Formal models of differential framing effects in decision making under risk.

Abstract

An intriguing finding in the decision-making literature is that, when people have to choose between sure and risky options of equal expected value, they typically take more risks when decisions are framed as losses instead of gains (Tversky & Kahneman, 1981).This framing effect is robust and has important implications for health, finance, and politics.However, theoretical debate exists on the origins of this effect.Moreover, pronounced task-related, individual, and developmental differences exist in the magnitude of the effect.These two issues-theoretical debate and differential framing effectscan be solved together, as an adequate theory of the framing effect should both describe the effect itself and describe differences therein.Therefore, we compare four theories on their capacity to describe differential framing effects: cumulative prospect theory (CPT), fuzzy trace theory (FTT), dual process theory, and a hybrid theory (HT) incorporating elements from lexicographic theory and fuzzy trace theory.First, in a theoretical analysis and empirical review, we build on recent advances in the fields of decision making, brainbehavior relationships, and cognitive development.Second, in an empirical study, we directly compare these theories by using a new experimental task and new analytic approach in which we use hierarchical Bayesian model-based mixture analysis of theories.Taken together, results indicate that differential framing effects are best described by the notion that the majority of decision makers decide according to the hybrid theory, and a sizable minority according to cumulative prospect theory and fuzzy trace theory.We discuss implications of these results for our understanding of the framing effect, and for decision making in general.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 16 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.