This paper addresses the impact of addiction and social interactions on cigarette demand, controlling for demographic and socioeconomic factors. A Box-Cox double-hurdle model for the simultaneous decisions of how much to smoke and whether to quit smoking is estimated on individual data from the 2000 Italian “Health Status and Use of Health Services” survey. The model incorporates the fixed costs of quitting and allows for the analysis of the effects of addiction and within-household interactions on smoking participation and cigarette consumption. Estimation results show that the duration of the smoking habit, used as measure of addiction, significantly increases the level of cigarette consumption and lowers the probability of quitting. Within-household social interactions affect individual’s attitude toward smoking. Participation decision is significantly influenced by the presence of other smokers and individual cigarette consumption increases as the consumption of the peer-group grows. Finally, gender differences are formally tested to verify whether male and female sub-samples can be pooled or should be separately analyzed. The hypothesis of equal consumption parameters is clearly rejected, suggesting the opportunity of distinguishing the consumption patterns of men and women.