Words or graphics to present a Discrete Choice Experiment: Does it matter?

Area of Science:

• Health economics
• Behavioral economics
• Decision science

Background:

• Choice-based conjoint (CBC) analysis is widely used to understand consumer preferences.
• The presentation format of attribute levels (words vs. graphics) may influence decision-making.
• Understanding optimal presentation is crucial for accurate preference elicitation.

Purpose of the Study:

• To compare the impact of presenting attribute levels in words versus graphics on interpretation, importance, and participation probabilities.
• To determine which format yields more reliable data in a choice experiment setting.

Main Methods:

• A choice-based conjoint (CBC) experiment was conducted with 959 parents of newborns.
• Participants completed a questionnaire with two versions of nine choice tasks, differing in attribute level presentation (words or graphics).
• Attributes related to rotavirus vaccination decisions were analyzed using mixed-logit models.

Main Results:

• Respondents showed a significant preference for attribute levels presented in words over graphics.
• Word-based presentations resulted in more consistent answer patterns and more valid attribute level estimates.
• Differences in data presentation led to variations in part-worth utilities, relative attribute importance, and predicted participation probabilities.

Conclusions:

• Words are a superior format for presenting attribute levels in CBC tasks compared to graphics, enhancing data quality.
• Graphics do not improve respondents' understanding of attribute levels and may introduce bias.
• Further research is needed to establish guidelines for optimal attribute level presentation in choice experiments.