Hierarchical Bayes estimation of response rates for an expenditure survey

Yong You, Philip Reiss

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Statistics Canada's Homeowner Repair and Renovation Survey (HRRS) is an annual supplement to the Canadian Labour Force Survey (LFS), measuring expenditures for additions, renovations, repairs and maintenance, and installation and replacement of household equipment. Currently the HRRS nonresponse is adjusted for by the same method as is used for LFS nonresponse: the sample is divided into response homogeneity groups (RHG), and weights within each RHG are multiplied by the inverse of its weighted response rate. This method produces somewhat unstable HRRS nonresponse adjustment factors, due to the small size of many of the RHG's. In this paper, a compound beta-binomial model is presented to smooth the direct survey estimate of the response rate of HRRS. A hierarchical Bayes approach and the Gibbs sampling method are employed to obtain the posterior estimate of the response rate, and the uncertainty of the estimator is measured by its posterior variance.
Original languageAmerican English
Title of host publicationProceedings of the Survey Methods Section
Subtitle of host publicationSSC Annual Meeting
Place of PublicationRegina, Canada
Number of pages6
StatePublished - Jun 1999
Externally publishedYes


Dive into the research topics of 'Hierarchical Bayes estimation of response rates for an expenditure survey'. Together they form a unique fingerprint.

Cite this