyze menu select New View From the analysis browser under Bivariate Plots select Scattergram and click Create Analy sis e Click OK to accept the default plotting parameters e Assign Group mean Upper and Lower to be Y variables assign Group to be an X variable e Use Draw tools to change plotting symbols Scattergram 4000 3800 H 3600 v e 3400 ar g 3200 4 v Low Upper 8 3000 4 S 2800 J 2600 4 2400 H 220
388. tial group names that are Group for and the integer values e g Group for 57 Underlying values indices are assigned according to increasing numeric value Long integers are changed to initial group names that are Group for and the long integer values e g Group for 57 689 Underlying values indices are assigned according to increasing numeric value Categories can have at most 255 levels Strings are converted to initial group names that are the unique values appearing in the variable Underlying values indices are assigned in alphabetic order Categories can have at most 255 levels Categories can have at most 255 levels Date time values are changed to initial group names that are Group for and the date time values e g Group for 5 10 63 Underlying values indices are assigned according to increasing numeric value Categories can have at most 255 levels Currency values are changed to initial group names that are Group for and the current representation of the currency values according to the Format and Decimal places settings e g Group for 1 234 Underlying values indices are assigned according to increasing numeric value Categories can have at most 255 levels 2 Datasets Variable attributes 77 Source Data source identifies the origin of data in the variable It can be user entered created by a formula or generated by an analysis User Entered Data you ente
389. tinuous Nominal Nominal 17104 1 234 ae For d a 12 1 87 1 1 0 Charlie 1 2 12 2 87 10 001 Bl i2z 3 e 20 001 2 12 4 87 30 001 z 1275 87 40 001 4 yellow Davis a red John 2 12 6 67 50 001 Z 1 ize Fs 60 001 red Coltrane 7 blue Roscoe 2 1278 67 70 001 z 1279 87 50 001 1 12710767 90 001 aE green Mitchell a T The following table shows how StatView assigns each variable type Variable Type Reason Column Integer All values are integer Column 2 Date Time All values are date time Column 3 Currency All values are currency Column 4 Real Although many values are integer the variable does contain one real value Column 5 Integer All values are integer The three missing values do not affect data type 3 Importing and exporting Older StatView products Macintosh only 105 Column 6 String Eight out of ten values are string There are repeated strings red and blue but we did not turn off the option to Import non numeric data as type string Column 7 String Nine out of ten values are string and none are repeated The number 23 7 is merely another string You can transfer more of the information from the text file by importing again with different options e Choose Open from the File menu e Choose Text file type e Select Text File Example and click Open e Uncheck
390. tions evaluate arguments or return results in radians there are 27 radi ans in 360 See Deg ToRad p 372 and RadToDeg p 405 of StatView Reference for help converting between radians and degrees II2 4 Managing data Formula 2 For logical functions 0 is false 1 or any nonzero nonmissing value is true and missing is missing You can use IsMissing 1s and 1sNor for special handling of missing values 3 For the statistical functions unless otherwise indicated the first argument must be a vari able name and remaining arguments can be any value Many functions let you specify which rows of the dataset to use in the calculation AllRows the default OnlyIncluded Rows or OnlyExcludedRows See Row inclusion p 325 of StatView Reference To insert a function in a formula definition double click it in the function browser Calculator keypad The calculator keypad offers the most frequently used operators and functions including the arithmetic operators and and the trigonometric functions sin cos and tan Button Function INV function HYP function INV HYP function sin sin arcsin sinh arcsinh cos cos arccos cosh arccosh tan tan arctan tanh arctanh and AND NOT AND LH r log log 10 In In e x y TA 24 12 or OR NOT OR 2 2 2 r tA r lt 2 gt
