SPSS Statistics

 View Only

 Propensity score matching leads to endless spiral of loading and all values turning into "?"

Tobias Holzer's profile image
Tobias Holzer posted Wed December 03, 2025 09:35 AM

Hello,

I'm new to using propensity score matching. My SPSS seems to have no problems with case control matching, but propensity score matching doesn't work. 

As soon as I start the calculations, every value turns to a "?". It shows that the programm is busy doing fuzzy grouping, but nothing happens after that. This happens even if I use a really small data set, for example 50 values. My processor is not being used extensively during that time (usage below 3 percent, so it doesn't seem to be a problem of computing power). 

Installing and reinstalling the Add-Ons for Propensity Score Matching or Python/R doesn't change the outcome. 

Anyone had this happen before? Would love to hear a solution! Thank you very much!

Jon Peck's profile image
Jon Peck IBM Champion posted Wed December 03, 2025 10:18 AM

Case-control matching and PSM use the same underlying code via the FUZZY extension command, but PSM first runs a logistic regression and uses the resulting probabilities as the matching variable.  (CCM also has a few options not available in PSM).  So PSM, which is actually the STATS PSM extension command, generates a FUZZY command and executes it using a Python call to the Submit function.

Unfortunately, the Submit function was broken, apparently in SPSS version 30, resulting in the behavior you are seeing.  The fix for this is in SPSS v31.0.1.0, which you can install from IBM Fix Central if you are already on V31.0.0.0.  Otherwise, the workaround would be to run the logistic regression directly, saving the predicted probabilities, and then using that with CCM (FUZZY)

Related Content