Emanuela Sala, Dipartimento di Sociologia e Ricerca Sociale, Universita degli Studi di Milano-Bicocca (Italy)
Jonathan Burton, Gundi Knies
Data linkage, consent rates and consent bias, interview features
Linkage of administrative data to survey data is becoming increasingly popular both in the UK and elsewhere. Major social surveys have linked their data with a wide range of administrative data including data on benefit receipt, adolescent’s school performance, health and morbidity. The reason why data linkage is so appealing to researchers lays in the potential to overcome some of the main challenges currently facing survey practitioners, for example by improving data quality, reducing survey costs in the longer term and easing respondent (and interviewer) burden.
Despite its wide spread use, there is very little methodological research on data linkage in household panel surveys. As asking for consent to link administrative records to survey data is often compulsory, obtaining respondents’ consent is key for the success of the data linkage procedure. Most of the survey design decisions are often based on anecdotal accounts and common sense rather than being driven by sound empirical evidence. Such decisions, however, may have strong implications as they are likely to have an impact on consent rates and bias.
This paper investigates the impact of survey design features (i. e., type of question and position of the consent question) on consent to datalinkage and provides a deeper understanding of the reasons why people consent or do not consent. Using a unique set of experimental data collected in Wave 4 of the Innovation Panel and recordings from survey interviewes, we show that, under certain conditions, interview features such as question format (dependent/independent questions) and placement of the consent question may have an impact on consent rates. We also find evidence that shows that a specific interviewer training and carefully drafted question wording may have a positive effect on consent rates. The paper also provides practical guidance to survey methodologists and survey agencies on the implementation of data linkage .