Information effects in discrete choice experiments: Does type of delivery matter for valuation estimates and attribute nonattendance?

We examine the impact of using combined text script and audio clip versus text script only to convey information on US low-skilled nonmigrant workers' willingness to accept agricultural field jobs amid COVID-19 pandemic. Using a discrete choice experiment, subjects were randomly assigned to one of the two information modalities. Respondents assigned to combined text script and audio clip were willing to accept more for several job attributes, especially when estimates were conditional on attribute attendance. Moreover, this information modality lowered the prevalence of attribute nonattendance. The findings have important implications for the design of information provision experiments.