This paper describes a model, derived from subjective expected utility ( SEU ) theory, which attempts to explain the effects of a fear-arousing communication about smoking and lung cancer on smokers' decisions to try to stop smoking. According to the model, the strength of the smoker's intention to try to stop smoking will depend on three factors: (1) the utility of lung cancer; (2) the perceived reduction in the risk of getting lung cancer that follows from successfully stopping smoking (probability difference); and (3) the subjective probability of successfully stopping smoking, or confidence. The usefulness of this model vis-á-vis one based on the amount of fear aroused by the communication was investigated in two experimental studies conducted in the field in which smokers watched either a videotape about smoking or one on a different health topic. The smoking videotape influenced not only attitudes and beliefs measured immediately after exposure but also subsequent self-reports of behavior (whether or not the subject tried to stop smoking or to cut down in the 3-month period following exposure to the videotape). In accordance with predictions, the three variables specified by the model together affected intention to try to stop smoking, which in turn affected behavior. The effect of confidence on intention was relatively large. There was no evidence for the multiplicative combination of utilities and subjective probabilities predicted by a strict SEU model. Contrary to the predictions of the model, but consistent with previous research, the amount of fear aroused by the videotape had an independent effect on intention, suggesting that explanations of the effects of fear-arousing communications will have to incorporate both cognitive and emotional mediation processes.