Tests of consequence monotonicity in decision making under uncertainty

Area of Science:

• Decision Theory
• Behavioral Economics
• Cognitive Psychology

Background:

• Consequence monotonicity is a fundamental principle in decision theory, stating that a gamble with a better consequence should be preferred if all other consequences are equal.
• Previous research indicated potential violations of this principle when gambles are evaluated using certainty equivalents, a subjective valuation of a gamble's worth.

Purpose of the Study:

• To investigate the apparent failures of consequence monotonicity when using certainty equivalents.
• To compare different judgment and choice procedures for assessing gamble preferences.
• To determine if observed violations are systematic or due to methodological factors.

Main Methods:

• Comparison of a judgment procedure with two variants of a choice procedure across multiple experiments.
• Systematic analysis of gamble pairs exhibiting nonmonotonicities.
• Testing a noise model to account for discrepancies in certainty equivalent estimations.

Main Results:

• Fewer nonmonotonicities were observed in one of the choice procedures compared to the judgment procedure.
• Overall violations of consequence monotonicity were reduced compared to previous studies.
• Experiments indicated that procedural or stimulus presentation differences did not account for the observed effects.
• A noise model suggested that estimation noise in certainty equivalents was the primary driver of observed violations.

Conclusions:

• The observed violations of consequence monotonicity in certainty equivalent judgments are likely attributable to noise in the estimation process.
• Consequence monotonicity remains a viable principle in decision-making, even when accounting for estimation noise.
• Choice procedures may offer a more robust method for assessing gamble preferences compared to judgment procedures.