Memory for Dichotic Pairs: Disruption of Ear Report Performance by the Speech‐Nonspeech Distinction

When dichotic pairs are presented in rapid succession, the best strategy is usually to segregate the items by ear of arrival. If the stimuli fall into distinguishable classes such as letters and digits, then report by class is roughly comparable to that obtained by the ear method. In the present experiment, a stimulus class distinction markedly interfered with listeners' ability to use the ear report method. The classes were speech (/ba, da, ga/) and nonspeech (500‐, 700‐, and 1000‐Hz tones). A trial consisted of three successive pairs of speech to one ear and nonspeech to the other. In separate blocks of trials, subjects reported the order of arrival for a given stimulus class (speech or nonspeech) or for a given ear (left or right). Performance was excellent for both types of report as long as all the speech went to one ear and all the nonspeech went to the other ear. However, when speech and nonspeech switched between the ears during a trial, class report remained excellent but ear report dropped substantially. The status of stimuli as speech or nonspeech, then, is a fundamental distinction, one that can reduce the effectiveness of the normally useful ear report method in dichotic memory tests.