3.5.2.4 The COMPUTE command 1
Contents
1. L MrT You have to do this four times to get to E Compute Variable Target Variable Numeric Expression tatguo2 J mv PB 4 VOB WOT e vare Type amp Label F Serial number of q a gt Function group 1 First Subject JW G 1 Second Subie JG Third Subject AW 1 Fourth Subjec a G2 Anticipated ac a 0 3 Anticipating F All Arithmetic COF amp Noncentral CDF Conversion Curent DateTime Date Anthmetic Date Creation a G 4 Expected Cor Date Extraction a 0 5 Anticipated Jc Functions and Special Yarables JW O 6 Likelihood of a L a Are girle as gi a Q 7b Are girls less a 0 17c Boys encou a 0 7d Cookery more a 0 8 First quality v A C 8 Second qualit optional case selection condition ok J Beste Reset Cancel Herr And click on to get COMPUTE Slatquo2 y275 V270 V277 vVA78 EXECUTE 11 As before statquo2 has been appended in the last row of the Data Editor i uml i fifth form sav DataSet1 SPSS Data Editor Sele File Edit wiew Data Transform Analyze Graphs Utilities Window Help e HAm tm fp E 2 ob amp Name Type REE a a a Missing 230 v430 Numeric 2 10 Monet 1 25 24 231 STATUO Numeric 1 G 34 Status quo count Mone 232 sexism Numeric 1 33 Sexism count None None 2a eat statquoz Numeric 8 Mone Mone z 4 Data View 4 Variable View lt SPS5 Processor is ready Transformations pending2. 6 CDF amp Noncentral CDF J G 1 Third Subject OnVerrsion tl 0 1 Fourth Subjec Current Date Time m G2 Anticipated a 0 3 Anticipating F a G4 Expected Cor JW G 5 Anticipated Jc Date Anthmetic Date Creation Date Extraction lL AF Functions and Special Vanables JW G 6 Likelihood of a Q 7a Are girls as gi H Q 7b Are girls less a G1 c Boys encou a Gd Cookery more mj EE First quality v J G 8 Second qualit optional case selection condition Scroll down to find the items o p q and r in Q 33 v275 to v278 type statquo2 in the Target Variable box and then highlight and transfer the variables one at a time with gt into the Numeric Expression box click on the button in between each one 10 E Compute Variable Target Variable Numenc E pression statquoz Type amp Label a Wade Many same A JEG 34d Wrong that a L342 Immigrant pi a Q 34f Increase spe a 0 34g Laws contre JEG 34h Mixed mari oi 0 34 War is huma ol G 34k Royal Wed a Q 341 Guns for poli Functions and Special Yarables JEG 34m Ban Hucles ol O 34n YOP a wast JI 8 340 Shouldn t be a Q 34p Contentmer J G 34q Social positi a L341 Happiness iz J 0 358 Fiddling Soc optional case selection condition Function group oog al nthmetic WEE CDF amp Noncentral COF Conversion Curent DateTime Lo JU Date Arithmetic Date Creation GEG08 Be a E E ie Date Extraction
3. but this time because of the automatic EXECUTE command the scores have also been calculated and entered Switch to Data View and press CTRL P to skip to the last column then scroll up and down to see them owt fifth form sav DataSet1 SPSS Data Editor File Edit wiew Data Transform Analyze Graphs Utilities Window Help SERN wb A EE BEERA amp Visible 233 o wi wi wg z wi 50 STATOUO sexism statquoz F 6 00 112 1 113 1 114 A l 6 D0 T15 1 E 3 00 a 2 00 Data View A variable View i i SPSS Processor is ready wt We still need to add a variable label and change the format to get rid of the decimal places Go back to Variable View za fifth form sav DataSet1 SPSS Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help eos m o F d Fy Name Type pvc Berra Label Values Missing 230 v430 Numeric 2 10 Mone 1 25 24 231 STATUO Numeric 1 C 34 Status quo count Mone 232 sexism Numeric 0 C135 Sexism count None None 233 statquoz Numeric 8 2 None None 234 4 Data View Wariable View lt SPS5 Processor is ready Transformations pending and write your own label in the Label column ua fifth form say DataSet1 SP55 Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help e HBE oe BP FE See E Name Type Width Decimals Label Values Missing 230
