SPSS does not have a conditional logistic regression procedure. Some conditional logistic regression models can be estimated via COXREG and NOMREG (you can Google to find the details), but I think they
may be limited to 1:1 matching.
Another option you might consider is using GENLIN with generalized estimating equations (GEE). Some years ago, a colleague wanted to do conditional logistic regression with 3:1 matching, so I used the clogit command in Stata to estimate the models. Later, after GENLIN came along in SPSS, I estimated the same models with GENLIN & GEE.
The basic GENLIN syntax for my models looked like this:
GENLIN maj_surg (REFERENCE=FIRST) WITH gp
/MODEL gp INTERCEPT=YES
DISTRIBUTION=BINOMIAL LINK=LOGIT
/REPEATED SUBJECT=mchgrp SORT=YES
CORRTYPE=UNSTRUCTURED ADJUSTCORR=YES
COVB=ROBUST MAXITERATIONS=100 PCONVERGE=1e-006(ABSOLUTE) UPDATECORR=1
/MISSING CLASSMISSING=EXCLUDE
/PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION (EXPONENTIATED).
* gp = an indicator for General Practitioner (vs specialist).
* mchgrp = match group.
The results from these models were very similar to the conditional logistic regression results from Stata--see below.
Comparison of Conditional Logistic Regression & GEE
----------------------------------------------------------
Outcome | Odds Ratio [95% Conf. Interval] METHOD
----------------------------------------------------------
maj_surg gp | 1.599 1.064 2.403 1
| 1.606 1.072 2.406 2
----------------------------------------------------------
all_maj gp | 1.626 1.126 2.346 1
| 1.631 1.128 2.358 2
----------------------------------------------------------
blood gp | 0.755 0.536 1.065 1
| 0.839 0.616 1.142 2
----------------------------------------------------------
maj_oth gp | 1.579 0.734 3.340 1
| 1.585 0.735 3.419 2
----------------------------------------------------------
Method 1: Conditional logistic regression (Stata)
Method 2: GEE (via GENLIN in SPSS)
HTH.
ariela wrote
We need to do a conditional logistic regression for a case control study that is matched 1:2. Any ideas how to do this in SPSS (Version 21)?
--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/"When all else fails, RTFM."
PLEASE NOTE THE FOLLOWING:
1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above.
2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (
https://listserv.uga.edu/).
Locked
