Objectives. Discrete choice experiments in which participants choose between alternatives differing on attribute levels are an important research method for preference elicitation. In such experiments choice stimuli is typically presented in tables with verbally described attributes, in tables with visual attributes, or as product mock-ups simulating realistic products as close as possible. So far little is known about how presentation formats affect visual attention patterns and choice behavior. This study addresses the question by analysing visual attention and part-worth utilities in choice experiments across three different presentation formats. Method. Participants’ visual attention was measured by means of eye tracking during a discrete choice experiment for yoghurt products varying on six attributes with two to four levels. The study used a mixed within-between subjects design in which the presentation format varied between a verbal information table, a table with visual attributes levels and a realistic product mock-up presentation. Results. A strong relationship between attention and choice was observed so that attributes with a higher importance for participant choices also received a higher share of attention. We also observed a quadratic relationship between fixation duration and part-worth utility. This top-down effect of preferences on attention was stronger for the verbal information table, where attributes were highly similar with regards to size, visual saliency and perceptual fluency. Strong heterogeneity effects were observed in part-worth utilities and this was reflected in individual attention biases towards the more important attributes. Finally, presentation format was found to exert a significant bottom-up effect on visual attention and subsequent choice. Choices in the product mock-up presentation required significantly fewer fixations and less decision time, and more within-alternative transitions were observed compared to the verbal and visual presentation formats. Attributes presented visually or at larger size had a higher impact on participants’ choices. These bottom-up effects caused an overestimation of the impact of certain attributes in the table formats compared to the product mock-up presentation. Conclusion. Visual attention and choice behaviour in discrete choice experiments are strongly affected by stimuli presentation format. These results suggest that the choice of presentation format has strong implications for the external validity of discrete choice estimates and should be considered carefully by researchers.
decision making; choice; visual attention; eye tracking; bottom up processing; top down processing
Main Research Area:
Workshop on Testing Theories of Choice Behavior, 2012