Probability-based online panels offer numerous advantages to the social science research community. For example, data on people’s opinions, attitudes, and behaviour can be gathered at short notice, at frequent intervals and across time, permitting researchers to detect and explore societal changes as soon as they occur. However, to allow researchers to be confident about their findings, the online panel must reflect society as a whole. To achieve this, the recruitment of the online panel is of key importance.
This project explores strategies for increasing cost efficiency, response, and representativeness in panel recruitment in three steps. First, we apply predictive analytics to predict which people are best addressed with which recruitment method. Second, we conduct a field study to experimentally test the estimated models and novel recruitment strategies in practice. Third, the project assesses data quality in the newly recruited panel sample and develops best practice recommendations.
