Covariation-Based Attribution: On the Ability to Assess Multiple Covariates of an Effect

Abstract

For logical and normative reasons, attributions have to follow the covariation rule. However, little is known about the cognitive process by which the covariation of an effect with multiple causes is extracted from empirical observations. Previous research used verbal quantifiers to manipulate covariation with subjects (consensus), objects (distinctiveness), and time (consistency). In contrast, the inductive assessment of covariation from a sequence of observations was the focus of the present experiments. Empirical findings demonstrate that accurate covariation judgment is usually confined to one focused factor. The role of the focal factor is augmented rather than discounted when another covarying factor is present. When attributers can actively search for information, memory for singular observations is strong, but attributional judgments deteriorate. Experimental search strategies improve event memory but interfere with covariation-based judgments. A unidimensional joint scale model provides a better account of underlying cognitive processes than the common factorial-design model.