Rutgers School of Public Health interim dean, professor, and director of the Center for Tobacco Studies, Dr. Cristine Delnevo, along with colleagues, found that survey item response design impacts the prevalence estimates of emerging tobacco product use among youth. While survey science has largely standardized the measurement of cigarette smoking, standardization for emerging tobacco products, such as e-cigarettes has been challenging. Moreover, the impact of differing measurement approaches on emerging tobacco product estimates is not well understood.
[Photo: Dr. Cristine Delnevo]
Using 2014 New Jersey Youth Tobacco Survey (NJYTS) data, Dr. Delnevo and colleagues, examined e-cigarette and hookah ever and current use using both Center for Disease Control and Prevention core “check-all-that-apply” items and New Jersey state-added “forced-choice” items. They found that survey item response design impacts the prevalence estimates of emerging tobacco product use among youth. Specifically, they found that prevalence estimates calculated using a check-all-that-apply approach were substantially lower than the estimates calculated for the forced-choice responses for both ever and current use of e-cigarettes and hookah pipes. Furthermore, the authors point out that a major surveillance system for tobacco use changed their survey design from 2011 and 2013 and that the large increases over time previously noted for youth e-cigarette use may be overestimated because the 2011 estimate was likely under-estimated given the survey format used at that time.
Additionally, the study found that the two measurement approaches had especially low agreement among males, Hispanics and blacks, and among current smokers of traditional cigarettes. Dr. Delnevo also points out that “use of non-cigarette tobacco products that use check-all-that-apply questions, such as hookah pipes, may still be underestimated among U.S. adolescents” and urges the “public health community to use caution when interpreting trends from repeated cross-sectional surveys that change methods over time.”
The research paper was published in the March issue of The American Journal of Epidemiology: https://www.ncbi.nlm.nih.gov/pubmed/28369184