Variance Estimation for Survey-Weighted Data Using Bootstrap Resampling Methods: 2013 Methods-of-Payment Survey Questionnaire

• Activity Number: 21
• Type: Topic Contributed
• Date/Time: Sunday, August 9, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Heng Chen* and Rallye Shen
• Companies: and Bank of Canada
• Keywords: variance estimation ; raking ; calibration ; resampling ; bootstrap

Abstract

In the 2013 Methods-of-Payment (MOP) survey, sampling units were selected through an approximate stratified random sampling design. To compensate for non-response and non-coverage, the observations are weighted through a raking procedure so that the weighted sample is representative of the population with respect to control variables. The variance estimation of weighted estimates must take into account both the sampling design and raking procedure. We therefore propose using bootstrap resampling methods to estimate the variance. We produce replicate raking weights for the questionnaire (SQ) portion of the survey using the bootstrap resampling method and use them to compute the variance of weighted estimates. We find that the variance is considerably different when estimated through the bootstrap resampling method than when estimated through Stata's linearization method, where the latter does not take into account the correlation between the control variables and the outcome variable.