The bench · equal acts, unequal verdicts
Omission Bias
People judge harmful actions as morally worse than equally harmful inactions, leading to a preference for passivity even when doing nothing produces equal or worse outcomes.
Both pans hold the identical decision. Only the result differs — yet the beam tips toward the outcome.
Why the verdict bends
Omission bias is generally explained as an overgeneralization of a normally useful heuristic: commissions typically cause more direct and salient harm than omissions. People associate action with causation and responsibility, while omissions are perceived as allowing natural outcomes to unfold. Additionally, anticipated regret may be stronger for actions that lead to bad outcomes than for inactions with the same outcome, because actions involve clearer counterfactuals (the actor could have done nothing).
Competing account — Connolly and Reb (2003) argue that the 'omission bias' in vaccination decisions reflects a normality bias (preference for the status quo/normal course) rather than a true preference for omissions over commissions. Patt and Zeckhauser (2000) have also documented action bias in certain environmental decisions.
Cases brought to the bench
Subjects were reluctant to vaccinate when the vaccine itself could cause death, even when this risk was much smaller than death from the disease. The effect was amplified when a 'risk group' for vaccine death existed, even without a test for membership.
Ritov & Baron · 1990 · n = 146 undergraduates across multiple experiments
Subjects rated harmful omissions as less immoral and less bad as decisions than harmful commissions. A few subjects were willing to accept greater harm to avoid action. The effect was associated with judgments that omissions do not cause outcomes.
Spranca, Minsk, & Baron · 1991 · Multiple experiments with undergraduate participants
Parents who did not or would not vaccinate their children were significantly more likely to exhibit omission bias (p = 0.004), holding constant their beliefs about vaccine danger. These parents also believed vaccinating was more dangerous than not vaccinating (p < 0.001).
Asch et al. · 1994 · n = 103 respondents (52% response rate) from 200 mailed surveys
Evidence — mixed. The core effect has been replicated in multiple laboratory studies across different domains (vaccination, medical decisions, moral judgments). However, Connolly and Reb (2003) challenged the generality of the phenomenon, particularly in vaccination contexts, arguing that existing research had not convincingly demonstrated a general reluctance to vaccinate and that methodological confounds (limited response scales, normality confounds) could explain many findings. Baron and Ritov (2004) responded with new data showing small but reliable omission bias even with improved methods, while acknowledging substantial individual differences.
Judged in the wild
medicine · 1994
Pertussis vaccination refusal among parents
Asch et al. (1994) surveyed parents about pertussis vaccination decisions and found that those who refused vaccination exhibited significantly higher omission bias scores. This paralleled historical pertussis vaccination declines in Britain during the 1970s that led to disease outbreaks.
law · 2001
FDA drug approval bias toward inaction
Baron, Bazerman, and Shonk (2001) documented that U.S. Food and Drug Administration policies exhibit institutionalized omission bias, where the harm from approving a dangerous drug is weighted more heavily than the harm from delaying approval of a beneficial drug.
To weigh the act, not the result
Explicitly compare expected outcomes of action versus inaction using the same metric, and reframe inaction as a choice with consequences rather than a neutral default.
Asch et al. (1994) found that when subjects were forced to acknowledge the consequences of inaction, some omission bias was reduced. Framing inaction as a deliberate choice rather than a passive state can attenuate the bias.
Catch it in the act
Watch for situations where people favor 'letting things happen' over 'making things happen,' even when the expected outcome is worse. Common signs: 'At least I didn't cause it,' vaccine hesitancy focused on side-effect risk while ignoring disease risk, and policy paralysis where maintaining the current course is seen as morally safer than changing course.
Related judgment errors
- Status Quo BiassiblingBoth involve preferring inaction, but status quo bias specifically concerns maintaining the current state, while omission bias applies even when there is no prior status quo.
- Action BiasoppositeIn some contexts (e.g., sports referees, environmental decisions), people prefer action over inaction, the opposite of omission bias.
- Default Effectmechanistically-linkedThe default effect exploits omission bias by making the preferred option the default/inaction path.