SYSTAT: AN OVERVIEW
Contents
1. EE CONSTANT 10s 151538 14 322667 5 940408 X 0 174294 0 124232 0 224356 1 000000 Analysis of Variance S ree SS df Mean Sguares BEYE E pave Lue a a eee 4 EE REG Fess LOM 924933923 1 ZN JIo OLS 0000001 Residual i 31 266475 18 Leto Tze xxx WARNING Case 5 is an Outlier Studentized Residual 2 740993 Zs do IgG O el DLA Durbin Watson D Statistic First Order Autocorrelation Information Criteria ALC il x69802 AIC Corrected 73 193825 schwarz s BIC Joa e6 602 1 Conclusion The estimates of the regression coefficients are 10 132 and 0 174 so the regression equation is Y 10 132 0 174X F ratio in the analysis of variance table is used to test the hypothesis that the slope is O or for multiple regressions that all slopes are 0 The F is large when the independent variable s helps to explain the variation in the dependent variable Here there is a significant linear relation between Y and X Thus we reject the hypothesis that the slope of the regression line is zero F ratio 53 502 p value P lt 0 0005 SYSTAT also outputs statistics and warnings for outlier detection and for testing the assumptions in linear regression methodology Logistic Regression Logistic regression describes the relationship between a dichotomous response variable and a set of explanatory predictor or independ2. Symbol and Label You can change the symbol type by using any of SYSTAT s 23 built in symbols Graph Properties mah Lines Options Font Smoother Symbol Surface Type Symbol e Getting Help SYSTAT uses the standard HTML Help system to provide information you need to use SYSTAT and to understand the results This section contains a brief description of the Help system and the kinds of help provided with SYSTAT The best way to find out more about the Help system is to use it You can ask for help in any of these ways 1 177 SYSTAT An Overview Click the button 2 na SYSTAT dialog box This takes you directly to a topic describing the use of the dialog box This is the fastest way to learn how to use a dialog box Right click on any dialog box item and select What s this to get help on that particular item Select Contents or Search from the Help menu For help on commands from the command prompt on the Interactive tab of the Commandspace type HELP command name References Afifi A A May S and Clark V 1984 Computer aided multivariate analysis 4th ed New York Chapman amp Hall Fisher R A 1936 The use of multiple measurments in taxonomic problems Annals of Eugenics 7 179 188 Hand D J Daly F Lunn A D McConway K J and Ostrowski E Editors 1996 A handbook of data sets London C
3. Pearson Chi square 12 645 6 000 0 049 Number of Valid Cases 256 Conclusion Subject to the reservation mentioned in the Warning message we see that there is some association between Education and Depression state p value only just less than 0 05 The association is neither strong nor is the direction of the association vis a vis Education 1s clear Fitting Distributions The Fitting Distributions feature enables you to assess whether the observed data can be modeled by a distribution from a parametric family of distributions with appropriately chosen parameter values Example Fitting of Normal Distribution The data in FOREARMI contains length of forearm in inches from Pearson and Lee 1903 A normal distribution may be an appropriate model to describe the data on the forearm length To fit a normal distribution from the menus choose Analyze Fitting Distributions Continuous Analyze Fitting Distributions Continuous DX Available variable s Selected variable s ARMLENGTH ARMLENGTH Available distributions Selected distributions Beta A Normal Cauchy Chi square Double exponenti v Crlawe gm gt a Parameters Location or mean mu Scale or SD sigmal Save frequencies E oa 1 154 SYSTAT An Overview Add ARMLENGTH to the Selected variable s list Select Distribution as Normal Click OK
4. Editor tab or double click the corresponding node in the tree formed in the Output Organizer gt To Edit Graph Axes For editing graph axes as well as editing the graph as a whole you can use Graph Properties Dialog Box in the Graph Editor To open the Graph Properties Dialog box double click on the Graph Editor You can also right click on the Graph Editor open a menu with item Properties at the top and click Properties to open Graph Properties dialog box Through the Graph Properties dialog box you can modify features of a graph frame axis legend and element lt Gi Startpage Untitled syo rcity syz Graphi VR 300 om Zo X Ye ya LI BIG MAC Side al lt Output Examp Dyna ss gt USE reity syz H gt Interactive Log Untitled x Invoke the Graph Properties dialog box for editing 4 graph HTM T OYR Ti NUM For editing graph axes select the Axes page of the Graph Properties dialog box The Axes dialog enables you to alter the axes of your graphs It has four tabs Display Font Option and Line Suppose now you find that X axis label WORKWEEK is difficult to comprehend and you want to make it more explanatory by changing the label using the Graph Properties dialog box Select the Display tab I 169 SYSTAT An Overview Display tab Graph Properties B iy E Display Font Options Line g Minimum Decimals Maximum F Reverse sc
5. Types Deletion Continuous data Pearson Listwise Distance measures Bray Curtis O Pairwise Rank order data Spearman Unordered data Phi C Save matrix Binary data Positive matching 52 gi OLK 1 142 SYSTAT An Overview The following output is displayed in the Output Editor Results for SPECIES 1 000 Number of Observations 50 Means SEPALLEN SEPALWID PETALLEN PETALWID Pearson Correlation Matrix SEPALLEN SEPALWID PE TALLEN PETALWID PETALLEN SEPALWID SEPALLEN PETALWID SEPALLEN SEPALWID 1 000 0 743 1 000 0 267 0 178 IN 2 2 Scatter Plot Matrix PETALLEN PETALWID LU Ore 6 4 1000 1 143 SYSTAT An Overview Results for SPECIES 2 000 Number of Observations 50 Means SEPALLEN SEPALWID PETALLEN PETALWID Pearson Correlation Matrix SEPALLEN SEPALWID PETALLEN PETALWID PETALLEN SEPALWID SEPALLEN PETALWID SEPALLEN SEPALWID PETALLEN PETALWID 1 000 0 526 1 000 0 754 0 561 1 000 0 546 0 664 0 787 1 000 Scatter Plot Matrix 1 144 SYSTAT An Overview Number of observations 50 Results for SPECIES 3 000 Number of Observations 50 Means SEPALLEN SEPALWID PETALLEN PETALWID Pearson Correlation Matrix SEPALLEN SEPALWID PETALLEN PETALWID a SEPALLEN 1 000 SEPALWID 0 45
6. The output is displayed in the Output Editor ARMLENGTH Normal Variable Name Distribution Estimated Parameter s Location or Mean mu Scale or SD sigma 50 20145 El 164466 Estimation of Parameter s Maximum Likelihood Method Test Results Lower Limit Upper Limit Observed Expected ir 4 LO O00 0 Tl i yel Ele EOYUUU Er 690000 12 124449753 17 690000 E8x220000 1 6 19 802248 184220000 18 750000 29 25 247070 1610000 19290000 22 25 802405 19 23 0000 19 80 00 0 24 2 kyk DOS 19 3810000 20 340000 L 15 8990095 20 340000 15 11 786478 140 140 000000 Chi square Test Statistic 3 849814 Degrees of Freedom 5 p value Oe LO Kolmogorov Smirnov Test Statistic 0 047870 Lillietors Probabilicy 0 554270 shapiro Wilk Test Statistic 2991739 p value Ue 90263 1 155 SYSTAT An Overview Fitted Distribution Count Jeg 13d uol odolg ARMLENGTH Conclusion The above analysis indicates that a normal distribution fits the data well In this case we let SYSTAT estimate the parameters of the normal distribution It is also possible to fit a normal distribution with parameters of your choice in that case you need to enter the values in the parameter edit boxes provided for them in the dialog box Analysis of Variance We used the t test for comparing the mean of one sample with a specified value or for comparing the means of two groups In many situations there is a need to compare several means and to test the significance
