0
$\begingroup$

I'm looking for a way to produce a matrix of predicted probabilities on data that went through SPSS's logistic regression test. I only use two ordinal variables with a range of 1-4 and 1-10 respectively. Many other examples where the first variable has scale properties are demonstrated all over across the web that involve simple logit computing.

The one that I've found to be extremely close to what I'm asking involves three ordinal variables but fails to explain thoroughly the logic behind SPSS's matrix commands.

http://www.ats.ucla.edu/stat/spss/dae/ologit.htm
$\endgroup$
0
$\begingroup$

Why don't you just save the predicted probabilities directly from LOGISTIC? And perhaps then crosstab these by your two variables?

$\endgroup$
  • $\begingroup$ I saved every single predictory variable in the UCLA example from the output options but none of them seems to match those generated by the equation. My only problem in regards to the link example is the part when x is computed; I can't quite comprehend the purpose of matrix commands within the cells of a greater matrix and how the whole process is being interpreted. That's basically going to help me transition this to my 13x4 example. $\endgroup$ – John May 1 '15 at 18:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.