IRT Modeling Lab

4. Write Subscales to Separate Files


All IRT models discussed in this tutorial are unidimensional. The SPSS data file contains two personality scales, which are relatively uncorrelated, so the files should be calibrated seperately.

To write the files in the ASCII format required by BILOG or MULTILOG, use the following syntax.

SPSS Syntax:

  *Select calibration data.

GET
FILE='C:\IRTtutorial\raw_reversed.SAV'.
EXECUTE .

FILTER OFF.
USE ALL.
SELECT IF(split =1).
EXECUTE .

*Writeout calibration files with no spaces between entrees.

WRITE OUTFILE = 'C:\IRTtutorial\con1_cal.dat'
/ id c1 c2 c3 c4 c5 c6 c7 c8 c9 c10.
Execute.

WRITE OUTFILE = 'C:\IRTtutorial\agr1_cal.dat'
/ id a1 a2 a3 a4 a5 a6 a7 a8 a9 a10.
Execute.

*Select validation data.

GET
FILE='C:\IRTtutorial\raw_reversed.SAV'.
EXECUTE .

FILTER OFF.
USE ALL.
SELECT IF(split =2).
EXECUTE .

*Writeout validation files with no spaces between entrees.

WRITE OUTFILE = 'C:\IRTtutorial\con1_val.dat'
/ id c1 c2 c3 c4 c5 c6 c7 c8 c9 c10.
Execute.

WRITE OUTFILE = 'C:\IRTtutorial\agr1_val.dat'
/ id a1 a2 a3 a4 a5 a6 a7 a8 a9 a10.
Execute.
 



If you ran the syntax above, SPSS would have created the files below. You may download these files for comparison.

  Agreeableness* Conscientiousness*
Calibration Sample agr1_cal.dat
agr2_cal.dat
con1_cal.dat
con2_cal.dat
Validation Sample agr1_val.dat
agr2_val.dat
con1_val.dat
con2_val.dat


*The data was split into agr1 and con1 for polytomous models and agr2 and con2 for dichotomous models.

Back To Main