Alternative Approaches to the Analysis of Complex Sample Survey Data: Applying State-of-the-Art Methods to NCSES Surveys

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

Many scientific fields conduct research using data collected from large-scale probability samples of individuals or establishments. Secondary analyses of these data can be complicated by the complex nature of the sampling strategies, which are employed for data collection efficiency. It is important that any secondary analyses employ appropriate analytic techniques that account for the complex sample designs. This concept can be described as examining analytic error within the broader framework of Total Survey Error (TSE), which is a very promising area of survey research. This project will begin by identifying and documenting potential analytic errors in several complex NCSES surveys. The next step will be to synthesize best practices for analyses of these data. In the principal focus of the project, these state-of-the-art methods will be compared and contrasted to existing analyses of NCSES data, and the implications that alternative approaches can have for the interpretation of the results. This research has the potential to improve the analytic uses of NCSES data, but also to make contributions to the more appropriate use of complex sample survey data in general.