Examining the relationship between interview duration and data quality in the 9th round of the European Social Survey

Abstract

Response rates, survey costs, and the motivation of the interviewers can all be affected by the length of a questionnaire. Nevertheless, the effect of interview duration or interview speed on data quality is a rarely discussed problem. It is assumed that there is an optimal speed of an interview for every respondent, and that any deviation from that speed may lead to measurement error. Previous studies analyzing data of the European Social Survey (ESS) have already found a link between the speed of interviews and data quality, but the main focus of those studies was the role of interviewers rather than data quality. The aim of this study was to explore the relationship between interview duration and data quality in more detail by using more complex data quality indicators. We use data from the 9th round of the ESS from 2018. We found a weak, but significant association between interview duration and data quality in nearly half of the countries even when controlling for sociodemographic variables. Careless responses (high item-nonresponse rates and straightlining) were associated with too fast an interview speed, whereas socially desirable responses were associated with too slow a speed. Our results highlight that monitoring interview duration is important, and the training of interviewers needs to emphasise how to handle such negative respondent behavior and follow the optimal speed during interviews.