418 – Current Research and Evaluation Topics in the American Community Survey
Enhancing Respondent Representativeness Through Responsive Design and External Benchmarks
Julia Shin-Jung Lee
University of Michigan
Conventional methods for adjusting nonresponse bias are implemented at the end of data collection, using weighting methods such as raking and poststratification. These nonresponse adjustment methods are employed under the assumption that respondents and nonrespondents are similar for a given weighting cell or weighting covariates. This assumption can not be verified in the absence of nonrespondent information. To obviate such assumption, we describe a model-based strategy that combines prediction and balancing using a benchmark to improve sampling and analyses of a current survey employing multi-phase data collection. Models predicting sample characteristics from frame and contextual information are fit to data from the benchmark survey, which shares the same frame and contextual information as the current survey of interest. The fitted models are used to predict sample characteristics for the current survey to guide sampling decisions aimed at obtaining samples that better represent the targeted population. This study provides a framework for a stochastic responsive design strategy that aims to simultaneously attenuate nonresponse bias and increase inference precision.