What is cognitive bias?

Definition

A cognitive bias is a consistent, systematic tendency to think in ways that depart from logic or probability. The word "systematic" is important: biases are not occasional slips but reliable patterns that show up again and again under the same conditions. They are usually understood as the by-products of heuristics — efficient mental shortcuts that work well in many settings but produce characteristic errors in others. The umbrella covers dozens of specific biases, including confirmation bias, anchoring, the availability heuristic, and hindsight bias.

The heuristics-and-biases programme

The modern study of cognitive bias began with the work of Amos Tversky and Daniel Kahneman, whose influential 1974 paper in Science set out three core heuristics — representativeness, availability, and anchoring-and-adjustment — and the biases they generate. Their programme showed that judgement under uncertainty follows describable rules that systematically diverge from the predictions of classical rationality.

This line of research reshaped psychology and economics, contributing to the field of behavioural economics and earning Kahneman the 2002 Nobel Memorial Prize in Economic Sciences. It reframed errors not as signs of stupidity but as windows onto how the mind actually works.

Examples and relevance to research

Well-known cognitive biases include confirmation bias (favouring confirming evidence), anchoring (over-weighting an initial value), the availability heuristic (judging frequency by ease of recall), and hindsight bias (seeing past events as predictable). In research, such biases threaten objectivity in how questions are framed, data are interpreted, and results are reported. They are a primary reason the scientific method relies on structured safeguards — controls, blinding, pre-registration, replication, and peer review — to keep individual judgement from distorting conclusions.

Significance for methods

Understanding cognitive bias is foundational to research methodology and to the design of decision aids and survey instruments. Because biases operate below awareness, they cannot be willed away; instead, methods are built to constrain the points at which they could enter. Awareness of bias also informs how findings should be communicated, encouraging precise reporting and discouraging the over-simplification that itself exploits human cognitive tendencies.