391. tions see date time functions time series functions Correlation sR363 Difference sR373 Lag sr383 MovingAverage SR391 times see multiplication date time titles us182 graphs us183 inclusion subtitles us29 move USI86 show vus188 tables us199 tool bar us221 Tool tips us221 us223 transform data see Formula Recode transformations BoxCox sR358 Difference sR373 Ln sr385 Log sr385 LogB sr386 LogOdds sr386 transpose axes USI88 page US213 rows and columns us68 tables us199 US201 tutorial example us23 trend SR373 triangle controls analysis browser Us141 compact variables uss7 us88 US95 USI43 Formula dialog box usi1o us111 templates us163 tutorial example us31 trigonometric functions USII2 ArcCos SR349 ArcCosh sR349 ArcCot sR350 ArcCsc SR35I ArcSec sR352 ArcSin R353 ArcSinh sr354 ArcTan sp355 ArcTanh sr356 Cos SR363 Cosh sR364 Cot SR364 Csc sR366 DegToRad sr372 example dataset sr400 RadToDeg sr4o5 Sec sR418 Sin sR419 Sinh sr4i9 Tan sr426 Tanh sr426 trimmed mean sr2 TrimmedMean sr428 troubleshoot us250 us256 formulas and criteria UsII6 US254 US256 general problems us250 us251 import US252 US254 print us254 Recode us255 system configuration US252 true SR338 SR345 Trunc sR429 truth tables sr339 sR340 t test see paired comparisons unpaired com parisons one sample analysis Tukey Kramer sr87 Turn Grid On Off us217 Index SR StatView Reference US Usi
392. tions were used in this computation 156 5 Analyses Save a view Correlation Matrix Split By Country Cell Other Turning Circle Displacement Horsepower Gas Tank Size Weight Turning Circle 1 000 448 313 564 603 Displacement 448 1 000 845 791 859 Horsepower 313 845 1 000 768 748 Gas Tank Size 564 791 768 1 000 786 Weight 603 859 748 786 1 000 37 observations were used in this computation Correlation Matrix Split By Country Cell USA Turning Circle Displacement Horsepower Gas Tank Size Weight Turning Circle 1 000 679 450 581 765 Displacement 679 1 000 BOT 658 793 Horsepower 450 801 1 000 432 545 Gas Tank Size 581 658 432 1 000 845 Weight 765 793 545 845 1 000 49 observations were used in this computation This is how easy it is to calculate statistics for different subgroups of your data It is a flexible and interactive process and it follows the way you think Best of all you do not need to touch your dataset Save a view You can save a view at any time From the File menu select Save or Save As e Specify a filename and folder location e Choose a file format e Click Save As a view When you save a view in the default format SVV for Windows or StatView View for Mac intosh you save all aspects of your work the analyses the variables used anything you have drawn etc When you reopen the
393. to missing Currency values are rounded to an exact integer number of seconds after the earliest possible date varies by platform Values outside the range 0 lt x lt 4 294 967 295 are converted to missing Category values that match valid date time formats see the Formats menu are interpreted accordingly Other string values are converted to missing Currency Reals are reformatted with no loss of data according to the currency format chosen Integers are reformatted with no loss of data according to the currency format chosen Long integers are reformatted with no loss of data according to the currency format chosen String values that match valid currency formats see the Formats menu are interpreted accordingly Other string values are converted to missing Date time values are interpreted as an exact number of seconds since the earliest possible date varies according to platform and are reformatted according to the currency format chosen Categories are converted to their underlying codes indices with the currency units chosen for Format e g 1 2 3 Category Reals are changed to initial group names that are Group for and the character representation of the real numbers according to the Format and Decimal places settings e g Group for 1 234 Underlying values indices are assigned according to increasing numeric value Categories can have at most 255 levels Integers are changed to ini
394. together all four parts of a complete analysis of variance The anova table is first The p value is well under 0 05 so it looks like we can reject the null hypothesis ANOVA Table for Saturated fat g Inclusion criteria Big Three from Candy Bars Data DF Sum of Squares Mean Square F Value P Value Lambda Power Brand 2 106 516 53 258 6 101 0044 12 202 878 Residual 48 419 023 8 730 The next part of the output is a means table showing that M amp M Mars has the smallest mean You may examine the bar chart of means and confidence intervals and the post hoc test for further details on the analysis An S for significant marks the Fisher s pisp p value for the Hershey M amp M Mars combination the p value is well under 0 05 Since the count for Nestle is considerably smaller than for the other two we shouldn t pay much attention to the other PLSD results Fisher s PLSD for Saturated fat g Effect Brand Significance Level 5 Inclusion criteria Big Three from Candy Bars Data Mean Diff Crit Diff P Value Hershey M amp M Mars 3 103 1 850 0015 S Hershey Nestle 2 270 2 664 0931 M amp M Mars Nestle 833 2 844 5585 Means Table for Saturated fat g Effect Brand Inclusion criteria Big Three from Candy Bars Data Count Mean Std Dev Std Err Hershey 29 8 103 3 434 638 M amp M Mars 16 5 000 2 338 585 Nestle 6 5 833 1 169 477
