193815_en-US

liliconcom

Adjusted predictive values from a logistic model

When an LSM analysis is run, the solution model provides means adjusted for the effects of the model as well as non-adjusted means for each categorical element in the model. After running a logistic model with multiple predictors or an interaction, is it possible to obtain "adjusted" predictive values with confidence intervals for a different combination of predictors? I would appreciate any guidance or a similar example that deals with this issue.

 

For example, a logistic regression with two continuous covariates where both Cov1 and Cov2 are significant. How do I obtain values (or visualize) the effect of Cov1 when Cov2 is held at its mean?

Many thanks!

Re: Adjusted predictive values from a logistic model

Have you used the Prediction Profiler yet with the fitted model?

Re: Adjusted predictive values from a logistic model

Thanks for the suggestion. I will try the profiler.