Association of E-cigarette Advertising with E-cigarette and Cigarette Use Among U.S. Adults

Abstract

Aims

To estimate the association of e-cigarette advertisement exposure with e-cigarette and cigarette use behavior among U.S. adults.

Design

Data from the 2013–14 National Adult Tobacco Survey (NATS) were linked to Kantar Media and National Consumer Study data to construct measures of e-cigarette advertisements on TV and in magazines. The relationship between advertisement measures and outcomes was estimated using logistic and Poisson regressions, controlling for sociodemographics, state cigarette taxes, and state and year fixed effects.

Setting

United States.

Participants/cases

A total of 98,746 adults ages ≥18 years who responded to the 2013–14 NATS.

Measurements

The independent variables of interest were the number of e-cigarette advertisements in magazines to which an adult was exposed in the past 6 months and the number of e-cigarette advertisements in TV to which an adult was exposed in the past 6 months. Outcomes were awareness of e-cigarettes, ever e-cigarette use, current e-cigarette use, current cigarette use, and number of cigarettes smoked per month.

Findings:

Exposure to one additional e-cigarette advertisement on TV was associated with a 0.18 percentage point, 0.13 percentage point and 0.03 percentage point increase, respectively, in awareness, ever use, and current use of e-cigarettes among all adults (p<0.05). This exposure also was associated with a 0.11 percentage point increase in current cigarette use among all adults, and an increase in cigarette consumption of 2.24 cigarettes per month among adults ages ≥45 (p<0.05).

Conclusions:

E-cigarette advertising appears to be positively associated with the use of e-cigarettes and cigarettes among adults of all ages, and with increased cigarette consumption among older adults.

INTRODUCTION

Introduced into the U.S. market in 2007, demand for e-cigarettes has substantially increased every year, with U.S. sales reaching an estimated $7.0 billion in 2018 (Internet and vape shops sales are excluded).^1,2 In 2017, an estimated 6.9 million U.S. adults (2.8%) were current e-cigarette users.³ Concurrent with the increase in e-cigarette sales and use, there has been a substantial increase in advertising expenditures (television, print, radio, and Internet) from $5.6 million in 2010 to $125 million in 2014,⁴ with the vast majority of advertisement spending devoted to magazines (59%) and TV (27%), both with national reach.⁴

Although TV and radio advertising for cigarettes have been prohibited since 1971⁵ and for smokeless tobacco since 1986, advertising for other tobacco products, including e-cigarettes, remains largely unregulated. The U.S. Surgeon General concluded that e-cigarettes are marketed using a wide variety of media channels and approaches that have been used in the past for marketing conventional tobacco products to youth and young adults.⁶ It is well-established that exposure to cigarette advertising can induce product initiation among never users, discourage quit attempts in current users, and encourage relapse in those trying to quit.⁶ However, little is known about the relationship between exposure to e-cigarette advertising and measures related to the use of both e-cigarettes and cigarettes.

To date, a few studies have found that self-reported e-cigarette advertising exposure is positively associated with e-cigarette use among youth,^7–9 which is of concern because youth use of nicotine can cause addiction and can harm brain development⁶ and because e-cigarette use has been linked to an outbreak of lung injury.¹⁰ Studies focusing on adult smokers examined association of e-cigarette advertising with desire to use e-cigarettes and intentions to smoke regular cigarettes.^11–12 Some experimental and self-report longitudinal cohort studies found that greater interest in e-cigarettes was associated with exposure to e-cigarette advertising that suggested e-cigarettes were healthier or cheaper than regular cigarettes, help smokers quit, or depict active use of e-cigarettes.¹³ Moreover, exposure to e-cigarette advertisements was associated with increased desire to use both e-cigarettes and regular cigarettes,^14–15 and that former smokers had reduced intentions to abstain.^15–16 In contrast, a study using longitudinal survey data on self-reported e-cigarette advertising exposure found no association with e-cigarette use, disapproval of smoking, or quitting cigarettes.¹⁷

However, less is known about the impact of exposure to e-cigarette advertisements on population-level e-cigarette use and conventional cigarette use and consumption among adults.^18–19 Existing studies have only assessed effects on the smoking cessation.¹⁹ As such, the present study seeks to add to the limited evidence on whether e-cigarette advertising on television and in magazines encourages adult smokers to use e-cigarettes, and how this relates to use of conventional cigarettes. Using repeated cross-sectional data for adults ages ≥18 years from the National Adult Tobacco Survey (NATS), this study aims to estimate:

1. The strength of association of e-cigarette advertisements on TV and in magazines with e-cigarette awareness, ever use, and current use among all adults.

2. Whether the strength of this association differs across age groups (by interacting age with advertisement exposure).

3. The strength of association between e-cigarette advertising exposure and cigarette use and consumption among all adults.

4. Whether the strength of this association differs across age groups (by interacting age with advertisement exposure).

Previous studies found that tobacco use varies considerably across age groups.³ Given that the vast majority (90%) have initiated smoking prior to age 18, and virtually no one has initiated after age 26,²⁰ we expected to find differential effects of advertising exposure across age groups due to potential differences in the respondent’s nicotine dependency and the intensity of attachment to cigarette smoking.²¹ Understanding how e-cigarette use and smoking behaviors among adults of all ages are associated with e-cigarette advertising can inform evidence-based tobacco control policy and planning.