395. u try examples shown in the manual your results might look a little different from ours Some differences you might see 1 In a dataset variable attribute settings type format number of decimal places etc can cause values to look different 2 We often resize or scroll windows to focus the readers attention on a specific item and we often change display attributes such as plotting symbols line types and colors to accom modate black and white printing 3 Illustrations show both Windows and Macintosh versions of StatView Interface elements such as title bars window sizing controls scroll bars combo boxes or pop up menus vary slightly between platforms If important interface elements differ we show both Windows and Macintosh illustrations side by side Windows first then Macintosh 4 International system configurations can cause numeric currency and date time formatting to differ We use a variety of formats in our examples 5 Be aware that StatView performs numeric calculations in the fullest precision of the machine you are using therefore results can differ slightly among platforms 6 StatView represents missing values by bullets and illustrations in these manuals use bullets for maximum visibility However international versions of StatView represent miss ing values in the dataset and elsewhere with periods Keyboard and mouse chords Often in StatView special keyboard and mouse chords let you perform
396. uccess fully assigned categories for the correct number of subgroups see the picture below Notice that we defined the categories in order gender then smoking and also defined their groups in 2 Datasets Compact vanables 93 order male then female smoker then nonsmoker This is simpler than it might seem just read from left to right in your columns e Click Compact Compact Variables Choose one or more categories to identify this compact variable Compact variable name Cholesterol readings Categories Chosen Unlilied a spipgi Gender bender Smoking status Smoking statis Remove Cells in compact 4 Variables selected 4 Femer Chokes Cancel Compact e Click the variable name Cholesterol readings to select all four columns e Click and drag the borders between group names to make the columns narrower e Double click the attribute pane control F to close the attribute pane Cholesterol readings Male Female Smoker Monsmoker Smoker Monsmoker oa z s e EE posi et Again the variable browser shows a single variable Cholesterol with a triangle indicating that it is a compact variable e Click the triangle to reveal the nominal variables within Variables Variables Show Show Compact Compact Expand Epa Data Complex Compact Variable Order Datazet ord
397. ula of Column 18 Total fat g Saturated fat g Cholesterol g Sodium mg gt Mathematical gt Probabilities gt Random Numbers gt Series gt Special Purpose gt Statistical Now at the right end of the dataset you should see a new variable with a boring name e Click that boring name Column 18 e Type a better name Bars per day e Press Enter or Return Now you have a brand new variable whose values tell you how many of each candy bar you could eat There s only one problem you can t see the Name column anymore Not to worry You can split the dataset window horizontally e Drag the horizontal split pane handle to the right It s the black bar to the left of the horizontal scroll bar Scroll Brand and Name into view on the left side Scroll the right side to the end so you can see Bars per day 18 I Tutorial Analyze data EOE Candy Bars Data SEEEN Brand Name Bars per day Input aami 1 Adams amp Brooks Cup O Gold 12 500 2 Annabelle Abba Zabba 3 000 3 Annabelle Big Hunk 8 696 4 Annabelle Look 10 526 5 Annabelle U No Elue 3 000 6 Annabelle U No Green 3 000 T Bit 0 Honey Bit 0 Honey 10 000 3 Brown amp Haley Almond Roca 7407 3 Charms Sugar Daddy 10 000 10 Hershey Sth Avenue 7 143 11 Hershey Almond Joy 2 696 12 Hershey Bar None 8
