Web Survey Bibliography

Title Computing weights for the American National Election Study survey data
Author Debell, M., Krosnick, J. A.
Year 2009
Access date 18.04.2013
Full text

PDF (51,6 kb)

Abstract

This report describes methods to calculate weights for ANES studies that account for the sampling design and match population benchmark for selected variables. ANES data are based
on complex sample designs and must be weighted to adjust for the sample design in order to generalize to the population. Weights for such analysis must adjust for unequal probability of
household selection and for respondent selection within households. Additionally, the method described here includes nonresponse adjustment and post-stratification raking when needed to
match known population benchmarks. Poststratification factors should be selected based on a comparison to benchmark statistics, to correct notable differences from benchmarks. The
selection of poststratification factors, coding categories, and trimming thresholds should be made to strike a balance between variance in the weights and accuracy of the estimates. In panel
studies, weights should be calculated for each combination of waves that will be analyzed together. Panel weights may also be developed to adjust for panel attrition. Details are described
at each step in the process.

Year of publication2009
Bibliographic typeReports, seminars
Print

Web survey bibliography - 2009 (631)

Page:
Page: