Objective: This study examines monthly variation in reports of recent alcohol consumption behavior.
Method: Telephone survey data collected by the Behavioral Risk Factor Surveillance System were used to measure self-reports of drinking and heavy episodic drinking in the 30 days before the interview. The sample (N = 57,758) comprised cases collected in six states in 5 selected years. Monthly variability in self-reported alcohol consumption was evaluated by logistic regression analysis, using "deviation from means coding" for the month of interview, with background variables and year of interview controlled and state-level clustering adjusted.
Results: Adults were significantly more likely to report past 30-day alcohol consumption when interviewed during the month of January and less likely to report consumption when interviewed during March. Respondents were also more likely to report heavy episodic drinking during the month of July. Despite large differences in alcohol consumption between male and female participants, these seasonal patterns were essentially the same for both genders.
Conclusions: The findings suggest that assessments of alcohol consumption that do not consider seasonal variation may be biased. In order to minimize this bias, it is recommended that researchers avoid collecting data only during these particular times of the year or extend the data collection interval and control for seasonal variability.