Cognitive biases
The fundamental attribution error: why we blame character, not circumstances
A car on the highway cuts sharply in front of you. “Jerk,” you think. The thought that the driver might be taking a child to the emergency room, or is driving this road for the first time and missed a turn, comes later. If it comes at all. The explanation is ready before any facts appear.
This is the fundamental attribution error: when we explain other people’s actions, we overestimate the role of their character and underestimate the role of their circumstances. Attribution is simply the explanation of causes, the answer to the question “why did they do that?”
Below, we look at the classic experiments, what turned out to be inaccurate in the popular version of this error, and how it gets in the way in a couple and at work. In the middle of the article you can play a quiz in which the error was once shown very clearly.
Contents
In short: how the fundamental attribution error works
What the fundamental attribution error is
In 1967, Edward Jones and Victor Harris ran an experiment that later made it into many textbooks. Students read a short essay about Fidel Castro: one pro, one anti. Some were told that the author had chosen the position himself. Others were told that the position had been assigned and he was simply carrying out the task. Then they had to guess what the author really thought about Castro.
When the author chose for himself, it is logical to think the essay reflects his views. But even when readers knew the position had been assigned, they still considered the author of the pro essay more likely to be a supporter of Castro. They did adjust for the situation, but not enough.
Ten years later, the psychologist Lee Ross named this tendency the fundamental attribution error. In scientific texts it is also called correspondence bias. Daniel Gilbert and Patrick Malone, in a 1995 review, described it like this: we draw conclusions about a person’s lasting traits from behavior that the situation fully explains.
What is inaccurate in the popular version. This error is often retold as a rule: “we explain other people’s actions by character and our own by circumstances.” The second half has not held up well. It is a separate idea, the actor–observer asymmetry. Bertram Malle, in a meta-analysis of 173 studies, found that on average there is almost no such difference. It appears in narrower cases: for example, when the topic is bad actions or close people. The reliable core is more modest: when explaining others, we underestimate the situation.
Culture. The error is not equally strong everywhere. According to a review by Incheol Choi and colleagues, in East Asian cultures people more often take circumstances into account and fall into this trap less often, above all when the pressure of the situation is easy to see. Even so, people in all cultures are inclined to explain behavior by character.
Try it yourself: the quiz
In 1977, Lee Ross, Teresa Amabile, and Julia Steinmetz set up a quiz. In each pair, one person asked the questions and the other answered. Chance decided who would be who. The person who asked made up the questions out of what they knew well. Play a similar game.
Step 1. Tom keeps bees on his farm. He asks you five questions, about bees, of course. He chose the questions himself.
Step 2. You rate how much Tom knows in general and how much you know.
Step 3. You swap places: now you ask Tom a question, out of what you know well yourself.
Why a quiz? Here the person who asks has a built-in advantage, and you know about it from the start: they pick questions from their own field, and you answer someone else’s. If they seem smarter even so, then we are poor at accounting for circumstances. That is exactly what was tested in 1977.
What usually happens. In the original study, the people who asked the questions were consistently seen as more knowledgeable than the people who answered. The ones who thought so most strongly were the answerers themselves and outside observers. Even though chance assigned the roles and the askers made up the questions themselves. The askers themselves fell for this error less.
Why this is an error. Tom’s questions say nothing about how much he knows in general. They say that he knows bees well and that he chose the questions. Once you swap places, Tom is the one who cannot answer your question. If you rated Tom above yourself, you are not less smart than he is. You did the same thing as the participants in the original experiment: you took the advantage of a role for a property of the person.
Why circumstances are hard to see
Imagine you are watching a play. The actor on stage is visible in full: voice, gestures, face. But the script, the director, and the prompter are not visible. The comparison is loose, but in life it is roughly like this: the person’s behavior is before our eyes, while the pressure acting on them stays off camera.
Gilbert and Malone described several causes. We may simply not know about the circumstances: the driver does not tell us where he is rushing to. We may underestimate how strongly a situation affects people: “I wouldn’t have done that in his place.” And even when we do know about the circumstances, the adjustment for them comes out incomplete. That is exactly what the Castro essay experiment shows: readers knew that the position had been assigned, and still did not fully account for it.
There is also simple convenience. The conclusion “he is irresponsible” is short and explains a lot at once. The conclusion “he may have been overloaded, and the email may have gone to spam” needs checking. The first is ready at once; the second still has to be obtained.
Where it gets in the way
Relationships. Your partner is late for dinner: “you don’t care about me anyway.” They forgot to buy bread: “you never listen to me.” One episode turns into a conclusion about the person, and the argument is no longer about bread but about character. Thomas Bradbury and Frank Fincham, in a review of studies of married couples, showed that spouses who are unhappy in the marriage more often explain their partner’s actions in a negative light. The data suggest that such explanations may themselves affect marital satisfaction, though the direction of the link is not fully clear.
Work. A coworker missed a deadline: “unreliable.” The manager answered sharply: “that’s just how they are.” Sometimes that is true. But often behind a missed deadline there are other people’s delays, an unclear assignment, or three parallel projects. If you judge by one episode, it is easy to hang a label on a person and miss the problem in the process.
News and social media. A video of a person being rude to a cashier lasts twenty seconds. The comments under it pass a verdict on their whole life. What happened a minute before the recording, nobody saw. And it is not visible in the frame.
How you see yourself. We sometimes hang similar labels on ourselves. “I have no willpower,” “I’m lazy” are also conclusions about character drawn from separate episodes. And the episodes could have had quite situational causes: lack of sleep, overload, an unclear task. The question “what was going on in the situation?” is useful here too.
How to see circumstances
The fundamental attribution error is one of the biases on the cognitive bias map, in the section on how we judge people. Next to it is the halo effect: one noticeable trait of a person colors everything else.
When to see a psychologist
This error is something everyone has, and by itself it is not a reason for a consultation. A reason appears when in a couple almost every episode turns into proof of “you don’t care about me” or “you’re always like this.” Then the arguments are no longer about events but about characters, and each person defends against a label instead of solving the problem.
If you recognize your couple in this, it is worth working through such conversations with a specialist. In a consultation you can see how you explain each other’s actions, and learn to talk about specific situations rather than about “what kind of person you are.” You can book a consultation online. If threats are made in arguments or there is violence, call 112 in case of immediate danger.
Frequently asked questions
So character does not matter?
It does. People really do differ, and behavior often says something. The error is not that we take character into account, but that we underestimate the situation and draw a conclusion too fast, from a single episode.
How is the fundamental attribution error different from the actor–observer asymmetry?
The fundamental attribution error is about how we explain other people’s actions: too much character, too little situation. The actor–observer asymmetry claims that we explain ourselves differently from others. The first is well supported, although it is expressed differently in different cultures. The second was, on average, hardly supported in Malle’s meta-analysis.
Is this error found in all cultures?
The tendency to explain behavior by character is found everywhere. But in East Asian cultures people more often take circumstances into account, above all when the pressure of the situation is easy to see. So the word “fundamental” should not be taken literally.
When does someone’s behavior really say something about their character?
One episode says almost nothing. If the same thing repeats in different situations and after you have talked about it, that is already information. Before that, name two possible causes in the circumstances and check at least one, for example by asking a question without blame.
