We studied 16 patients with long-standing irritable bowel syndrome of moderate severity using a controlled, double-blind crossover method. Five sedative-anticholinergic drug combinations and a placebo were tested. The subjective response was assessed with four subjective methods to include an increasing number of response variables. The patients preferred 30 mg phenobarbital plus 8 mg belladonna (P & B) to placebo (P = 0.02). Five of ten patients were helped "some" or "a lot" with placebo, while ten of 15 were helped "some" or "a lot" with P & B (P = 0.07). The ten prominent-symptoms method revealed that subjective symptoms such as nervousness, sleep difficulties, and tiredness were experienced as greater problems than diarrhea. The factor analysis method documented a strong placebo response. Simpler evaluation methods such as drug preference and a five-choice method appear more likely to show a positive drug effect, while the inclusion of a larger number of variables appears to emphasize the placebo portion of the response. These observations may help explain some of the apparent discrepancies between the conclusions of some controlled clinical trials and subsequent clinical experience.