Abstract
We propose an incomplete-data, quasi-likelihood framework for estimation and score tests that accommodates both dependent and partially observed data. The motivation comes from genetic association ...