Drinking and Smoking: a Field Study of their Association

Alcohol and tobacco seem to go together: individuals who drink tend to smoke, and vice versa. Users of both substances also seem to use them together; in particular, drinking seems to cue smoking. However, this relationship has not been studied in real-world contexts. We examined the situational association between drinking and smoking using Ecological Momentary Assessments (EMAs). Using a palm-top computer, 57 subjects reported the circumstances surrounding smoking occasions; as a comparison, subjects were assessed at random times when they were not smoking. A total of 4,020 observations were gathered over one week. Drinking was associated with smoking (odds ratio = 2.1). The effect was greater among heavier drinkers. This link persisted even when situational factors, such as time of day and engagement in leisure activity, were controlled, weakening conditioning explanations of the association. The data also did not support pharmacological compensation models: the degree of cognitive impairment when drinking did not predict the likelihood of smoking. The data were also inconsistent with a model based on disinhibition of smoking restraint, in that the relationship was not stronger among more restrained subjects. The data suggested that questionnaire reports of smoking patterns were invalid: the actual, observed link between drinking and smoking was not related to subjects' characterization of this relationship.