Electronic cigarette and moist snuff product characteristics independently associated with youth tobacco product perceptions

INTRODUCTION Tobacco product characteristics convey product attributes to potential users. This study aimed to assess independent contributions of specific e-cigarette and smokeless tobacco product characteristics to adolescents’ perceptions about these products. METHODS In 2019–2020, students (N=1003) attending a convenience sample of 7 high schools in California (USA) were individually randomized to one of two discrete choice experiments, featuring either electronic cigarettes (e-cigarettes) or moist snuff. Participants were presented like-product pairs of randomlygenerated hypothetical tobacco products differing in device type, flavor, vapor cloud, and nicotine amount (for e-cigarettes) or differing in brand, flavor, cut, and price (for moist snuff). Within pairs, participants were asked about which product they were more curious, was more dangerous, would give a greater ‘buzz,’ and would be easier to use. Conditional logistic regression was used to quantify independent associations of product characteristics to participants’ choices. RESULTS Each e-cigarette and moist snuff characteristic was independently associated with multiple product perceptions. All non-tobacco flavors were associated with more curiosity and perceived ease-of-use but lower perceived danger. Tank and pod-type e-cigarettes were viewed as easier to use and garnered more curiosity than ‘cigalike’ or ‘drip-mod’ devices. Smaller vapor cloud e-cigarettes and lower-price moist snuff were viewed as less dangerous, less buzz-inducing, and easier to use. Product ever users held stronger perceptions than never users about device type (e-cigarettes) and brands (moist snuff), while product naïve participants more strongly associated flavor with danger and buzz. CONCLUSIONS Tobacco product characteristics convey product attributes to adolescents that may increase appeal. Restricting specific characteristics, including flavors, could reduce positive perceptions of these products among youth.


E-cigarette characteristics and levels displayed to discrete choice experiment participants
Legend: This image was displayed to survey participants in the e-cigarette discrete choice experiment to demonstrate the product characteristics and their possible levels that would appear in subsequent randomlygenerated hypothetic e-cigarette composite products.

Figure A.3. Example questionnaire item displayed to participants
Legend: Discrete choice items were displayed to participants in the format above. Participants saw either ecigarette items (above) or moist snuff smokeless tobacco items (not shown). Levels of product characteristics were randomly generated to construct each hypothetical product composite. Figure is simulated to enhance resolution: actual screen images differed slightly in presentation.
Are you more curious about?
Is more dangerous to health?
Would be easier to use?
Would give a bigger "buzz" or "head rush"?
Of the following products, which one…?  Table reports coefficients from conditional logistic regression models in a discrete choice experiment. Interaction terms were added to assess differences by gender. Coefficients indicate how much the attribute level in question (e.g., flavor: fruit) increased or decreased the log-odds of a displayed e-cigarette product being chosen relative to the reference level (e.g., flavor: tobacco), adjusted for all other displayed e-cigarette attributes. For each displayed pair of e-cigarettes, participants were asked about which they were more curious, and which was more dangerous, would give a greater buzz, and would be easier to use. Positive values indicate characteristics (relative to reference) that independently contributed to greater curiosity or perceived danger, buzz, and ease of use, respectively. Abbreviation: CI = Confidence interval  Table reports coefficients from conditional logistic regression models in a discrete choice experiment. Interaction terms were added to assess differences by having ever used an e-cigarette. Coefficients indicate how much the attribute level in question (e.g., flavor: fruit) increased or decreased the log-odds of a displayed e-cigarette product being chosen relative to the reference level (e.g., flavor: tobacco), adjusted for all other displayed e-cigarette attributes. For each displayed pair of e-cigarettes, participants were asked about which they were more curious, and which was more dangerous, would give a greater buzz, and would be easier to use. Positive values indicate characteristics (relative to reference) that independently contributed to greater curiosity or perceived danger, buzz, and ease of use, respectively. Abbreviation: CI = Confidence interval  Table reports coefficients from conditional logistic regression models in a discrete choice experiment. Interaction terms were added to assess differences by gender. Coefficients indicate how much the attribute level in question (e.g., flavor: fruit) increased or decreased the log-odds of a displayed moist snuff product being chosen relative to the reference level (e.g., flavor: tobacco), adjusted for all other displayed product attributes. For each displayed pair of moist snuff products, participants were asked about which they were more curious, and which was more dangerous, would give a greater buzz, and would be easier to use. Positive values indicate characteristics (relative to reference) that independently contributed to greater curiosity or perceived danger, buzz, and ease of use, respectively. Abbreviation: CI = Confidence interval    Table reports coefficients from conditional logistic regression models in a discrete choice experiment. Interaction terms were added to assess differences by having ever used conventional smokeless tobacco. Coefficients indicate how much the attribute level in question (e.g., flavor: fruit) increased or decreased the log-odds of a displayed moist snuff product being chosen relative to the reference level (e.g., flavor: tobacco), adjusted for all other displayed product attributes. For each displayed pair of moist snuff products, participants were asked about which they were more curious, and which was more dangerous, would give a greater buzz, and would be easier to use. Positive values indicate characteristics (relative to reference) that independently contributed to greater curiosity or perceived danger, buzz, and ease of use, respectively. Abbreviation: CI = Confidence interval