Analysis of Longitudinal Complex Survey Data Using Parametric Bootstrap

• Activity Number: 105
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Snigdhansu Chatterjee*+ and Partha Lahiri
• Companies: University of Minnesota and University of Maryland
• Keywords: Small Area Estimation ; Mixed Models ; Resampling

Abstract

We develop a general theory for parametric bootstrap method in the context of a general mixed model and illustrate its usefulness in analyzing a variety of longitudinal complex survey data. As a special case, we consider the problem of estimating small area characteristics using time series and cross-sectional model that combines data from previous time points of a longitudinal survey and relevant auxiliary variables. We use our general methodology to demonstrate the flexibility of our parametric bootstrap method to produce highly accurate mean squared error estimates of complex estimators with calibration and winsorization adjustments.