Actor-Observer Bias

Who first named this

Edward E. Jones & Richard E. Nisbett, 1971 — The Actor and the Observer: Divergent Perceptions of the Causes of Behavior

The counsel

The original account (Jones & Nisbett, 1971) attributed the asymmetry to two factors: differential perceptual focus (the actor sees the situation/environment in front of them, while the observer's attention is focused on the acting person, making the actor figural) and differential information (actors know their own history, intentions, and behavioral variability across situations, whereas observers do not). Storms (1973) operationalized the perceptual-focus account by reversing visual point of view with videotape.

Malle, Knobe & Nelson (2007) argue the disposition/situation framing is itself wrong: a folk-conceptual theory holds that any genuine self-other differences come from impression management, asymmetric knowledge, and co-presence rather than a global tendency to favor situational vs. dispositional causes. The Malle (2006) valence finding suggests much of the apparent asymmetry is actually a self-serving bias (taking credit for positive outcomes, deflecting blame for negative ones) confounded with the actor-observer manipulation.

The evidence answers

• In the standard orientation, actors made more situational attributions for their own behavior and observers more dispositional attributions for the actor, consistent with the hypothesis. Showing actors a videotape of themselves shifted (reversed) their attributions toward more dispositional self-explanations, supporting a perceptual-salience mechanism.

  Michael D. Storms, 1973 · not reported (original paper reports ANOVA results, not standardized d); later work (Martin & Huang, 1984) failed to cleanly reproduce the reorientation effect · college students in dyads (small-N laboratory study)

• Across studies the asymmetry was effectively nonexistent overall; it emerged only when the actor was idiosyncratic, events were hypothetical, actor and observer were intimates, or free-response explanations were coded. It held for negative events but reversed for positive events.

  Bertram F. Malle, 2006 · average d = -0.016 to 0.095; ~0 after correction for publication bias and valence · 173 published studies (cumulative literature)

• Found no support for the traditional dispositional/situational actor-observer hypothesis, but systematic support for reformulated asymmetries driven by impression management, knowledge differences, and co-presence.

  Bertram F. Malle, Joshua M. Knobe, Sarah E. Nelson, 2007 · not reported as a single pooled d (multiple studies with varied measures) · six studies, varied contexts and methods

The record

Jones and Nisbett (1971) proposed that actors attribute their own behavior to situational causes while observers attribute the same behavior to the actor's dispositions, and the idea became a textbook staple. But Malle's (2006) meta-analysis of 173 published studies found the overall asymmetry was essentially zero (average d between -0.016 and 0.095, ~0 after correcting for publication bias). The effect appears only under narrow conditions (e.g., negative events, idiosyncratic actors, intimates) and reverses for positive events, so the classic "robust, pervasive" framing is not supported by the cumulative evidence.