The irritable bowel syndrome (IBS) is a common and poorly understood chronic condition that is treated with a great variety of drugs and other therapies without notable enduring success. As there are no objective markers of improvement, and because there may be a very large placebo response, potential treatments for IBS are difficult to assess. Probably the only method that can reliably evaluate IBS therapies is the randomized, double-blind, placebo-controlled treatment trial. The purpose of this review is to critically examine issues central to establishing the efficacy of treatments for IBS in such trials. These include the definition of IBS, measures of efficacy, the placebo response, trial length, maintaining blindedness, the crossover design, ability to generalize, and statistical considerations. With this background, all published IBS treatment trials are examined. It is concluded that not a single study offers convincing evidence that any therapy is effective in treating the IBS symptom complex. Well-designed and executed IBS treatment trials are urgently needed; suggestions are given for essential features of such trials.