ABSTRACT

In order to adjust individual-level covariate effects for confounding due to 
unmeasured neighborhood characteristics, we have recently developed conditional 
pseudolikelihood and generalized linear mixed model methods for use with complex 
survey data. The methods require sampling design joint probabilities for each 
within-neighborhood pair. For the conditional pseudolikelihood methods, the 
estimators and asymptotic sampling distributions we present can be conveniently 
computed using standard logistic regression software for complex survey data, 
such as SAS PROC SURVEYLOGISTIC.  For the generalized linear mixed model 
methods, computation is straightforward using Stata's GLLAMM macro. We 
demonstrate validity of the methods theoretically, and also empirically using 
simulations. We apply the methods to data from the 2008 Florida Behavioral Risk 
Factor Surveillance System survey, in order to investigate disparities in 
frequency of dental cleaning both unadjusted and adjusted for confounding by 
neighborhood.

Coauthors: Zhuangyu Cai, Zhulin He, Hao Zheng, Amy Dailey.