Join Dr. Ana Diez-Roux, for SERforum Live!
May 1, 12pm EST


Do you ever find yourself struggling to figure out a question about epidemiologic methods, or other topics in epidemiology, and don’t know who to ask? The SERforum allows for individuals to answer questions that come up in our daily work around substantive and methodological topics in epidemiology.
All topics may be viewed, but to read and post comments, SER membership is required. If you are member, login! Not a member, join us!

You need to log in to create posts and topics.

Non-collapsible measures and log-likelihood/AIC

Hi everyone,

I'm doing a re-analysis of matched case-control data that involved large number of predictors for a relatively small sample size. In the original analysis, the final multivariable model was selected using forward stepwise selection and conditional logistic regression (evaluated 37 predictors with 73 matched pairs). Model selection was based on AIC, rather than the magnitude of the OR. Beyond the other (bigger) issues with forward stepwise selection, I'm wondering if non-collapsibility of the OR could have any bearing on log-likelihood and thus AIC?

Many thanks,