Abstract
It was suggested that the three basic attributional patterns outlined by Kelley (1967) serve to organize memory for incoming effects information. Utilizing a 6 x 6 Greco-Latin square design, 36 subjects were presented with a series of sentences that defined a 3 x 3 x 3 Kelley cube. On each trial, the sentences were precisely consistent with either a person or an entity pattern of effects (pattern variable) and had zero, four, or eight sentences randomly deleted from the ideal pattern (missing variable). Memory for the sentences was assessed using a recognition paradigm. Main effects were observed for the pattern and missing variables. As the number of missing sentences increased, poorer recognition occurred as indicated by decreases in d'. The entity pattern produced poorer recognition than the person pattern.