From 1973 to 1989, 1318 subjects ranging in age from 20 to 96 years completed auditory reaction time (RT) tasks as part of the extensive test battery of the Baltimore Longitudinal Study of Aging. In addition to central tendency, higher moments of reaction time frequency distributions (variability and skewness) were analyzed to (1) determine changes in the shape of the frequency distribution not revealed by analysis of central tendency alone, which could explain a portion of the slowing effect, and (2) examine differences in cross-sectional and longitudinal approaches. Cross-sectional results revealed significant age increases for median simple reaction time (SRT) and median disjunctive reaction time (DRT), as well as SRT and DRT variance. Males were significantly faster and less variable than females for both SRT and DRT tasks. SRT distributions were much more positively skewed than DRT distributions, and DRT distributions became more positively skewed with age. Increased variability appears to account for a sizable portion of the age-related performance slowing. Longitudinal analyses produced slightly different results, the most distinctive being the fading of age and sex differences over longer periods of time for median RT, variance, and skewness. Findings from this research have human factors implications for task design, personnel selection, accident analysis, and demographic norms.