Assessing the impact of opt-out definitions in choice experiments: a case study on food safety

In the current study, we explore the potential impact of using different opt-out definitions in a case study concerning consumer preferences for quality and food safety attributes in chicken fillets. Specifically, we focus on the impact on opt-out effects. Using three different opt-out definitions in split samples we find a significant impact on the deterministic part of the opt-out effect whereas we find no impact on the stochastic part of the opt-out effect. Furthermore, to test for respondents' potential use of a simplifying heuristic in terms of resorting to the opt-out when choice tasks are difficult, we incorporate the utility balance between the experimentally designed alternatives in the likelihood function. We find that increased utility balance increases the opt-out effect in terms of increased likelihood of a opt-out choice when choices are difficult. Interestingly, we find that this heuristic has a larger impact when the opt-out is defined as respondents own typical (perceived) purchase than when defined as a none-of-these or a provided hypothetical alternative. We consequently stress the importance for researchers to be aware of these possible implications related to the definition of opt-out alternatives in choice experiments.