Defensive attribution hypothesis
Also known as: Defensive attribution
We blame victims more, or less, depending on how much we resemble them.
What it means
The defensive attribution hypothesis holds that how observers assign responsibility for an accident is shaped by two self-protective motives, not the evidence alone. Harm-avoidance pushes observers to hold the person who caused the accident responsible, so they can believe such outcomes are controllable and avoidable, while blame-avoidance pushes them to exonerate a responsible party who is similar to themselves, since blaming a similar other implies they too could be blamed. Developed by Walster (1966) and Shaver (1970) alongside Lerner's just-world work, it treats personal similarity and outcome severity as moderators. Consequently, the more severe the outcome and the less the observer resembles the person responsible for the accident, the more responsibility that person tends to be assigned. Victim-blaming is a related but distinct application, closer in spirit to the just-world account.
What the evidence shows
Walster's 1966 experiment reported that people assigned a driver more responsibility as the damage from his runaway car grew worse, which she read as evidence that a frightening outcome makes us want to believe someone could have prevented it. The finding proved fragile. Walster failed to reproduce it in follow-ups, and Shaver, also unable to replicate it, proposed that similarity was the missing variable. Burger's 1981 meta-analysis of twenty-two studies settled the picture. The raw severity effect, more blame for worse outcomes, was statistically real but weak. The robust pattern was an interaction: when observers resembled the person who caused the accident, greater severity lowered the blame they assigned; when they did not, greater severity raised it. Severity alone predicts little; severity crossed with similarity predicts the reversal.
Personal similarity versus situational relevance
Shaver drew a distinction the single word similarity hides. Personal similarity is resemblance to the person who caused the harm, same age, job, or temperament, and it drives blame-avoidance: condemning someone like me implies I could be condemned for the same lapse. Situational relevance is the sense that I could plausibly end up in that position, whether or not I resemble anyone in it. The two often move together but can be pulled apart, and they sometimes work against each other. A person unlike the careless driver but who drives the same icy commute feels the situation is relevant while feeling personally dissimilar. Keeping them separate matters, because they license opposite readings of the same case, and studies that collapse both into one measure blur the very mechanism the hypothesis rests on.
Where it shows up
The pattern recurs wherever fault is apportioned after something goes wrong. Salminen interviewed the parties to ninety-nine serious workplace accidents in Finland and found the injured worker reaching for external causes, faulty equipment or impossible conditions, while foremen and coworkers, who needed to believe the job was safe for them, located the cause in the worker's own carelessness. Juries and the public do much the same when the responsible party is unlike them, one strand in the studied tendency to judge sexual-assault complainants who do not fit a juror's self-image more harshly. Insurers, road-safety researchers, and hospital reviewers all confront it: the identical crash or clinical error reads as bad luck or bad judgement depending on who the observer takes the responsible person to be.
Limits and where it breaks down
The hypothesis carries a lot of moving parts, outcome severity, personal similarity, situational relevance, and later a split between causal and moral responsibility, and that flexibility is a weakness as much as a strength. With so many moderators, almost any result can be fitted after the fact, which makes the theory hard to falsify. The severity main effect that launched the field is weak and inconsistent. Rival accounts explain some of the same data without a self-protective motive: observers may simply confuse causal contribution with blameworthiness, or infer that a bad outcome signals a bad decision. The safe reading is narrow. Similarity reliably shifts blame in the self-serving direction, but the fuller machinery of competing motives is better treated as a useful frame than a settled mechanism.
Examples
Observers blame a careless driver more as the crash grows more severe — but blame a driver less when they share their background and circumstances.
Hearing a colleague's diagnosis, people hunt for what she did wrong — did she smoke, did she skip the screening — because a controllable cause means the same thing cannot come for them.
When a project fails under a manager much like themselves, colleagues cite bad luck and impossible timelines; when the manager is nothing like them, the identical failure becomes poor judgement.
When a surgeon's error kills a patient, reviewers who work in the same specialty and face the same pressures tend to read the death as an unforeseeable complication, while colleagues from other fields, given the identical chart, call it avoidable negligence.
After a fund manager's failed bet wipes out a client's pension savings, analysts who run the same kind of strategy write the loss off as an unforeseeable market shock, while those who invest differently judge the very same trade reckless.
First described in Kelly Shaver (1970); Walster (1966).
Key references
- Salminen, S. (1992). Defensive attribution hypothesis and serious occupational accidents. Psychological Reports, 70(3), 1195-1199. doi.org/10.2466/pr0.1992.70.3c.1195
- Burger, J. M. (1981). Motivational biases in the attribution of responsibility for an accident: A meta-analysis of the defensive-attribution hypothesis. Psychological Bulletin, 90(3), 496-512. doi.org/10.1037/0033-2909.90.3.496
- Shaver, K. G. (1970). Defensive attribution: Effects of severity and relevance on the responsibility assigned for an accident. Journal of Personality and Social Psychology, 14(2), 101-113. doi.org/10.1037/h0028777
- Walster, E. (1966). Assignment of responsibility for an accident. Journal of Personality and Social Psychology, 3(1), 73-79. doi.org/10.1037/h0022733