0

I want to use NOMREG of SPSS (by GUI from "Regression --> Multinomial Logistic Regression") for my matched data. However, I don't know where to insert the strata variable (the matching variable) into the GUI or syntax.

On a side note, I have a question on conditional logistic regression in R that have posted it to the programming branch of the StackExchange because the last time I sent a code-related question here, mods fried me (although I see other people send their R code here). Anyways, the address for my R question is this (LINK).

Community
  • 1
  • 1
Vic
  • 167
  • 1
  • 10
  • I believe this question may belong on SO, too--it's not appropriate here because it is solely about how to control your particular stats program. You can find good SPSS links at http://meta.stats.stackexchange.com/a/807. – whuber Aug 28 '13 at 18:48
  • Removed [R] tag and changed the title to reflect the content of this question. – IRTFM Aug 28 '13 at 18:59
  • A quick [search of google](https://www.google.com/search?q=conditional+logistic+regression+SPSS&oq=conditional+logistic+regression+SPSS) brings up a variety of answers. It appears it is simplest to just use `COXREG` with the `STRATA` sub-command. – Andy W Aug 28 '13 at 19:17
  • You're allowed to post R code at CrossValidated if the fundamental question is about statistics and not the code per se. Also, this question could get closed for being too broad. It's really two. – John Aug 28 '13 at 19:22
  • Andy W, of course I could do so and I had googled it and had found it. This is exactly why I clearly stated using multinomial logistic regression, not cox regression. But thanks anyway. – Vic Aug 28 '13 at 20:09
  • John, I considered asking the same question once for R and once for SPSS double posting. So I wanted to avoid it. – Vic Aug 28 '13 at 20:11
  • Whuber, thanks for the link – Vic Aug 28 '13 at 20:11

0 Answers0