Hi all,

I recently upgraded from SPSS version 25 to version 29 and noticed that the GENLIN output gives me somewhat different results. I have already verified that the data set is the exact same and that no filters were used differently. The syntax I used both times is:

genlin Q6Effective_Binary (reference = first) by Item10_Cat (order = descending)
/model Item10_Cat intercept = yes
distribution = binomial link = logit
/repeated subject = Participant_ID sort = yes corrtype = unstructured
adjustcorr= yes covb = robust
/print cps descriptives modelinfo fit summary solution(exponentiated).

What I noticed is that the exp(B) differ somewhat. Does anyone have a clue about what might have changed between versions 25 and 29?

The differences observed in the GENLIN output between SPSS version 25 and version 29, despite using the same dataset and syntax, could be attributed to various factors. One possibility is that SPSS updated the default settings for the GENLIN procedure in version 29, which could affect convergence criteria, optimization algorithms, or other aspects of the analysis. Another factor could be algorithm enhancements in version 29, leading to more accurate and precise estimates. Additionally, bug fixes may have been implemented in version 29, resulting in more reliable and consistent results.