7. syc or cmd Log In the Log tab you can view the record of the commands issued during the SYSTAT session through Dialog or in the Interactive mode By default the tabs of Commandspace are arranged in the following order Interactive Log Untitled You can cycle through the three tabs using the following keyboard shortcuts CTRL ALT TAB Shifts focus one tab to the right CTRL4 ALT SHIFT TAB Shifts focus one tab to the left 1 135 SYSTAT An Overview SYSTAT Data Command and Output files Data files You can save data files with SY Z extension Command files A command file is a text file that contains SYSTAT commands Saving your analyses in a command file allows you to repeat them at a later date These files are saved with SY C extension Output files SYSTAT displays statistical and graphical output in the output Editor You can save the output in SYO Rich Text format RTF and HyperText Markup Language format HTM The Data Editor The Data Editor is used for entering editing and saving data Entering data is a straightforward process Editing data includes changing variable names or attributes adding and deleting cases or variables moving variables or cases and correcting data errors SYSTAT imports and exports data in all popular formats including Excel ASCII Text Lotus BMDP Data SPSS SAS StatView Stata Statistica JMP Minitab and S Plus as well as from any ODBC compliant app
8. 000 0 108 31 0 165 0 108 1 000 Conclusion We see that only RACE is significant The likelihood ratio statistic of 10 683 is chi squared with two degrees of freedom and a p value of 0 005 Graphs SYSTAT offers a wide variety of graphical analysis tools that enable better visualization of the data The editing options in SYSTAT allow you to fine tune and change the display of the graph To create Summary charts Density displays Plots click on the graph toolbar menu or select the icon from the Graph toolbox Note Graph menus are available when a data file is in use Example Simple Scatter Plot Let us create a simple scatter plot Consider the following data file In various international cities how long must people work to earn enough to buy a Big Mac How does this time relate to the length of a typical work week We plot BIG_MAC the working time in minutes to buy a Big Mac against WORKWEEK the length of the work week in hours The data are in the RCITY file that has 46 cases one for each city Open the RCITY SYZ data file from DATA folder of main SY STAT directory Note By default the file location is C Program Files SYSTAT 12 Data You can also change the default path To do so from the menus choose Edit gt Options Select the File Locations tab Select the radio button Set custom directories Change the path for Open data To plot Big Mac against WORKWEEK from the menus choose 1 167 SYS
9. 000 1 000 Binary LOGIT Analysis Dependent Variable LOW Input Records 189 Records for Analysis 189 1 165 SYSTAT An Overview Sample Split Category Choices BS SPS pe ee Ems ee E ees 0 REFERENCE 1130 1 RESPONSE 1 59 Total 189 Log Likelihood Iteration History Log Likelihood at Iteration 131 005 Log Likelihood at Iteration2 112 159 Log Likelihood at Iteration3 111 995 Log Likelihood at Iteration4 111 995 Log Likelihood at IterationS 111 995 Log Likelihood 111 995 Information Criteria AIC 229 989 Schwarz s BIC 239 715 Parameter Estimates 95 Confidence Interval Parameter Estimate Standard Error Z p value Lower Upper Beet eee x e E a ETE 1 CONSTANT 1 535 0 380 4 043 0 000 2 278 0 791 2 RACE 1 0 263 0 176 1 501 0 133 0 081 0 607 3 LWD 0 982 0 366 2 681 0 007 0 264 1 700 Odds Ratio Estimates 95 Confidence Interval Parameter Odds Ratio Standard Error Lower Upper ER ar S a e e a ee ee ae la ae a ar 2 RACE 1 301 0 228 0 923 1 836 3 LWD 2 671 0 978 1 302 5 476 Log Likelihood of Constants only Model LL 0 117 336 2 LL N LL O 10 683 df 2 p value 0 005 McFadden s Rho squared 0 046 Cox and Snell R sguare 0 055 Naglekerke s R sguare 0 077 1 166 SYSTAT An Overview Covariance Matrix i l gt 2 eee ee l 0 144 21 0 058 0 031 31 0 023 0 007 0 134 Correlation Matrix i 1 2 3 SEE E E 1 1 000 0 867 0 165 21 0 867 1
10. 0000 75 0000 iz 3 0000 44 0000 13 1 0000 53 0000 ME 2 0000 64 0000 15 3 0000 16 ar 18 19 9 The variable MEASURE is the typing speed using three types of machines The levels 1 2 and 3 correspond to machines ELECTRIC WORD PROCESSOR and PLAIN OLD respectively in the TRIAL column Of course you might like to rename Trial as Equipment and Measure as Speed using the Variable Properties dialog Now let us do one way analysis of variance using the wrapped data To perform One Way ANOVA from the menus choose Analyze Analysis of Variance Estimate Model 1 157 SYSTAT An Overview x FA Analysis of Variance Estimate Model Repeated Measures Options Resampling Dependent s MEASURE Available vanablelzl TRIAL MEASURE Add gt Factors TRIAL Add gt Missing values Lovanatelel Add gt lt Remove Cl Save Residuals Add MEASURE as the Dependent variable Add TRIAL as the Factor Click OK The output is displayed in the Output Editor Effects coding used for categorical variables in model The categorical values encountered during processing are Variables E E eee TRIAL 3 levels 1 case s Levels 1 000000 2000000 3 000000 are deleted due to missing data Dependent Variable MEASURE N 14 Multiple R 992206 Squared Multiple R 0 906811 Analysis of Variance gourc Type IIL 55 df
11. Palette apart from the 48 predefined 1 175 SYSTAT An Overview colors you can access more than 16 million colors using Define Custom Colors Simply specify the RGB Red Green Blue or Hue Sat Lum Hue Saturation Luminosity values use the slider on the right to adjust the shading and press Add to Custom color Suppose you want to highlight the points for SETOSA SPECIES Select Setosa from the drop down list of labels Go to the Elements page Click the Symbols tab m Select suitable options ze Advanced Onik Access Window bip lolx Graph Properties m GE gt al Startpage Unttled syo iris syz Graphi 7 7 SF rare NE a Lines Options Fot T 44 a 5 4 z x iat z Size 1 5 w lt Fil Bound LLI 3k l oun g SPECIES T Color oor O Setosa width 2 i 2 lt Versicolor 0 Verginica SEPALLEN Frome OGRAPM HTM MM TI er m To change the color of the elements in the graph select the option Select color m Select a color from the Color drop down list for each of the y variables m Select the fill pattern from fill tab m Select the symbols from symbol tab 1 176 SYSTAT An Overview Fill To change the fill pattern for the elements in a graph select the option Select fill Select a fill pattern from the Fill Pattern drop down list for each of the y variables Smoother
