White Papers

Fitting Longitudinal Mixed Effect Logistic Regression Models With the NLMIXED Procedure

• Tags:
• models,
• procedure,
• parameter,
• equations

Overview The NLMIXED procedure fits nonlinear mixed models; it is also useful for fitting linear mixed models having non-Gaussian error distributions. The authors lately used the NLMIXED procedure to perform longitudinal logistic regressions in which a random intercept was included to induce a compound symmetry covariance structure for repeated measures on individual subjects. They devised a macro to automate fitting this model. The NLMIXED procedure requires writing out regression equations, declaring parameter names, and providing initial parameter estimates. To get accurate initial parameter estimates for the NLMIXED models the authors fitted Generalized Estimating Equations (GEE) models with the GENMOD procedure.

• 
• Did you find this white paper useful?

• Subscribe to the IT White Papers Newsletter
• White Paper RSS