Procedure for dichotomous outcome and multilevel data

[wfung]

We are trying to decide the best procedure to use for dichotomous outcomes (whether students reached academic standard/score or not). The data we have is multilevel (students nested within schools). Does anyone have recommendation on the best procedure to use? Can PROC LOGISTIC handle multilevel analysis? 

[Ksharp · Accepted solution]

PROC LOGISTIC can NOT handle multilevel analysis.

For nested effect data, I suggest use GEE model (PROC GEE or PROC GENMOD + REPEATED)

or GLMM model(proc glimmix) .