The effect of sleep loss on reaction time (RT) performance varies as a function of age, with RTs of older subjects typically showing less decrement (relative to rested baseline) than those of younger subjects. In the current paper, we examined the nature of this relationship in a 7-day sleep restriction study. The number of repeated measures made it possible to model both intra-individual trajectories over days and individual differences in these trajectories. Results revealed (a) consistent individual differences in RT patterns over time after controlling for experimental design effects; (b) less cumulative RT decline among older individuals regardless of the degree of sleep restriction; and (c) consistent individual variability in performance patterns even after accounting for the effects of age.