Analysis of Complex Sample Survey Data

Abstract

Social survey data are often collected using complex sample designs that include unequal selection probabilities, stratification, and clustering; and this departure from simple random sampling assumptions requires special methods for variance estimation and inferential analysis. This article reviews the issues and techniques for analysis of complex survey data; and the application of special techniques is illustrated, based on various complex survey data. Discussion centers on methodological and some practical issues in data analysis and the need for proper analysis of complex survey data.