Character is the first explanation offered and the least examined

The explanation that arrives first

Watch somebody act badly and an explanation arrives before you have asked for one. They are careless, or rude, or dishonest. That explanation concerns the person, it’s available instantly, and it requires no information beyond what you have just seen. Explanations involving the circumstances — what they had been told, what kind of day it was, which options were actually open — arrive later if they arrive at all, because they depend on details a bystander almost never has.

The laboratory version of this is one of the older findings in social psychology and one of the more durable. People read an essay defending some position, are told plainly that the writer was assigned that position and had no say in it, and still conclude that the writer holds roughly the view expressed. The information that should defeat the inference is sitting in the instructions, and it does not defeat it. The design has been run in many variations, and the basic pattern reproduces.

Why the view from inside looks different

Your own behaviour comes with its situation permanently attached. You know what the message said before you replied, you know what you had already been dealing with, and you know which alternatives you had ruled out. Nobody watching has any of that, so they work from what is visible, which is the behaviour and the person performing it.

The tidy version of this observation — that we explain ourselves by circumstance and everybody else by character — has fared less well than the underlying effect. It is called the actor-observer asymmetry, and a substantial review of the evidence found it far weaker than the textbook account implies, close to absent under many conditions, and appearing mainly when the behaviour was negative or the actor well known. So the tendency to over-read character in others is solid; the neat mirror-image story usually told about it is not.

The cross-cultural qualification

The effect is often described as a basic feature of human reasoning, and comparative work complicates that. Studies comparing samples from different cultural settings have repeatedly found differences in how much weight situational information receives, with participants in some East Asian samples attending more to context and offering fewer purely dispositional explanations than participants in North American ones.

Those findings deserve to be held loosely for the usual reasons. The samples are mostly students, the cultural categories are extremely coarse, the differences are matters of degree rather than presence and absence, and whether the underlying process differs or only the attention paid to context is unresolved. What the comparative work does establish is narrower and still worth having: the strength of the tendency isn’t fixed, which means it’s responsive to how people have been trained to look.

Where the cost actually lands

It lands hardest wherever somebody is judged from a single observation. A candidate seen once, a stranger in a brief transaction, a colleague encountered on the worst afternoon of their year. In each case the observed behaviour is a sample of one, drawn from a situation you can’t see, and it gets converted into a stable claim about a person.

It lands again in how organisations diagnose failure. When the same role produces the same problem under three different people, the explanation is almost certainly structural, and it is nonetheless common to replace the person a third time. The disposition account is available immediately and needs no investigation, which is exactly why it wins arguments in rooms where nobody has time.

The correction is a question, not a resolution

Deciding to be fairer does not work, in the same way that deciding to be unbiased about anything else does not work. What does something is a specific question asked out loud: what situation would produce this behaviour from an ordinary, reasonable person? The question forces a search in a region the mind was not searching, and it usually returns candidates.

A second question is more diagnostic still. Do other people in this position behave similarly? Consistency across people in the same slot points at the slot; behaviour that stands out from everybody in the same conditions points at the person. That comparison is the closest thing available to a control group in everyday judgement, and it costs a few seconds.

What this does not license

It does not follow that character is fiction. People differ in reasonably stable ways, behaviour has genuine consistency across situations, and a person who has been unreliable in six unrelated contexts is telling you something. The claim is about weighting, not about existence: the situational explanation is systematically underweighted, not systematically correct.

The over-correction is worth naming because it is fashionable in some places. If every failure is structural then nobody is responsible for anything, which is both untrue and convenient. The useful position sits in the middle and is harder to hold: assume the situation is doing more than it appears, keep asking what it is doing, and revise towards the person only when the behaviour survives a change of circumstances.