Some methodological issues in the conduct of caffeine research.

The caffeine literature of the past 10 years shows the frequent recurrence of several weaknesses--the use of weak or erroneous hypotheses in experimental design, design flaws (e.g. the use of excessively high doses) that prevent the derivation of any meaningful implication for humans, the violation of ordinary rules of logic in extrapolating experimental results to human situations (e.g. the application of findings in naive animals to caffeine-tolerant human subjects) and the biased selection of literature citations to support a hypothesis and ignore opposing evidence. Workers who review the literature dispassionately, who use dosages relevant to human caffeine consumption when they wish to extrapolate from animal studies to man, who distinguish carefully between acute and chronic effects and who include proper controls in feeding studies to prevent confusion of the effects of dietary deprivation with the pharmacological effects should succeed in advancing our understanding of caffeine's effects in the human body.