Skip to main content
. Author manuscript; available in PMC: 2008 Oct 30.
Published in final edited form as: Ann Epidemiol. 2006 Oct 4;16(11):850–859. doi: 10.1016/j.annepidem.2006.04.013

TABLE 2.

Latent class logistic regression results for a two-class model of underage drinking using a GEE2-type approach with naive, within-subcluster–and within-cluster–adjusted covariance estimators

Naive independence
 Robust within-subcluster
 Robust within-cluster
Estimate SE p SE p SE p
Male 0.435 0.062 < 0.001 0.122 < 0.001 0.125 0.001
Age 0.397 0.032 < 0.001 0.062 < 0.001 0.062 < 0.001
Urban −0.104 0.066 0.115 0.129 0.420 0.148 0.482

GEE2 = second-generation generalized estimating equation; SE = standard error.