Studies of smoking relapse and temptation episodes have relied on retrospective recall and confounded between- and within-subject variability. Real-time data on temptations and lapses to smoke were gathered using palm-top computers. We made within-subject comparisons of the initial lapse, a temptation episode, and base rate data obtained through randomly scheduled assessments. Negative affect discriminated all three situations, with lapses worse than temptations, and temptations worse than random situations. Participants attributed lapses to negative mood and smoking cues, whereas temptations were more often attributed to behavioral transitions. Participants were 12 times more likely to report coping in temptations than in lapses. However, only cognitive (vs. behavioral) coping strategies were effective. Lapses (vs. the other situations) were more likely to occur when smoking was permitted, when cigarettes were easily available, and in the presence of other smokers. The results have clinical implications, and the computerized monitoring methods may be applicable to an array of clinical research problems.

We report the development of a new multidimensional questionnaire to measure nicotine dependence, based on Edwards's syndromal conceptualization of dependence. We present three studies. In study 1, we administered the Nicotine Dependence Syndrome Scale (NDSS) to 317 smokers in a smoking cessation study. Factor analysis of the NDSS revealed five factors: Drive (craving and withdrawal, and subjective compulsion to smoke), priority (preference for smoking over other reinforcers), tolerance (reduced sensitivity to the effects of smoking), continuity (regularity of smoking rate), and stereotypy (invariance of smoking). A single overall score based on the first principal component, NDSS-T, was retained as a single core measure of dependence. The NDSS showed promising psychometric properties: NDSS-T and factor scores showed strong associations with dependence-relevant measures, even when we controlled for scores on the Fagerström Tolerance Questionnaire (FTQ); and the NDSS predicted urges when smoking, withdrawal in acute abstinence, and outcome in cessation. The five factor scores showed differential patterns of correlations with external validators, supporting the multidimensionality of the measure. In study 2, we revised the NDSS to expand some subscales and administered it to 802 smokers in a cessation study. The same five factors were extracted, the internal reliability of some subscales was improved, and the factor scores again showed associations with dependence-relevant validators, which were largely maintained when we controlled for FTQ scores. In study 3, with 91 smokers in a cessation trial, we established that the test-retest reliability of the subscales was adequate. Thus, the NDSS presents a valid multidimensional assessment of nicotine dependence that may expand on current measures.

Smokers who recently quit (N = 214) monitored smoking urges for up to 26 days after quitting. Computers administered 4-5 assessments daily at random times; participants rated urges on waking and when they experienced temptation episodes. Urge intensity after cessation did not generally exceed urges reported during baseline ad lib smoking. Urge intensity and temptation frequency consistently declined over the quit period. Controlling for urge intensity at baseline, all daily urge intensity measures predicted lapse the following day in proportional hazards survival analyses. Average duration of temptation episodes also predicted lapses; frequency of temptation did not. To isolate the effect of day-to-day variations in urges, participants' nicotine dependence and urge intensity on quit day were controlled for. Only urge intensity at waking still predicted lapse risk; this was not because of this measured being closer in time to the day's lapses. Among lapsers, urge intensity at waking and in temptations rose preceding a lapse.

The authors assessed the association between smoking and situational cues, including affect, in real-world contexts. Using ecological momentary assessment, 304 smokers monitored ad-lib smoking for 1 week, recording each cigarette on palm-top computers. Generalized estimating equations contrasted 10,084 smoking and 11,155 nonsmoking situations. After controlling for smoking restrictions, smoking was strongly related to smoking urges and modestly related to consumption of coffee and food, the presence of other smokers, and several activities. Smoking was unrelated to negative or positive affect or to arousal, although it was associated with restlessness. Thus, in daily life, affect appears to exert little influence over ad-lib smoking in heavy smoking adults.

Recent research suggests that retrospective coping assessments may not correspond well with day-to-day reports. The authors extended this work by examining the correspondence between short-term (within 48 hr) retrospective coping reports and momentary reports recorded via a palm-top computer close in time to when the stressor occurred. There was relatively poor correspondence between the 2 assessments. Some reports of momentary coping were not reported retrospectively, and some coping reported retrospectively was not reported at the time the stressor occurred. Cognitive coping was more likely to be underreported retrospectively; behavior coping was overreported. Participants were consistent in their discrepancies, but there was no correspondence between discrepancy rates and demographic or personality variables.