Tenure/Promotion Colloquium, G. Zhang
Title: Generalized Variance Functions for Longitudinal Data
Abstract: Generalized variance functions (GVFs) are used to produce convenient published
estimates of variances for a number of large surveys such as the Current
Population Survey (CPS). The GVF pooled together one year or some certain
time period of data and treated the population total as a constant. However, as
the size of population changes over time, standard error of the estimators can
actually change substantially. In this research, we propose longitudinal
generalized variance functions (LGVFs) by incorporating time effect into modeling.
Asymptotic properties of the estimators that are linear combination of cluster
means from stratified two-stage cluster samples are investigated. Implementation
of the methods to CPS show that LGVFs are efficient in producing the standard errors.
Contact Name: Guoyi Zhang
Contact Email: email@example.com