4. v430 Mumeric 2 0 10 None 1 23 24 231 STATUO Mumeric 1 L154 Status quo count Mone 0 232 sexism Numeric 0 L135 Sexism count Mone Mone 2 G 34 Status quo compute Mone Mone 233 statquoz Mumeric E 234 4 4 Data View A Variable View SP55 Processor is ready 12 and amend the format by changing 2 to 0 in the Decimals column us fifth form sav DataSet1 SPSS Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help e HA od ab A EE Beka y ae Mame Type Width Decimals Label Values Missing 230430 Numeric 2 0 i Z 0 None 1 23 24 ___231 STATQUO Numeric O 934 Status quo count None oM 232 sexism Numeric a 0 33 Sexism count Mone None 233 statquo2 Numeric T p FE a34 Status quo compute None i None SPS5 Processor is ready Now for the frequency count You should be able to do this by yourself by now so go back to the Data Editor Analyze Descriptive statistics aan Frequencies Highlight statquo2 E Frequencies a vd27 Varable s od v428 JE 429 JE 430 amp L4 Status qua c G 0 33 Sexism cour 4 G 34 Status guo i Display frequency tables _ transfer it with gt into the Variable s box E Frequencies oll v426 Yarable s 42r d 0 34 Statys qua comp T JE p428 JE p429 JE p40 F G 34 Status quo e amp 0 33 Sexism cour Display freq
5. disagree 1 23 24 q Data View h Variable View lt SPF55 Processor is ready Now you know why positional variable names and informative variable labels are so useful You can work straight from the questionnaire to the data set and vice versa The syntax way If you want to do this using the drop down menus feel free but it s cumbersome and takes forever It s So much easier to open a new syntax file es Syntax SPSS Syntax Editor A led File Edit View Data Transform 4nalyze Graphs Utilities Bun Window Help eHAARrR ib A gt Es SP55 Processor is ready and type in your command line ea Syntax SPSS Syntax Editor Sele File Edit View Data Transform Analyze Graphs Utilities Bun Window Help Hahtobhak A gt Fz comp statquod vAr SAA baile AO 5 a SP55 Processor is ready Leave the cursor in or on the new line and press CTRL R or Run Current This creates a new variable STATQUO2 to measure attachment to status quo but this time the resultant score will lie in the range 4 to 16 because we are adding together 4 items with values in the range 1 to 4 COUNT ignores any missing values and will produce a score for every case whereas with one important exception COMPUTE will automatically assign the system missing value dot if one or more items in the list is missing Nothing appears to happen but if you go back to the Data Editor and scroll down to the last row you wi
6. 1 2 3 4 76 life than those who are always trying to improve their position We are all born to our various social positions and it won t 2 3 4 77 do to change then r The greatest source of happiness o in life is to be satisfied with 3 4 79 whatever you have On the questionnaire all items appear to be loaded in the same direction so the values can be added together to yield a simple score but it s always best to check We did this in a previous tutorial Data checks 1 Status quo so there s no need to do it again here In this edition of the file the values of all four variables tally exactly with those in the user manual and we can confidently proceed to create a new variable using the COMPUTE command COMPUTE creates new variables or new values for existing variables through an arithmetic expression or by combining existing variables according to one or more functions Only one transformation may be requested per COMPUTE command The general format is COMPUTE lt computed variable gt lt arithmetic expression gt in which blue indicates required SPSS syntax and lt plum gt things you have to type in yourself e g COMPUTE STATQUO2 V275 V276 V27 7 V278 This command will create a new variable STATQUO2 by adding together the values of the four variables v275 v276 v277 and v278 lf any of the four variables have missing values then the value for the resultant variable STATQUO2 will be set to system missing
