Beschreibung

"Many survey organizations attempt to reduce response burden in establishment surveys by augmenting the collected interview data with external record information. However, not all responding establishments actively grant survey organizations permission to access and link their external record data. The lack of full consent has implications for the usefulness of the linked data and the accuracy of any inferences derived from them - issues that have only been scantily explored in the establishment survey literature. In this article, we test a strategy for optimizing linkage consent rates and minimizing the risk of consent bias in a web survey of establishments by experimentally manipulating the placement of the consent question in the questionnaire. We consider three approximate consent question placement locations: beginning, middle, and end. In addition to examining the role of question placement on the consent rate, we also examine the magnitude of consent bias in estimates derived from a set of commonly linked administrative variables. In short, we find that placing the linkage consent item at the beginning of the questionnaire yields a higher consent rate than placing it at the middle or end of the questionnaire. Furthermore, we show that the magnitude of linkage consent bias - which is generally small across a range of linked administrative variables - is largest when the consent question is placed at the end of the questionnaire. We conclude with a discussion of the implications of these findings for survey practice." (Author's abstract, IAB-Doku) ((en))