398. ult but effectively assign it to several If you used the Group command in the Layout menu to group graphs and tables selecting one of the objects effec tively selects them all The next action you take in the view then applies to all objects in the group See Group objects p 215 Random crashes If the program appears to crash somewhat randomly try throwing away the StatView Library file see StatView Library p 233 Take note that doing this restores all default preference settings Windows only If discarding the Library file does not help try restarting with the plainest possible configuration Boot with a generic set of start up files and be sure to check whether any recent changes to your hardware or network configuration could be involved Macintosh only If discarding the Library file does not help try restarting without loading system extensions This will reveal whether the difficulty is simply an incompatibility between StatView and an extension or control panel in your system To restart without extensions choose Restart from the Special menu then hold down the Shift key until the start up screen appears with the message Extensions off Also be sure to check whether any recent changes to your hardware or network configuration could be involved Problem solving techniques If you encounter a problem using StatView or do not understand something about the pro gram we recommend that you follow these
399. uous a Format Free Format Fixed b Dec Places Now we can enter the values e Click the first input cell Type the first male value 127 e Press Enter Continue entering all the values Your dataset should look like this i Integer Integer Source User Ente User Ente Continuous Continuous 127 193 88 2 Datasets Compact vanables Since we have fewer female values than male the female column has missing values at the bot tom Now we can compact the two columns into one variable e Shift click both variable names to select both columns e Click the Compact button in the top left corner of the window Or you can select the two names in the variable browser and then click the Compact button in the variable browser Criteria e In the Compact Variables dialog box type a variable name Cholesterol readings e Click Compact Compact Variables Name the new compact variable Cholesterol readings More Choices Cancel a e Double click the attribute pane control to close the attribute pane Cholesterol readings That s all there is to it Now take a look at the variable browser Notice that it shows a single variable Cholesterol with a little triangle in front of it The triangle indicates that Choles terol is a compact variable e Click the triangle
400. up labels can be as long as 255 characters They appear in the cells of your dataset and as sub headings below compact variable names and they are used to label analysis and graph results for their groups Type the name exactly as you want to see it in the dataset and in anal ysis results e Enter Dogs in the Group label box and click Add or press Return Dogs appears in the scrolling list e Type a name for the category Category names can be as long as 255 characters The name you give appears only in the Choose Category dialog box it is never visible in your dataset or view windows When you are through the Edit Category dialog box shows your list of group labels Edit Category Create a new category Category name Animals Group labet Dogs K D To replace or rename a label select the group label in the scrolling list type a new name in the Group label box and click Replace To delete a group label select the group label in the scroll ing list and click Delete e Click Done to save the category definition 2 Datasets Categories 83 Enter category data You have two ways to enter data in a column with type category 1 Type the first letter of a group label If the first letter is unique StatView completes the label for you If several labels begin with the same letter type as many letters as necessary for a unique match If no
