Cognitive Bias

Type: Concept

Category: Psychology

Difficulty: Beginner

Reading time: 6 min

Definition

A cognitive bias is a systematic, predictable pattern of deviation from rational judgment or accurate perception. Unlike random errors, cognitive biases consistently push judgments in a specific direction — driven by identifiable mental mechanisms. The term was introduced by Kahneman and Tversky in the 1970s as part of their heuristics-and-biases research program, which documented over 100 named biases across decision-making, probability judgment, social perception, and memory.

The key word is systematic: biases are not noise. They are the predictable output of cognitive shortcuts (heuristics), motivational distortions, and information processing constraints operating in conditions where they diverge from the statistically or logically correct answer.

Why it matters

Biases are patterned, not random

Because cognitive biases are systematic, they can be predicted, documented, and in some cases corrected. The research program established that:

• Specific biases have specific causes (availability bias is caused by the availability heuristic; anchoring bias is caused by insufficient adjustment from an initial value)
• Biases apply across domains but vary in magnitude depending on the domain’s structure
• Expertise and knowledge reduce but do not eliminate most biases — even experts show availability bias, anchoring, and overconfidence in their domains

This makes cognitive bias research practically useful: understanding which mechanism drives which bias points toward specific corrections.

Major bias categories

Heuristic biases (from substitution of easy proxy for hard target):

• Availability bias: judging frequency by ease of recall, producing overestimation of vivid/recent events
• Representativeness bias: judging probability by prototype similarity, producing base rate neglect
• Affect bias: judging risk/benefit by emotional response, producing affect-driven risk perception

Framing and reference-point biases (from prospect theory):

• Loss aversion: treating losses as roughly twice as painful as equivalent gains
• Framing effects: logically equivalent descriptions producing different choices via gain/loss frame
• Anchoring: insufficient adjustment from an initial value

Judgment confidence biases:

• Overconfidence: believing one’s judgments are more accurate than they are
• Hindsight bias: believing after the fact that the outcome was predictable
• Planning fallacy: underestimating time, cost, and risk of future actions

Social and narrative biases:

• WYSIATI (What You See Is All There Is): treating available information as complete
• Halo effect: one attribute coloring all others
• Narrative fallacy: constructing causal stories from what are random sequences

The dual-process explanation

Most cognitive biases are explained by the dual-process framework: System 1 produces a fast, associative, pattern-matching response that System 2 fails to override — either because it is not triggered, or because System 2 is lazy and endorses the System 1 output without scrutiny.

Key takeaways

Key takeaways

• Cognitive biases are systematic and predictable deviations from accurate judgment — they push in identifiable directions driven by identifiable mechanisms.
• They are the output of heuristics (mental shortcuts) operating in conditions where the shortcut diverges from accurate reasoning — not random noise.
• Major categories: heuristic biases (availability, representativeness, affect), framing/reference-point biases (loss aversion, anchoring), confidence biases (overconfidence, hindsight), and narrative biases (WYSIATI, halo effect).
• Expertise reduces but does not eliminate biases — experts show domain-specific biases and generalized biases alike.
• The dual-process explanation: System 1 generates a fast, plausible-but-biased answer; System 2 fails to override it because it is lazy or not triggered.
• Correction requires identifying which mechanism drives the bias — statistical training helps with base rate neglect; structured processes help with planning fallacy; awareness alone is insufficient for most biases.

Mental model

click to zoom

The cognitive bias landscape — major categories organized by their generating mechanism.

Read it as: All four major categories of cognitive bias share a common mechanism: System 1 produces a fast, associative response that feels correct, and System 2 endorses it without adequate scrutiny. Each category differs in which System 1 mechanism drives the error — substitution of proxy questions, reference-point distortion, confidence miscalibration, or narrative coherence. Knowing the mechanism points toward the correction.

Related lessons

Related

Chapter 7: A Machine for Jumping to Conclusions WYSIATI and the halo effect — how System 1's narrative construction produces biased conclusions from incomplete information. Chapter 9: Answering an Easier Question Heuristic substitution — the generating mechanism behind availability, representativeness, and affect biases. Heuristics The mental shortcuts that produce cognitive biases — what they are, how they work, and when they fail.