Panel conditioning: Types, causes and empirical evidence of what we know so far
Jul 26, 09:00
Panel conditioning is an umbrella term summarizing learning effects that can occur due to the participation in surveys, resulting in different answers given by respondents who have already taken part in a survey from the answers that these respondents would have given if they were participating for the first time
Panel conditioning can be advantageous or disadvantageous for data quality. Advantageous panel conditioning refers to an increased quality of reporting, for example, due to improved understanding of the questionnaire and the surveying procedure. Disadvantageous panel conditioning refers to decreases in data quality such as learning to answer filter questions in a way that would allow avoiding follow-up questions. Panel conditioning can result in respondents 1) changing their reports, 2) changing actual behaviour, attitudes, or knowledge, or 3) both.
The objective of this paper is to provide an overview of findings about the mechanisms, presence, and magnitude of the effects of panel conditioning and to offer practical guidelines regarding survey design. We review the literature, with purposeful selection of studies, including those whose design allows drawing inferences about the phenomenon and development of the recommendations for the survey practice.
We first provide a review of the mechanisms from a perspective of the framework of survey response. We illustrate the mechanisms of panel conditioning providing empirical evidence regarding the conditions under which questions are susceptible to panel conditioning and survey features have been found to influence the undesirable effects of panel conditioning. We conclude by providing some tips on how to minimize the negative effects of panel conditioning on data quality and by giving advice on optimal designs on how to study panel conditioning. We also point out areas for future research that have received little attention to date.