Patients in clinical practice and participants in clinical trials are warned about side effects that may result from their treatment. Such warnings could lead to placebo-induced side effects if they create an expectation of these effects. We used an experimental model to test this possibility. Undergraduates reporting sleep difficulty received placebo treatment disguised as a hypnotic for one week and were warned about either one or four bogus side effects. Placebo treatment significantly improved sleep difficulty relative to a no treatment control group, as indicated by self-report and by objective outcomes. At the end of the treatment week participants who had been warned about a single side effect showed better recall of this effect than those warned about four side effects. Most importantly, participants tended to report experiencing a side effect they had been warned about, with a trend towards a larger effect in participants warned about one side effect. This evidence for placebo-induced side effects may need to be considered when interpreting data on side effects from clinical trials.