The sessions provided psychoeducation and support (e.g., Brown, 2003) and were run by doctoral students in clinical psychology. A half-day trial abstinence period during which participants were instructed to briefly attempt abstinence occurred on Day 9 (Shiffman, sellckchem Ferguson, et al., 2006); accordingly, we excluded that day from analysis. Measures At baseline, participants reported a variety of demographic measures, including age, years of education, gender, income, ethnicity, years smoking, and smoking rate. They also completed the FTND (Heatherton et al., 1991), and the NDSS (Shiffman et al., 2004). Each time they smoked, participants recorded this on ED. On approximately five randomly selected smoking occasions per day, ED administered an assessment of craving and situational variables.

Craving Craving was assessed (the prompt was ��Rate cigarette craving��) on a single Visual Analog Scale (VAS) with anchors at either end: 0 = ��no craving,�� 10 = ��maximum craving.�� Despite the limitations of single-item scales (see Sayette et al., 2000; Tiffany, Carter, & Singleton, 2000), this assessment was deemed well suited for an EMA study; a single item allows the measure to reflect craving at the particular moment of the report, to minimize the intrusiveness of the report itself, and to decrease the probability that completing the measure will impact the reported craving level (Sayette et al., 2000). Moreover, in a prior study (Shiffman et al, 2002), where both ��craving�� and ��urge�� were assessed, the two correlated r = .80, demonstrating very high reliability for the single items.

That is, since reliability is the upper limit of validity, the reliability of each must be at least .80 to achieve the correlation of .80 (see Wanous, Reichers, & Hudy, 1997). Situations Situational variables were treated as binary��present or absent��in the analyses. Subjects reported any smoking restrictions in the place that they first decided to smoke. In cases where they had changed locations in order to smoke, they were also asked about restrictions in the location in which they actually smoked. All other variables referred to the situation where they first decided to smoke. Throughout this section, the actual text for questions assessing the smoking situation is included in parentheses along with a description of the situational variable.

Subjects�� reports of smoking restrictions (��Smoking regulations��Smoking allowed? Forbidden, Discouraged, Allowed��) were dichotomized by comparing episodes in which smoking was ��forbidden�� or ��discouraged�� to those in which smoking was ��allowed.�� Because smoking restrictions may Anacetrapib influence smoking behavior (Chandra et al., 2007; Shiffman et al., 2002) and covary with other situational variables (e.g., being at work), we controlled for the presence of restrictions in all subsequent analyses.