Nonlinear Decision Weights or Moment-Based Preferences? A Model Competition Involving Described and Experienced Skewness

The predictive power of Cumulative Prospect Theory (CPT) and Expected Utility Theory (EUT) is typically compared using decisions from description (DfD), wherein lotteries' outcome values and probabilities are explicitly stated. In decisions from experience (DfE), individuals sample (without cost) from the return distributions to learn outcome values and their relative frequencies; therein CPT and EUT require the calculation of probabilities from experience. Individuals, however, may be more attuned to the experienced moments of outcome distributions. We therefore test the Mean-Variance-Skewness (MVS) model, and estimate the proportion of EUT (over income), CPT and MVS populations using a latent-class hierarchical Bayesian model across six large datasets. For simple lotteries (with 1-2 outcomes), we find a mixture of CPT and MVS populations in both DfD and DfE. For more complex lotteries (with 2-3 outcomes), all participants are classified as CPT types in DfD, but as MVS types in DfE. This suggests that in decisions from experience with more complex return distributions, preferences for skewness are more predictive than nonlinear probability weighting