Applying Post-Stratification Raking Adjustments to Survey Weights Using the High School Longitudinal Survey of 2009 (HSLS:09)

• Activity Number: 79
• Type: Contributed
• Date/Time: Sunday, July 31, 2016 : 4:00 PM to 5:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Austin Lasseter* and Jonathan Phelan
• Companies: Summit Consulting, LLC and American Institutes for Research
• Keywords: raking ; weights ; survey ; post-stratification ; bias ; coverage error

Abstract

This study examines how post-stratification raking adjustments to survey weights can reduce bias due to non-coverage in order to facilitate a cross-cohort comparison between the High School Longitudinal Study of 2009 (HSLS) and the NAEP HSTS survey. HSLS is a nationally representative study of 23,000 students from 944 schools who were 9th graders in 2009. HSLS survey weights are calculated for each student, representing the inverse probability of being sampled, with adjustments for nonresponse. The study begins with a brief overview of the rationale and methodology, including coverage error, measuring bias from non-coverage, and the use of post-stratification raking adjustments. The study then examines how post-stratification raking adjustments were applied to survey weights for three distinct subsets of the full HSLS sample: public school students, public and private school students, and those students who participated in the HSLS-NAEP overlap sample. Results indicate that raking adjustments substantially reduced bias due to non-coverage in all three subsets of the data.