Patient on-off diaries are used in clinical trials, but a method to assure agreement between patient and examiner has never been developed. We tested whether a patient-teaching tape increased the rate of agreement between patient diary ratings and simultaneous neurologic assessment by a trained professional. A total of 32 consecutive patients who had Parkinson's disease with motor fluctuations independently completed a 4-h on-off diary (nine ratings) at the same time as an examiner. Those with <80% agreement with the examiner (n = 20) were randomized to view either a training tape that showed motor fluctuations (experimental group) or another videotape of general patient educational material (control group). All patients then underwent the same 4-h assessment of motor fluctuations. To test for long-term retention, they returned 1 month later and, without reviewing the videotape, underwent a final 4-h correlation assessment. After the training tape, the experimental group showed significant improvement, whereas the control group showed no improvement. Furthermore, another month later, the improvement in the experimental group was retained. Based on these findings, we suggest that future clinical trials assessing motor fluctuations incorporate this tape into their basic methodology.