This study was undertaken to determine the relationship between extended-high-frequency (EHF) hearing losses and wideband middle ear impedance in children with chronic otitis media (OM) histories. Children with OM histories were selected from a prospective study cohort if they had normal tympanograms, no air-bone gaps, and no otoscopic evidence of active OM at the time of testing. OM subjects were divided into two groups, those with Better Hearing in the EHF range and those with Worse Hearing in the EHF range. The OM groups were compared with an age-matched, healthy Control group that had no more than five documented episodes of OM and no more than two in any 1 yr. All children were 9 to 16 yr of age. Subjects were tested by standard audiometric methods in the conventional audiometric range (0.25 to 8.0 kHz) and the EHF range (8 to 20 kHz). Middle ear impedance and reflectance were measured with an experimental system over the frequency range 0.25 to 10.08 kHz. The Worse Hearing OM group had slightly poorer hearing in the conventional audiometric frequency range compared with the other two groups. The Better Hearing OM group and the Control group had nearly identical EHF hearing. The Worse Hearing OM group had significantly poorer EHF hearing compared with the other two groups, the difference increasing exponentially with frequency. Middle ear impedance differences among groups were confined to low frequencies (<2 kHz). The Control group had significantly higher negative reactance than the two OM groups. There were no significant group differences in impedance or reflectance in the high frequencies (2 to 10 kHz). The results of this study confirm those of previous reports that children who have recovered from chronic OM have significantly poorer hearing in the EHF range compared with children without significant OM histories. The EHF hearing losses that occur in children with OM histories are strongly frequency dependent, suggesting a preferential effect on the base of the cochlea. Middle ear impedance and reflectance differences do not account for the EHF hearing losses observed in children with OM histories. The results support the hypothesis that OM-related EHF hearing losses are cochlear in origin.