← The Casebook

The bench · equal acts, unequal verdicts

Outcome Bias

Outcome bias is the tendency to evaluate the quality of a decision based on its eventual result rather than on the information and reasoning process available at the time the decision was made.

Both pans hold the identical decision. Only the result differs — yet the beam tips toward the outcome.

Why the verdict bends

Norm theory (Kahneman & Miller, 1986) provides a leading account: people implicitly assume that outcomes reflect the foreseeability of events, and that others who made the 'correct' choice would have achieved the favorable outcome. The 2023 replication found that perceived social norms explained approximately 15% of the variance in outcome bias. A second mechanism involves affective reactions to outcomes that color subsequent evaluations independently of rational analysis.

Competing account — Some researchers distinguish outcome bias from hindsight bias (the tendency to see outcomes as having been more predictable after the fact), though they frequently co-occur. Outcome bias specifically concerns evaluation of decision quality, whereas hindsight bias concerns probability estimates.

Cases brought to the bench

Case 01Medical decision vignettesd = 0.77 (from 2023 replication; original reported F-values)

Decisions leading to patient survival were rated significantly more positively (M=0.85) than identical decisions leading to death (M=-0.05) on a 7-point scale. Main effect of outcome was significant at p<.001.

Baron & Hershey · 1988 · n = 20 undergraduates (original); n = 692 (2023 between-subjects replication)

Case 02Anesthesiologist outcome biasnot reported

Not only the harshness of judgments but also the willingness to make judgments at all increased when there was a severe outcome. Physicians were more likely to criticize identical clinical decisions when the patient outcome was poor.

Caplan, Posner & Cheney · 1991 · Practicing anesthesiologists

Case 03Replication with extensionsCohen's d = 0.77 (patient decisions), d = 1.10 (physician decisions)

Successful outcomes rated higher in decision quality (M=1.78) than failures (M=0.68); F(1,688)=152.37, p<.001, Cohen's f=.47. Outcome also affected perceived responsibility (F(1,688)=32.36, p<.001).

Aiyer, Kam, Ng, Shi, Young & Feldman · 2023 · n = 692

the beam tips reliably

Evidence — robust. The effect has replicated consistently across multiple studies and domains. The 2023 pre-registered replication (n=692) confirmed a moderate-to-large effect size in a between-subjects design, addressing concerns about the original within-subjects methodology. A 1991 study with anesthesiologists confirmed the effect with medical experts, not just lay participants. Effect sizes remain stable across medical, financial, and legal decision contexts.

Judged in the wild

  • law · 2009

    Medical malpractice litigation and expert witness bias

    Medical experts reviewing clinical cases for negligence claims are systematically influenced by knowledge of patient outcomes. A 2009 study found that experts who knew an adverse clinical outcome were less likely to support a physician's management decisions than those blinded to the outcome. In one analysis, physicians won 79% of cases rated 'defensible' but still lost 21% of cases with poor outcomes despite reviewers rating the care as appropriate.

  • medicine · 2009

    Radiological missed diagnosis cases

    Radiologists reviewing x-rays for missed diagnoses show severe outcome bias when they know the patient suffered a serious outcome (e.g., quadriplegia from a missed cervical fracture). The known outcome makes the abnormality appear 'obvious' in retrospect, leading experts to conclude negligence when the original reading was reasonable under normal clinical conditions.

To weigh the act, not the result

Use structured process-focused evaluation frameworks (e.g., pre-mortem analysis, consider-the-alternative strategies) that force evaluators to assess the quality of the decision-making process independent of the outcome. Explicitly instructing evaluators to consider what information was available at the time and whether alternative outcomes were plausible can reduce the bias.

Consider-an-alternative debiasing strategies have been shown to reduce related biases (hindsight bias, overconfidence) by forcing simulation of multiple possible outcomes. The 2023 replication found that perceived norms partially mediated outcome bias, suggesting making evaluation criteria explicit and transparent helps.

Catch it in the act

Watch for situations where someone rates a decision as 'good' because it succeeded or 'bad' because it failed, even when the decision-maker used the same process and information as they would have in the opposite scenario. Common in performance reviews, investment post-mortems, and malpractice assessments.

Related judgment errors

File your own case

Open the same case on your own draft.

Paste a memo, a research draft, or a strategy argument. It is scored against all 175 cards, and the strongest two or three risks come back with the evidence quoted and one practical next check.

Open a case on your draft →