The development of formal models of decision making under risk has largely focused on monetary outcomes. Cumulative prospect theory (CPT), the most prominent model, describes monetary choices well but performs less well for nonmonetary, nonnumerical outcomes (e.g., medications with side effects). We suggest that affective processes, which are omitted in CPT, play a larger role in nonmonetary than in monetary choices, and propose two psychologically motivated modifications: (a) using affect ratings rather than monetary equivalents to represent the subjective value of nonmonetary outcomes (affective valuation); and (b) allowing the probability weighting of an outcome to depend on the amount of affect triggered in a choice problem (affective probability weighting). In a model comparison involving four datasets (N=240), we show that the modifications improve model performance relative to standard CPT for the nonmonetary, but not for the monetary choices. A complementary eye-tracking study on nonmonetary choice (N=68) further supported the predictions of affective probability weighting: increasing risk aversion and decreasing attention to probabilities, as the affective value of the worst outcome in a choice problem increases. Our work integrates previous ideas on how affect modulates decision processes in naturalistic nonmonetary contexts and offers a refined framework for modeling such decisions.
Fuławka et al. (Wed,) studied this question.