Mnemonic distortion following behavior observation: Interactions of negative versus positive behavior valence, diagnostic label, and relative behavior frequency

Abstract

To study how diagnostic labels, positive or negative behavior valence, and high versus low behavior frequencies affect distortion of memory for behavior, 72 undergraduates used an observation system to record occurrences of 10 negative and 10 positive behaviors shown on one of two videotapes of a child. Pilot testing had verified that one tape showed many negative but few positive behaviors, and that the other tape showed few negative but many positive behaviors. Just before viewing the tape, subjects were told either that the child was “behaviorally disturbed” or “normal,” or had been given no label. Exaggeration of observed frequencies at a 1-week recall was shown by significant and separate interactions of behavior valence and frequency, and of behavior valence and label. Specifically, negative behavior frequencies were exaggerated significantly at recall when the “behaviorally disturbed” or no label was present for both high and low frequencies of the negative behaviors. However, recalled frequencies of negative behaviors were exaggerated for the “normal” label only when they were low in frequency relative to positive behaviors. Positive behavior frequencies were exaggerated significantly in memory only if the “normal” label was used and then only if the positive behaviors had occurred at low frequencies relative to negative behaviors. No significant underestimation of previously observed behavior frequencies was found for either positive or negative behaviors. Implications of these findings for cognitive theories of memory and behavior, and for minimization of mnemonic stigma, are noted.