This article reviews efforts to assess the effectiveness of television programs on changing health behaviors. The relatively limited literature to date has featured primarily correlational studies which permitted no causal conclusions. The few experimental studies available present a mixed picture. Numerous methodologic problems beset any effort at assessment of effectiveness. Perhaps the greatest difficulty is operationalizing a design that permits true experimental manipulation on a relevant sample without contamination. The several methodologic problems and some potential means of surmounting them are discussed.