Moving from CAPI to Sequential CATI-CAPI Data Collection in a Household Panel Survey: Experiences from the UKHLS Innovation Panel

Abstract

This paper assesses the effect of changing the mode of data collection in the UKHLS Innovation Panel on different aspects of data quality. Wave 1 of the Innovation Panel was conducted with face-to-face interviews (CAPI). At wave 2, households were randomly allocated to either CAPI or telephone interviewing (CATI). In the CATI group, non-respondents were further followed up in CAPI. The paper assess the effects of the mode treatment on individual and household-level response rates, item non-response and response distributions.