Between Task Complexity and Question Sensitivity: Nonresponse to the Income Question in the 2008–2018 European Social Survey

Abstract

Apart from unit nonresponse, failure to obtain a response to a particular survey question is a crucial component of the Total Survey Error. It decreases the sample size and boosts the risk of significant nonresponse bias, which warrants further investigation if many respondents abstain from answering a specific question. In this article, we take as a case study item nonresponse to the question about a household’s total net income in the European Social Survey (2008–2018), in which the income question received the highest item nonresponse across all survey rounds. We explore two mechanisms responsible for the likelihood of nonresponse to the income question: task complexity and question sensitivity. Additionally, we broaden this framework by including age and gender (mis)match between respondents and interviewers and consider how interviewer dissimilarity from respondents increases task complexity and question sensitivity when answering the income question. In so doing, we apply multilevel multinomial regressions to model income nonresponse ‘don’t know’ and ‘refuse to say’ answers (refusals). We find that both tested mechanisms operate at the individual, interviewer and country levels, indicating that item nonresponse is a multi-layered phenomenon in the survey process.