The effect of incentives in survey experiments on panel stability and sample composition
Jul 25, 09:55
We examine the effect of differential respondent incentive payouts to panelists as a result of survey experiments on panel stability and sample composition.
Incentives have been successfully used as a method to increase or stabilize participation in longitudinal studies. Incentives have been paid out to all sample members or have been targeted towards a specific subgroup such as reluctant respondents.
In this presentation however we examine the effects of incentives that are used as rewards in survey experiments conducted as part of a panel study. Survey experiments are increasingly used by social scientists to generalize laboratory findings. Little is known about the methodological effect of conducting such experiments in general population surveys. In this study we focus on the use of respondent incentives in such experiments.
In the German Internet Panel, an online panel of the general population based on a probability sample, three survey experiments have been conducted in subsequent waves offering monetary incentives to a select group of respondents. Payouts went to a random group of respondents and their value varied between €5 and €80. This setting creates the unique opportunity to examine the effects of such differential incentives on panel attrition and sample composition controlling for incentive value. Receiving incentives could boost panel participation, not being among the winners could increase attrition. In addition, the effect of incentives could be different across specific subgroups of the sample and thus inducing a bias in sample.
Results will add to the body of knowledge on the use of differential incentives and conducting survey experiments in longitudinal studies