12. mie El workweek vs bigmack N O O Font Zoom Rotate Layout Options Font Algerian v Style Bold underline na Size 10 v Rotation jo E Color Ez C Background color C Strikethrough Case UPPERCASE C Apply all BIG_MAC Oo O 30 35 40 45 50 Average working hours per week QGRAPH HTM ECHO Graph Editor NUM Thus the Graph Properties Dialog box enables you to edit graphs in various modes Example Fisher s IRIS Data We again use the famous IRIS data set from Fisher and explore it graphically We have already found that SEPALLEN and PETALLEN have the strongest correlation for SPECIES 3 1 e Virginica Now you may want to know are these two variables vary substantially for different species I 171 SYSTAT An Overview Let us try to answer this question graphically Open IRIS from the data folder From the menus choose Graph Scatterplot E Graph Scatterplot Fill Symbol and Label act Available variablels variable s SPECIES rere SEPALLEN SEPALLEN SEPALWID PETALLEN Y variablels PETALWID rove PETALLEN Z variablels Add hemove Grouping variable s Aud SPECIES C Univariate density display on border Histogram Overlay multiple graphs into a single frame Main Options Smoother Residuals C
13. of differences between three or more means from independently sampled populations Example One Way ANOVA This example uses a one way design to compare average typing speeds of three groups of typists Fourteen beginning typists were randomly assigned to three types of machines and given speed tests The following are their typing speeds in words per minute Electric Word processor Plain old 52 67 52 47 73 43 51 70 47 49 75 44 53 64 Does the equipment influence typing performance Ho The average speeds of the three machines are the same H The average speeds of the three machines are not all the same To carry out analysis of variance using the above data we need to reorganize the data in a form suitable for SYSTAT This is done by using the Reshape feature and wrapping the columns as follows Wrapping puts the group variable in one column and the measurement 1 156 SYSTAT An Overview variable in another column Thus we need to wrap the data in two columns for which from the menus choose Data Reshape Wrap Unwrap Data Reshape Wrap Unwrap S Wrap Available yvarable s Wrap vanable s ELECTRIC ELECTRIC WORD_PROC WORD PROC The data file looks as follows TRIAL MEASURE VAR00003 il 1 0000 52 0000 eal 2 0000 67 0000 3 3 0000 52 0000 4 1 0000 47 0000 2 0000 73 0000 3 0000 43 0000 7 1 0000 51 0000 3 2 0000 70 0000 Ta 3 0000 47 0000 10 1 0000 49 0000 11 2
14. 0 157 000 104 000 105 000 4 M0 167 OMI 147 000 114 000 101 L 4 6 6 000 178000 145 000 101 000 35 00 i 7 O00 15 0 168 000 121 U0 Meni TG a B 206 000 180 000 1344 000 105 000 g 4 p0 1ra 000 147 000 115 000 103 0 10 10 000 145 000 136 000 102 000 ll 11 11 000 17400 151 000 SH UD a0 17 12 000 201 000 160 000 119 000 aaow 13 14 000 188 0 1F8 000 105 000 110 i 1 000 14 000 1 00 107000 103 000 14 14 000 144 10 131 000 100 0100 Ez 0M 16 if 18 13 T H ii Z H 7 aT 7 l a a ni Data Vanable Par Heli pres Fl Dr aa HIM NUM The null hypothesis is Ho ua 0 i e there is no difference in the systolic blood pressure of the patients before and after the administration of the drug The alternative hypothesis is H ua gt 0 i e there is positive difference in the systolic blood pressure of the patients between before and after the administration of the drug indicating that the drug has the desired effect To perform paired t test from the menu choose Analyze Hypothesis testing Mean Paired t test Hypothesis Testing Mean Paired t Test PIR Main Resampling Available yarable s Selected variables PATIENT ID A S y SBP_BEFORE SYSBP_BEFOR T AE SYSBP AFTER fg WS DIABP BEFORE erative type tert we ie maz Bonferroni Confidence C Qunn Sidak ORK AddSYSBP BEFORE and SYSBP_AFTER in the Selected variable s list From t
15. 0 300 1 000 5 000 3 400 1 500 0 200 1 000 4 400 2 900 1 400 0 200 1 000 4 900 3 100 1 500 0 100 1 000 5 400 3 700 1 500 0 200 1 000 4 800 3 400 1 600 0 200 1 000 4 800 3 000 1 400 0 100 1 000 4 300 3 000 1 100 0 100 1 000 5 800 4 000 1 200 0 200 1 000 5 700 4 400 1 500 0 400 1 000 5 400 3 900 1 300 0 400 1 000 5 100 3 500 1 400 0 300 1 000 5 700 3 800 1 700 0 300 1 000 5 100 3 800 1 500 0 300 1 000 5 400 3 400 1 700 0 200 v si A Aja m gt Outpu Examp Dynam Data Variable gt USE iris syz i gt 5 Interactive Log Untitled x For Help press F1 QGRAPH HTM NUM gt Statistical Analyses through SYSTAT Descriptive Statistics Descriptive Statistics offers basic statistics and stem and leaf plot for columns as well as rows The basic statistics are number of observations N minimum maximum mean sum trimmed mean geometric mean harmonic mean standard deviation variance coefficient of variation CV range median standard error of mean etc Besides the above options you can perform the Shapiro Wilk test for normality If you have chosen more than one variable you can also compute multivariate statistics like multivariate skewness and multivariate kurtosis and carry out the Henze Zirkler multivariate normality test Example We will use the IRIS data to compute descriptive statistics This data set consists of four measurements made on 50 random samples of Iris flowers from each of the three species
16. 0 4 700054 5 000000 33 500000 5254620 AS00000 18 166667 4 290543 6 000000 Ico rs sie 55 6 067744 32000000 4 400000 4 700054 5 000000 8 500000 53254020 4 000000 13 600000 A 100054 S 000000 124833333 4 290543 6 000000 ae 200 000 Au fOO054 54000000 Ze Sy ol 22 oS Lou SYSTAT An Overview Conclusion In two way ANOVA begin the analysis by looking at the interaction effect The DRUG DISEASE interaction is not significant p 0 396 so shift your focus to the main effects The DRUG effect is significant p lt 0 0005 but the DISEASE effect is not p 0 164 Thus at least one of the drugs differs from the others with respect to blood pressure change but blood pressure change does not vary significantly across diseases Note Along with ANOVA table SYSTAT also displays the Estimates of the model parameters To get the estimates you need to select LONG as the PLENGTH option To do so from the menus choose Edit Options Select the Output tab From the Output results select Length as Long Linear Regression Regression analysis is used to investigate a predictive relationship between a response variable and one or more predictors Example Let us study the relationship between noise exposure predictor or independent variable and hypertension dependent or response variable The following data were collected on Y blood pressure rise in millimeters of mercury and X sound pressure level in decibels DANIOCIM