401. ust be worth a thousand statistics Let s look at a box plot of these variables 30 I Tutorial Analyze data We can do this a number of ways We could create a box plot analysis and add variables to it just as we did to create our first descriptive statistics table Or we could select a table and then adopt its variable assignments for a new analysis e Select the first table by clicking it e From the analysis browser select Box Plot e Click Create Analysis StatView creates a box plot analysis object and then automatically adds the Calories variable StatView also assigns Brand as a Split By variable again so that this box plot is the graphical equivalent of the statistics table Box Plot Split By Brand Inclusion criteria Big Three from Candy Bars Data 500 450 4 a L 400 5 H EJ Hershey M amp M Mars Nestle 350 4 F E Foo 4 200 4 a 150 Calories Quiz Which brand offers the widest variety of calories per serving in its candy bars Examine the box plots The box and whisker for Hershey is more than twice as wide as the other boxes Which brand has the highest calorie per serving candy bar Again in the box plot notice that Hershey s highest point is well above the maxima for the other two brands Which brand has the lowest calorie per serving candy bar M amp M Mars takes the honors here its lowest point is right at the bottom of the graph Is the fat variation similar to
402. val descending sort in that order Generally the dataset should be sorted first by whatever variable will be on the horizontal X axis of your step plot and second by the variable that will be on the vertical axis Sort by group time and survival in that order if you want separate survival function plots for different groups Copy the entire Time variable in the new dataset Insert an empty row before the first row in the new dataset Paste the Time variable back over itself by clicking on the name of the Time variable then choosing Paste from the Edit menu In a view window plot Time as an X variable and Cum survival and confidence limits variables if desired as Y variables on a bivariate line chart Change the point size to for the symbols associated with the lines This removes the sym bols from the plot and gives a clean step function showing only lines Troubleshooting Here we provide tips and troubleshooting information to help you use StatView more effi ciently and solve problems you might have with the application Answers to many questions can be found in the Hints window and these manuals General problems Results appear incorrect 1 Results might seem incorrect if you have too few significant digits displayed in your tables Change the number of decimal places displayed with the Preferences command in the Manage menu Choose Table click Modify and change the default number
403. val curves US249 survival regression analyses sR188 univariate plots sR217 unpaired comparisons sR37 connect lines us188 constants R324 SR325 T SR400 SR375 consultant us162 Index SR StatView Reference US Using Statview 263 contingency coefficient SRI13 contingency tables sri data requirements sRII4 SRII6 dialog box sri14 discussion sRIII exercise SRIIZ results sRIIG templates sRi16 continuous data class us78 control charts SR283 SR284 lines sR266 SR267 control limits 3 sigma rule sR258 QC subgroup measurements analysis sR262 violations sR25I SR252 convert data types sR331 convert values see Recode coprocessor US73 Copy Us65 US66 USIO4 USI8I USI82 as text and picture US233 US236 unusual selection shapes us68 us70 copy analysis with new variables see clone copy variable assignments see adopt corner center control us205 correct errors R330 Correlation sR362 correlation SR3I R43 SR53 data requirements SR47 dialog box sr46 discussion SR43 exercise USI55 SR35 SR48 factor analysis SRI31 Kendall rank sri21 matrix US53 results sR47 Spearman rank sRI21 templates SR47 Cos SR363 Cosh sR364 Cot SR364 Count SR365 count Us6o Covariance SR365 covariance SR43 SR45 data requirements SR47 dialog box sr46 discussion sR43 exercise USI55 SR48 matrix US53 results sR47 templates sr47 covariates SR78 proportional hazards models sr169 survival functions us245 U
404. variable is incorrect for that particular slot e Click OK Horsepower 166 6 Templates Use templates A new view shows this table of basic descriptive statistics Descriptive Statistics Mean Std Dev Std Error Count Minimum Maximum Missing Weight 2957 629 535 664 49 735 116 1695 000 4285 000 0 Turning Circle 38 586 3 132 291 116 32 000 47 000 0 Displacement 158 310 60 409 5 609 116 61 000 350 000 0 Horsepower 130 198 39 822 3 697 116 55 000 278 000 0 Now we use a regression template to determine whether a linear relationship exists between gas tank size and the weight of the car Because a dataset is already open we won t need to choose a new one From the Analyze menu select Regression Regression Simple e Drag Gas Tank Size to the Dependent slot e Drag Weight to the Independent slot e Click OK Assign Variables for Regression Simple Please double click or drag the desired variables into the proper slots in the template Template Variables Dependent Variable Data Car Data Gas Tank Size Order Dataset order Independent Variable Model Deight Country Turning Circle Displacement Horsepower Gas Tank Size kii H Our view is still active so the re