METHODS

Data

Data came from the 2012–2013 and 2013–2014 waves of NATS. NATS is a stratified, national, annual, landline, and cell phone survey of non-institutionalized adults ages ≥18 years residing in the 50 states or D.C.^22–23 Response rates were 47.2% for landlines and 36.3% for cell phones in 2012–2013, and 47.6% for landlines and 17.1% for cell phones in 2013–2014.^22–23 We obtained data on TV watching and magazine reading habits from the Simmons National Consumer Study (NCS, http://www.simmonssurvey.com ) and data on advertising placements for e-cigarettes in these media from Kantar Media (https://www.kantarmedia.com/us), for 2012 through 2014. The NCS is a national survey of American consumers ages ≥18 years. It provides comprehensive information on media habits, products, services, and in-depth consumer demographic and lifestyle characteristics. Kantar Media provides information on the date and time that e-cigarette advertisements aired, what brand of e-cigarette is advertised, and on what channel and during what program the advertisement is aired. Kantar Media also provides the issue and date of each specific magazine that carried e-cigarette advertisements.

Measures

Measurement of Advertising

We constructed exogenous measures of exposure to e-cigarette advertising in magazines and TV instead of relying on self-reported data, which might be subject to endogeneity or recall bias. Endogeneity could result from current or former users, or individuals with a higher unobserved propensity to use cigarettes or e-cigarettes, being more likely to recall such advertisements, leading to overestimation of the effect of advertising on use.²⁴ Bypassing these sources of potential bias, we applied the method used in Dave et al. (2018)¹⁹ to construct a measure of exposure to e-cigarette advertisements for each person in the NCS, by matching advertisement placement data from Kantar to respondents in the NCS based on the programs that they watched and magazines that they read.

The method applied by Dave et al (2018) is summarized as follows. A respondent in the NCS was considered to have been potentially exposed to an e-cigarette advertisement if she reported watching a program, channel, and time slot where an electronic cigarette advertisement aired in the past six months, based on advertisement placement data from Kantar. The respondent also reported the frequency of watching the program. Hence, total TV advertisement exposure for each person in the NCS was computed as a weighted sum of all advertisements to which a respondent was exposed over the past 6 months over all shows that they reported watching, with weights assigned depending on the frequency of viewing a program. For instance, with respect to a daily show that they watch, respondents reported whether they view it once a week, twice a week, three times a week, four times a week, or five times a week; this logically translates into weights of 0.2, 0.4, 0.6, 0.8, or 1, respectively, each capturing the likelihood that the respondent watched a given episode of that show on the channel and time slot during the week. For weekly shows, weights of 0.25, 0.5, 0.75 or 1 capture whether the respondent watches it (on a specific channel and time slot) once a month, twice a month, thrice a month, or four times a month. These viewing patterns pertain to the past week (for instance, for daily shows) or the past month (for instance, for weekly shows or other cable programs). In constructing advertisement exposure measures over the past 6 months, viewing patterns in the past week or past month were assumed to be representative of those in the past six months. Magazine advertisement exposure was measured as the weighted sum of the number of advertisements that appeared in all magazines in the past six months that the respondent has read, weighted by the frequency with which the respondent reads each magazine. Specifically, the weights for reading a magazine are 0.1, 0.25, 0.5, 0.75, and 1, reflecting whether the respondent read one out of the past four, two out the past four, three out of the past four, or all four out of the past four issues of the magazine, respectively.

Recognizing that advertising may have cumulative or persistent effects on behaviors, the analyses were based on advertisement exposure over the past 6 months. The main advertising measures therefore captured the number of e-cigarette advertisements in magazines to which an adult was exposed in the past 6 months and the number of e-cigarette advertisements on TV to which an adult was exposed in the past 6 months. Advertisement exposure over a 6-month recall period assumes that exposure does not depreciate over time until 6 months after exposure when it depreciates fully. This assumption is supported in various reviews of the consumer advertising literature.^24–25

As the NCS are nationally representative, we aggregated advertisement exposure from the NCS to the state/year/quarter/demographic group (age group, gender, race/ethnicity, education) level. These collapsed measures of advertisement exposure then were linked to NATS based on the same characteristics, yielding more detailed individual-specific measures of advertisement exposure for each person in NATS than those aggregated just at the state level. Intuitively, these measures of advertising capitalize on the fact that certain individuals and groups may be more exposed to e-cigarette advertisements based on their media consumption patterns, because they resided in states and watched specific TV programs during certain time slots, as well as read specific magazines that carried more of these advertisements. These measures of advertising are plausibly exogenous to the respondent. The variation in advertisement exposure across individuals is driven only by variation in advertising across areas and groups due to differences in media consumption patterns and the placement of advertisements across programs and magazines consumed by the respondents.

