Hilda Project Discussion Paper No 1/12, April 2012

Abstract

Computer-assisted personal interviewing (CAPI) offers many attractive advantages over paper and pencil interviewing (PAPI) and most large ongoing face-to-face surveys using PAPI methods are likely to eventually shift to CAPI. Its advantages include improved and more complex routing, automated checking of inconsistencies in the data with the respondent as they occur, simultaneous data entry, reduction in lags between interviews and data delivery, and capture of system information, such as section timestamps, during the interview (de Leeuw, Hox and Snijkers, 1995). Further, in panel surveys, CAPI facilitates much greater use of data collected in previous waves, thereby improving data quality (Jäckle, 2008). Despite these apparent advantages, it is important to have a good understanding of the impacts that a move from PAPI to CAPI has on survey outcomes, both to enable survey managers to make appropriate implementation and budgeting decisions, and to ensure appropriate interpretation of the resulting data.