Online Program

Representativeness of response to establishment surveys and registers
*Barry Schouten, Statistics Netherlands 
Natalie Shlomo, University of Southampton 

Keywords: Nonresponse, Business surveys

The response rate is an indicator of quality of response to a survey; it limits the maximal impact of selective response. However, the response rate presents an incomplete picture as it is the contrast between respondents and nonrespondents that underlies nonresponse bias. In recent literature, this deficit of the response rate has been mentioned and various alternative indicators have been proposed. In this presentation, we present one such indicator, the R-indicator. R-indicators are based on a definition of representative response; response is representative with respect to some set of characteristics when response propensities are constant for these characteristics. Along with the R-indicator comes a set of additional indicators, so-called partial R-indicators, that allow for detailed analyses of the representativeness of response. The partial R-indicators are available at the variable-level and at the category-level, and measure the impact of specific variables or categories of these variables. In the presentation, we illustrate and discuss the various purposes and features of the indicators with applications to establishment surveys and VAT registers.