Our measure of advertising exposure parallels Target Ratings Points (TRPs)²⁶ but also offers some distinctions that are particularly well-suited to this research question and matching with NATS. Similar to TRPs, our measure captures both the frequency (number of advertisements aired) as well as the reach (fraction of the demographic segment who “views” an advertisement, or the probability that a given individual in this segment has viewed the advertisement). We measure exposure from the micro-level NCS, and are thus able to define narrow segments (by state/year/quarter/age groups/gender/race/ethnicity/education). These sociodemographic cells maximize the variation in advertising exposure due to differences in TV viewing and magazine reading habits and also maximize the salience of the advertising exposure measure. By matching advertising exposure within these finely defined cells exploiting time, geography, and detailed socio-demographics, we also minimize targeting bias, endogeneity, and measurement error.

Outcomes Measures

E-cigarette-related outcomes included awareness, ever use, and current use of e-cigarettes. Awareness of e-cigarettes was defined as an affirmative response to the survey question “Before today, had you ever heard of electronic cigarettes or e-cigarettes?” Ever use was defined as an affirmative response to the previous question and to this question: “Have you ever used an electronic cigarette, even just one time in your entire life?” Current e-cigarette use was defined as affirmative responses to the previous two questions and answering “every day” or “some days” to this question: “Do you now use electronic cigarettes every day, some days, rarely, or not at all?”

Cigarette-related outcomes included current cigarette use and number of cigarettes smoked per month. Current cigarette use was defined as smoking at least 100 cigarettes in lifetime and answering “every day” or “some days” to this question: “Do you now smoke cigarettes every day, some days, or not at all?”. Number of cigarettes smoked per month was computed for daily smokers by multiplying the number of cigarettes smoked per day “On average, about how many cigarettes do you now smoke each day?” times 30. Among respondents who smoked somedays, the number of cigarettes consumed per month was computed by multiplying the number of days smoked in the past 30 days “On how many of the past 30 days did you smoke cigarettes?” times the number of cigarettes smoked per day “On the day that you smoked, how many cigarettes did you smoke?”

Covariates

Covariates included age (18–24, 25–44, ≥45 years), sex (male, female), race/ethnicity (non-Hispanic whites, non-Hispanic blacks, others), annual household income (<$30,000, $30,000-$49,999, $50,000-$99,999, ≥$100,000, unknown), marital status (married, unmarried, unknown), education (high school or less, some college, college or higher), and state cigarette excise taxes. The latter was obtained from the CDC State Tobacco Activities Tracking and Evaluation (STATE) website.²⁷

Analysis

Association of e-cigarette advertisement with e-cigarette awareness, ever e-cigarette use, current e-cigarette use, and current cigarette smoking status were assessed using multivariable logistic regression. Association between e-cigarette advertisement and number of cigarettes smoked per month, conditional on current smoking status, was assessed using robust multivariable Poisson regression²⁸ (similar results were obtained using Negative Binomial regression). The independent variables of interest were the number of e-cigarette advertisements in magazines to which an adult was exposed in the past 6 months and the number of e-cigarette advertisements in TV to which an adult was exposed in the past 6 months. All models controlled for age, sex, race/ethnicity, annual household income, marital status, education, and interaction terms between age and e-cigarette advertisements, allowing the marginal effect of exposure to both TV and magazine advertisements to vary by age group.

We also controlled for state cigarette excise taxes as well as fixed effects for years (to capture unobserved national trends) and for states (to capture unobserved time-invariant state-specific heterogeneity). Standard errors were clustered on states to adjust for any arbitrary correlation in the error term across individuals and over time within states, and also adjust for any overdispersion in the Poisson model. Analyses were weighted to yield national estimates.

Means and 95% confidence intervals of all outcomes, advertisement measures, and covariates were reported overall and by age groups (Table 1). Statistical differences between subgroups were tested using standard chi-square tests. Regression outputs showing coefficients and confidence intervals of all independent variables (including age interactions) are shown in Table 2. Additionally, average marginal effects (in percentage points) of advertisement exposure from the multivariable regressions were reported overall and at each age group (Table 3). The marginal effects represent the change in tobacco use resulting from exposure to one additional e-cigarette advertisement, calculated using the regression coefficients of advertisement measures, age, and interaction between advertisement measures and age. Using marginal effects allows us to place the magnitude of the effects in context relative to the population size (by dividing marginal effects by prevalence of the outcome measures in Table 1 and then multiplying by 100). In addition to the adjusted marginal effects shown in Table 3, unadjusted marginal effects were also reported in Table A1 in the Supplementary File. P-values <.05 were used to determine statistical significance.

Table 1.

Mean E-cigarette Advertisement Exposure, Overall and by Sociodemographic Variables among US Adults, 2012–2014

