The causal role of outcome expectancies in smoking motivation and behavior was tested in a laboratory experiment. Two expectancy manipulations-Health Consequences Expectancy Increase (HC) and Mood Management Expectancy Challenge (MM)-were crossed in a 2 x 2 factorial design. One hundred twenty-three heavy smokers were randomly assigned to the conditions and shown videotaped interviews of smokers and exsmokers followed by an interview that personalized the information. Both manipulations produced changes in targeted expectancies on at least one of two expectancy instruments. As predicted, only the HC manipulation produced increases in motivation to quit smoking and a reduction in smoking 3 months later. This study provides initial evidence that smoking expectancies play a causal role in smoking motivation, consistent with previous research with alcohol.