I am using the extension command SPSSINC RASCH to fit a set of items to the Rasch model and save person fit statistics, using the following commands:
/VARIABLES ITEMS=aitem1 aitem2 aitem3 aitem4 aitem5 aitem6 aitem7 aitem8 aitem9 aitem10
/OPTIONS MISSING = LISTWISE EXECUTE=TRUE
/PRINT SUMMARY ITEMFIT
/PLOT DESCRIPTIVES FACTORSCORES ICC IIC
The person fit data is to be saved in the new data set subj_fit, as specified in the /SAVE subcommand. The data set subj_fit is created and has the expected person fit variables in the dictionary. Also, the number of cases in that data set reflects the number of observed item response combinations. However, the values for all of the variables are system missing for all of the cases. The Rasch output in the Statistics Viewer (item coefficients and fit statistics, graphs, etc.) appear reasonable and there are no error messages. Why is the person fit data set full of system missing values?
The empty person-fit file was due to a defect in an early version of the SPSSINC RASCH command that caused all the values in the dataset to be missing. This defect has been fixed in the current version of the SPSSINC RASCH command that is available in the SPSS Community Extension commands page.