Suppressing myside bias in civil litigation.

Abstract

Objective: Myside bias-the tendency to evaluate and generate evidence as well as test hypotheses in a manner biased toward prior beliefs-causes disputants in litigation to harbor overconfident expectations of judicial awards and reduces odds of settlement. Two studies tested three interventions to suppress myside bias in civil litigation settings.

Hypotheses: I predicted that the participants in the baseline conditions would exhibit myside bias in award estimates and argument ratings and that the interventions would attenuate it.

Method: Two between-subjects experimental studies using students of law (n = 164, Mage = 24.21 years, 53% female; n = 181, Mage = 20.89 years, 61% female) compared the participants' award estimates and argument ratings in a simulated civil dispute. The interventions (a) manipulated the advocates to think they represented the opposing side during initial information processing (side-switch condition), (b) required the participants to generate and evaluate arguments for both sides (dialectical condition), and (c) affected the participants' motivations by threatening dismissal in case of estimation error (goal states condition).

Results: Baseline groups in both studies displayed significant myside bias in award estimates (all ds ≥ 1.12) and argument ratings (all ds ≥ 1.29). In Study 1, the side-switch intervention eliminated bias in argument ratings (d = 0.73 and 0.72) but only reduced (d = 0.35) rather than eliminated bias in award estimates. In Study 2, the dialectical intervention reduced bias in argument ratings (d = 0.74 and 0.58) but did not eliminate it; it also failed to reduce bias in award estimates. The goal states intervention suppressed myside bias in both argument ratings (d = 0.76 and 0.82) and award estimates (d = 0.78).

Conclusions: Myside bias in litigation settings is robust and difficult to suppress. Accountability interventions show potential as bias-attenuating strategies. (PsycInfo Database Record (c) 2025 APA, all rights reserved).