Participants watched online survey-imbedded promotional videos featuring influencers. Influencer age appearance was manipulated using AI-based age filter.

Responses to one question adapted from PATH, "Based on the videos you just watched, do you think using e-cigarettes for vaping nicotine is harmful to your health?" The outcome was assessed on the 5-point ordinal scale with the following response choices: Not at all harmful, slightly harmful, somewhat harmful, very harmful, extremely harmful.

Responses to the following questions: using e-cigarette is: not cool/cool, unattractive/attractive were assessed on the 7-point semantic differential scale with the word pairs anchored at each end. The scale was adapted from prior research. The items were combined (by summing all the non-missing values of the items) into one variable (α=0.94).

Susceptibility to e-cigarette use was measured (among never-users of e-cigarettes), using the validated three-item scale (indicating peer influence, intention, and curiosity about e-cigarette use) adapted from PATH, and combined into one variable (α=0.93). Consistent with prior research, the measure was dichotomized with responses "definitely not" to all items being coded as "not susceptible" and responses "probably not," "probably yes", or "definitely yes" being coded as "susceptible.

Perceived influencer credibility (i.e., honesty, trustworthiness, knowledge) were assessed using a 0 (e.g., dishonest) -100 (e.g., honest) scale that has been validated in prior research.