SPSS .sav data is missing new variables?

Hi, I've set up my study and all is running smoothly, EXCEPT for some reason medialab is not recording my participants' responses to the message recall items or to a second thought-listing task. both are marked to record text, so i'm not sure what's going on, expecially since everything else is running so smoothly. any thoughts or suggestions to help figure this out would be MUCH appreciated!

thanks,
Sonia

p.s. i've attached a sample questionnaire so that you can see what i am talking about -- the issue is with items #56 and #170. thanks!

Hi Sonia,

Did you add the items after some initial test runs? And are you looking at the SPSS .sav data file? If so, remember that the .sav file can not have variables added to it after it has been created. So if you add items, try clearing out your data folder and then run a test session. If this is the problem, you should now find that everything is writing OK. This same problem creates trouble for .csv and .txt dat files but only in so far as the data won't match the origional file headers.

Bottom line is that if you've made some changes to your experiment structure, try clearing your data folder and doing a fresh run. If that doesn't fix it, please let me know.

-Blair

Blair,
I've deleted the data folder and that worked to fix the initial problem (so now all my responses are being recorded!); however, i want the thought-listing and message-recall string items to record more that 50 characters, but i don't know how to change this so that it records it properly to the SPSS.sav file (since when I change the SPSS file to increase the number of characters recorded it no longer records those items). Any other place i can change the # of characters recorded (and the variable names to be listed in SPSS)?

thanks!
Sonia

The text responses in the automatically generated SPSS .sav file are truncated at 50 characters. The full text values are available in the other data file formats (e.g., [I].csv[/I]., [I].txt[/I]). If you use the SPSS syntax to manually read in the [I].txt[/I] data, you can change the default [I]A50[/I] values to [I]A255[/I] or however many characters you would like SPSS to use before truncating.

Hi Blair,

we have basically the same problem. We record string variables (Ss generate descriptions of a target person), but they are truncuated. However, when I try to modify (or even attempt to use the syntax file without modifying it) nothig happens (or I get error messages). This happens for a long time now (using different ML versions as well as SPSS versions). I just tried to switch the language in the regional settings, but it does not work either.

SPSS Error message:
[QUOTE]>Error # 4135 in column 13. Text: 1
>You must not specify the column location for variables read in FREE or LIST
>format.
>This command not executed.[/QUOTE]

The only thing I changed was the path.

Thx in advance,
Rene

Can you send me or post the syntax and a sample data file? I'll try to have SPSS read it here and see if I can tell what's wrong.

Blair,

attached you'll find one single data file and a complete data file (both byvariablename) as well as an original syntax file, as generated by ML.

The error replicates on my work PC. SPSS Version 14 (work) and 15 (home) were used on both machines. (Although the Version generating the syntax was 12!).

Wrting the above I followed up on different version. Trying to read a data file using a syntax generated by an actual SPSS version (15) I got no error message but a unuseable SPSS matrix. Switching regional settings did not alter the outcome (see SPSS15test files). BTW, this is the data of the current experiment where some variables were not written (see [url]http://www.empirisoft.com/Support/showthread.php?t=789)[/url].

Best
Rene

I see what you mean. Two things. First the issue with the syntax. You had a space embedded in one of your variable names ([SIZE=1]rel_imp 1). You also have some variable names greater than 8 characters which can be a problem for earlier versions of SPSS. I've attached a revised syntax file which should run ok for you.[/SIZE]
[SIZE=1][/SIZE]
[SIZE=1]Unfortunately I don't think that will solve the problem as SPSS will still limit you to 255 characters. Maybe this limit has been increased in more recent releases than I have. I believe it has been increased in the most recent version of Excel--perhaps you could try opening the CSV formatted file in Excel v2007 if you can get your hands on a copy.[/SIZE]
[SIZE=1][/SIZE]
[SIZE=1]Bottom line, is that I agree ML needs some way of dealing with custom items that produce so much data. What do you think of adding a parameter that would act as a switch to write the data as a seperate file?[/SIZE]
[SIZE=1][/SIZE]
[SIZE=1]
[/SIZE]

Blair,

thanks for your reply. Unfortunately, the altered syntax does not work. Like I said earlier, we've have that problem for a long while know, using all kinds of SPSS and ML versions. I believe that SPSS supports longer varnames since Version 12. Anyway, I used the original syntax, I altered it, I tried to vary the string variable length and it simple won't work (using data generated by the latest ML and using actual (14 or 15) version of SPSS).

The result (which basically looks the same every time) is attached. What I don't get is why it works if it is run by ML or whatever you do that generates the data files. The ML generated sav-files all look perfectly fine (despite the embedded space; the only consequence is a spss generated varname var0001 or so...). BUT if I try to do it myself (after I merged txt-files from different PCs), it won't work.

As a result I have to add data/cases via SPSS. It would be much more comfortable, if it was possible just to read in the (already merged) raw data!

You see, the read in/syntax and the string variable problems are independant...

Thanks again,
Rene

The sytnax file I sent you does not eliminate this error you were getting?
[QUOTE]
Error # 4135 in column 13. Text: 1
You must not specify the column location for variables read in FREE or LIST format. This command not executed.
[/QUOTE]

There are two issues I see here, one is getting the SPSS syntax to work on reading in the data. The other issue is getting it to read the data in properly when these big strings are involved. The syntax file I sent is only addressing the first issue. The "Text: 1" that it's referring to is your variable "[SIZE=1]rel_imp 1" which, when changed to "rel_imp1" fixed this error and allowed the data to be read in. However, the problems to the string data remain. Are we on the same page with that?[/SIZE]