Survey introductions and data quality

While measurement error arising from the interaction between respondent and interviewer during the survey interview has been the focus of much research, the impact of the interaction during the introductory request for participation on subsequent data quality has received less attention. This article attempts to link these two sets of interactions using data from the 1990 National Election Study. At the end of each contact with a sample person, interviewers completed a series of structured items, providing information on what was said by the sample person and interviewer, and other characteristics of the contact. These data, combined with the substantive data for the interviewed cases, allow us to explore the relationship between respondent statements during the introduction and the content and quality of information provided by them during the interview. We find that the statements made by respondents during the introduction are associated with differences in both data quality and substantive responses. We discuss the implications of these findings for both nonresponse and for response error.