Believing you've had caffeine boosts performance, even when you haven't

Caffeine is one of the most widely used performance aids in sport, and a long line of research treats its benefits as pharmacological. But a separate body of work has shown that simply telling someone they have taken caffeine, while giving them an inert placebo, can also improve how they perform. If belief alone moves performance, that matters twice over: it means some of what athletes attribute to caffeine may be expectation, and it means caffeine trials that fail to account for placebo effects may overstate the drug's true contribution. This review set out to measure how big that belief effect actually is.

The authors systematically searched for studies, from database creation to July 2026, that sampled healthy adults and deceptively administered a placebo described as caffeine before a sport performance test. Seventeen studies met the criteria, covering outcomes such as bench press, cycling time trials and repeated sprints. For each, the standardised mean difference between the placebo-caffeine condition and baseline was calculated, and results were combined using random-effects models. Study quality was judged using Cochrane's risk of bias assessment, and meta-regressions explored what features of the studies predicted larger or smaller effects.

The pooled result showed a small-to-moderate benefit of placebo caffeine on performance, with a standardised effect of 0.49 and a confidence range of 0.28 to 0.69. Meta-regressions suggested effects were more likely to be greater when the researchers themselves were blinded, and tended to be smaller in studies that sampled more males. The quality picture was poor: nine of the 17 studies were judged at high risk of bias and the remaining eight were unclear, and the overall certainty of the evidence was graded as low.

This review consolidates a scattered literature into a single estimate and confirms that expectation effects around caffeine are not trivial. It also complicates the interpretation of standard caffeine trials, since part of what those trials measure may be belief rather than pharmacology. The finding that effects looked larger when researchers were blinded points to how easily these studies can be distorted by the people running them. The authors call for stronger designs with blinded administration and concealed allocation before the size of the placebo effect can be pinned down with confidence.

This is a reason to respect the power of ritual and expectation in your own training. If you feel sharper after your pre-workout drink, some of that lift may come from believing in it, not just from what is in the cup. None of this shows that caffeine itself does not work; the studies here tested belief, not the compound. But it is a useful lens on any supplement that feels like it works instantly, and a reminder that low-quality studies can make both drugs and beliefs look better than they are.