Journal of Official Statistics, Vol.18, No.1, 2002. pp. 1–11
Evaluating Socio-economic Status (SES) Bias in Survey Nonresponse
John Goyder Keith Warriner and Susan Miller
Abstract:Recent work by Groves and Couper reraises the issue of how typical upper SES bias is in survey response. Using data from the Kitchener-Waterloo area of Ontario, Canada, this issue is addressed through a set of local area surveys record-linked to sampling frame information. An innovation is the use of status-ratings of photographs of households in a pretest sample. These are used to generate regression weights for integrating three pieces of information - property value, ownership and dwelling type - into an index of residential SES. This index is then used for evaluating SES bias across seven of the local area surveys. Consistently, and across three modes of contact, the "middle class bias" to be expected from much previous work on nonresponse bias is present.
Keywords:Survey nonresponse; middle class bias.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group