The reactive effects of three self-monitoring procedures were evaluated with respect to compliance with the self-monitoring requirements and session attendance. A total of 50 smokers who agreed to participate in a smoking cessation program were randomly assigned to one of five conditions: three self-monitoring conditions and two no self-monitoring control conditions. Results indicate that groups were initially equal on the variables of age, sex, pack year history, estimated rate of consumption, and alveolar carbon monoxide level. Drop-out rates for the no self-monitoring control groups were equal at the second session (10% each) and significantly lower than those of the self-monitoring groups (40%, 40%, and 60%). This significant difference was still present at the sixth treatment session. The implications of increased drop-out rates in smoking cessation programs resulting from self-monitoring are discussed.