7. As well as a full range of arithmetic operators COMPUTE also has available some special functions There s a selection of these listed in the appendix page 18 For a fuller explanation see the SPSS User Guide or use the help button A useful one is SUM n which calculates the sum of all variables in the expression but with optional n requires a minimum number of valid values otherwise the value for the new variable will be set to system missing e g COMPUTE STATQUO2 SUM 4 V275 TO V278 which requires all four questions to be answered before calculating a score Another is the truncation function TRUNC which knocks off the figures after the decimal point after division Thus if you have a variable like actual income in or or age in whole years you can generate quick groupings by dividing by ten or a thousand etc eg COMPUTE AGEGROUP TRUNC AGE 10 COMPUTE INCGROUP TRUNC INCOME 1000 These would not normally be used in reports but can be helpful when exploring a new data set The other functions are far too advanced for beginners and very rarely used anyway except as intermediate stages Attachment to status quo First we need a name for the new variable We ve already got STATQUO so let s call this one STATQUOZ2 Going back to the general format COMPUTE lt computed variable gt lt arithmetic expression gt in which blue indicates required SPSS syntax and lt plums gt things you have to t
8. S and the same table and histogram as before Statistics Q 34 Status quo compute N Valid Missing Mean Sid Deviation Histogram Frequency Mean 5 75 Sid Dev 1 956 N 106 told you it was quicker in syntax Next tutorial s 3 5 2 5 Data checks 2 Sexism Feedback on ease of understanding and use of tutorial please to johnfhall orange fr 17 Appendix The COMPUTE command COMPUTE creates new variables or new values for existing variables through an arithmetic expression or by combining existing variables according to one or more functions Only one transformation may be requested per COMPUTE command The general format is COMPUTE lt computed variable gt lt arithmetic expression gt The arithmetic operators available are addition division subtraction exponentiation ie to the power of multiplication together with brackets e g COMPUTE STATQUO V2754 V276 V277 V278 If any of the variables included in the expression have missing values then the value for the resultant variable is also declared missing except for statistical functions as shown below As well as a full range of arithmetic operators COMPUTE has available some special functions There are seven classes of functions within the SPSS COMPUTE command Arithmetic Logical Statistical Random Missing Value Date Time functions Cross case For a fuller explanation of these facilities see the SPSS Us
9. Survey Analysis Workshop Copyright 2011 John F Hall Block 3 Analysing two variables and sometimes three Derived variables 3 5 2 4 The COMPUTE command 1 Attachment to status quo Screenshots SPSS15 last modified 24 Aug 2011 previous version 1 October 2009 Exemplar Fifth form survey fifthx sav Variable to be derived Attachment to status quo Source variables v275 tov278 Q340 p q r Previous tutorial 3 5 2 3 The COUNT command 1 Attachment to status quo For the Fifth Form survey we have already used the SPSS command COUNT to create a new variable ST ATQUO a measure of teenagers Attachment to status quo by counting the number of agreements with items o p q andr in Q 34 A much more accurate and useful score can be obtained by combining all the responses for each item so that disagreement can be taken into account as well as agreement and also the level of agreement or disagreement To generate this much superior measure we shall use the SPSS command COMPUTE Remember four items in question Q 34 0 p q r are replicated from a scale developed by Himmelweit to measure attachment to status quo among teenagers Q 34 Do you agree or disagree with the following statements Ring the numbers Disagree Tend to Tend to Agree Strongly Disagree Agree Strongly It is best to be like the others and not to stand out from the 1 2 3 4 75 rest People who are content with what they have will have a better
