Understanding Participant Representativeness in Deliberative Events: A Case Study Comparing Probability and Non-Probability Recruitment Strategies

Abstract

Deliberative event participants often differ in meaningful ways from the population they are intended to represent; however, less is known about whether various recruitment methods influence participant representativeness. Furthermore, a better understanding of where in the recruitment process lack of representation occurs is needed. We present a framework for understanding why event attendees might not represent the target population and then compare two different recruitment strategies using this framework. Specifically, we consider a Deliberative Poll that used a random-digit-dial telephone recruitment survey and a deliberative event that used a convenience sample web recruitment survey. For two stages in the recruitment process, we calculate nonresponse errors for statistics assessing demographic characteristics and confidence in local government. Notably, both recruitment methods resulted in event attendees that were older and better educated than the population they were intended to represent providing evidence that probability recruitment methods do not necessarily outperform nonprobability methods. Additionally, we demonstrate that aspects of the recruitment process other than the recruitment survey sampling method used can influence participant representativeneess. We conclude by discussing adjustments to the recruitment process that might improve the representativeness of event attendees.