Improving Surveys with Paradata: Analytic Uses of Process Information

CHAPTER 12

MODELING CALL RECORD DATA: EXAMPLES FROM CROSS-SECTIONAL AND LONGITUDINAL SURVEYS

GABRIELE B. DURRANT and JULIA D’ARRIGO

Southampton Statistical Sciences Research Institute (S3RI), University of Southampton, UK

GERRIT MÜLLER

Institute for Employment Research (IAB), Nuremberg

12.1 INTRODUCTION

One class of paradata is interviewer call record data, sometimes referred to as call history data (Sangster and Meekins, 2004; Bates et al., 2008; Lee et al., 2009; Durrant et al., 2010; Blom et al., 2010; Durrant et al., 2013). Call record data is collected in interviewer-administered surveys, including both face-to-face and telephone surveys. Such data contain, for example, information about the day and time of the call to a sample unit and the outcome of the call. Further characteristics of the call may be captured, such as voice characteristics and initial responses from the householder or observational data about the household in face-to-face surveys. Call data may be recorded automatically via a computerized system or via the interviewer. In recent years, many survey agencies have started to collect basic call history data on a routine basis. Survey researchers hope to employ such data to inform and improve responsive and adaptive survey designs (Groves and Heeringa, 2006; Laflamme et al., 2008), that is, to monitor the survey data collection process and to inform intervention strategies. For example, the information may be used to identify more difficult cases with regard ...