405. variable name or any group label defined in a category You may break both rules if you need to but StatView warns you with an error message For a complete list of function names see Formulas p 315 of StatView Reference For more information about category group labels see Categories p 80 Set attributes Each column in the dataset has specific attributes that describe the variable you enter in the column These attributes are data type source class format and number of decimal places Each of these is discussed in detail in the section Save datasets p 70 You set variable attributes in a part of the data window called the attribute pane The first five lines of this area are pop up menus for each variable attribute The input column of a new dataset begins with default attributes Input Column Class Continuous a Dec Places 3 e Position the cursor over the cell of the attribute you want to change The cursor changes to a pop up menu icon pzd 2 Datasets Enter data 59 Click and hold the mouse button From the pop up menu select the correct attribute Release the mouse button To change the attributes of several columns at once e Select all the columns you want to change Shift click or click and drag to select adjacent columns Control click Windows or Com mand click Macintosh to select nonadjacent columns e Make changes to any one of the selected
406. variables You can restrict your analyses to a subset of your dataset by including and excluding rows or by defining criteria for which rows to include These techniques apply equally to results created with templates A result is a result no matter how you create it You can mix both templates and your own analyses in a single view Finally we discuss how to print views and how to save views as documents or templates There are several ways to build analyses in the view 1 Create an empty analysis object with the analysis browser then use the variable browser to assign variables to the empty object Or assign variables first and then create the analysis 2 Adopt the variable assignments from one analysis to use with a new analysis you choose from the analysis browser 3 Clone a completed analysis to analyze new variables you choose from the variable browser 132 5 Analyses Overview Exercise A simple exercise shows how this works This diagram shows the basic sequence of steps Create an analysis by choosing an analysis from the analysis browser 3 and a variable a from the variable browser Adopt a variable from the histogram for a different analysis from the analysis browser C Clone the box plot with a different variable from the variable browser Untitled View 1 E mi es Recalculate Edit Analysis Edit Display
407. view using Open from the File menu you can resume your work right where you left off If you save a view whose dataset is untitled because you have not saved it StatView cannot find the dataset when you later reopen the view For that reason Stat View warns that you should save the dataset first If you close a dataset whose variables are used in open views those variables are removed from the results StatView warns you and gives you a chance to cancel closing the dataset 5 Analyses Reopen your work 157 As a template You can use any view you save as a template See the next chapter Templates p 161 If you plan to use the view as a template consider saving it in the Template folder Windows or Stat View Templates folder Macintosh or a subfolder of that folder You can then access it from the Analyze menu A view stored elsewhere can be opened as a template through Open View As from the File menu see Assign different variables p 158 To save disk space you might prefer not to save calculated results with a view you plan to use as a template See View preferences p 232 As a text file You can save views as plain ascii text files for use with word processors such as Microsoft Word Text files contain all the text and data from tables but only the titles and notes of graphs As a PICT file Macintosh only You can also save views as pict files for use in drawing or painting applica tion
