An assessment of the utility of a Bayesian framework to improve response propensity models in longitudinal data

Abstract

Response propensity (RP) models are widely used in survey research to assist in targeting fieldwork resources. With a good RP model, likely nonrespondents can be targeted using responsive and adaptive strategies with the aim of increasing response rates, improving survey efficiency, and potentially reducing survey costs. Generally, however, RP models exhibit low predictive power, which limits their effective application in survey research to improve data collection. This paper explores whether the use of a Bayesian framework can improve the predictions of response propensity models in longitudinal contexts. In this approach coefficients from a previous wave RP model are used to specify informative priors for the next wave model. We apply this approach using the first five waves of the Understanding Society panel survey. Our findings indicate that conditioning on previous wave data does not produce worthwhile improvements in the predictive accuracy of the response propensity models.

More information

Identifiers

ORCID for Gabriele B. Durrant: orcid.org/0009-0001-3436-1512

ORCID for Olga Maslovskaya: orcid.org/0000-0003-3814-810X

ORCID for Patrick Sturgis: orcid.org/0000-0003-1180-3493

Catalogue record

Export record