Analysis Using Sampling Weights and Appropriate Standard Error Estimation With NCES Data

Session Type: Professional Development Course

Abstract

This course will cover weighting and sampling design issues in several National Center for Education Statistics (NCES) large-scale datasets (e.g., ELS, ECLS-K, PIRLS, and SASS), focusing on the specialized statistical procedures required to analyze the datasets. This course will enable participants to: explain typical sampling designs used in data collection; describe differences among the many weights available (e.g., sampling weights, panel weights, replicate weights); and undertake basic statistical analysis. The course will focus on obtaining unbiased parameter estimates and standard errors by using sampling weights and specialized variance estimation techniques including both linearization and replication. Lecture material will be woven into hands-on exercises throughout the course. Participants should bring their own laptop with SPSS (including the “Complex Samples” module) or SAS, as well as the free Mplus demo.

Sub Unit

• Professional Development and Training Committee

Instructor

• Laura M. Stapleton, University of Maryland