Syntax for creating a composite Variable
Locked
Syntax for creating a composite Variable
 
|
Interested in growing your business? Find out how with Yahoo! Search Marketing! |
Re: Syntax for creating a composite Variable
|
|
Hi Johnny:
First, you could simplify your syntax: COMPUTE SEXAGE = 9*(SEX=2) + AGE. Will replace all the IF lines. Second: With Richard Ristow's permission (big smile), add EXE at the end. COMPUTE SEXAGE = 9*(SEX=2) + AGE. EXE. Or, instead of EXE, use any procedure that reads your data: COMPUTE SEXAGE = 9*(SEX=2) + AGE. VALE LABELS .............(add the rest of the command)... LIST SEXAGE /CASES=FROM 1 TO 10. The values aren't actually computed until EXE or any procedure that reads the data is run. HTH, Marta Johnny Amora wrote: > I run the following syntax to create a composite demographic variable > SEXAGE. The SEXAGE was created but showed empty. What is wrong or > lacking with the syntax? > > IF (SEX=1 AND AGE=1) SEXAGE= 1. > IF (SEX=1 AND AGE=2) SEXAGE= 2. > IF (SEX=1 AND AGE=3) SEXAGE= 3. > IF (SEX=1 AND AGE=4) SEXAGE= 4. > IF (SEX=1 AND AGE=5) SEXAGE= 5. > IF (SEX=1 AND AGE=6) SEXAGE= 6. > IF (SEX=1 AND AGE=7) SEXAGE= 7. > IF (SEX=1 AND AGE=8) SEXAGE= 8. > IF (SEX=1 AND AGE=9) SEXAGE= 9. > IF (SEX=2 AND AGE=1) SEXAGE=10. > IF (SEX=2 AND AGE=2) SEXAGE=11. > IF (SEX=2 AND AGE=3) SEXAGE=12. > IF (SEX=2 AND AGE=4) SEXAGE=13. > IF (SEX=2 AND AGE=5) SEXAGE=14. > IF (SEX=2 AND AGE=6) SEXAGE=15. > IF (SEX=2 AND AGE=7) SEXAGE=16. > IF (SEX=2 AND AGE=8) SEXAGE=17. > IF (SEX=2 AND AGE=9) SEXAGE=18. > VALUE LABELS > SEXAGE > 1 'M<12' > 2 'M12-14' > 3 'M15-17' > 4 'M18-20' > 5 'M21-29' > 6 'M30-39' > 7 'M40-49' > 8 'M50-64' > 9 'M65+' > 10 'F<12' > 11 'F12-14' > 12 'F15-17' > 13 'F18-20' > 14 'F21-29' > 15 'F30-39' > 16 'F40-49' > 17 'F50-64' > 18 'F65+'. > > Thank you. > Johnny > > > ------------------------------------------------------------------------ > Interested in growing your business? Find out how with Yahoo! Search > Marketing! > <http://searchmarketing.yahoo.com/en_SG/arp/internetmarketing.php?o=SG0147> -- For miscellaneous SPSS related statistical stuff, visit: http://gjyp.nl/marta/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Syntax for creating a composite Variable
|
|
In reply to this post by Johnny Amora
Open emails faster. Yahoo! recommends that you upgrade your browser to the new Internet Explorer 8 optimized for Yahoo!.Get it here! (It's free) |
Re: Syntax for creating a composite Variable
|
|
Johnny Amora wrote:
> I dont understand the right side of your equation > > "COMPUTE SEXAGE = 9*(SEX=2) + AGE." > (SEX=2) is a logic expression that will return 0 if the expression is false (cases where sex=1 will return 0) and 1 if it is true (cases where sex=2). For cases with sex=1, the expression reads as: sexage = 9*0 + age = age (if age=1, sexage= 1; age=2, sexage= 2, and so on...). For cases with sex=2, the expression reads as: sexage = 9*1 + age = 9+age (if age=1, sexage=10; age=2, sexage=11, and so on...) > > > The syntax was exactly written in the spss training manual. I wonder > why it does not work. > Try adding EXE at the end of your syntax (since you don't trust mine, although I can assure you it works). Marta -- For miscellaneous SPSS related statistical stuff, visit: http://gjyp.nl/marta/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Syntax for creating a composite Variable
|
|
|
Re: Syntax for creating a composite Variable
|
|
In reply to this post by Johnny Amora
Design your own exclusive Pingbox today! It's easy to create your personal chat space on your blogs |
Re: Syntax for creating a composite Variable
|
|
Johnny Amora escribió:
> It is not that I do not trust you, Marta. > I was joking, of course.Perhaps I shuold have added an emoticon to make that clear... > I threw a follow-up question because I dont like to pretend that I > understood the logic. At the first glance I did not understand where > did you get the "9" :). Now I got the logic. > I admit it is complicated stuff. I got some off-line questions concerning it. > Thank you so much. > Always welcome. > > By the way, the syntax in the spss manual does not work even an EXE. > is added in it. > Strange, it should. Does mine work? If it fails too, then you should take a look at your data. You have already been given a hint concerning missing values for one or both variables. Check that. Do you get any error messages? Just a silly question (no offense intended, please): is SEX variable really coded as NUMERIC with 1=males and 2=females?. I had problems in the past when people handled me SEX variables coded as STRING, but with values of 1 & 2 ('1' and '2', to be more precise), and value labels attached to the numeric strings (a mess). That would cause any of the two versions of the syntax (long&short) fail. If that is the case, then this modified version should work: COMPUTE SEXAGE = 9*(SEX='2') + AGE. EXE. Check AGE too, if it is also a STRING variable instead of NUMERIC, the code might fail too, unless modified: COMPUTE SEXAGE = 9*(SEX='2') + NUMBER(AGE,F8). EXE. HTH, Marta -- For miscellaneous SPSS related statistical stuff, visit: http://gjyp.nl/marta/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Syntax for creating a composite Variable
|
|
In reply to this post by Johnny Amora
Get your preferred Email name! Now you can @ymail.com and @rocketmail.com. |