10. e label by writing your own text in the Label column vai fifth form sav DataSet1 SPSS Data Editor mB File Edit wiew Data Transform Analyze Graphs Utilities Window Help e HAR oo ah Fe Sek 2 Name Type Roy EE Label Values Missing 2310 v430 Numeric 10 None 1 25 24 251 STATQUO Numeric 1 34 Status quo count None 232 sexism Numeric 0 1 35 Sexism count None None 233 statquoz Numeric 8 2 2 34 Status quo compute Mone Mone 234 4 4 Data View A Variable View 4 SP55 Processor is ready and the format by changing 2 to 0 in the Decimals column ut fifth form sav DataSet1 SP55 Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help eHO wi BER 2 Mame Type pain ferra Label Values Missing 230 v430 Numeric 2 0 10 Mone 1 25 24 251 STATQOUO Numeric 0 1 34 Status quo count Mone 232 sexism Numeric 0 9 33 Sexism count Mone Mone 233 statquo2 Numeric jo D 0 34 Status quo compute Mone Mone 234 4 Data View A Variable View 4 SP55 Processor is ready Alternatively you can add variable labels and format commands to the syntax file ES Syntax1 SPSS Syntax Editor File Edit View Data Transform Analyze Graphs Utilities Run Window Help Ge amp Ha amp H sb AP wu h compute statquod W275 WATO vere WATO freq statquoz var lab statquoz 1 54 Status quo compute format statquoz sexism 2 0 sud SP55 Proc
11. eady to yield Q 34 Status quo compute TT Cumulative Frequency Percent Valid Percent Percent Valid 0 1 7 9 9 1 3 2 1 2 8 3 8 2 2 1 4 1 9 5 7 3 6 4 2 5 7 11 3 4 13 9 2 12 3 23 6 5 18 12 7 17 0 40 6 6 26 18 3 24 5 65 1 i 17 12 0 16 0 81 1 8 14 9 9 13 2 94 3 9 4 2 8 3 8 98 1 10 2 1 4 1 9 100 0 Total 106 74 6 100 0 Missing System 36 25 4 Total 142 100 0 You could also use compute statquo2 sum 4 v275 to v278 4 but emphatically do not use compute statquo2 sum v275 to v278 4 because it will calculate a score even if one or more constituent items are missing Of the 36 missing cases in the above table 18 are missing all four items another 18 are missing at least one item but would still be given a score Because STATQUO2 is a genuine interval scale also now a ratio scale it is legitimate to produce descriptive statistics and a histogram overlaid with a normal curve As an option we can also suppress the frequency table format notable and just ask for selected Statistics e g statistics mean stdev and or graphics e g histogram normal ES Syntax1 SPSS Syntax Editor SEE File Edit View Data Transform Analyze Graphs Utilities Run Window Help e H amp R o Hanh gt Ky gP compute statguoz w275 Verb AT W270 freg statquoz var lab statuo 1 34 Status quo compute format statquoz f2 0 freq statquoz compute statuo statquoz 4 freq statquoe freq stat
12. er Guide or use the help button Some of the most common functions you are likely to use in survey analysis are TRUNC truncate SD n std deviation of values across list SUM n sum of values across variable list MIN n minimum of values across list MEAN n mean of values across variable list MAX n maximum of values across list Where the optional n indicates for statistical functions the minimum number of arguments required to be valid before the expression is declared missing A useful one is SUM n which calculates the sum of all variables in the expression but with optional n requires a minimum number of valid values otherwise the value for the new variable will be set to system missing e g COMPUTE STATQUO SUM 4 V275 TO V278 which requires all four questions to be answered before calculating a score Another is the truncation function TRUNC which knocks off the figures after the decimal point after division Thus if you have a variable like actual income in or or age in whole years you can generate quick groupings by dividing by ten or a thousand etc eg COMPUTE AGEGROUP TRUNC AGE 10 COMPUTE INCGROUP TRUNC INCOME 1000 18
13. essor is ready and run the frequency count again ES Syntax SPSS Syntax Editor File Edit View Data Transform Analyze Graphs Utilities Pun Window Help e amp Hba amp o hnaeb MP ir compute statquozd y275 W276 ver WATO freq statquoz var lab statquaz Ll 34 Status quo compute format statuo sexism f2 0 freq statguoz SP55 Processor is ready which produces Q 34 Status quo compute Valid Cumulative Percent Percent Percent Total Missing System Total This table shows a range of 4 to 14 out of a theoretical 16 We can improve it by adjusting the score so that it becomes a ratio scale with a true 0 point and a theoretical range of 0 12 instead of 4 16 Todo this we subtract the number of items in the scale 4 using a further SPSS command compute statquo2 statquo2 4 This is the equivalent of the original items having been coded 0 to 3 instead of 1 to 4 and avoids the error common in many reports especially when comparing group means of discussing means as though they come from a ratio scale Do freq again ES Syntax1 SPSS Syntax Editor File Edit View Data Transform Analyze Graphs Utilities Run Window Help e HAm ee Btb Mm Pp Kg P compute statquo2 w275 WAB AT W O freq statquoz var lab statquoz 4 34 Status quo compute format statquoz f2 0 freq statquoz compute statquo statquoz 4 freq statquaz zjm gt SP55 Processor is r
