Looking for documentation for propensity score matching in Version 27 under "data".

Tried to find through Help --> SPSS support as well as drop down Help -->documentation through PDF format (English) and couldn't wade through each doc to find this subject.

Either found old info or info about the "Fuzzy not defined error" which I did encounter and used the workaround method so the script would run (running another extension like FUZZY/help).

However, is there full documentation?

See the dialog box help. And for more information, see the syntax help for FUZZY (Run FUZZY /HELP or press F1 on an instance of the command.) These are not covered in the Syntax Reference, because they are extensions.

If you have further specific questions, post them here, and I'll respond. I am the author of these items.

Thanks for your response. What do you mean by the "Diaglog box help"? Are you referring to the information in the drop down GUI boxes that appears in the corner "This procedure runs a logistic regression..." etc. as well as the info that appears when you hover over the fields?

I'm looking for documentation that explains what the match tolerance is referring to- Is this the same as a caliper?

What is the Maximize execution performance under Options?

I'm also encountering an error after the syntax runs and noticed a temporary variable from the syntax is still in the dataset along with the new variables. Everything looks in order, but I get error

LINE 44 Which is the end of the program line "end program."

COMMAND: DATASET ACTIVATE

INFORMATION: Expecting word or =. Found DataSet 1.

Also, another error:

Text: M0.9096370428527336 Command: DELETE VARIABLES

An undefined variable name, or a scratch or system variable was specified in a variable list which accepts only standard variables. Check spelling and verify the existence of this variable.

Execution of this command stops.

In the output after the above is repeated:

Traceback (most recent call last): 

 File "<string>", line 39, in <module> 

 File "C:\PROGRA~1\IBM\SPSS\STATIS~1\27\PYTHON3\Lib\site-packages\spss\spss.py", line 1508, in Submit 

  raise SpssError(error) 

spss.errMsg.SpssError: [errLevel 3] Serious error.

The dialog box help says

Match Tolerance Specify the tolerance for the score in matching cases and controls. A control is eligible to match a case if the absolute value of the difference in the propensity scores is less than or equal to this value. A value of 0 means exact matches only while a value of 1 means any control would match any case. Smaller values produce closer matches but may increase the number of unmatched cases.

• The output tables produced by this procedure show the distribution of the number of eligible matches given the tolerance and can be useful in choosing a tolerance value.

Regarding execution performance, the FUZZY syntax help says this:

By default, memory usage is minimized in picking supplier dataset candidate match cases (all eligible cases have an equal probability of selection). This requires an extra data pass. If the possible number of matching cases for a demander case is small or the supplier dataset is not large, specifying MINIMIZEMEMORY=FALSE may improve performance by eliminating the extra data pass. In the case of 1-1 matching, this is recommended. MINIMIZEMEMORY=TRUE cannot be combined with EXACTPRIORITY=TRUE.

When MINIMIZEMEMORY=TRUE, each supplier case is eligible for assignment at random to just one demander case for which it is within tolerance. When FALSE, the case is eligible for all cases within the tolerance, and one is picked at random from all of those. This means that when TRUE, a usually small number of demander cases may go unmatched when a match could have been found, but when FALSE, the matching tables can become very large, especially in the PSM case.

As for the error, please send me the spv file that shows the complete output (jkpeckSupport Member).