We share these data to provide timely information on the total prevalence of adult tobacco consumers in the U.S. as well as in the major tobacco product categories: cigarettes, cigars, smokeless tobacco and e-vapor products. Data from nationally representative surveys that measure adult tobacco use prevalence are generally not available for at least a year after data collection. In contrast, we will release quarterly ATCT results based on 12 month moving averages the quarter following data collection. We hope that these data will be useful to regulators and researchers in the consideration and evaluation of policies intended to reduce the harm from use of combustible tobacco products.
Data Collection Procedures
ATCT uses a mixed landline and cell phone random digit dial (RDD) approach. This dual-mode approach is also currently used by two well-established government surveys that track tobacco consumer behaviors: the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS). The ATCT utilizes many of the study design parameters of BRFSS in order to maintain consistency of estimated measures.
The dual-mode telephone design is used, given its following benefits:
- response rate stability,
- prevalent use of dual-mode telephone by experienced survey research organizations,
- use of dual-mode telephone methods by government tobacco consumer trackers, and
- ability to more flexibly manage long term implementation and use.
To achieve a sample representative of the U.S. adult civilian, noninstitutionalized population, ATCT uses a random sample design (by U.S. Census region) in the data collection process. If the final sample of completed interviews was permitted to fall naturally (in terms of demographics), the sample would not be representative of total adults in the U.S. due to differential reachability (by phone) and response rates across various demographic subgroups. This is a sampling issue that affects all survey research design and execution to different degrees. To mitigate this, survey researchers commonly establish respondent sample quota guidance then weight the resultant data to assure that the sample is representative of the general study population. While this does not represent a perfectly representative sample design, this approach significantly mitigates sampling biases and improves weighting efficiency. We use this approach for ATCT, and quota the sample to avoid extensive data weighting and ensure adequate numbers of adult respondents within all weighted subgroups.
Criteria for Inclusion/Exclusion
Due to legal requirements for purchasing tobacco products and to correctly estimate prevalence at the national level, respondents must satisfy the following criteria at the time of screening to be enrolled into the study:
- Before August 2019, the respondent must be of legal age to purchase tobacco or older, which is 18+, 19+, or 21+, as defined by the state, county or locality where a respondent lives, regardless of current tobacco product use status. Starting from August 2019, respondents must be 21 years of age or older to be qualified for the study.
- The respondent must be a civilian member of the noninstitutionalized population of the U.S., including residents of non-institutional group quarters such as college dormitories, group homes, shelters, rooming houses, and civilians dwelling on military installations.
From the inception of ATCT in July, 2013 until August, 2015 the monthly sample of adult respondents was 1,838. The number of adult respondents was increased to 3,676 from August, 2015 to December, 2016. From December, 2016 through the present, ATCT surveys 2,400 respondents per month.