17. 00 40 000 2 000 3 000 1 000 15 000 1 000 0 000 1 000 C 37 000 2 000 33 000 2 000 4 000 1 000 19 000 4 000 0 000 0 000 1 39 000 1 000 59 000 2 000 2 000 1 000 23 000 4 000 0 000 0 000 C 40 000 1 000 42 000 3 000 5 000 1 000 23 000 4 000 1 000 0 000 C 41 000 1 000 19 000 1 000 3 000 1 000 11 000 4 000 0 000 0 000 C 42 000 1 000 32 000 2 000 7 000 1 000 23 000 4 000 0 000 0 000 Cm PT ER r ann j a a RE PE a PE EA NN mi Data Variable For Help press Fi QGRAPH HTM ECHO SEL BY WGT FRQ ID CAT OVR o NUM To study the relationship between depression and education label the EDUCATN and CASECONT into categories using the Label dialog box To open the Label dialog box from the menus choose Data Label Data Value Labels Available yarable s ID W Value s Label SEX Le Dropout AGE B 3 HS grad MARITAL 4 5 college EDEMEN 6 EMPLOY INCOME RELIGION mi m 2 e Select EDUCATN as the variable Type the value s that require labels Type the label for each specified value Click OK Repeat the process for the variable CASECONT and label the value l as depressed and 0 as normal 1 152 SYSTAT An Overview To tabulate from the menus choose Analyze Tables Two Way 8 Analyze Tables Two Way Main Measures Cell Statistics Resampling Available yariable s Row variables SEX AGE MARITAL EDUCATN EMPLOY lt s amp ID a i lt
18. 2 3 4 2 1 1 44 30 2 3 16 3 1 1 36 31 3 1 1 4 1 1 13 32 3 1 29 5 1 1 19 33 3 1 19 6 1 1 22 34 3 2 11 7 1 2 33 35 3 2 9 8 1 2 26 36 3 2 7 9 1 2 33 37 3 2 1 10 1 2 21 38 3 2 6 11 1 3 31 39 3 3 21 1 159 SYSTAT An Overview 12 1 3 3 40 3 3 1 13 1 3 25 41 3 3 9 14 1 3 25 42 3 3 3 15 1 3 24 43 4 1 24 16 2 1 28 44 4 1 9 17 2 1 23 45 4 1 22 18 2 1 34 46 4 1 2 19 2 1 42 47 4 1 15 20 2 1 13 48 4 2 2 21 2 2 34 49 A 2 12 22 2 2 33 50 4 2 12 23 2 2 31 51 4 2 5 24 2 2 36 22 4 2 16 25 2 3 3 53 4 2 15 26 2 3 26 54 4 3 22 2 2 3 28 55 4 3 7 28 2 3 32 26 4 3 29 o 4 3 5 58 4 3 12 To perform Two way ANOVA from the menus choose Analyze Analysis of Variance Estimate Model FA Analysis of Variance Estimate Model Model Repeated Measure Options Resampling Available varable s Dependent s DRUG SYSINCH DISEASE Lai SYSINCR eral DISEASE fe lt Remove Missing values Covariates a lt Remove Select SYSINCR as the Dependent variable Add DRUG and DISEASE in the Factor list box Click ie 1 160 SYSTAT An Overview The output is displayed in the Output Editor Effects coding used for categorical variables in model The categorical values encountered during processing are Variables DRUG DISEASE Dependent Variable N Multiple R Squared Multiple R a ee SS ee SS O O 4 levels 3 levels Analysis of Variance Source DRUG DISEASE
19. 3 000 3 000 3 000 9 000 1 000 0 000 0 000 C 5 000 2 000 33 000 4 000 3 000 1 000 35 000 1 000 0 000 0 000 C 6 000 1 000 24 000 2 000 3 000 1 000 11 000 1 000 0 000 0 000 C 7 000 2 000 58 000 2 000 2 000 5 000 11 000 1 000 2 000 1 000 1 8 000 1 000 22 000 1 000 3 000 1 000 9 000 1 000 0 000 1 000 z 10 000 1 000 30 000 2 000 2 000 1 000 35 000 4 000 0 000 0 000 C 12 000 2 000 57 000 2 000 3 000 2 000 24 000 1 000 0 000 0 000 C 13 000 1 000 39 000 2 000 2 000 1 000 28 000 1 000 1 000 1 000 C 15 000 2 000 23 000 2 000 3 000 1 000 15 000 2 000 0 000 0 000 C 18 000 2 000 55 000 4 000 2 000 3 000 19 000 1 000 1 000 0 000 1 19 000 2 000 26 000 1 000 6 000 1 000 15 000 1 000 0 000 0 000 C 21 000 2 000 44 000 1 000 3 000 1 000 6 000 2 000 0 000 0 000 C 22 000 2 000 25 000 2 000 3 000 1 000 35 000 1 000 0 000 0 000 C 24 000 2 000 61 000 2 000 3 000 1 000 19 000 2 000 0 000 0 000 C 25 000 2 000 43 000 3 000 3 000 1 000 6 000 1 000 0 000 0 000 C 26 000 2 000 52 000 2 000 2 000 5 000 19 000 2 000 1 000 2 000 1 27 000 2 000 23 000 2 000 3 000 5 000 13 000 1 000 0 000 0 000 C 28 000 1 000 73 000 4 000 2 000 4 000 5 000 2 000 0 000 1 000 z 30 000 2 000 34 000 2 000 3 000 1 000 20 000 1 000 0 000 0 000 C 32 000 2 000 31 000 2 000 4 000 1 000 45 000 4 000 1 000 1 000 C 33 000 1 000 60 000 2 000 3 000 1 000 35 000 1 000 0 000 0 000 C 34 000 2 000 35 000 2 000 3 000 000 23 000 1 000 0 000 1 000 35 000 2 000 56 000 2 000 3 000 2 000 23 000 1 000 0 000 0 000 C 36 000 1 0
20. 7 1 000 PETALLEN 0 864 0 401 L000 PETALWID 04201 Us 026 Ore 1 000 Scatter Plot Matrix PETALLEN SEPALWID SEPALLEN PETALWID Quick Graphs Quick Graphs are graphs which are produced along with numeric output without the user invoking the Graph menu A number of SYSTAT procedures include Quick Graphs The Quick Graphs above are automatically generated when you request correlations with the Quick Graphs options on If you want to turn off the Quick Graph facility Under Edit menu click Options In the Global Options dialog select the Output tab Turn off the Display statistical Quick Graphs option Or you can turn off the Quick Graph facility using the QGRAPH tab in the status bar at the bottom of user interface 1 145 SYSTAT An Overview The above Quick Graphs in this example are in the scatterplot matrix SPLOM In each SPLOM there is one bivariate scatterplot corresponding to each entry in the correlation matrix that follows A univariate histogram for each variable 1s displayed along the diagonal and 75 normal distribution based confidence ellipses are displayed within each plot For species 3 e Virginica the plot of SEPALLEN and PETALLEN has the narrowest ellipse and thus the strongest correlation which is 0 864 Hypothesis Testing SYSTAT provides several parametric tests of hypotheses and confidence intervals for means variances proportions and correlations This section provides examples of th
21. BRWODARAUNWNAHRHKEUNHK OH CO n 100 To perform Linear Regression from the menus choose 1 162 SYSTAT An Overview Analyze Regression Linear Least Squares Regression Linear Least Squares Model Estimation Options Predict Resampling varlablelz D ore Independents Include constant Save Residuals Select Y as the Dependent variable Select X as the Independent variable Click OK The output is displayed in the Output Editor Eigenvalues of Unit Scaled X X 1989028 0010972 Condition Indices 1 000000 15 4603799 Variance Proportions 1 2 L L m ml a CONSTANT 0 005486 0 994514 X 0 005486 0 994514 Dependent Variable 1 Y N 20 Multiple R l 0 865019 Squared Multiple R 0 748257 Adjusted Squared Multiple R 0 734271 Standard Error of Estimate deol 96S Regression Coefficients B X X 1 xX yY 1 163 SYSTAT An Overview Std Effect Coefficient Standard Error Coefficient Tolerance t p value a a 4 CONSTANT 10 151556 1I4994900 0 000000 m O20 0 000078 X 0 174294 0 023829 0 865019 1 000000 7 314472 0 000001 Confidence Interval for Regression Coefficients 95 0 Confidence Interval Effect Coefficient Lower Upper VIF ny eee cee a eee a
22. DRUG DISEASE Error Type L997 io TOS SOU Least Squares Means Factor Level DRUG toa DRUG i DRUG bo DRUG 4 Least Squares Means Factor Level DISEASE 1 21 DISEASE 2 19 DISEASE 3 LS Least Squares Means Factor DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE DRUG DISEASE Level Durbin Watson D Statistic First Order Autocorrelation Information Criteria ATC ke sU 8 be AIC Corrected 458 291085 Schwarz 6 BIC 76 0041127 1 161 00 0 000081 Qa L6 3156 Ue 395546 Levels OOO 0i0 2000000 34000000 4 0000 1 000000 2 000000 3 000000 SYSINCR 58 0 675296 0456024 III SS AE Mean Squares F ratio 471860 3 DIY LS 126 046033 8 7 3046 2 ZU 3025 14882287 2600259 6 im e ob deals 961225 816667 46 LU 152556 S Mean Standard Error N 994444 2 751008 I 000000 lt 3 596 Ze EROI 15 000000 744444 3100558 T2 000000 544444 2 i o ENEAS 16 000000 LS Mean Standard Error N a a a MERE E EE a EE EE 816667 Z 492580 t9 000000 J 15553 2 445986 19 000000 son over oul 2 374380 20 000000 LS Mean Standard Error N 2904333333 4 290543 6 000000 ZO 42250000 5254620 4 000000 20 400000 4 700054 54000000 28 00000
