Poor working memory predicts false memories

Abstract

Two studies investigated whether individual differences in simple span verbal working memory and complex working memory capacity are related to memory accuracy and susceptibility to false memory development. In Study 1, undergraduate students (N=60) were given two simple span working memory tests: forward and backward digit span. They also underwent a memory task that is known to elicit false memories of nonpresented words, the Deese/Roediger-McDermott (DRM) paradigm. Poor simple span working memory (as reflected by suboptimal backward digit span scores) was related to elevated levels of false recognition. Study 2 (N=65) replicated this finding, in that suboptimal backward digit span performance was found to be predictive of false recognition. However, complex working memory capacity (operation span) was not related to false recognition. This pattern suggests that even in a homogenous sample of undergraduates, poor working memory is associated with the susceptibility to recollect words never presented.

This study was supported by a grant from the Dutch organisation for scientific research NWO grant number 452-02-006. We would like to thank Dr. Jason Watson, associate editor Professor Nelson Cowan, and two anonymous reviewers for their constructive comments on an earlier version of the manuscript, and Kevin Sijstermans and Benny Gorski for their help in recruiting and testing participants.

Notes

¹To correct for possible ceiling effects in our critical lure data, corrected critical lure false recognition (false recognition of critical lure – false recognition of unrelated lure; M=.81, SD=.16, range=.65) was calculated and related to the digit span scores. Only backward digit span score was negatively related to corrected false recognition for critical lures indicating a similar correlation as in the initial analyses, r=−.39, p<.01.

²As in Study 1, corrected false recognition of critical lures was calculated (M=.63, SD=.16, range=.75) to control for possible ceiling effects in false recognition. In line with Study 1, and the described analyses in Study 2, when correlations were calculated between the corrected false recognition of critical lures and the working memory indices, only backward digit span significantly correlated to corrected false recognition, r=−.50, p<.01.