MBS 93-02
Tests of Consequence Monotonicity in Decision Making Under Uncertainty
Detlof von Winterfeldt, Ngar-Kok Chung and Duncan Luce

---

Elimination of dominated alternatives is a fundamental principle of rationality. In decision making under risk, dominance has been shown to be violated in experimental tests of the Allais paradox and the common ratio effect. However, these tests mixed assumptions about structural equivalences (e.g., reduction of compound lotteries) with pure dominance tests. A more basic form of dominance is consequence monotonicity: If a gamble is altered by replacing one consequence by a more preferred consequence, the modified gamble should be preferred. Previous tests of this assumption showed that when subjects compared the dominating and the dominated gambles directly, they recognized the dominance relationship and chose the dominating gamble. However, when eliciting judged certainty equivalents for each gamble separately, systematic violations of consequence monotonicity were found for some pairs of gambles. We hypothesized that this violation of consequence monotonicity is due to anchoring and insufficient adjustment processes in the judged certainty equivalent response mode. If true, the violations should disappear when proper choice procedures are used to elicit certainty equivalents. to test this hypothesis, we conducted three experiments. In all three experiments we found moderate violations of consequence monotonicity when using judged certainty equivalent procedures. A procedure that elicited certainty equivalents through repeated choices produced fewer violations, and, to the extent that they occur, they can be explained by the noise in the data. A third procedure that mimicked a decision analyst's elicitation process produced about the same number of violations as they judged certainty equivalent procedure. We conclude that consequence monotonicity is descriptively valid, when using proper choice stimuli and response modes.