14. he command to get Statistics statquo2 N Valid 106 Missing 36 statquo2 Cumulative Percent Valid Percent Percent 4 00 5 00 6 00 7 00 8 00 9 00 10 00 11 00 12 00 13 00 14 00 Total Missing System Total File Edit View Data Transform Analyze Graphs Ulkilities Window Help e HA amp im Pp A EE BELEK amp 1 statquaz Visible 233 of 23 W425 wie W425 v42 vfa a TATOUO sexism statquaz var A 10 00 4 1 10 00 v 4 Data View A Variable Vie j lt gt SP S5 Processor is ready 1 0 2 Scroll down and you will see that this time although STATQUO2 scores have now been generated some cells are still empty This is because some pupils were missing on one or more of the four items and SPSS has automatically set their score to system missing Go back to Variable View ut fifth form sav DataSet1 SPSS Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help e HA m tm fp FE ES ei Fy Name Type Ene Label Values Missing 2310 v430 Numeric 10 None 1 25 24 251 STATQUC Mumeric 1 1 34 Status quo count None 232 sexism Numeric 0 G33 Sexism count None None ae statquo2 Numeric 8 2 Mone Mone z 4 Data View Variable View lt SP55 Processor is ready Transformations pending We still need to give STATQUOZ2 a label and and get rid of the decimal places Both can be done directly in the Data Editor th
15. ll see that statquo2 has been appended cas fifth form say DataSet1 SPSS Data Editor File Edit Yiew Data Transform Analyze Graphs Utilities Window Help gt c HA m tm i E BTA Name Type Width Decimals Values Numeric 2 O None 231 STATQUO Numeric 1 Q 34 Status quo count None sexism Numeric 1 Q 33 Sexism count None statquo2 Numeric 8 None 234 4 gt Data View A Variable View lt SPSS Processor is ready Transformations pending However if you switch to Data View and press CTRL to skip to the last column cas fifth form say DataSet1 SPSS Data Editor File Edit Yiew Data Transform Analyze Graphs Utilities Window Help gt HA amp wb A EFE BERK amp 1 statquo2 Visible 233 of 23 v426 w427 w428 w429 v430 STATQUO sexism var a 1 1 i 3 1 lt Data View Variable view lt gt SPSS Processor is ready Transformations pending you will see that no scores have yet been entered for STATQUO2 Scroll up and down the column to confirm This is because SPSS is waiting for an EXECUTE command or for a statistical procedure before it calculates any scores We need a frequency count for STATQUQ2 anyway so ES Syntax1 SPSS Syntax Editor File Edit View Data Transform Analyze Graphs bilities Run Window Help e HAr eo Hane A s va compute statquoz2 v2 5 Verb yer WATO freq statquoz am 5P55 Processor is ready Run t
16. quoz tor not sta mean stdey fhis nor am 2 5P55 Processor is ready Run to get Statistics Q 34 Status quo compute Valid Missing Mean Std Deviation Histogram Frequency Mean 5 75 Std Dev 1 956 N 106 Nice one to show your statistics tutors they love normal distributions NB There s a label on the output but it has not copied across The point and click way You can use the drop down menus if you prefer but it takes a lot longer Go back to the Data Editor and find the four items we need eas fifth_form 1 sav DataSet5 SPSS Data Editor File Edit wiew Data Transform Analyze Graphs Utilities Window Help cHe amp tm EP F ES i Fy 2 Name Type width Decimals Label 138 275 Numeric Z 0 1 340 Shouldnt be conspicuous 11 Strongly disagree 1 23 24 139 v276 Numeric 2 0 C34 Contentment brings better life 11 Strongly disagree 1 23 24 140 v2F7 Numeric 2 0 Q 34 Social positions fixed 11 Strongly disagree 1 25 24 141 278 Numeric 2 0 Q 34r Happiness is satisfaction 1 strongly disagree 1 25 24 4 Data view Variable View lt SP55 Processor is ready Values Missing Click on Transform Compute variable W Compute Variable aa Varnable Numernc Expression ppe amp Label Serial number of qu gt Function group r G 7 First Subject F a All 0 1 Second Subje Arithmetic L4 5
