
GESIS Panel Research Spotlight: Saskia Bartholomäus - Mitigating Attrition Bias in Political Variables by Improving the Survey Experience of Politically Disengaged Panelists
Free
- Length
- 1 hr
- Platform
- Zoom
- Topic
- Politics & Policy
About this webinar
Politically disengaged panelists are less likely to continue participating in panel surveys covering political topics, which progressively biases estimates of political variables. To overcome this problem, I examine whether improving the survey experience of politically disengaged panelists can mitigate these biases. I present two studies. In the first study, I use data from the GESIS Panel to examine how political interest and survey experience relate to participation in subsequent waves. The results suggest that a positive survey experience may increase the participation propensities of politically disengaged panelists, aligning them with those of politically engaged panelists with a less positive survey experience. A data simulation indicates that such improvements can reduce attrition bias in political variables. Building on these findings, the second study tests how the survey experience of politically disengaged panelists can be improved in practice. Specifically, I use a data simulation based on a survey experiment fielded in the GESIS Panel to compare a short questionnaire, an interesting questionnaire, and an additional monetary incentive in fixed and adaptive survey designs. Targeting politically disengaged panelists with an interesting questionnaire reduces the difference in participation propensities between politically engaged and disengaged panelists about as much as an additional incentive for all panelists. It is also the only strategy that reduces attrition bias in political variables consistently across waves. Panel surveys covering political topics can thus retain politically disengaged panelists by tailoring part of the survey content to their interests.



