Feasibility and reliability of a mobile tool to evaluate exposure to tobacco product marketing and messages using ecological momentary assessment

- Commonly used measures of tobacco marketing exposure are subject to recall bias.
- Ecological momentary assessment (EMA) reduces retrospective recall error.
- EMA is a feasible and valid method to measure tobacco marketing exposure.
- Recall measures might be improved with clearer response categories or with priming.

IntroductionExisting measures of tobacco marketing and messaging exposure are limited, relying on recall, recognition, or proxy measures. This study aimed to determine the feasibility and reliability of a mobile application for the measurement of tobacco and e-cigarette marketing and message exposure using ecological momentary assessment (EMA).MethodsYoung adults from Austin, TX (n = 181, ages 18-29) were instructed to use a mobile application to record all sightings of marketing or social media related to tobacco (including e-cigarettes) in real-time for 28 days (Event EMAs). Tobacco product use and recall of message encounters were assessed daily using an app-initiated EMA (Daily EMAs).ResultsThe mobile app was a feasible and acceptable method to measure exposure to tobacco messages. The majority of messages (45.0%) were seen on the Internet, and many were user-generated. Thirty-day recall of messages at baseline was poorly correlated with messages reported via Event EMA during the study period; however, the correlation between post-study 30-day recall and Event EMA was much stronger (r = 0.603 for industry-sponsored messages, r = 0.599 for user-generated messages). Correlations between Daily EMAs and 30-day recall of message exposure (baseline and post-study) were small (baseline: r = 0.329-0.389) to large (post-study: r = 0.656-0.766).ConclusionsThese findings suggest that EMA is a feasible and reliable method for measuring tobacco message exposure, especially given the prevalence of messages encountered online and on social media. Recall measures are limited in their ability to accurately represent marketing exposure, but might be improved by a period of priming or clearer response categories.