Impact of Prepaid Incentives in Face-to-Face Surveys: A Large-Scale Experiment with Postage Stamps

Abstract

Nonresponse is a general problem in survey research. Although response rates and the nonresponse trends differ between countries, the general trend is that response rates have been declining over the years (De Leeuw & De Heer, 2002). A high nonresponse rate is a problem as it reduces the number of respondents and consequently the precision of estimates. In addition, nonresponse can be selective. This occurs when nonrespondents differ systematically from respondents as to the survey objectives. As a result, the survey estimates of the key indicators may be biased. Nonresponse is not only a statistical problem, but it is also a financial problem as declining response rates cause increasing survey costs. In the nineties, a response rate of 55 percent in face-to-face surveys conducted by Statistics Netherlands was not unusual (De Heer, 1999). Due to a substantial reorganization of the fieldwork department and the raise of the minimum number of contact efforts from three to six, the response in Statistics Netherlands’ face-to-face surveys gradually increased to 65 percent in 2005 (De Bie & Luiten, 2005). In order to realize a further increase in the response rate, we decided to focus on the refusals. As Statistics Netherlands does not use a refusal conversion policy—which is a common practice in many other fieldwork organizations, including National Statistical Institutes—we had to consider other measures. One option was the use of incentives. Prepaid cash incentives are known to be most effective, but sending cash by mail is not allowed in the Netherlands. Therefore, postage stamps are chosen as a prepaid incentive, which are expected to be an opportunity to bridge the gap between cash and material incentives. In this article, the effect of stamps as a prepaid incentive in face-to-face surveys on response is tested in a large-scale field experiment embedded in the Dutch Labor Force Survey.