23. Files Did you know Recent Output Files a SYSTAT Output E Untitledi syz You can enable Quick Graphs which display graphs automatically in the process of analysis using the Edit gt 0ptions gt 0utput menu and checking the Display statistical Quick graphs box h ae i Themes x classic systheme Next Tip Default systheme Scratchpad IntrotoStatistics systheme E J Manuals Dut Exa Dyn V Show at startup xis Interactive Log Untitled x For Help press F1 QGRAPH HIM ECH SEL BY WGT FRQ ID CAT OVR NUM I Viewspace has the following tabs Output Editor Graphs and statistical results appear in the Output Editor You can edit print and save the output displayed in the Output Editor Data Editor The Data Editor displays the data in a row by column format Each row is a case and each column 1s a variable You can enter edit view and save data in the Data Editor Graph Editor You can edit and save graphs in the Graph Editor Startpage Startpage window appears in Viewspace as you open SYSTAT It has five sub windows i Recent Files ii Tips iii Themes iv Manuals v Scratchpad You can resize the partition of the Startpage or you can close the startpage for the remainder of the session If you want to view the Data Editor and the Graph editor simultaneously click Window menu or right click in the toolbar area and select Til
24. Mean Squares EEE LO p value a A A EE TRIAL 1469 357143 2 T34 678571 554017051 0 000002 Brror 1512000000 11 Loe he tO Least Squares Means Factor Level LS Mean Standard Error N PE 4 EE TRIAL 1 50 400000 1 656941 5 000000 TRIAL 2 69 800000 1 656941 5 000000 TRIAL 3 46 500000 s Y 4 000000 1 158 SYSTAT An Overview Se 192 519 2676026 Durbin Watson D Statistic First Order Autocorrelation Information Criteria AIC OLU 5594 AIC Corrected 85 469838 Schwarz s BIC 83 581623 Least Sguares Means EQUIPMNT Conclusion We reject the hypothesis as the p value is small The Quick Graph illustrates this finding Although the typists using electric and plain old typewriters have similar average speeds 50 4 and 46 5 respectively the word processor group has a much higher average speed Example Two Way ANOVA Consider the following data from a two factor Drug amp Disease experiment from Afifi and Azen 1972 cited in Neter et al 1996 The dependent variable SYSINCR is the change in systolic blood pressure after administering one of four different drugs to patients with one of three different diseases Patients were assigned randomly to one of the possible drugs The data are stored in the SYSTAT file AF IFT S no DRUG DISEASE SYSINCR S no DRUG DISEASE SYSINCR 1 1 1 42 29
25. Remove EDUCATN Add lt Remove Column variable Add CASECONT Tables Counts _ Percents Row percents Column percents Options C List layout C Save OKK C Counts and percents C Expected counts Deviates C Standardized deviates C Include missing values J Select EDUCATN as the Row variable s and CASECONT as the Column variable Below the Tables check the Counts and the Row percents boxes by CASECONT columns depressed Total Click OK Counts EDUCATN rows normal Dropout 3 Dropout 35 HS grad 80 college 42 college cts Degreet 14 Degreet 7 Total aLla Row Percents EDUCATN Dropour Dropout HS grad college college Degreet De reet Ows normal depressed Lo DU 4 1 29s lr t 1 6 367 1 6 667 1 epe e UM T I Tenno a E Lp S8 t I 44 000 by CASECONT columns Total N 00 000 3 000 00 000 47 000 00 000 92 000 00 000 45 000 00 000 41 000 00 000 14 000 00 000 8 000 00 000 2036 UUU 1 153 SYSTAT An Overview xxx WARNING More than One fifth of the fitted Cells are sparse Frequency lt 5 Significance Tests computed on this table are Suspect Chi square tests of association for EDUCATN and CASECONT Test Statistic ME a ec A GE A R M LE a
26. SYSTAT AN OVERVIEW T Krishnan Cranes Software International Limited Mahatma Gandhi Road Bangalore 560 001 krishnan t systat com 1 Introduction SYSTAT was designed for statistical analysis and graphical presentation of scientific and engineering data In order to use this tutorial knowledge of Windows 95 98 2000 Nt XP would be helpful SYSTAT provides a powerful statistical and graphical analysis system in a new graphical user interface environment using descriptive menus toolbars and dialog boxes It offers numerous Statistical features from simple descriptive statistics to highly sophisticated statistical algorithms Taking advantage of the enhanced user interface and environment SYSTAT offers many major performance enhancements for speed and increased ease of use Simply pointing and clicking the mouse can accomplish most tasks SYSTAT provides extensive use of drag n drop and right click mouse functionality SYSTAT s intuitive Windows interface and flexible command language are designed to make your research more efficient You can quickly locate advanced options through clear comprehensive dialogs SYSTAT also offers a huge data worksheet for powerful data handling SYSTAT handles most of the popular data formats Excel SPSS SAS BMDP MINITAB S Plus Statistica Stata JMP and ASCII All matrix operations and computations are menu driven The Graphics module of SYSTAT 12 is an enhanced version of the existing graphics mod
27. TAT An Overview Graph Plots Scatterplot W Graph Scatterplot Fill symbol and Label surface and Line Style Residuals Coordinates i i i i Available varnablelel varlablelz W REWEEK REGION REGION g lt Remove WORKWEER T varnlablelz BIG MAC i m BIG_MAC Repeated trials LIVECOST Add gt EARNINGS lt Remove ENNE lt Remove E varlablelz Las Matrix columns Display as fo Remove Grouping varlablelel m Remove MultiPlat Univariate density display on border Histogram Overlay multiple graphs into a single frame Select WORWEEEK as the X variable s Select BIG_MACK as the Y variable Click OK The Output Editor displays the following graph SYSTAT Graphi e y ay R _ Startpage Untitied syo roty sy3 Graphi 200 Zoom of O fm pe 1 ve o Ossa f 5 8 Output Exam Dyna USE rcity oys gt interactive Log Untitled x Ge ogh Eder QAH HM 55 ii By Wat FRG ID CA ene m 1 168 SYSTAT An Overview Customization of an existing graph Once you have created a graph you can use the Graph Editor tab to change many of its features without recreating the graph Using the Graph menu you can change the properties such as color axes labels symbols titles and graph size Note To view the graph in the Graph Editor either double click on it or click the Graph
28. ale Tick mark intervals Labeled Tick Unlabeled Pip To enter the new label for the x axes select bottom from the drop down list Change the WORKWEEK in the X axis label to Average working hours per week Click Ok Now the X axis label will be changed into AVERAGE WORKING HOURS PER WEEK If you want to change the labels of other axes also proceed in a similar way Note Using the same dialog box you can specify suitable ranges for different axes using the Minimum and Maximum boxes For a better specification you can specify the number of Tick Mark Intervals you want using the labeled Tick and Unlabeled pip boxes You can also give a title for the graph using the same dialog box Go to the Graph page Click Options tab Check the Title box Enter a new title for your graph say WORKWEEK vs BIG_MACK For a better presentation you may want to color the graph Check Color box and select a suitable color 1 170 SYSTAT An Overview Graph Properties mn gt ZoomRotate Layout Options workweek vs bigmack Background color Coordinate system Rectangular Pa Projection type You can also select a suitable font for the graph title by using the Font option See this graph as an example which is Algerian bold underline uppercase size 10 MEJEJ Startpage Untitled syo V reity syz VGraph1 WORKWEEK VS BIGMACK 300 Graph Properties
