Using Cari Tools To Understand Measurement Error

Abstract

Computer audio-recorded interviewing (CARI) has been used for more than a decade to detect field-interview fabrication, but it has the potential for much more. On the 2007 National Home and Hospice Care Survey, an innovative CARI system combined with behavior coding assessed problematic questions, monitored general compliance with protocols, and evaluated individual interviewer performance. This article discusses the benefits of CARI for recording large samples of interviews in a systematic, objective manner at reduced cost and less operational burden than traditional audio recording. The increased objectivity and volume of recordings allowed identification of potential data-quality issues, specifically measurement error due to the questionnaire, the interviewer, and their interaction. CARI's potential as a tool for estimating the magnitude of the error in survey data is also discussed.