Skip to content
PaperFren

Concept

Cognitive Bias

6 studiesEvidence last moved Sep 20, 2026

A cognitive bias is a systematic departure of judgement from a defined normative standard. The term covers a collection of distinct measured effects with different causes and different moderators — not a single trait some people have more of.

Treating bias as one quantity licenses two mistakes that the studies here contradict: that a group showing a bias differs in kind from a group that does not, and that a bias measured on one judgement predicts bias on another. Both matter for how clinical and educational claims about 'debiasing' get read.

Studies

6

Findings

4

5 supporting · 0 challenging · 0 qualifying citations

Open tensions

1

Latest change

Concept page published

Cognitive Bias

Currently

What we know

  1. The clinical group was further along a shared distribution, not on a different one.
  2. The bias was in what the rating scale actually captured, not in the raters' effort.
  3. Simplification moved felt competence, and the study did not show judgement improved with it.
  4. Where an effect appeared, it depended on intensity — and both trials used passive controls.

Largest unresolved question

The exercise trials here cannot be pooled into a single 'exercise reduces bias' claim. They differ in intensity, duration, outcome measure and control condition — a single session of walking or cycling against a within-subject baseline versus a multi-week table-tennis programme against a passive control that matched neither activity nor social contact.

Common misconceptions

  • Cognitive bias is one underlying tendency that varies between people.

    The effects measured here have different moderators: illusion of control tracked gambling status, complexity ratings tracked arousal, credibility judgements tracked processing ease, and contingency judgements varied with individualist versus collectivist values. Nothing in these studies shows they index a common factor.

  • People with a gambling problem think about chance in a fundamentally different way.

    Both gamblers and controls showed the illusion of control on the same task. The gambler group's ratings were higher (70.61 versus 57.20), a difference of magnitude within a shared effect.

  • Making research easier to read helps people evaluate it better.

    Simplified summaries raised both perceived credibility and self-rated ability to evaluate the work. The study measured confidence, not accuracy, so it shows the easiness effect rather than improved judgement.

Related