29. bel Write Setosa in the Change to box Select 2 from the drop down list and write Versicolor Select 3 from the drop down list and write Virginica 1 174 SYSTAT An Overview P N _ m DE Options Display legend Title SPECIES Wert o E ce G Pa L4 Label 1 l lt 3 Change to Setosa a Location Layout Horizontal Rows SPECIES 5 vertical o Columns Setosa Versicolor Verginica SEPALLEN Prana QRA HIM Co L a Ee Cal DUE MUM In the Graph Editor the legend labels are changed accordingly Note that if you do not want to display legends just uncheck the Display legend checkbox You can also choose the symbols for different SPECIES gt To Edit Appearance of the Graph We have already customized some aspects of the appearance of a graph Here are some more aspects The Variable Properties dialog box will enable you to customize some more aspects Using the Graph Properties dialog box you can change font color symbol style fill pattern etc EEE ee Custom colors JIB BB Hue 120 Hed ME KE ee eee pe Sat 240 Green 255 Define Custom Colors gt gt Colorsolid m Blue Add to Custom Colors SYSTAT allows you to set color for fonts symbol fill symbol boundary tick marks axes lines and elements by choosing a color from the color palette that pops up by pressing of the corresponding color picker button In the Color
30. d One sample t test of DL_READING with 20 Cases Ho Mean 100 00 vs Alternative less than Mean gt O 95 00 Confidence Bound 95 617 Standard Deviation 14 904 t i 3 044 df i 19 p value U 003 1k cx wW O OG ok OO O CO O O cio 60 T 80 30 00 110 120 130 DL READ NG Graph Edir QAP MH Conclusion We observe that the one sided p value is 0 003 which is highly significant Clearly the mean DL u reading for current smokers is significantly lower than 100 DL Paired t test The paired t test assesses the equality of two means in experiments involving paired measurements Example Paired t test To illustrate the paired t test we use the data from Hand et al 1996 The data were collected on the systolic blood pressure of 15 patients MacGregor et al 1979 The interest is to see if there is any difference in the systolic blood pressure of the patients before and after the administration of a drug called captopril The BP data file gives the supine systolic and diastolic blood pressures mm Hg for 15 patients with moderate essential hypertension immediately before and two hours after administering the drug 1 148 SYSTAT An Overview Seartnage Ubvikled sya bp syr Graphi SYEEP BEFORE SYSBP AFTER D ABP BEFORE DABP AFTER 210 000 2071 000 120 000 125 000 7 163 000 165 000 122 000 121 000 3 2 000 187 000 186 000 14 000 1 000 4 4000 160 00
31. e one sample t test and the paired t test One Sample t test The one sample t test is used to test if the mean of the population from which the data set form a sample is equal to a hypothesized value Example One Sample test Let us study the effect of cigarette smoking on the carbon monoxide diffusing capacity DL of the lung Ronald Knudson Walter Klatenborn and Benjamin Burrows found that current smokers had DL readings significantly lower than those of exsmokers or nonsmokers Let us find out if the data indicate that the mean DL u reading for current smokers is significantly lower than 100 DL The carbon monoxide diffusing capacities for a random sample of n 20 are entered in the Data Editor 1 146 SYSTAT An Overview 103 765 08 602 T3003 123 056 91 052 92 295 61 675 gg 6rT a 023 6 014 100 615 a8 017 71 210 02 115 09 222 102 754 106 574 73 194 106 755 90 479 Data Variable To perform one sample t test from the menus choose Analyze Hypothesis testing Mean One Sample t test Mean One Sample t Test Available yvarable s Selected varablelel DL_AEADING DL_READING Add gt Men Alternative type e f J Bonferroni Confidence Dunn Sidak Add DL_Reading to the Selected variable s list Enter Mean 100 From the drop down list select the alternative type as less than Click OK 1 147 SYSTAT An Overview The following output is displaye
32. e or Tile vertically 1 134 SYSTAT An Overview a TAT Gubput Lirvithed rez Apel an karma irsta la Sethe Pist SYSTAT 12 m Did you know Session Start Friday February nd 2007 Fie Untitled aye Fue Yorke an heomallastall older Dalal ey Hate Ra af Va ariah l s m GRAPH VR NUM II Workspace has the following tabs Output Organizer The Output Organizer tab helps primarily to navigate through the results of your statistical analysis You can quickly navigate to specific portions of output without having to use the Output Editor scrollbars Examples The Examples tab enables you to run the examples given in the user manual with just a click of mouse The SYSTAT examples tree consists of folders corresponding to different volumes of user manual and nodes You can also add your own example Dynamic Explorer The Dynamic Explorer can be used to rotate 3 D graphs apply power transformations to values on one or more axes and change the confidence intervals ellipses and kernels in scatter plots By default the Dynamic Explorer appears automatically when the Graph Editor tab is active HI Commandspace has the following tabs Interactive In the Interactive tab you can enter commands at the command prompt gt and issue them by pressing the Enter key Untitled The Untitled tab enables you to run the commands in the batch mode You can open edit submit and save SYSTAT command file
33. e provides a display of joint frequencies of categorical or discrete data to study relationships between two or more variables Using Crosstabulation you can analyze Pared Sanples t test on SSP OORE ve SSOP ATTER with 15 Cages Abe hath eater Mari Mian BYSAP_AEFORE Mean SYSOP_ArTER Mean Deterrence 25 00 ioniidente Bound Standard Deviahon ol merenie I di pwale m m Btartpaps Untied y BP yr Graphi and save frequency tables that are formed by categorical variables Example Contingency Table This example uses questionnaire data from a community survey Afifi et al 2004 The survey was conducted to study depression and help seeking behavior among adults The CESD depression index was constructed by asking people to respond to 20 items The SURVEY2 data file includes a record case for each of the 256 subjects in the sample The data set consists of following variables ID INCOME SAD ENJOY MIND DRINK CHRONIC SEX RELIGION FEARFUL BOTHERED TALKLESS HEALTHY MARITAL AGE BLUE FAILURE NO_EAT UNFRNDLY DOCTOR SEX 1 151 MARITAL DEPRESS AS GOOD EFFORT DISLIKE MEDS AGE EDUCATN LONELY HOPEFUL BADSLEEP TOTAL BED_DAYS EDUC hale oc DIE ga EMPLOY CRY HAPPY GETGOING CASECONT ILLNESS SYSTAT An Overview LONEL 1 000 2 000 45 000 2 000 3 000 1 000 28 000 1 000 0 000 0 000 C 4 000 2 000 50 000
