ABSTRACT Purpose: to investigate the variability of the Dichotic Sentence Test through the test and retest in normal-hearing adults. Method: We evaluated thirty-six individuals aged 19 to 44 years old, right-handed and with normal hearing thresholds. We performed the basic audiological evaluation and then we applied the Dichotic Digit Test and Dichotic Sentence Test. The test and retest had two sessions, with an interval from 30 to 40 days, in the same shift. Results: In the integration task, there was an advantage of the right ear in both evaluation sessions. There was no significant difference between the measures obtained in the right ear in the two evaluation sessions, while in the left ear, we found a significant difference. In the analysis of the differences in ears between the test and the retest, we found that 64% of the individuals kept the same result in the right ear, while in the left one, only 36% of the individuals kept the same result in both stages and 44% showed a 10% difference between the two evaluations. We observed moderate positive correlation for both the right ear (r=0.420) and the left ear (r=0.550), with a tendency to improve retest scores. In the separation task, there was a small variability only in the left ear, also with improved retest scores. Conclusion: There was a significant difference between the measures obtained in the test and retest only in the integration task in the left ear, but there was a moderate positive correlation between the measures obtained in the two evaluation sessions, showing a tendency to improve scores in the second evaluation session.