Speech discrimination was assessed in three groups of hearing‐impaired children fitted with three different sensory aids: (1) single‐channel House 3M cochlear implants; (2) 22‐channel Nucleus cochlear implants; and (3) Tactaid II devices. Nine subtests of a speech discrimination battery were constructed, which focused both on suprasegmental and segmental speech contrasts. A change/no change task was used to assess discrimination. A change trial consisted of four standard stimuli followed by four contrast stimuli; the no‐change trial had eight standard stimuli. Hits and false alarms were tallied, and P(C)max was computed for each subject for each subtest. Results indicated that, in general, subjects with Nucleus devices had the highest mean P(C)max across subtests, followed by subjects with 3M House and Tactaid II devices. However, intersubject variability was high for each device. In addition, differences between devices frequently varied on individual subtests. The ability of this test battery to reveal differences between children's speech perception with different sensory aids will be discussed. [Work supported by NIH (NS24875).]