Hi!
I got the 'spss script for analyzing IAT data' from the IAT template scripts page, adapted it and am trying to apply it to my data. However, it is not working yet. I have two questions that I hope someone here will be able to help me with:
1) I do not understand why I should run the following command:
AGGREGATE OUTFILE = * / BREAK = SUBJECT ORDER TEST / M1 M2 ERR1 ERR2 N1 N2 SD1 SD2 = FIRST (M1 M2 ERR1 ERR2 N1 N2 SD1 SD2).
As far as I understand, this leads to me just looking at the first test block (i.e. the one with 20 trials and not the one with 40) of a kind (i.e. compatible or incompatible) for each respondent. Is that correct? Why would that be the case?
2) The following command does not work when I try to use it:
COMPUTE D_asis_num = M2 - M1.exe.
Which makes sense, as far as I can see, because M2 and M1 are still in different lines in my data file; each respondent (still) has 2 lines. Did I do something wrong earlier on in the script? Am I misunderstanding something?
I am seriously stuck because of this, so again: I hope someone will be able to help (despite the fact that this is not really an Inquisit problem)!
Lieselotte
LieselotteBlommaert: 1) I do not understand why I should run the following command: AGGREGATE OUTFILE = * / BREAK = SUBJECT ORDER TEST / M1 M2 ERR1 ERR2 N1 N2 SD1 SD2 = FIRST (M1 M2 ERR1 ERR2 N1 N2 SD1 SD2). As far as I understand, this leads to me just looking at the first test block (i.e. the one with 20 trials and not the one with 40) of a kind (i.e. compatible or incompatible) for each respondent. Is that correct? Why would that be the case?
(1) The 20 trial blocks are the practice blocks. Only the test blocks contain 40 trials. The "improved" scoring algorithm (Greenwald, Nosek & Banaji, 2003) includes the data from practice trials in computing the D-measure.
(2) About the particular syntax snippet you posted: In this step you go from a data file with *4* lines per subject (first pairing practice, first pairing test, second pairing practice, second pairing test) to a data file with *2* lines per subject:
Before:
-----------------------------------------------------------------------------------------------------------subject blocknum ORDER PAIRING TEST MEAN_LAT NTRIALS-----------------------------------------------------------------------------------------------------------
1,00 First pairing practice compatible first compatible 1st combined block 555,85 20 1,00 First pairing test compatible first compatible 2nd combined block 558,33 40 1,00 Second pairing practice compatible first incompatible 1st combined block 563,20 20 1,00 Second pairing test compatible first incompatible 2nd combined block 581,88 40 2,00 First pairing practice incompatible first incompatible 1st combined block 561,10 202,00 First pairing test incompatible first incompatible 2nd combined block 555,00 402,00 Second pairing practice incompatible first compatible 1st combined block 602,20 202,00 Second pairing test incompatible first compatible 2nd combined block 560,48 40-----------------------------------------------------------------------------------------------------------
After:
----------------------------------------------------------------------subject ORDER TEST M1 M2----------------------------------------------------------------------
1,00 compatible first 1st combined block 555,8500 563,20001,00 compatible first 2nd combined block 558,3250 581,87502,00 incompatible first 1st combined block 602,2000 561,10002,00 incompatible first 2nd combined block 560,4750 555,0000----------------------------------------------------------------------
As you can see from the color-highlighted bits, the first column for each subject now contains the mean latencies achieved in practice trials (M1 = compatible practice; M2 = incompatible practice). The second line of data for each subject contains the mean latencies achieved in test trials (M1 = compatible test; M2 = incompatible test). Any further questions?
LieselotteBlommaert: 2) The following command does not work when I try to use it: COMPUTE D_asis_num = M2 - M1.exe. Which makes sense, as far as I can see, because M2 and M1 are still in different lines in my data file; each respondent (still) has 2 lines. Did I do something wrong earlier on in the script? Am I misunderstanding something?
Nope, doesn't make sense because -- as can be seen from the data provided above -- M1 and M2 should not be in different lines at this point. If this really is the case for your data set, it probably means you've introduced some syntax error(s) when modifiying the script. I quickly screened through your the file you attached and indeed there are some minor mistakes which might cause some trouble (a missing '.' in one of the AGGREGATE commands, missing brackets later on). I suggest you go over your syntax again (the newer SPSS versions have neat syntax highlighting which is helpful in catching such errors), correct these mistakes and run your data through it again.
Hope this helps,
~Dave
"To understand recursion, you must first understand recursion." - Unknown Zen Master
Hi Dave,
Turns out I made a mistake in computing the 'test' variable and that led to problems with aggregate commands later on. Your tips definitly helped to find the mistake, so thanks a lot!