A Laboratory Approach to Measuring the Effects on Survey Participation of Interview Length, Incentives, Differential Incentives, and Refusal Conversion
Robert M. Groves, Eleanor Singer, Amy D. Corning, and Ashley Bowers
A laboratory-based experimental method is used to study influences on survey participation. The experiment tested specific hypotheses concerning the combined effects of person-level attributes and survey design features on cooperation with a survey request. The laboratory protocol presented subjects with videotaped vignettes portraying interviewer persuasion attempts, and then elicited their reactions to them.
The vignettes were constructed so that three factors were systematically varied in a factorial design: (1) disclosure of differential payment to other sample households by the interviewer; (2) refusal conversion efforts by the interviewer; and (3) burden, as reflected in the length of interview. These factors were completely crossed in the design. In addition, two other factors were partially crossed: offering a rationale for the differential payment versus no rationale; and offering an incentive prior to the interview versus offering no incentive.
This article presents a description of the methodology as well as analyses testing the effects of the experimental factors.
Nonresponse; burden; laboratory; experiment; vignettes; incentives