Context matters.

Area of Science:

• Behavioral economics
• Experimental economics
• Decision theory

Background:

• Accurately measuring risk aversion is critical in experimental economics.
• Various methods exist for eliciting risk aversion, but their comparability is debated.
• Understanding method-specific biases is essential for reliable experimental design.

Purpose of the Study:

• To compare the efficacy of different risk aversion elicitation methods.
• To assess the influence of the chosen method on estimated risk aversion levels.
• To determine whether method or preference functional has a greater impact on results.

Main Methods:

• Conducted an experiment with 96 subjects across four distinct elicitation methods: Holt-Laury price lists, pairwise choices, Becker-DeGroot-Marschak, and allocation questions.
• Analyzed data by fitting four preference functionals: Expected Utility and Rank-Dependent Expected Utility, each with CRRA and CARA utility functions.
• Employed a within-subjects design to directly compare method impacts.

Main Results:

• Inferred risk aversion levels were significantly more sensitive to the elicitation method than to the choice of preference functional (e.g., CRRA vs. CARA).
• Differences in estimated risk aversion were substantial across the four methods investigated.
• The experimental context demonstrated a pronounced effect on subject responses.

Conclusions:

• The choice of risk aversion elicitation method is a primary determinant of experimental outcomes.
• Researchers should prioritize careful consideration of the elicitation context over solely focusing on theoretical preference models.
• Findings suggest a need for standardized or context-aware approaches to risk aversion measurement in experiments.