34. eature computes correlations and measures of similarity and distance Example In the previous example we computed basic statistics for SEPALWID We will now compute the correlations between the four variables Often we may want to compute certain statistics separately for each group defined by certain variable s in the data set In this case we may want to examine 1f the correlations are of the same magnitude in the three species SYSTAT facilitates such computations by its By Groups feature Let us use By Groups in the Data menu to request separate results for each level of SPECIES grouping variables From the menus choose Data By Groups 1 141 SYSTAT An Overview Data By Groups Available yvanable s Selected varlablelz EPEL EE SFECIES SEPALLEN SEPALWID PETALLEN PETALWID Exclude missing Turn off als In the By Groups dialog box select SPECIES as variable Click OK Return to the Simple Correlations dialog box Select all the four variables and add it to the Selected variable s list Click OK To compute correlations between pairs of the four variables SEPALLEN SEPALWID PETALLEN and PETALWID from the menus choose Analyze Correlations Simple BH Analyze Correlations Simple Main Options Resampling Available variablels Selected variablefs SPECIES SEPALLEN SEPALLEN gt EN SEPALWID PETALWID PETALLEN PETAL WID
35. ent variables The explanatory variables may be continuous or dummy variables discrete Example Binary Logistic Regression To illustrate the use of binary logistic regression we consider this example from Hosmer and Lemeshow 2000 The purpose is to analyse low infant birth weight LOW as a function of several risk factors I 164 SYSTAT An Overview For the present analysis we are considering only mother s weight during last menstrual period LWT and race RACE 1 white RACE 2 black RACE 3 other The dependent variable is coded 1 for birth weights less than 2500 gms and coded 0 otherwise Instead of considering LWT itself we are taking LWD a dummy variable coded 1 if LWT is less than 110 pounds and coded 0 otherwise Our model is simple regression of LOW on a constant LWD and RACE To perform Logistic regression from the menus choose Analyze Regression Logit Estimate Model Regression Logit Estimate Model Model Category Discrete Choice Options Results Available varnablelzl Dependent DE gene ge lt lt Remove Independenti Cross gt ge lt Remove Conditionalls Add gt Cro lt Remove Include constant Confidence oK Select FALL as the Dependent variable Select DIFFICULTY and SEASON as the Independent variables The categorical values encountered during processing are Variables i Levels PETROES bam ae LOW 2 levels 1 0
36. hapman amp Hall Hosmer D W and Lemeshow S 2000 Applied logistic regression 2nd ed New York John Wiley amp Sons Neter J Kutner M H Nachtsheim C J and Wasserman W 1996 Applied linear regression models Homewood IL Irwin Pearson K and Lee A 1903 On the laws of inheritance in man I Inheritance of physical characters Biometrika 2 357 462 1 178
37. he drop down list select the alternative type as greater than Click OK 1 149 SYSTAT An Overview The output is displayed in the Output Editor Paired Samples t test on SYSBP_BEFORE vs SYSBP_AFTER with 15 Cases Alternative greater than Mean SYSBP_BEFORE gt 6 SO Mean SYSBP AFTER 155 000 Mean Difference 18 955 95 00 Confidence Bound 14 828 Standard Deviation of Difference Fea t x 8 123 Ci 14 p value 0 000 Paired t test 120 SYSBP AFTER SYSBP BEFORE ndex of Case From the above graph it is seen that the systolic blood pressure has decreased after the administration of the drug captopril The test results mean difference 18 933 p 0 000 indicate that the drug captopril reduces the systolic blood pressure You can do the same testing using the Example tab of Workspace as this is already included as an example in Hypothesis testing of Statistics I So for running this example using the Examples tree which is collapsible first click the example tab in Workspace then click Statistics Statistics 1 Hypothesis Testing Paired t Test Then you just double click or right click and select Run 1 150 SYSUAT Uintitbed SYSTAT An Overview gt Halli PS Statistics gt Basir appi ad Sapi i Clnsfication and egr Trees i ul rg rnzer SUSE iris sys H il F EYE ae J Interactive aE recente fr ef Cer Leg United R x C Contingency Table A contingency tabl
38. hing like this SEE X Workspace Startpage Untitled syo Untitled2 syz TA x a L SYSTAT Output Bi Untitledi syz ei CALORIES B Untitled2 syz 240 000 2 Weight Yate 220 000 6 000 3 Healthy Cho 250 000 3 000 4 Stouffer 370 000 19 000 5 Gourmet 440 000 26 000 6 Tyson 330 000 14 000 7 Swanson 300 000 12 000 3 9 10 11 a N j iili gt a Out Exa Dyn Data Variable x gt 7 So v el 5 Interactive Log Untitled os For Help press F1 QGRAPH HTM ECHO SEL BY WGT FRQ ID CAT OVR NUM SERL For saving the data from the menus choose File Save As Importing Data To import IRIS xls data of Excel format from the menus choose File Open Data File name iris Files of type Microsoft Excel EN Cancel From the Files of type drop down list choose Microsoft Excel m Select the IRIS xls file m Select the desired Excel sheet and click OK The data file in the Data Editor should look something like this 1 139 SYSTAT An Overview TAT Output Untitled syz SEPALLEN SEPALWID PETALLEN PETALWID VAFA Untitled2 syz pelican karma InstallFolder 3100 3 300 1 400 0 200 4 900 3 000 1 400 0 200 1 000 4 700 3 200 1 300 0 200 1 000 4 600 3 100 1 500 0 200 1 000 5 000 3 600 1 400 0 200 1 000 5 400 3 900 1 700 0 400 1 000 4 600 3 400 1 400
39. lication Data can be entered or imported in SYSTAT in the following way Entering data Consider the following data that has records about seven dinners from the frozen food section of a grocery store Brand Calories Fat Lean Cuisine 240 5 Weight Watchers 220 6 Healthy Choice 250 3 Stouffer 370 19 Gourmet 440 26 Tyson 330 14 Swanson 300 12 To enter these data into Data Editor from the menus choose File gt New gt Data This opens the Data Editor or clears its contents if it is already open SYSTAT Untitled2 syz File Edit view Utilities Graph Analyze Advanced Quick Access Window Help a Sb S EBD Ee gt AP Startpage Untitled syo Untitled2 syz yk Workspace x E SYSTAT Output H Untitledi syz E Untitled2 syz Gi Interactive Log Untitled e For Help press F1 QGRAPH HTM NUM SCRL cg I 1 136 SYSTAT An Overview Before entering the values of variables you may want to set the properties of these variables using Variable Properties Dialog Box To open Variable Properties Dialog Box form the menus choose Data Variable Properties Or right click VAR in the data editor and select Variable Properties Or you can use CTRL SHIFT P Data Variable Properties Variable name BRAND Variable label Variable type Display options O Numeric sti C Categorical Characters ting Numeric display options Decimal places Comme
