Analyzing Data from Complex Sampling Designs: An Overview and Illustration
2014-2015 Methodology Applications Series
Simple random sampling (SRS) is generally an impractical means for collecting data in education research. More complex sampling strategies, including stratification, clustering and unequal probability sampling, are often used to achieve valid statistical inferences. Conventional analytic approaches that assume SRS are not appropriate. This presentation will provide an overview of the theory and application of statistical methods appropriate for analyzing data from complex sampling designs. Model-based, design-based, and hybrid frameworks will be discussed.
Topics such as sampling weights and alternative variance estimators will be included in the discussion of these frameworks. Several data analysis examples will be provided to illustrate practical application using software packages such as Mplus, SAS and R.