The actor-observer asymmetry in attribution: A (surprising) meta-analysis

The effect everyone cites to explain why we judge others by character averages out to roughly nothing across 173 studies.

Malle, B. F. (2006). The actor-observer asymmetry in attribution: A (surprising) meta-analysis. Psychological Bulletin, 132(6), 895–919.

Type

Journal article

Find it

doi.org/10.1037/0033-2909.132.6.895

Checked

Against the abstract, 16 September 2026

Why it is on the list

The actor–observer hypothesis says people explain their own behaviour by the situation and other people’s by the person. Malle notes it has been described as robust, firmly established, and pervasive, then reports a meta-analysis of 173 published studies with average effect sizes from d = −0.016 to d = 0.095. The asymmetry held in specific conditions — a highly idiosyncratic actor, hypothetical events, actor and observer who are intimates, free-response explanations that were coded — and it held for negative events while reversing for positive ones. Across valence, the abstract concludes, no asymmetry exists.

Week 2 does not use a blanket actor–observer bias to justify the asleep test. The test is the course’s own rule of thumb: a rule about an exchange can be tested in another exchange; a verdict about a person is harder to check.

The result for negative events is relevant when a journal entry records a conflict or a failed exchange. It does not mean every entry needs to concern one, and negative events were not the only condition in which the review found an asymmetry.