Graphic instructions for the collection of clean-voided urine specimens were posted in outpatient clinic bathrooms. These signs were intended to supplant clinic staff verbal instructions and to improve patients' ability to perform this procedure correctly. To assess the impact of the signs, the contamination rates were determined for urine culture specimens for the 6-month periods immediately preceding and following their introduction. The contamination rate for 306 consecutive specimens submitted before the signs were hung (12 per cent) was significantly lower (p = 0.004) than for the 360 specimens submitted afterward (20.6 per cent). Further, the incidence of true bacteriuria was significantly higher before the signs were posted (21.8 per cent) than afterward (15.2 per cent) (p = 0.02). There was no reported change in the provision of verbal instructions by clinic personnel, and the signs provoked some negative reactions from patients and visitors. This unexpected adverse outcome of a wellintended patient education effort is reported so that others may profit from the authors' mistakes.