Brand equity was assessed using the truth Brand Equity scale. The detailed measures in this scale were originally adapted from Aaker [21
] and construction of the scale has been described elsewhere [22
]. The scale includes items related to the perceptions of the truth brand within four constructs: brand loyalty, leadership/popularity, brand personality, and brand awareness. These four constructs were determined based on: (1) a quantitative assessment of audience perceptions of truth advertising messages; (2) a content analysis of truth initiative social media posts on Twitter, Instagram, and Facebook in which the campaign was actively pushing out messages for three 2-week periods of time in September 2014, January 2015, and March 2015; and (3) previous research that indicated these four constructs formed a higher-order brand equity factor (the full Brand Equity scale) that mediated the effects of exposure to the original truth campaign on smoking outcomes [22
]. Items used to measure the brand loyalty construct included: (1) I’d like to help truth end smoking in my generation; (2) I’d defend truth on social media if someone were putting it down; (3) I’d follow truth on social media; and (4) I would be part of a movement to end smoking. Leadership/popularity items included: (1) truth is helping my generation end smoking; and (2) truth is for people like me. Brand personality was assessed with the following items: (1) If truth was a person, truth would be...(a) inspired, (b) powerful, (c) in control of their own decisions, (d) independent, (e) honest, and (f) innovative; (2) People that follow truth are just like me; and (3) People that follow truth are like the friends I hang out with. Finally, brand awareness was measured with the item: When you think of truth, you think, (1) fewer and fewer young people today smoke cigarettes; (2) tobacco companies lie; (3) the tobacco industry tried to get young people to smoke other products like hookah; and (4) tobacco company ads are a joke. All items were listed as statements with a 5-point agreement scale, which was coded so that −2 = strongly disagree, −1 = disagree, 0 = neither, 1 = agree and 2 = strongly agree. Weighted responses for each item are included in Table 1
. Mean response scores were then calculated for each construct. Based on prior confirmatory factor analysis results, all items loaded satisfactorily to one of the brand equity constructs, and each construct demonstrated adequate model fit [22
Smoking Status. The outcome of interest in this study was self-reported past 30-day cigarette use at wave 5. This outcome was defined using the following item: “During the past 30 days, on how many days did you smoke cigarettes (even 1 or 2 puffs)”? Those who reported smoking on one or more days were categorized as current smokers, and this variable was modeled as categorical for this analysis (1 = current smoker, 0 = never smoked or had smoked, but not in the past 30 days).
Intention to Quit Smoking. An additional outcome of interest was intention to quit smoking at wave 5. This variable was defined using the following items: “During the past 30 days, on how many days did you smoke cigarettes (even 1 or 2 puffs)”? and “Do you want to completely stop smoking cigarettes”? Only participants who self-reported smoking on one or more days in the past 30 days saw the second item. Those who answered “yes” to the second item were defined as intending to quit smoking cigarettes (coded as 1). Those who answered “no” were categorized as not intending to quit smoking (coded as 0).
Control variables were identified based on existing literature on predictors of tobacco use or predictors of campaign exposure among youth and young adults. Detailed descriptions of each control variable are included in Table 2
and Appendix A
. As not all control variables included in the model were measured at wave 3, measures of these variables at participants’ baseline assessment were included.
Demographics and psychographics. Demographics included age (continuous), gender (male, female), race/ethnicity (White, Black, Hispanic, other), and parent education (higher than high school, HS/GED or less). Psychographics included participants’ media use (low/medium/high), school achievement (much better/better than average, average/below average), and sensation seeking.
Several covariates were included to measures cigarette and other tobacco/substance use, participants’ immediate environment, and perceptions of smoking. A Smoking Intensity measure, adapted from prior work examining progression of youth smoking [36
], was used at baseline to account for differences between those who may have smoked only a few times and those who are more regular or daily smokers. The measure included five levels of smoking progression: (1) Closed to smoking; (2) Open to smoking; (3) Low-intensity non-daily smoking; (4) High-intensity non-daily smoking; and (5) Daily smoking. The first two levels classified those who reported no past 30-day cigarette use. The remaining three levels classified participants who reported past 30-day smoking, and was based on a combination of number of days smoked and the number of cigarettes per smoking day (CPD). Other health-related variables, including participants’ past 30-day use of other tobacco products, household e-cigarette and/or combustible use (no one I live with smokes, someone I live with smokes cigarettes/cigars and/or e-cigarettes), and peer cigarette use.
An 18-item multi-dimensional anti-tobacco scale (ATS) was developed to reflect the five attitudinal constructs that exemplify the themes of the truth campaign: anti-smoking imagery, disapproval of social smoking, support for an anti-tobacco social movement, anti-tobacco industry sentiment, and independence. All attitude items were measured before assessing any campaign awareness at baseline. For each of the 18 items, participants were asked to what degree they agree or disagree with each statement (1 = strongly disagree, 2 = disagree, 3 = neither, 4 = agree, and 5 = strongly agree). Scales were created for each of the five attitudinal constructs, and an average score across the five attitudinal scales was calculated. The ATS scored within the excellent range on reliability (α = 0.90) [29
Ad Awareness. Awareness of truth advertisements was assessed by calculating an index of the self-reported cumulative number of individual ads seen, as well as the estimated frequency of seeing each ad from baseline to wave 3. Awareness of other national anti-tobacco campaigns, such as the Food and Drug Administration’s (FDA) Real Cost Campaign and the Centers for Disease Control and Prevention’s (CDC) Tips from Former Smokers campaign, was assessed at baseline.
Policy. DMA-level smoking prevalence and tobacco-related policy variables, including county-level smoke free air laws, state-level tobacco control expenditures, and state-level cigarette tax, were included based on participants’ residential address at baseline.