View source for Ecological rationality

'''Ecological rationality''' is a framework in cognitive science and decision research, developed primarily by [[Gerd Gigerenzer]] and the ABC Research Group at the Max Planck Institute for Human Development. It proposes that the rationality of a decision strategy should be evaluated not by its internal coherence or adherence to normative axioms (as in classical rational choice theory), but by its '''fit to the structure of the environment''' in which it operates. A heuristic is rational, on this view, if it exploits the statistical structure of a particular decision environment to achieve accurate, fast, and frugal outcomes.

The framework is a direct response to the [[Heuristics and Biases|heuristics and biases]] program associated with [[Daniel Kahneman]] and [[Amos Tversky]], which documented systematic deviations between human judgment and the norms of probability theory and expected utility theory. Gigerenzer's program does not deny these deviations, but it reframes them: rather than treating heuristics as cognitive shortcuts that introduce error, ecological rationality treats them as adaptive tools that are well-suited to specific environmental structures. The error, on this view, lies not in the mind but in applying the wrong tool to the wrong environment.

The central concept is the '''[[less-is-more effect]]''': in environments with noncompensatory cue structures — where a single strong predictor dominates all combinations of weaker predictors — heuristics that ignore most information can outperform strategies that integrate all available information. The [[Take-the-best]] heuristic is the paradigmatic example. It searches cues in order of validity and stops at the first discriminating cue, ignoring all remaining information. In certain environments, this produces higher accuracy than [[Multiple Regression|multiple regression]] or [[Machine Learning|machine learning]] models that use the full cue set.

Critics argue that the ecological rationality framework redescribes well-known statistical phenomena — particularly the bias-variance tradeoff — in cognitive terms without adding explanatory power. A model with fewer parameters has lower variance; in environments where variance dominates bias, it wins. This is not a new theory of rationality but a special case of regularization and model selection. The framework's emphasis on environmental