		% (95% CI)
	Sample Size	Whole sample	Age 18–24 years	Age 25–44years	Age ≥45 years
Outcome Measures
Heard of e-cigarettesa	98,746	86.0 (85.7, 86.3)	91.4 (90.4, 92.5)	88.4 (87.8, 89.0)	83.6 (83.2, 83.9)
Ever e-cigarette useb	98,724	15.8 (15.5, 16.1)	31.6 (29.9, 33.3)	21.5 (20.8, 22.2)	9.6 (9.3, 9.9)
Current e-cigarette usec	98,709	2.6 (2.5, 2.7)	4.1 (3.4, 4.8)	3.4 (3.1, 3.7)	1.8 (1.7, 2.0)
Current cigarette smokingd	98,503	17.0 (16.6, 17.3)	17.7 (16.3, 19.1)	20.9 (20.1, 21.6)	14.4 (14.1, 14.8)
No. cigarettes smoked/monthe	12,361	403.4 (394.1, 412.7)	304.4 (266.9, 341.9)	374.7 (358.6, 390.9)	449.5 (439.3, 459.7)
E-cigarette Ad Measures f
No. of e-cigarette ads in magazines	98,882	3.0 (2.9,3.0)	2.7 (2.6, 2.9)	3.3 (3.2, 3.4)	2.8 (2.8, 2.8)
No. of e-cigarette ads on TV	98,882	1.1 (1.1, 1.1)	1.2 (1.0, 1.4)	1.2 (1.2, 1.3)	1.0 (1.0, 1.1)
Covariates
Sex
Male	41,088	47.1 (46.7, 47.6)	52.0 (50.1, 53.8)	47.8 (46.9, 48.6)	45.9 (45.5, 46.4)
Female	57,794	52.9 (52.4, 53.3)	48.0 (46.2, 49.9)	52.2 (51.4, 53.1)	54.1 (53.6, 54.5)
Age (years)
18–24	4,214	9.6 (9.2, 9.9)	100.0 (100.0, 100.0)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)
25–44	21,922	34.4 (34.0, 34.8)	0.0 (0.0, 0.0)	100.0 (100.0, 100.0)	0.0 (0.0, 0.0)
≥45	72,746	56.0 (55.6, 56.5)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)	100.0 (100.0, 100.0)
Race/Ethnicity
White, Non-Hispanic	81,668	71.5 (71.1, 72)	59.2 (57.4, 61.0)	63.3 (62.5, 64.2)	78.7 (78.2, 79.2)
Black, Non-Hispanic	5,989	8.6 (8.3, 8.8)	8.3 (7.3, 9.3)	8.9 (8.4, 9.4)	8.4 (8.1, 8.7)
Others	11,225	19.9 (19.5, 20.3)	32.5 (30.7, 34.3)	27.8 (26.9, 28.6)	12.9 (12.5, 13.3)
Education
≤High school	30,099	41.5 (41.1, 42)	57.7 (56.0, 59.5)	35.2 (34.3, 36.1)	42.6 (42.1, 43.1)
Some college	27,656	30.5 (30.1, 30.9)	33.1 (31.4, 34.8)	31.1 (30.4, 31.9)	29.6 (29.2, 30.1)
College or higher	41,127	28.0 (27.7, 28.3)	9.2 (8.4, 10.0)	33.7 (33.0, 34.4)	27.8 (27.4, 28.1)
Annual Household Income
<$30,000	16,185	17.3 (17.0, 17.7)	17.2 (15.8, 18.5)	16.3 (15.6, 17.0)	18.0 (17.6, 18.4)
$30,000-$49,999	17,074	18.2 (17.8, 18.5)	19.5 (18.1, 21.0)	19.0 (18.3, 19.7)	17.4 (17.1, 17.8)
$50,000-$99,999	26,209	26.1 (25.7, 26.4)	22.0 (20.4, 23.5)	28.3 (27.6, 29.1)	25.4 (25.0, 25.8)
≥$100,000	19,896	18.8 (18.5, 19.1)	12.3 (11.1, 13.5)	20.5 (19.9, 21.1)	18.8 (18.5, 19.2)
Unknown	19,518	19.7 (19.3, 20.0)	29.0 (27.3, 30.7)	15.9 (15.2, 16.5)	20.4 (20.0, 20.8)
Marital Status
Married	56,411	59.1 (58.7, 59.6)	19.9 (18.5, 21.3)	64.7 (63.9, 65.5)	62.4 (61.9, 62.9)
None married	41,968	40.5 (40.0, 40.9)	79.7 (78.2, 81.1)	35.0 (34.2, 35.8)	37.1 (36.7, 37.6)
Unknown	503	0.4 (0.4, 0.5)	0.4 (0.2, 0.6)	0.3 (0.2, 0.4)	0.5 (0.4, 0.5)
State cigarette excise taxes ($)	98,882	1.59 (1.58, 1.60)	1.62 (1.58, 1.66)	1.58 (1.56, 1.60)	1.59 (1.58, 1.60)

^aHeard of e-cigarettes was dichotomized as respondents who reported hearing of e-cigarettes versus otherwise.

^bEver e-cigarette use was dichotomized as respondents who reported hearing of e-cigarettes and ever using e-cigarettes (even just one time in their entire life) versus otherwise.

^cCurrent e-cigarette use was dichotomized as respondents who reported hearing of and ever using e-cigarettes and were currently using e-cigarettes “everyday” or “somedays versus otherwise.

^dCurrent cigarette smoking was dichotomized as respondents who had smoked at least 100 cigarettes in their entire life and were currently smoking cigarettes “everyday” or “somedays versus otherwise.

^eNumber of cigarettes smoked/month was computed for respondents who smoked cigarettes everyday as the number of cigarettes smoked per day times 30, and among respondents who smoked somedays, as the number of days smoked in the past 30 days times the number of cigarettes smoked per day.

^fNumber of e-cigarette ads to which an adult was exposed in the past 6 months.

Table 2.

Association of Exposure to E-cigarette Advertisements with E-cigarette and Cigarette-related Outcomes among US Adults, 2012–2014 (Regression Coefficients)^g

Dependent variables	Heard of e-cigarettesa	Ever e-cigarette useb	Current e-cigarette usec	Current cigarette smokingd	No. cigarettes smoked/monthe
Main independent variables
- No. of e-cigarette ads in magazinesf	0.03 (0.01, 0.05)	0.01 (−0.00, 0.01)	−0.02 (−0.04, −0.01)	0.02 (0.01, 0.02)	−0.00 (−0.00, 0.00)
- No. of e-cigarette ads on TVf	0.01 (0.00, 0.03)	0.02 (0.01, 0.03)	0.02 (0.01, 0.04)	0.02 (0.01, 0.03)	0.00 (0.00, 0.01)
Age interactions with magazine ads
- Age (18–24) # No. of ads in magazines	−0.04 (−0.08, −0.01)	−0.01 (−0.03, −0.00)	−0.01 (−0.07, 0.05)	−0.03 (−0.06, −0.01)	0.00 (−0.01, 0.01)
- Age (25–44) # No. of ads in magazines	−0.03 (−0.06, 0.00)	0.00 (−0.01, 0.01)	0.04 (0.02, 0.06)	−0.02 (−0.02, −0.01)	0.00 (−0.01, 0.01)
- Age (≥45) # No. of ads in magazines	ref	ref	ref	ref	ref
Age interactions with TV ads
- Age (18–24) # No. of ads in TV	0.03 (−0.03, 0.10)	0.00 (−0.02, 0.01)	−0.02 (−0.05, 0.01)	0.00 (−0.02, 0.01)	0.00 (−0.01, 0.01)
- Age (25–44) # No. of ads in TV	0.00 (−0.03, 0.03)	−0.02 (−0.03, −0.01)	−0.02 (−0.04, 0.01)	−0.02 (−0.03, −0.01)	−0.01 (−0.02, −0.00)
- Age (≥45) # No. of ads in TV	ref	ref	ref	ref	ref
Covariates
Sex
Male	ref	ref	ref	ref	ref
Female	−0.55 (−0.6, −0.51)	−0.34 (−0.42, −0.26)	−0.28 (−0.44, −0.12)	−0.39 (−0.47, −0.32)	−0.23 (−0.26, −0.19)
Age (years)
18–24	1.26 (1.11, 1.42)	1.39 (1.29, 1.49)	0.73 (0.47, 1.00)	0.08 (−0.08, 0.24)	−0.43 (−0.52, −0.35)
25–44	0.68 (0.56, 0.80)	1.11 (1.04, 1.17)	0.64 (0.52, 0.76)	0.68 (0.59, 0.77)	−0.18 (−0.25, −0.12)
≥45	ref	ref	ref	ref	ref
Race/Ethnicity
White, Non-Hispanic	ref	ref	ref	ref	ref
Black, Non-Hispanic	−0.65 (−0.81, −0.49)	−0.65 (−0.88, −0.41)	−0.99 (−1.4, −0.57)	−0.12 (−0.31, 0.06)	−0.46 (−0.53, −0.39)
Others	−1.14 (−1.23, −1.06)	−0.48 (−0.59, −0.36)	−0.50 (−0.83, −0.16)	−0.34 (−0.5, −0.19)	−0.27 (−0.33, −0.21)
Education
≤High school	ref	ref	ref	ref	ref
Some college	0.70 (0.58, 0.82)	−0.01 (−0.09, 0.07)	0.09 (−0.08, 0.27)	−0.25 (−0.3, −0.20)	−0.07 (−0.10, −0.03)
College or higher	0.41 (0.31, 0.52)	−0.91 (−1.00, −0.83)	−1.11 (−1.23, −0.98)	−1.36 (−1.42, −1.29)	−0.27 (−0.32, −0.21)
Annual Household Income
<$30,000	ref	ref	ref	ref	ref
$30,000-$49,999	0.30 (0.21, 0.39)	−0.09 (−0.21, 0.03)	−0.01 (−0.2, 0.17)	−0.22 (−0.32, −0.12)	0.00 (−0.05, 0.06)
$50,000-$99,999	0.60 (0.44, 0.77)	−0.18 (−0.32, −0.05)	0.01 (−0.21, 0.23)	−0.45 (−0.54, −0.36)	−0.05 (−0.09, −0.01)
≥$100,000	0.69 (0.49, 0.90)	−0.43 (−0.66, −0.21)	−0.21 (−0.42, −0.00)	−0.77 (−0.93, −0.60)	−0.12 (−0.18, −0.05)
Unknown	−0.09 (−0.24, 0.07)	−0.36 (−0.5, −0.21)	−0.24 (−0.4, −0.09)	−0.41 (−0.53, −0.30)	−0.03 (−0.08, 0.01)
Marital Status
Married	ref	ref	ref	ref	ref
None married	0.03 (−0.10, 0.17)	0.28 (0.21, 0.34)	0.25 (0.14, 0.36)	0.34 (0.27, 0.40)	0.02 (−0.02, 0.06)
Unknown	−0.19 (−0.45, 0.07)	0.34 (−0.05, 0.73)	−0.77 (−1.76, 0.23)	0.22 (−0.14, 0.57)	0.11 (−0.08, 0.31)
Year Dummies
2012	ref	ref	ref	ref	ref
2013	0.47 (0.37, 0.56)	0.18 (0.1.00, 0.26)	0.54 (0.24, 0.84)	−0.12 (−0.19, −0.05)	−0.01 (−0.06, 0.03)
2014	1.40 (1.31, 1.49)	0.52 (0.41, 0.64)	1.02 (0.68, 1.36)	−0.11 (−0.20, −0.01)	0.01 (−0.05, 0.06)
State cigarette excise taxes ($)	0.07 (−0.03, 0.17)	−0.02 (−0.27, 0.23)	0.42 (−0.6, 1.43)	−0.06 (−0.31, 0.19)	0.07 (−0.01, 0.15)
State fixed effects	Yes	Yes	Yes	Yes	Yes
Sample size