|
|
In reply to this post by Johnny Amora
Hello,
Can someone advise me how I would recode all values in a file that are greater than 1 the value of 1? I can do this for individual variables, however wanted to make the change to all variables, Any assistance is appreciated, Bob ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: recoding all values greater than 1 in a file
|
|
No time to test--
does the ALL keyword work in RECODE? How about the keyword TO?-- "firstvar TO lastvar" --jim -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Bob Green Sent: Thursday, April 23, 2009 4:22 PM To: [hidden email] Subject: recoding all values greater than 1 in a file Hello, Can someone advise me how I would recode all values in a file that are greater than 1 the value of 1? I can do this for individual variables, however wanted to make the change to all variables, Any assistance is appreciated, Bob ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: recoding all values greater than 1 in a file
|
In reply to this post by bgreen
Is something like this what you mean?
recode firstvar to lastvar (1 thru hi=1) (else=copy) into newvar001 to newvar999. Art Kendall Social Research Consultants Bob Green wrote: > Hello, > > Can someone advise me how I would recode all values in a file that > are greater than 1 the value of 1? I can do this for individual > variables, however wanted to make the change to all variables, > > Any assistance is appreciated, > > Bob > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD > > ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
Re: Syntax for creating a composite Variable
|
In reply to this post by Marta Garcia-Granero
At 03:29 AM 4/23/2009, Marta García-Granero wrote:
>First, you could simplify your syntax: > >COMPUTE SEXAGE = 9*(SEX=2) + AGE. > >Will replace all the IF lines. Exactly. However, I'd use COMPUTE SEXAGE = 10*(SEX=2) + AGE. Marta's syntax gives the coding the poster requested, and mine doesn't. But I'd recommend the coding, so the first (decimal) digit is sex, the second one age, and the 'age' digit has the same meaning regardless of sex. >Second: With Richard Ristow's permission (big smile), add EXE at the end. Marta, for you anything, even that. -Smiling, Richard ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
In reply to this post by bgreen
Working with v 16.0.1--networked. I have a dataset with almost 4200 cases and about 150 fields. Almost all are designated as strings although many contain only numbers. With at least one field, statistics such as frequencies and crosstabs miss picking up several valid values. An example: I have a string field for employee status that is a value of 'E' or 'S' or 'G'; there are no missing values. I have a second string field for Occupational Activity Code: '10', '20', '30', '40', '50', '60', '70'. Again, no missing values. Both are nominal measures. When I look at the data file and sort by status and OACode I can count down the rows and see that I have 341 'E's with an OACode of '10' and 2 's's with a code of '10'. When I crosstab, I shouls see these as observed cells. There are NO filter variables in the dataset. There is NO filter turned on. When I run crosstabs, I get the following: Occupational Activity Category * Personnel Act Crosstabulation Personnel Act E G S Total OACode 0 880 0 880 10 3 0 0 3 20 1242 0 0 1242 30 346 0 286 632 40 0 0 201 201 50 0 0 501 501 60 0 0 116 116 70 0 0 277 277 Total 1929 880 1383 4192 Note the 'E' column. It shows a sum of 1591 'E' in cells but there are 1929 'E's in the dataset. Looking at Frequecies the same thing occurs: Occupational Activity Category Freq Perc Valid Cum 880 21.0 21.0 21.0 10 3 .1 .1 29.2 20 1242 29.6 29.6 58.8 30 632 15.1 15.1 73.9 Valid 40 201 4.8 4.8 78.7 50 501 12.0 12.0 90.6 60 116 2.8 2.8 93.4 70 277 6.6 6.6 100.0 Total 4192 100.0 100.0 If you sum down the frequency column, you will see that 340 cases are not being counted. An added rub: I CAN turn on the filter and a filter out cases where OACode is not '10' and it recognizes all of the ten values by filtering all but the 343 valid cases. Personnel Act Freq Per Valid Cum E 341 99.4 99.4 99.4 Valid S 2 .6 .6 100.0 Total 343 100.0 100.0 As you can see above, all 343 non-filtered cases are there, yet they aren't registering when I run frequencies. Occupational Activity Category Freq Per Valid Cum Valid 10 3 .9 .9 100.0 Total 343 100.0 100.0 Again, I just opened this dataset. There are no filters on. I rerun it after Select Cases = all and I get the same discrepancy. I get no error messages in the log. This file is my bread and butter and was CREATED using SPSS but now seems to be unt rustworthy in SPSS. Has anyone else seen this issue? *************************************************************************************************************************************************************** Mark A. Davenport Ph.D. Senior Research Analyst Office of Institutional Research The University of North Carolina at Greensboro 336.256.0395 [hidden email] 'An approximate answer to the right question is worth a good deal more than an exact answer to an approximate question.' --a paraphrase of J. W. Tukey (1962) |
|
|
If you send me a .sav file with 3 variables, an ID you can match back
to, the status variable, and the Activity variable, I'll take a look at
it.
