The Certainty Effect for Gains and Losses Emerges in Joint Evaluation but Not Always in Separate Evaluation

Abstract Six studies investigated people's attitudes toward uncertainty. Participants rated the attractiveness of pairs of gambles in either a gain or a loss frame. We varied the level of uncertainty, the monetary outcomes, and the evaluation mode of the gambles (i.e., joint versus separate evaluation). Experiments 1a and 1b compared a sure gain (loss) to a risky gain (loss), with both gambles having identical expected value. Experiments 2a and 2b included an almost sure (i.e., 98%) gain (loss) and risky gain (loss). When gambles entailed gains, a risky gamble became less attractive when evaluated in joint than in separate evaluation. The opposite pattern emerged when gambles entailed losses. The difference between a risky and a sure (or almost sure) gamble was weaker (or eliminated) in separate evaluation. Experiments 3a and 3b presented a risky gamble alone or with other gambles with varying probability and outcomes to be gained or lost. When gambles entailed gains (losses), a risky gamble became less (more) attractive and was chosen less (more) frequently when paired with gambles offering a higher probability of gaining (losing) smaller amounts. Overall, affective reactions and preferences for uncertain gambles depend on the decision context, and the certainty effect can disappear in separate evaluation.