�� Controlling thoroughly for whether or not individuals moved to smoke and whether smoking was restricted when they decided to smoke did not affect these results. All other analyses controlled for the presence of smoking restrictions. Table 1. Differences in craving by situational variables Subjects reported similar levels of craving for cigarettes smoked at home versus at work. Nor was craving affected by interacting with others or being at a bar/restaurant. With regard to drinking alcohol or being at a bar or restaurant, results did not change when we limited the analysis to participants who reported smoking while drinking alcohol during the study (n = 170). Craving was 0.05 points higher when participants were with other people versus when alone (p < .05) and was 0.

16 points higher when people were with others in a group versus only with others in view (p < .001). Overall, craving was similar whether or not others were smoking nearby or whether the participant was in a group with others or simply had others in view. Nor was there a significant interaction between these two variables. Craving was 0.07 points lower for cigarettes smoked when people were inactive (vs. active; p < .05) and was also 0.07 points lower when people were smoking while engaged in leisure activities (vs. work; p < .05). Craving was 0.26 points higher for cigarettes smoked while consuming food or drink (p < .0001). When food and drink consumption were analyzed separately, both had a similar effect on craving individually (eating vs. not: p < .0001; drinking vs. not: p < .001).

Drinking alcohol or caffeinated beverages was unrelated to craving. Finally, craving was 0.125 points higher (p < .001) for the first assessed morning cigarette (average time of assessment: about 8:30 a.m.) than those smoked later on (average time of assessment: about 4:15 p.m.). Affective correlates of craving are summarized in Table 2. Negative affect had a curvilinear relationship with craving, such that craving rose with negative affect up to a value of 1.0 (1 SD above the mean) and then decreased thereafter (p < .01; see Figure 1). In contrast, there was a positive linear relationship between craving and positive affect (p < .0001; see Figure 1) and craving and arousal (p < .0001; see Figure 1). There was an interaction between negative affect and arousal, such that negative affect was associated with greater increases in craving when arousal was high (p < .

01; see Figure 2). Table 2. Affective predictors of craving Figure 1. Negative affect, positive affect, and arousal predict cigarette craving based on modeled regression curves. Figure 2. Negative affect (NA) interacts with arousal to predict cigarette craving. Lines indicate relationship between arousal and craving at three values of NA, based on regression Cilengitide model.