Art Kendall Social Research Consultants Mark A Davenport MADAVENP wrote:
Art Kendall
Social Research Consultants |
Re: Strange behavior
|
|
In reply to this post by Mark A Davenport MADAVENP
Thanks to Art and Jon. Jon wins the cookie. Your solution didn't seem to work but your diagnosis may have been right on the money. I created a new variable manually and recoded it to to equal the old OACODE. I now have my 341 missing cases. I thought I had done this recode before with this data before. Perhaps not. I have had a great deal of trouble with this data set (hundreds of valid values suddenly turn into '?', etc.) It has been coded and recoded a hundred times. I guess it is getting tired. Thanks again guys! Mark *************************************************************************************************************************************************************** Mark A. Davenport Ph.D. Senior Research Analyst Office of Institutional Research The University of North Carolina at Greensboro 336.256.0395 [hidden email] 'An approximate answer to the right question is worth a good deal more than an exact answer to an approximate question.' --a paraphrase of J. W. Tukey (1962)
Strange behavior, indeed. I have a strong hunch about what’s wrong. Some of your cases may have some junk in the Activity Code variable past the second byte. To check this, you could just try inserting temporary. compute OACode = OACode. in front of your CROSSTABS command. Jonathan Fry SPSS Inc. From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Mark A Davenport MADAVENP Sent: Monday, April 27, 2009 12:20 PM To: [hidden email] Subject: Strange behavior Working with v 16.0.1--networked. I have a dataset with almost 4200 cases and about 150 fields. Almost all are designated as strings although many contain only numbers. With at least one field, statistics such as frequencies and crosstabs miss picking up several valid values. An example: I have a string field for employee status that is a value of 'E' or 'S' or 'G'; there are no missing values. I have a second string field for Occupational Activity Code: '10', '20', '30', '40', '50', '60', '70'. Again, no missing values. Both are nominal measures. When I look at the data file and sort by status and OACode I can count down the rows and see that I have 341 'E's with an OACode of '10' and 2 's's with a code of '10'. When I crosstab, I shouls see these as observed cells. There are NO filter variables in the dataset. There is NO filter turned on. When I run crosstabs, I get the following: Occupational Activity Category * Personnel Act Crosstabulation Personnel Act E G S Total OACode 0 880 0 880 10 3 0 0 3 20 1242 0 0 1242 30 346 0 286 632 40 0 0 201 201 50 0 0 501 501 60 0 0 116 116 70 0 0 277 277 Total 1929 880 1383 4192 Note the 'E' column. It shows a sum of 1591 'E' in cells but there are 1929 'E's in the dataset. Looking at Frequecies the same thing occurs: Occupational Activity Category Freq Perc Valid Cum 880 21.0 21.0 21.0 10 3 .1 .1 29.2 20 1242 29.6 29.6 58.8 30 632 15.1 15.1 73.9 Valid 40 201 4.8 4.8 78.7 50 501 12.0 12.0 90.6 60 116 2.8 2.8 93.4 70 277 6.6 6.6 100.0 Total 4192 100.0 100.0 If you sum down the frequency column, you will see that 340 cases are not being counted. An added rub: I CAN turn on the filter and a filter out cases where OACode is not '10' and it recognizes all of the ten values by filtering all but the 343 valid cases. Personnel Act Freq Per Valid Cum E 341 99.4 99.4 99.4 Valid S 2 .6 .6 100.0 Total 343 100.0 100.0 As you can see above, all 343 non-filtered cases are there, yet they aren't registering when I run frequencies. Occupational Activity Category Freq Per Valid Cum Valid 10 3 .9 .9 100.0 Total 343 100.0 100.0 Again, I just opened this dataset. There are no filters on. I rerun it after Select Cases = all and I get the same discrepancy. I get no error messages in the log. This file is my bread and butter and was CREATED using SPSS but now seems to be unt rustworthy in SPSS. Has anyone else seen this issue? *************************************************************************************************************************************************************** Mark A. Davenport Ph.D. Senior Research Analyst Office of Institutional Research The University of North Carolina at Greensboro 336.256.0395 [hidden email] 'An approximate answer to the right question is worth a good deal more than an exact answer to an approximate question.' --a paraphrase of J. W. Tukey (1962) |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |