Despite the large number of studies on caffeine and attention, interpretation is often difficult because of methodological weaknesses. In the present study, use of a small battery of tests with four key outcome measures, combined with an appropriate sample size, addresses many of these problems. This methodology was used to examine whether effects of caffeine (a dose of 2 mg/kg) could be explained in terms of reversal of the effects of caffeine withdrawal. This was achieved by examining effects in non-consumers (N = 35), who could not be withdrawn, and also in a group of consumers (N = 35) who had undergone withdrawal for a week and no longer reported symptoms of withdrawal. The results showed no effect of short-term withdrawal on the performance measures, even though subjective reports showed an increase in symptoms after withdrawal. In contrast, the caffeine challenge carried out on Day 8 showed that ingestion of caffeine was associated with faster simple reaction time, fewer long responses, greater detection of targets in the cognitive vigilance task, and faster encoding of new information. These results suggest that it is important to continue to investigate mechanisms underlying these effects of caffeine and to further evaluate the practical implications of such effects.