Notes: Boldface indicates statistical significance (p<0.05).

^aHeard of e-cigarettes was dichotomized as respondents who reported hearing of e-cigarettes versus otherwise.

^bEver e-cigarette use was dichotomized as respondents who reported hearing of e-cigarettes and ever using e-cigarettes (even just one time in their entire life) versus otherwise.

^cCurrent e-cigarette use was dichotomized as respondents who reported hearing of and ever using e-cigarettes and were currently using e-cigarettes “everyday” or “somedays versus otherwise.

^dCurrent cigarette smoking was dichotomized as respondents who had smoked at least 100 cigarettes in their entire life and were currently smoking cigarettes “everyday” or “somedays versus otherwise.

^eNumber of cigarettes smoked/month was computed for respondents who smoked cigarettes everyday as the number of cigarettes smoked per day times 30, and among respondents who smoked somedays, as the number of days smoked in the past 30 days times the number of cigarettes smoked per day.

^fNumber of e-cigarette ads to which an adult was exposed in the past 6 months.

^gRegression coefficients for e-cigarette awareness (heard of e-cigarettes), ever e-cigarette use, current e-cigarette use, and current cigarette smoking were estimated using multivariable logistic regression. Regression coefficients for the number of cigarettes smoked per month were estimated using robust multivariable Poisson regression. Standard errors were clustered on states. Analyses were weighted to be nationally representative.

Table 3.

Association of Exposure to E-cigarette Advertisements^a with E-cigarette and Cigarette-related Outcomes among US Adults, 2012–2014 (Marginal Effects)

Outcomes	Sample Size	Marginal Effectg of E-cigarette Ads on TV (95% CI)	Marginal Effect of E-cigarette Ads in Magazines (95% CI)
Overall
Heard of e-cigarettesb	98,746	0.18 (0.04, 0.32)	0.19 (0.03, 0.36)
Ever e-cigarette usec	98,724	0.13 (0.05, 0.22)	0.06 (−0.01, 0.12)
Current e-cigarette used	98,709	0.03 (0.01, 0.06)	−0.02 (−0.05, 0.02)
Current cigarette smokinge	98,503	0.11 (0.02, 0.20)	0.07 (0.00, 0.15)
No. cigarettes smoked/monthf	12,361	0.82 (−0.39, 2.02)	−0.27 (−1.13, 0.58)
Age 18–24 years
Heard of e-cigarettes	98,746	0.28 (−0.08, 0.65)	−0.08 (−0.25, 0.09)
Ever e-cigarette use	98,724	0.30 (0.13, 0.48)	−0.16 (−0.37, 0.04)
Current e-cigarette use	98,709	0.02 (−0.07, 0.11)	−0.12 (−0.30, 0.06)
Current cigarette smoking	98,503	0.16 (0.01, 0.32)	−0.22 (−0.47, 0.03)
No. cigarettes smoked/month	12,361	0.69 (−2.16, 3.54)	−1.27 (−3.84, 1.30)
Age 25–44 years
Heard of e-cigarettes	98,746	0.13 (−0.10, 0.36)	0.00 (−0.24, 0.25)
Ever e-cigarette use	98,724	0.03 (−0.13, 0.19)	0.15 (−0.01, 0.30)
Current e-cigarette use	98,709	0.02 (−0.03, 0.08)	0.06 (0.00, 0.11)
Current cigarette smoking	98,503	−0.05 (−0.21, 0.12)	0.00 (−0.14, 0.14)
No. cigarettes smoked/month	12,361	−2.32 (−4.84, 0.20)	−0.10 (−2.12, 1.91)
Age≥45 years
Heard of e-cigarettes	98,746	0.18 (0.01, 0.36)	0.37 (0.16, 0.58)
Ever e-cigarette use	98,724	0.16 (0.07, 0.25)	0.04 (−0.02, 0.11)
Current e-cigarette use	98,709	0.04 (0.02, 0.06)	−0.04 (−0.07, −0.01)
Current cigarette smoking	98,503	0.19 (0.08, 0.30)	0.18 (0.11, 0.24)
No. cigarettes smoked/month	12,361	2.24 (0.61, 3.87)	−0.25 (−1.35, 0.85)

Notes: Boldface indicates statistical significance (p<0.05).

