Current studies are inconclusive regarding specific patterns of gender differences in age-associated hearing loss. This paper presents results from the largest and longest longitudinal study reported to date of changes in pure-tone hearing thresholds in men and women screened for otological disorders and noise-induced hearing loss. Since 1965, the Baltimore Longitudinal Study of Aging has collected hearing thresholds from 500 to 8000 Hz using a pulsed-tone tracking procedure. Mixed-effects regression models were used to estimate longitudinal patterns of change in hearing thresholds in 681 men and 416 women with no evidence of otological disease, unilateral hearing loss, or noise-induced hearing loss. The results show (1) hearing sensitivity declines more than twice as fast in men as in women at most ages and frequencies, (2) longitudinal declines in hearing sensitivity are detectable at all frequencies among men by age 30, but the age of onset of decline is later in women at most frequencies and varies by frequency in women, (3) women have more sensitive hearing than men at frequencies above 1000 Hz but men have more sensitive hearing than women at lower frequencies, (4) learning effects bias cross-sectional and short-term longitudinal studies, and (5) hearing levels and longitudinal patterns of change are highly variable, even in this highly selected group. These longitudinal findings document gender differences in hearing levels and show that age-associated hearing loss occurs even in a group with relatively low-noise occupations and with no evidence of noise-induced hearing loss.