Propensity Score Matching SPSS 23.0.0
Hello,
i have the following Problem.
We want to do a propensity score matching (PSM). Now we have installed the following issues. Normally PSM is integrated in this SPSS version, but in our company it was excluded.
Some Facts: - we are working with Windows 7 32-Bit. -Installations: - IBM SPSS Statistics 23.0.0.0 (32-Bit) - Python is installed Plug-in for Python - Plug-in for R - Plug-in for Java - R for Windows 3.1.0 (32-Bit) - SPSS_Statistics_REssentials_23.0-FP2_win32 - PSMATCHING3.04.SPE as Administrator in SPSS is installed.
Now we have in SPSS two options to do the PSM: SPSS -> Data (see Details under Path I) SPSS-> Analyse (Path II)link text
Both analysis show different outputs and partly with error reports. The two outputs you can see attached and the Analysis with variables selectionlink text.
In the Analysis option "SPSS-> Analyse-> PSM": you can see error messages (e.g. $.data.frame)- can you say me if the analysis is working nevertheless? What is the Problem? Because this analysis with more options to choose (and the possibility to create graphs) is our favorite one.
I hope anyone can help me with this problems.
Kind regards, JuWe
You can find a short, nontechnical discussion of PSM here: psm
After that, you can go a little deeper with this article.
Thank you for replying. I really appreciate your help. could you please guide me to a video tutorial where I can do step by step? or other kinds of tutorials?
I am sorry, but I don't have additional references to suggest.
thank you for your help before, I worked through and I am still feeling my way around this. I have rearranged my data to be able to generate PSM output but after going through the wizard for propensity score, how can I get the output in a graph? in the output, there is: logistic regression, block 0, block 1, and case-control matching. I don't know how to generate the graph
These are two unrelated procedures. I would not expect them to give identical results, although both may be valid.
The messages from your preferred procedure appear to be harmless and can be ignored. If you have any doubts, you can check the matches outside the procedure to be sure that they make sense.
Thank you for your reply. This sounds good! Can you please explain what do you mean exactly with check matches outside the procedure? Thanks a lot!
@JuWe
I meant that if you look at individual paired case and control records, you can see how closely they match.