Item analysis of a single response dichotic listening test

Ten left-handed and eight right-handed subjects were tested on five replications of a single-response dichotic listening test. Nine CV dichotic pairings of the six stop consonants plus /a/ were used. Computer synthesis guaranteed that these pairs differed on only a single phonemic feature, thus permitting fusion. Mean laterality scores were significantly higher (greater right ear advantage) for right-handers than for left-handers. Selective attention instructions did not alter laterality scores. Test-retest correlation coefficients were consistently high (greater than .70 for all adjacent replication comparisons). Error rates were low (averaging 5.5%) and 39% of these were psychoacoustic confusions rather than true (guessing) errors. These confusions were found for only two of the nine pairs. Stimulus dominance, as opposed to ear dominance, was considered separately. Two dichotic pairs had such strong stimulus dominance for all subjects that little ear advantage information was available from those pairs. It was recommended that the three dichotic pairs prone to psychoacoustic confusions and/or overwhelming stimulus dominance be removed from the test.