40. nts Different dinner brands available in the food section of a grocery lt i amp Save changes while navigating Type BRAND for the name The dollar sign at the end of the variable name indicates that the variable is a string or a character variable as opposed to numeric variable Note Variable names can have up to 256 characters Select String as the Variable type Enter the number of characters in the Characters box In the Comments box you can give any comment or description of the variable if you want As here the variable BRAND is explained Click OK to complete the variable definition for VAR To type CALORIES as Variable name again open the dialog box in the same way Select Numeric as the Variable type Enter the number of characters in the Characters box The decimal point is considered as a character Select the number of Decimal places to display Click OK to complete the variable definition for VAR_2 Repeat this process for the FAT variable selecting Numeric as the variable type or you can do the same in another way Double click VAR or click the Variable tab in data editor to get Variable Editor With Variable Editor you can edit variables directly 1 137 SYSTAT An Overview Fal SWETAT Guipa aripi e Lintitied so Untitled yer s te Unbitledi sya po Untithed2 yz Sting 2 CALORIES CALOQRE Humariz nder
41. of Setosa Versicolor and Virginica coded as 1 2 and 3 respectively The four measurements are Sepal length Sepal width Petal length and Petal width in cm This is a famous data set from Fisher 1936 To calculate basic statistics for the iris data from the menu choose Analyze Basic Statistics 1 140 SYSTAT An Overview Analyze Basic Statistics Main N amp P Tiles Resampling Available yariable s Selected variable s SPECIES SEPALWID SEPALLEN Add gt SEPALWID L_Add gt _ PETALLEN PETALWID Options J ll options N Geometric mean GM C Range Minimum C Harmonic mean HM Variance Maximum C Trimmed mean Skewness C Sum SE of skewness Arithmetic mean AM Median Kurtosis L SE of AM SD SE of kurtosis C Cl of AM 0 95 cv Shapiro Wilk normality test C Anderson Darling normality test Multivariate normality assessment Mardia skewness Mardia kurtosis Henze Zirkler test Save statistics ef ORICA Choose SEPALWID and add it to the Selected variable s list Select N Mean SD Minimum Maximum To check for normality select the Shapiro Wilk normality test option Click OK The following output is displayed in the Output Editor SEPALWID N of cases 150 Minimum 2 000 Maximum 4 400 Mean 3057 Standard Dev 0 436 SW Statistic 0 985 SW P Value 0 101 Correlation The Correlation f
42. oordinates X Axis Y Axis 7 Ay C Mirror Dual MultiPlot oaa Select SEPALLEN as the X variable s and PETALLEN as the Y variable s Select SPECIES as the Grouping variable s Click OK The Output Editor displays the following graph Gt View Dita Eiti Graph Anshan Advanced Quick A rmik Startpage Unmled eya iris syz raphi PETALLEN i PETALLEN J i SEPALLEN e x Praia O QRAM HM Suppose you want to enter a title for individual frames SPECIES 2 1 172 e g add a ti tle Versicolor for SYSTAT An Overview Click the scatterplot for SPECIES 2 Open the Frame page of Graph Properties dialog box Click Options tab Check Title box s Write VERSICOLOR ri 1 T T T I i we a lw EL 7 o i f 4 i i J L Pl LER 3 L r SemALLEN VERSICOLOR T T T gis dal T ama d endi Wat f e m Sag Eal W EL Li P i i F SEPALLEN GAH H R Mih MUFI Graph Properties m r B w Ml Zoom Rotate Options VERSICOLOR Title Background color T Graph type Coordinate system Rectangular 4 Projection type Now from the graph it appears that PETALLEN and SEPALLEN vary substantially for different SPECIES For getting a better impression it may be useful to plot them on a common graph For thism from the menus choose Graph Scatterpl
43. ot B Graph Scatterplot AllAxes Layout Legend Color Fil Symbol and Label Surface an Main Options Smoother Residuals Coordinates X Axis Y Axis Z Axis Www X varlablels SEPALLEN Available variablefs SPECIES SEPALLEN Add SEPALWID PETALLEN Y variablels Pelee Add PETALLEN Z variablels Grouping variable s SPECIES Ke lt Remove C Univariate density display on border Histogram Overlay multiple graphs into a single frame Repeated trial Mirror Dual MultiPlot CRI 1 173 SYSTAT An Overview Select SEPALLEN as the X variable s and PETALLEN as the Y variable s Select SPECIES as the Grouping variable s Check the Overlay mode Click OK The Output Editor displays the following graph i File Edt view Data Utiities Graph Analyze Advanced Quick Access Window Help SX _ Startpage X Untitled syo Iris syz gt Graphi ae gt 6 5 LI 4 2 UJ S A 2 SPECIES O 1 x2 0 2 T 3 4 9 6 8 SEPALLEN Graph Editor QGRAPH HTM NUM Now from the graph it is clear that PETALLEN and SEPALLEN vary significantly from one species to another Now if you want to label the SPECIES go to the Legend page of the Variable Properties dialog box Note that in the Overlay mode Legend tab is activated Select 1 from the drop down list of La
44. ule of SYSTAT 11 This module has better user interactivity to work with all graphical outputs of the SYSTAT application Users can easily create 2D and 3D graphs using the appropriate top tool bar icons which provide tool tip descriptions of graphs Graphs could be created from the Graph top tool bar menu or by using the Graph Gallery which facilitate accomplishing complex graphs e g global map with contour 3D surface plots with contour projections etc with point and click of a mouse Simply double clicking the graph will bring up a dialog to facilitate editing most of graph attributes from one comprehensive dynamic dialogue Each graph attribute such as line thickness scale symbols choice etc can be changed with mouse clicks Thus simple or complex changes to a graph or set of graphs can be made quickly and done exactly as the user requires 2 Getting Started With SYSTAT 2 1 Opening SYSTAT for Windows To start SYSTAT for Windows NT4 98 2000 ME and XP gt Choose Start gt All Programs SYSTAT 12 SYSTAT 12 Alternatively you can double click on the SYSTAT icon to get started with SYSTAT SYSTAT An Overview 2 2 User Interface The user interface of SYSTAT is organized into three spaces I Viewspace II Workspace HI Commandspace The Screenshot of startpage of SYSTAT 12 SYSTAT Startpage Startpage Untitled syo Untitledi syz vx SYSTAT 12 Recent Data Files Recent Command
45. verishie s irora Dutpa Exame Dynami pata Variable Zi Interactive Log Linttled x For Melo mers FL OFA HMO ICH ii War Tegi IG TAT MM You can specify the properties of FAT variable in the same way in the third row Now after setting the variable properties you can start entering data by clicking the Data tab in Data Editor Click the top left data cell under the name of the first variable and enter the data To move across rows press Enter or Tab after each entry To move down columns press the down arrow key Note To navigate the behavior of the Enter key in the Data Editor From the menus choose Edit Options Data Edit Options Output Output Scheme Graph Defaut font x 9 Default date and time format hd Add yyy dd M bibi rn dd MB yy wey ddd MMM yy hh mm tt Data Editor cursor Enter key moves right 8 Enter key moves down Save category variable information to data file Save D variable information to data file File Locations Century range for 2 digit pears 20th century Begin year 1900 End year 1959 21st century Begin year 2000 End year 2099 8 Custom Trim leading and trailing spaces for string variable data C Allow multiple data file view 1 138 Click either of the two radio buttons below Data Editor cursor SYSTAT An Overview Once the data are entered in the Data Editor the data file should look somet
Download Pdf Manuals
Related Search