17. s Cut points for hia equal groups _ Percentile s P Add Change Remove Dispersion Minimum Maximum S E mean Variance Range then click on E Frequencies amp senal number of qu A JEG First Subject 0 1 Second Subje JEG Third Subject 0 1 Fourth Subjec a Q Anticipated ar 0 3 Anticipating F diot Exmected Corg Display frequency tables Central Tendency Feon Median L Med Mode Sum Values are group midpoints Distribution Skewness Kurtosis Variable s amp G 34 Status gua comp click on 15 then check the button for Histograms under Chart Type and the box With normal curve Frequencies Charts Chart Type None Bar charts Fie charts C Histograms Chart Values Frequencies W Frequencies g Serial number of qu A Varlable s Wa First Subject 0 1 Second Subje 0 1 Third Subject 0 1 Fourth Subjec 0 2 Anticipated ar 0 3 Anticipating F 0 4 Expected Cor Wa c aeia Display frequency tables MS Frequencies F Seral number of qg wenei l 0 1 First Subject E amp G 34 Status guo comp 0 1 Second Subje 0 1 Third Subject 0 1 Fourth Subjec G Anticipated ar 0 3 Anticipating F O 4 Expected Cor WM ee a Display frequency tables press 16 SPSS generates the following syntax FREQUENCIES VARIABLES statquo2 FORMAT LIMIT 5 STATISTICS MEAN MEDIAN HISTOGRAM NORMAL ORDER ANALYSI
18. uency tables and click to run the job and get the same frequency table as on page 8 13 To suppress the frequency table and get only selected statistics and a histogram Analyze Descriptive statistics penne Frequencies Highlight statquo2 ME Frequencies al 427 Wariable s ol p428 H 425 Faste ol p430 g 0 34 Status quo c D 0 33 Sexism cour 20 34 Status quo i Display frequency tables and transfer it with into the Variable s box MS Frequencies Vanable s Pi 0 34 Status qua camp F 0 34 Status quo c FEE Sesim cour Display frequency tables Uncheck Display frequency tables and you ll get the following warning SPSS 15 0 for Windows generate no output 1 You have turned off all output Unless you request Display Frequency Tables Statistics or Charts FREQUENCIES will Click but you must now ask for something to display This warning does not appear if you select Statistics or before unchecking Display frequency tables 14 E Frequencies amp senal number of qu A 0 1 First Subject 0 1 Second Subie JEG Third Subject JEG Fourth Subjec a Le Anticipated ar fd 0 3 Anticipating F a W 4 Expected C Ol Aam ce saio Display frequency tables Variable s 34 Status quo come click on Statistics and check the boxes for Std deviation under Dispersion and Mean under Central Tendency Frequencies Statistics Percentile Values Quartile
19. ype in yourself A small table helps COMPUTE STATQUO2 v275 v276 v277 v278 We can therefore write COMPUTE STATQUO2 V275 V276 V277 V278 Because the variables are adjacent in the file you can also use COMPUTE STATQUO2 SUM 4 V275 TO V278 Do not use COMPUTE STATQUO2 SUM V275 TO V278 It gives a different answer because it calculates a score even if one or more items are missing Start by downloading the saved file fifthx sav Don t worry if you haven t got immediate direct access to SPSS just follow the tutorial anyway You should be able to do this yourself by now so Scroll down to find variables v275 to v278 containing the data for items Q340 to Q34r use the mouse to drag the column separators around to make the display easier to read then reduce the pane horizontally to include only essential information and reduce the pane vertically to show only these variables a fifth form sav DataSet1 SESS Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help e HAm tim i FE SER a Name Type Width Decimals Label Values Missing 138 275 Numeric 2 0 340 Shouldnt be conspicuous 1 Strongly disagree 1 25 24 139 v276 Numeric 2 0 C134 Contentment brings better life 1 Strongly disagree 1 23 44 140 277 Numeric 2 0 G 34gq Social positions fixed 1 Strongly disagree 1 23 24 141 275 Numeric 2 0 1 54r Happiness is satisfaction 1 Strongly
Download Pdf Manuals
Related Search