^aNumber of e-cigarette ads to which an adult was exposed in the past 6 months.

^bHeard of e-cigarettes was dichotomized as respondents who reported hearing of e-cigarettes versus otherwise.

^cEver e-cigarette use was dichotomized as respondents who reported hearing of e-cigarettes and ever using e-cigarettes (even just one time in their entire life) versus otherwise.

^dCurrent e-cigarette use was dichotomized as respondents who reported hearing of and ever using e-cigarettes and were currently using e-cigarettes “everyday” or “somedays versus otherwise.

^eCurrent cigarette smoking was dichotomized as respondents who had smoked at least 100 cigarettes in their entire life and were currently smoking cigarettes “everyday” or “somedays versus otherwise.

^fNumber of cigarettes smoked/month was computed for respondents who smoked cigarettes everyday as the number of cigarettes smoked per day times 30, and among respondents who smoked somedays, as the number of days smoked in the past 30 days times the number of cigarettes smoked per day.

^gMarginal effects were calculated in percentage points using the regression coefficients of advertisement measures, age, and interaction between advertisement measures and age (from Table 2). All models controlled for age, sex, race/ethnicity, annual household income, marital status, education, state’s cigarette tax rates, interaction terms between age and e-cigarette ads, and state and year fixed effects. The first panel, shows average marginal effects overall. The next three panels show marginal effects at age group. Standard errors were clustered on states. Analyses were weighted to be nationally representative.

RESULTS

Prevalence of E-Cigarette Advertising Exposure and Tobacco Product Use

During 2012–2014, 48.0% and 85.8% of US adults were exposed to e-cigarette advertisements on TV and in magazines, respectively (data not shown). Mean exposure to e-cigarette advertisements increased from 2012 to 2014, from 0.4 to 1.8 for advertisements on TV, and from 0.8 to 4.1 for advertisements in magazines. Adults ages 25–44 years were exposed to more of these advertisements compared to other age groups.

During 2012–2014, 86.0% of US adults reported having heard of e-cigarettes; 15.8% reported ever using e-cigarettes, 2.6% reported currently using e-cigarettes, and 17.0% reported current cigarette use (Table 1). On average, each current smoker smoked 403.4 cigarettes per month. Variations in tobacco use were observed by age groups. E-cigarette use was more pronounced among young adults ages 18–24 years compared to older adults ≥45 years (P<0.05). Furthermore, adults ages ≥45 years were less likely to smoke cigarettes but they smoked more cigarettes on average compared to younger adults (P<0.05).

Association between E-Cigarette Advertising Exposure and E-Cigarette Awareness and Use

Exposure to e-cigarette advertisements on TV was associated with increased awareness and use of e-cigarettes among all adults. Specifically, exposure to one additional e-cigarette advertisement on TV was associated with increases of 0.18 percentage points in awareness of e-cigarettes (0.2% increase relative to the mean), 0.13 percentage points in ever use of e-cigarettes (0.8% relative increase), and 0.03 percentage points in current e-cigarette use (1.2% relative increase) (Table 3). We found no evidence that the association of TV advertisements with awareness or current use of e-cigarettes varied across age groups; however, a significant variation was observed with respect to ever-e-cigarette use (Table 2). Specifically, TV advertisement exposure had a greater association with ever e-cigarette use among adults ages 18–24 years (0.30 percentage points; 0.95% relative increase) and adults ages ≥45 years (0.16 percentage points; 1.67% relative increase). These effects are statistically different from zero (Table 3); however, there was no evidence of a significant difference between them (Table 2). We found no evidence of association between TV advertisement and ever-e-cigarette use among adults ages 25–44 years.

Furthermore, exposure to e-cigarette advertisement in magazines was associated with some e-cigarette-related behaviors, though these effects were substantially weaker for others. Specifically, exposure to one e-cigarette advertisement in magazines was associated with a 0.19 percentage point increase in overall e-cigarette awareness (0.22% relative increase); however, there was no evidence of association with ever or current e-cigarette use. These effects were larger among adults ages ≥45 years. Specifically, exposure to one e-cigarette advertisement in magazines was associated with a 0.37 percentage point increase in overall e-cigarette awareness (0.44% relative increase) and with a 0.04 percentage point decrease in current e-cigarette use (2.22% relative decrease) among adults ages ≥45 years.

Association between E-Cigarette Advertising Exposure and Cigarette-Related Outcomes

Exposure to one additional e-cigarette advertisement on TV was associated with a 0.11 percentage point increase in current cigarette use (0.65% relative increase) among all adults. This association was smaller (with no evidence of being statistically significant) among adults ages 25–44 years. However, one additional TV advertisement was associated with 0.16 (0.90% relative increase) and 0.19 (1.32% relative increase) percentage point increases in cigarette use among adults ages 18–24 years and adults ages ≥45 years, respectively. We found no evidence of a significant difference between these estimates (Table 2). Furthermore, we found no evidence of an association between TV advertisement and total number of cigarettes smoked per month among all adults. However, among adults ages ≥45 years, exposure to one TV e-cigarette advertisement was associated with a 2.24 cigarette increase in monthly consumption (0.50% relative increase).

