The Effect of Targeted Incentives on Response Rates and Representativeness: Evidence From the Next Steps Age 32 Survey

This paper analyses the effect on response rate and non-response bias of tailoring the level of monetary incentives offered to participants in a longitudinal study, depending on sample members’ response propensity. Specifically, we test whether offering higher value incentives to prior wave non-respondents (and lower value incentives to prior wave respondents) would lead to overall higher response rates and lower non-response bias, compared to offering to all sample members the same monetary incentive. In order to test these hypotheses, we use large-scale experimental data from the Next Steps Age 32 survey. Next Steps is a longitudinal cohort study, following the lives of approximately 16,000 participants born in 1989-90. We find that offering higher incentive to prior wave non-respondents did not significantly increase response. We also do not find support for targeted monetary incentives being particularly effective at boosting response amongst particular population subgroups. The use of targeted incentives has to date been relatively rare, particularly in the UK context, but in this study the approach was not found to be successful.