c © European Survey Research Association The Effect of Incentives on Response Rates and Panel Attrition: Results of a Controlled Experiment

Steadily declining response rates lead to an increased usage of incentives in all kind of surveys. As for mail surveys, much empirical evidence supporting the usage of unconditional incentives is available (Berk, Mathiowetz, Ward and White 1987; Church 1993). For face-to-face stud-ies, however, fewer results are available, and even less often in experimental settings (Martin, Abreu and Winters 2001; Singer, Hoewyk, Gebler, Raghunathan and McGonagle 1999). We conducted a three wave panel CAPI survey (N=600), where respondents were randomly assigned to three experimental groups: the first group received an unconditional incentive in form of a 10 e voucher, the second group was promised a 10 e voucher conditional upon participation, the third group was a control group that received no incentive. The allocation to a given experimental group was constant across the three waves and interviewers were kept blind to the experimental conditions. This research adds to the methodological literature in several ways: firstly, we compared the effectiveness of prepaid and conditional monetary incentives in a face-to-face interview in cross-sectional and longitudinal perspective, and found that conditional incentives performed significantly better in retaining respondents till the third wave. Secondly, we controlled for accurateness of the interviewers ’ work on the field, and found evidence that improving selec-tion and training of interviewers remains mandatory, since poor quality in the interviewer pool cannot be compensated by incentives