Exposure to one additional e-cigarette advertisement in magazines was associated with a 0.07 percentage point increase in cigarette use (0.41% relative increase) among all adults. Similarly, this association was larger among adults ages ≥45 years (0.18 percentage points; 1.25% relative increase). We found no evidence of overall or subgroup association between e-cigarette advertisement in magazines and the total number of cigarettes smoked per month.

DISCUSSION

The findings suggest that greater exposure to e-cigarette advertising is associated with greater ever and current use of e-cigarettes among U.S. adults. There is presently limited evidence on e-cigarette advertisements and e-cigarette use, with the few prior studies having focused largely on youth and relying on self-reported advertising exposure information.^7–9 The current study adds to this limited evidence base by focusing on adults, using detailed information on actual advertisement placement in magazines and TV to yield plausibly exogenous measures of advertising exposure.

These results are consistent with prior studies that used self-reported measures of advertising exposure, ^7–9 suggesting that recall and reporting bias resulting from use of self-reported measures is likely minimal. Our findings suggest similar patterns for adults as those reported for youth. Specifically, we found that increased exposure to e-cigarette advertisements on TV was associated with increased awareness and use of e-cigarettes among all adults.

Furthermore, among older adults ages ≥45 years, increased exposure to e-cigarette advertisements on TV was associated not only with increased e-cigarette use, but also with higher proportions of current cigarette smokers and increased consumption among smokers. To place these results in context, the average adult is exposed to about one e-cigarette advertisement on television and about three advertisements in magazines over our sample period. Increasing TV advertisement exposure by one additional advertisement leads to a 0.16 and 0.19 percentage point increases in ever use of e-cigarettes and current use of cigarettes, respectively, among adults ages ≥45 years. Applied to the US adult population ages ≥45 years in 2014, these estimates translate into an increase of about 208,000 e-cigarette users and 247,000 cigarette users.

Our findings regarding the relationship between e-cigarette advertising exposure and conventional cigarette use are concerning given the serious health consequences of smoking even a small number of cigarettes per day.²⁹ This is consistent with prior research finding that e-cigarette advertising can trigger cravings to smoke.³⁰ The increase in cigarette consumption among older adults may reflect a higher nicotine dependency, potentially leading some adults to use e-cigarettes to deliver nicotine in situations and places where they may not be able to smoke. Furthermore, the vast majority of adult e-cigarette users are current or former smokers.³¹ Our results suggest that at least in the short term, exposure to e-cigarette advertisement on TV may influence dual use of e-cigarettes and cigarettes. Future research should use longitudinal data to investigate how e-cigarette advertising exposure impacts the various detailed margins of e-cigarette use and cigarette use dynamically.

E-cigarettes are implicitly promoted among adults as smoking cessation aids,³² and many individuals who use e-cigarettes believe they will help them quit smoking cigarettes.³¹ However, current evidence is insufficient to recommend e-cigarettes for smoking cessation. ^33–39 Long-term follow-up data could help assess whether adult smokers, exposed to more advertisements, who are using more e-cigarettes and cigarettes, continue this pattern of use (contemporaneous complements) or if they transition at some point to lower consumption of cigarettes or successfully quitting smoking (dynamic substitutes). Furthermore, future research might help explore differential effects of exposure to different types of e-cigarette advertisements (those that tend to glamorize it vs. those that make health or cessation aid claims).

E-cigarette advertising is currently subject to limited restrictions in the U.S. Although several groups support extending regulations for conventional cigarettes to e-cigarettes, potential legal barriers exist to this approach, including commercial speech rights of e-cigarette companies.⁶ The Food and Drug Administration (FDA) now requires that e-cigarette labels bear the following warning label statement: “WARNING: This product contains nicotine. Nicotine is an addictive chemical.”⁴⁰Additionally, federal agencies have exercised authority to prohibit e-cigarette companies from marketing to children, especially in light of the recent outbreak of lung injury associated with use of e-cigarettes.⁴¹ The findings of this study provide evidence to inform potential regulatory actions on e-cigarette advertising, and whether to place stricter evidentiary standards on promoting e-cigarettes as a cessation aid.

This study has some limitations. First, NATS was last conducted in 2013–2014, and so may not reflect more recent trends. Of particular interest are the increased e-cigarette advertisements on social media associated with the introduction of JUUL in 2015.⁴² However, exploring the association between e-cigarette advertising with e-cigarette and cigarette use before increased awareness of e-cigarettes on social media might help reduce confounding factors in this relationship. Second, exposure to e-cigarette advertisements for respondents in NATS was predicted from the information in the NCS and Kantar based on similarities in sociodemographic characteristics, state, year, and quarters. These advertisement measures might be subject to measurement errors if exposure to advertisements varied by other unobserved characteristics that also predict tobacco use. Finally, response rates from NATS were lower than those from comparable surveys, such as the 2014 National Health Interview Survey (73.8%).⁴³ However, tobacco prevalence estimates from NATS align with other national surveys.^44–45

CONCLUSIONS

E-cigarette advertisement on TV is associated with the use of e-cigarettes, including among young adults. TV advertisement exposure is also linked to increased cigarette use and consumption. Further research on the factors that influence this association could help inform regulatory strategies to maximize potential benefits of these products and minimize risks.