408. w number 1 of the dataset then row number 2 then 3 and so on If the Age value in row number 1 is 10 the resulting formula for row number 1 is 5 10 divided by 2 Logical formulas can be created with the if then else function such as if Age gt then 10 else Age The result of this formula is 10 for each row of the Age variable with a value greater than 10 Otherwise the result is the existing value of the Age variable in that row Consult the Formulas p 315 of StatView Reference for more examples and further discus sion Here we create a new formula variable to average two variables Triglycerides and Trig 3yrs e Open Lipid Data from the Sample Data folder e From the Manage menu select Formula e Start typing a variable name Triglycerides 4 Managing data Formula 115 Shortcuts As soon as you type Trig the rest of the variable name is filled in You type as few characters as are needed to distinguish a variable from all others e Click on the keypad Triglycerides ae In the Formula window s variable browser double click Trig 3yrs The selected is replaced by the variable name The formula now reads Triglycerides Trig 3yrs Variable names dates and nominal levels of a category appear in quotation marks if they are not purely alphabetic or contain spaces For rules see Quotation marks p 325 of StatView Reference e Select the entire formula e Click t
409. w or hide levels of detail as seen in the picture below For example Frequency Distribution analysis can produce summary tables histograms Z score standardized histograms and pie charts QC Subgroup Measurements is a more complex example It has four categories of measure ments Xbar R S and cusum Statistics and a Summary Table Each category produces sev eral types of results line charts needle charts bar charts point charts and results tables Descriptive Stati gt Frequency Distrib gt Percentiles One Sample Analy Unpaired Compari Correlation Cova gt Regression p gt ANOVA gt Contingency Table Nonparametrics gt Factor Analysis gt Survival Nonpar gt Survival Regres gt Univariate Plots p gt Bivariate Plots gt Cell Plots Box Plot Compare Percenti gt QC Subgroup Mea gt QC Individual Mea DOC PNP poccu D Pareto Analysis Let s work with Frequency Distribution to create a summary table and a histogram of Calo Paired Comparisons Descriptive Stati 4 Iyfrequency Distrib Percentiles fine Garanla nat Lompare Fercent OC Subgroup Mea C Individual Mpa oC PAN bac c u gt Pareto Anaidsis PGC Subgroup Mea Parar Statistics AE Statistics S Statistics p CUSUM Statisti Summary Table Descriptive Stati p Frequency Distrib Summary Table Histogram Score Histog Pie Chart
410. with just one decimal place e Control click Windows or Command click Macintosh the four variable names to select all four columns Calories Integer User Entered Continuous From the Decimal Places pop up menu in any one of the columns select 1 Click and hold the 3 cell in one of the columns select 1 and release the mouse button 10 I Tutorial Manage data Let s also make the Name column wide enough for its values e Click any value in an unselected column to deselect the four columns e Click and hold the border between Name and Serving pkg e Drag the border to the right and release the mouse button Name SeBving String Reali User Entered User Enter Informative Continuous Free Form i Snickers Pean Cookies n Mint Cadbury Dairy Let s close the attribute pane and save the dataset e Double click the pane control to close the attribute pane Brand Name Serving pkg Oz pkg Cal 1 z Hershey Cookies n Mint z Hershey Cadbury Dairy Mik STO 4 5 From the File menu select Save e Specify a filename Candy Bars First 5 e Click Save Let s close this dataset From the File menu select Close Candy Bars First 5 Next we learn how to import data Import data Often you have data already entered in another application StatView can read Excel files directly and it can read plain text asczz files exported
411. wlncluded NOT true Mathematical Abs Average AveragelgnoreMissing Ceil Combinations CumProduct CumSum CumSumSquares Difference Div DotProduct e Erf Factorial Floor Lag Ln Log LogB Mod MovingAverage Norm Percentages Permutations Pi 7c Remainder Round Sqrt Sum SumlgnoreMissing Trunc Probabilities ProbBinomial ProbChiSquare ProbF ProbNormal Probt ReturnChiSquare Returnf ReturnNormal ReturnT Random Numbers RandomBeta RandomBinomial RandomChiSquare RandomExponential Random RandomGamma RandomNormal RandomPoisson Random RandomUniform RandomUniformlnteger Series BinomialCoeffs CubicSeries ExponentialSeries FibonacciSeries GeometricSeries LinearSeries QuadraticSeries QuarticSeries RowNumber Special Purpose 0 EL lt lt or lt gt or 2 gt ChooseArg VariableElement Statistical BoxCox CoeffOfVariation Correlation Count Covariance GeometricMean Groups HarmonicMean LogOdds MAD Maximum Mean Median Minimum Mode NumberMissing NumberOfRows OneGroupChiSquare Percentile Range Rank StandardDeviation StandardError StandardScores SumOfColumn SumOfSquares TrimmedMean Variance Text Concat Find Len Substring Trigonometric ArcCos ArcCosh ArcCot ArcCsc ArcSec ArcSin ArcSinh ArcTan ArcTanh Cos Cosh Cot Csc DegToRad RadToDeg Sec Sin Sinh Tan Tanh Note that 1 Trigonometric func
