Children with dichotic left ear deficits received intensive training in phase I and phase II clinical trials designed to establish the efficacy of directly training dichotic listening. Dichotic verbal material was presented in the sound field with intensity adjusted separately for each speaker. Output from the right-sided speaker was initially 20-30 db HL lower than for the left-sided speaker, resulting in excellent performance in the left ear. Intensities were adaptively adjusted throughout training in 1, 2, and 5-dB steps in order to keep performance high across dichotic tasks. In both phase I (n=8) and phase II (n=13) trials, children demonstrated significant gains in dichotic left ear performance after training. In phase II, children also demonstrated significant gains in right ear performance. Overall results from the two trials support the feasibility of this training approach for improving a larger than normal interaural asymmetry on dichotic listening tasks. Significant improvements in language comprehension and word recognition in phase II suggest that this type of training may also facilitate language skills in some children.