Biofeedback has been advocated as first-line therapy for patients whose symptoms of mild to moderate fecal incontinence have not responded to simple dietary advice or medication. Three main modalities have been described: (1) use of an intra-anal electromyographic sensor, a probe to measure intra-anal pressure, or perianal surface electromyographic electrodes to teach the patient how to exercise the anal sphincter; (2) use of a 3-balloon system to train the patient to correctly identify the stimulus of rectal distention and to respond without delay; and (3) use of a rectal balloon to retrain the rectal sensory threshold, usually with the aim of enabling the patient to discriminate and respond to smaller rectal volumes. Although a systematic review found that biofeedback eliminated symptoms in up to one half of patients and decreased symptoms in up to two thirds, these studies suffered from methodological problems, a lack of controls, and a lack of validated outcome measures. No studies have compared different exercise instructions, measured patient compliance with those instructions, or determined any trends in symptom response to the exercises prescribed. A recent study by the author suggests that patient-therapist interaction and patient coping strategies may be more important in improving continence than performing exercises or receiving physiological feedback on sphincter function. Better-designed randomized, controlled trials are needed to evaluate different exercise programs and different elements of biofeedback. Development and validation of outcome measures are important, and predictors of outcome and effects in patient subgroups, especially elderly and neurologically impaired patients, should also be investigated.