The present study compared test-retest reliability for dichotic listening (DL) performance to consonant-vowel (CV) syllables under three different attentional instructions. Previous studies of reliability in dichotic listening have shown great variability which most likely reflects the large variations in stimuli, procedures, and techniques used in the different studies. Sixteen subjects were tested twice, 2 weeks apart, with the same dichotic listening procedure, including a divided attention condition and two focused attention conditions on each test occasion. The results showed test-retest correlations ranging from .61 to .86 for the three attentional conditions, with the smallest correlation coefficient for the divided attention condition.