How Heuristics Handle Uncertainty

Abstract

Traditionally, it is assumed that a trade-off exists between effort and accuracy: The more effort we put in, the more accurate our inferences. This chapter shows how this trade-off does not hold in general, or even typically, by explaining how simple heuristics that ignore information can outperform more sophisticated inference strategies. Explanations of such “less-is-more” effects are given using statistical learning theory applied to situations where organisms face what is referred to as the bias–variance dilemma. Under conditions of uncertainty, it is argued, heuristics can both be more accurate and consume fewer resources than typical “rational” models of cognitive processing.