Effects of questionnaire and fieldwork characteristics on call outcome rates and data quality in a monthly CATI survey.

摘要

The goal of this study is twofold. First, it aims to examine the effects of questionnaire design and fieldwork characteristics on outcome rates in a monthly RDD survey conducted by the Center for Survey Research (CSR) at The Ohio State University. Second, the study also attempts to uncover possible ways the fluctuations in outcome rates may be influencing data quality. The analysis is based primarily on data gathered from CSR administrative records, as well as information generated by the CATI program. The main variables of interest are the response, cooperation, refusal and contact rates calculated based on the final disposition codes recorded for each month, and the possible determinants of these rates. The monthly values for additional measures are calculated from the individual level data, specifically, the answers provided by respondents to media use and demographic items included in the survey.;The study finds that the main factors influencing outcome rates are the presence of newly hired interviewers on the survey, questionnaire length, and the percentage of completed interviews that are the result of refusal conversions. Examining the effects of fluctuations in the outcome rates on the demographic representativeness of the surveys reveals that higher nonresponse does not always amount to less representative data. However, the results also show that study characteristics, including outcome rates, may influence not only the marginals obtained from a survey, but also the relationships between variables, in this case the correlations between media use and demographic measures.