Binding of Independent Contexts in Source Memory

Abstract

Within the present study the binding between two independent source dimensions in a multidimensional source memory paradigm was investigated. Specifically, I examined stochastic dependence between the retrieval of each source dimension. Previous work has defined stochastic dependence as the higher probability of correct source retrieval for one dimension contingent on previous correct retrieval of a second source dimension, versus when contingent on incorrect retrieval of the second source dimension. Evidence is mixed as to whether item information within source dimensions must be encoded simultaneously in order to demonstrate eventual stochastic dependence. The present study tested the binding of two cross-modal source attributes (visual [left or right side of screen] and auditory [male or female voice]) over time by manipulating the lag between each independently-encoded source dimension. Source dimensions were encoded simultaneously or separately by two intervening encoding trials (separate condition). Based on the behavioral measures of binding, source dimensions presented more closely in time resulted in stochastic dependence in the context of remembering and not knowing. However, a multinomial model of response frequencies produced evidence of joint retrieval in the context of remembering for both the simultaneous and separate conditions. Because source dimensions were presented over time in some conditions, working memory capacity was measured as a potential predictor of stochastic dependence. However, working memory measures did not correlate with source memory performance. Future directions in examining this separate condition, stochastic dependence and individual differences in working memory capacity are discussed.