Consistency bias

Biological Parallel

Animals smooth environmental fluctuations in memory because consistent models enable faster prediction than tracking constant change. Clark's nutcrackers cache thousands of seeds across hundreds of locations and relocate them months later with 89% accuracy when aiming for cache centers, relying on spatial memory of large structural landmarks rather than tracking micro-scale changes. Gray wolves remember territorial boundaries as stable zones maintained through scent marking, even though actual boundaries shift with resource availability and neighbor interactions. Honeybees following waggle dances combine the dance vector with cognitive memory of landmark locations—when familiar landmarks appear, routes become direct and efficient; when expected landmarks are absent, searches turn exploratory and imprecise. The computational advantage is clear: stable spatial models compress information and accelerate movement. Consistency bias in humans exploits the same principle: retrospectively imposing consistency on past attitudes and beliefs reduces memory load and simplifies forecasting what happens next.