412. xample 160 is the mean calories per serving for the bars made by Adams amp Brooks of which there is only 1 according to the Count statistic whereas all the different M8 M Mars bars average out to 236 calories per serving The top row of the table shows statistics for the total or for candy bars from all brands combined Clone an analysis with different variables We re still interested in fat so let s clone this table into a new one using the Total fat g variable e Make sure the analysis is still selected e In the variable browser select Total fat g e Control Shift click Windows or Command Shift click Macintosh the Add button I Tutorial Analyze data 27 Descriptive Statistics elit By Brand Total fat g Total otal fat g Adams amp Brooks Total fat g Annabelle otal fat g Bit O Honey Total fat g Brown amp Haley otal fat g Charms Total fat g Hershey otal fat g Just Born Total fat g Leaf otal fat g M amp M Mars Total fat g Myerson otal fat g Nabisco Total fat g Nestle otal fat g Pearson Total fat g Sherwood otal fat g Standard Total fat g Tobler otal fat g Tootsie i Total fat g Weider B Mean Std Dew Std Error Count Minimum Maximum 11 873 5 728 661 TS 0 000 23 000 2 000 1 8 000 8 000 9 600 6 341 3 053 5 3 000 17 000 4 000 1 4 000 4 000 13 000 1 13 000 13 000 2 500 1 2 500 2 500 14 603 6 140 1 140 23 1 500
413. y major graph components Title Major ticks Minor ticks Graph frame axes 150 i Plot points F Legend symbols ota ee e E oe Vo goe So 9 Se S o amp 6 oe o Plot lines g 22 egend text 3 ystolic BP 5 lt Diastolic BP gt a a BS 80 So _ gt eae e Axis labels SS Ee VA gt gt Grid lines 8 ag 2g oL 70 E aiun l 0 j Axis oe 180 200 220 240 Notes eight eee Systolic BP 117 445 038 Weight R 2 025 Diastolic BP 70 272 047 Weight R 2 019 This table explains each component Title The text above a graph that identifies the analysis variables and criteria or row inclusion in effect Frame The set of axes that border a graph a closed rectangle or L shape Interior The area inside a graph but outside the plot X axis The horizontal or X axis abscissa Y axis The vertical or Y axis ordinate Axis values The numbers or words along an axis that label points on the axis Tick marks Small hatch marks along an axis that identify intervals of the scale Axis labels Text below the X axis and left of the Y axis that indicate what variable or quantity is plotted on that axis Plot The points bars lines or boxes that represent a single variable or group in a graph Grid lines Vertical and horizontal dotted lines at each major tick interval Plotted lines Regression lines confidence b
414. you use the latter two arguments be aware that any Include Row Exclude Row and Criteria commands you use will cause these formula variables to recalculate Include and Exclude vs Criteria Include Row and Exclude Row commands are most useful when you are exploring your data For example you might temporarily exclude a few cases that seem to be outliers You can eas 4 Managing data Formula 109 Formula ily adjust which rows are included until you begin to make sense of your data At that point or when you want to work with subsets systematically according to logical rules e g all rows with Weight values between 100 and 300 it is better to use criteria commands Criteria are also preferable when you need to preserve subsets for future use Formula creates variables through algebraic definitions called formulas For example you can create simple formulas to sum two variables or to log a variable Or you can write a complex formula with many arguments Formula variables are dynamically linked to the variables used to define them so if those vari ables change the formula variable automatically updates You can turn this link off by switching the source from Dynamic Formula to Static Formula Warning if you change the names of variables used in formulas you must fix the formulas yourself e From the Manage menu select Formula e Use the browsers and keypad to build an expression Or type an expression directly into the text

Using StatView

Contents

Download Pdf Manuals

Related Search

Related Contents