People express more bias in their predictions than in their likelihood judgments.

Abstract

refers to when people's expectations about an uncertain event are biased by outcome preferences. Prior work has provided limited evidence that the magnitude of this motivated bias depends on (is moderated by) how expectations are solicited-as discrete outcome predictions or as likelihood judgments expressed on more continuous scales. The present studies extended the generalizability and understanding of the moderating process. The authors proposed that solicitations of predictions and likelihood judgments have different connotations that ultimately affect how much bias is expressed; this varies from a prior account that attributed the moderation effect to response scale differences (dichotomous vs. continuous). Study 1 confirmed the connotation difference, with predictions being viewed as more affording of hunches. Studies 2-4 directly tested the moderation effect, and unlike prior work focusing on expectations for purely stochastic events, the present studies involved more naturalistic events for which likelihood information was not supplied or directly knowable. Before viewing scenes from a basketball game (Study 2) or an endurance race (Studies 3 and 4), participants were led to prefer one contestant over another. After viewing most of the closely fought contest, they made either a prediction or likelihood judgment about the outcome. Participants' tendency to forecast their preferred contestant to win was significantly stronger among those making predictions rather than likelihood judgments. In support of the proposed account, this effect persisted even when both types of solicitations offered only dichotomous response options. Broader implications for measuring and understanding people's expectations or forecasts are discussed. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

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 — 7 citations exist upstream.

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