Analyzing Interviewer Call Record Data Using a Multilevel Multinomial Modeling Approach to Understand the Process Leading to Cooperation or Refusal

• Activity Number: 383
• Type: Invited
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Julia D'Arrigo*+ and Gabriele B. Durrant and Fiona Steele
• Companies: University of Southampton and University of Southampton and University of Bristol
• Keywords: paradata ; interviewer call data ; multilevel multinomial logistic regression

Abstract

In interview-based household surveys, effective interviewer calling behaviors are critical in achieving cooperation and reducing the likelihood of refusal. This paper aims to analyze best times to gain cooperation from sample members in a range of face-to-face surveys. Of particular interest is to what extend the best times may vary with the type of household, such as a single household or a household with children. We use an unusually rich dataset, the UK Census Link Study, which combines paradata from six UK surveys, including detailed call record data, interviewer observations about the household and information about the interviewer-household interaction, which was linked to information about the household from the UK Census. The data have a multilevel structure with households nested within a cross-classification of interviewers and areas. A multilevel multinomial logistic regression approach which jointly models the different types of outcomes at each call is used to predict the likelihood of interview or refusal. The findings may have important implications for survey practice for determining best times of interviewer calls.