The effects of caffeine on levels of task difficulty were examined with a digit-symbol substitution task. In a double-blind study, 37 undergraduates were randomly assigned to one of four doses: 0, 200, 400, or 600 mg of oral administration of caffeine. Subjects attempted the task 10 and 90 min post-drug. Performance increased over time and declined across increasing levels of task difficulty. In general, neither caffeine nor task difficulty showed a systematic influence on the task. The exception was at the first test time where the 600 mg dose decreased performance at a moderate level of task difficulty. As expected of the digit-symbol substitution task which had a speed component, males were more faster and yet maintained accuracy in the task compared to females; initially, at 10 min post-drug, males also showed fewer errors than females. Post-hoc analysis with history of caffeine use was done; subjects were grouped as lower or higher caffeine users. In general, higher caffeine users had higher performance than lower users, especially at the moderate levels of task difficulty. Overall, the male higher caffeine users showed the highest performance.