|| Checking for direct PDF access through Ovid
Quantifying cigarette consumption is fundamental to smoking research. We examine the correspondence among 3 methods of capturing cigarette consumption in nondaily smokers. In the study, 232 nondaily smokers recorded cigarette consumption over 2 weeks of ad libitum smoking (total = 3,303 days) using 3 methods: (a) interactive voice response (IVR; calling an automated telephone line when they smoked), (b) butts (storing and returning the butts of cigarettes smoked), and (c) time-line follow-back (TLFB; retrospectively reporting daily consumption). Analyses examined relationships among the measures (cigarettes per day and proportion of days abstinent) for the average over 2 weeks and for each day. In averaged data, the methods were highly correlated for both quantity smoked and percentage of abstinent days (rs > .95); the average was very reliable (α = .99). All 3 methods showed similar, very strong relationships to urinary cotinine (rs > .70). Estimates of the percentage of days abstinent differed modestly but significantly by method (butts [32%] > TLFB [31%] > IVR [28%]). For individual days, there was no significant difference in cigarettes per day by method, the 3 methods correlated highly (rs = .75–.96), and agreement on abstinence was very high (≥90%; κs ≥ .95). In nondaily smokers reporting smoking by IVR, butts, and TLFB, the resulting estimates of cigarette consumption and abstinence are highly concordant and equivalent. The composite (average) of the 3 methods yields a highly reliable estimate of cigarette-smoking behavior in this population. In contrast to past studies, this study suggests these 3 methods are equivalent and implies that any 1 can be used by nondaily smokers.