Visual versus Text Attribute Representation in Choice Experiments

We examine the impact of attribute representation format in choice experiments on willingness to pay (WTP), and serial stated attribute non-attendance. Subjects in an online choice experiment were randomly assigned to a Visual or Text Treatment. Participants in both treatments were given the same choice sets but those in the Visual Treatment were shown pictures of the product alternatives in the choice sets as they would appear in a grocery store, while those in the Text Treatment were presented with text descriptions. Results from random parameters logit models accounting for serial stated attribute non-attendance show that more attributes were significant in the Text Treatment than the Visual Treatment, and among variables significant in both treatments, attribute WTP was higher in the Text Treatment than in the Visual Treatment. Hence, the attribute representation format should be considered when designing choice experiments.