Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Evaluating social networks and impact of micro-influencers who promote e-cigarettes on social media
(USC Thesis Other)
Evaluating social networks and impact of micro-influencers who promote e-cigarettes on social media
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Evaluating social networks and impact of micro-influencers who promote e-cigarettes on social
media
Julia Vassey
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(HEALTH BEHAVIOR RESEARCH)
August 2024
Copyright, 2024 Julia Vassey
ii
TABLE OF CONTENTS
List of Tables………………………………………………………………………………….iii
List of Figures………………………………………………………………………………...iv
Abstract …………………………………………………………………………………….....vi
Introduction …………………………………………………………………………………...1.
Chapter 1: Study 1. The impact of healthy lifestyle imagery alongside e-cigarettes
in social media influencer marketing on perceptions of e-cigarettes among adolescents:
a survey-based experiment……………………………………………………………………...14.
Chapter 2: Study 2. The impact of social media influencer marketing on perceptions of
e-cigarettes among adolescents: a survey-based experiment……………………………………38.
Chapter 3: Study 3. Worldwide connections of micro-influencers who promote
e-cigarettes on Instagram and TikTok: a social network analysis………………………………65.
Discussion………………………………………………………………………….................94.
References …………………………………………………………………………………..101.
iii
List of Tables
Study 1
Table 1. Experiment outcome measures...................................................................................29.
Table 2. Participant characteristics for adolescents (N = 664).................................................30.
Table 3. Multivariable analyses to assess the effects of healthy lifestyle activities
performed by micro-influencers in e-cigarette promotional videos on harm
perceptions and susceptibility to e-cigarette use among adolescents………………………...31.
Supplementary table 1. Influencer credibility perception scores per each influencer
featured in the stimuli in the treatment and control condition………………………………..37.
Study 2
Table 1. Experiment outcome measures....................................................................................55.
Table 2. Participant characteristics for adolescents (N = 3,221)...............................................56.
Table 3. Multivariable analyses to assess the effects of healthy lifestyle activities
performed by micro-influencers in e-cigarette promotional images
on harm perceptions and susceptibility to e-cigarette use among
adolescents………………………………………………….....................................................57.
Supplementary table 1. Influencer credibility perception scores per each influencer
featured in the stimuli in each experimental
condition………………………………………………………................................................64.
Study 3
Table 1. Individual node centrality measures and network measures of directed
networks of 104 influencers and 55,622 commenters on Instagram and 100
influencers and 68,673 commenters on TikTok………………………………………………85.
Supplementary table 1. Normalized centrality measures and network measures
assessed on the directed network of 104 influencers on Instagram and 100
influencers on TikTok…………………………………………………………………………91.
iv
List of Figures
Introduction
Figure 1. Connection of the Aims………………………………………………………..........6.
Figure 2. A young-looking micro-influencer (perceived age: < 21)……………………….....7.
Figure 3. A micro-influencer who promotes e-cigarettes on Instagram while
working out.………………………………………………………………………....................7.
Figure 4. A micro-influencer who promotes a flavored disposable e-cigarette
device while streaming a video game on Twitch……………………………………………...7.
Figure 5. A micro-influencer promoting a disposable e-cigarette device and
a disposable cannabis Delta-8 THC device……………………………………………………8.
Study 1
Figure 1. Experimental conditions………………………………………………………………33.
Figure 1. Interaction plots……..………………………………………………………………...34.
Supplementary figure 1. Profile of a Randomized Clinical Trial………………………….........36.
Study 2
Figure 1. Experimental conditions………………………………………………………………59.
Figure 2a. Interaction plots…………………………………………………...............................60.
Figure 2b. Interaction plots……………………………………………………...........................61.
Supplementary figure 1. Influencer perception score by experimental condition………………62.
Supplementary figure 2. Profile of a Randomized Clinical Trial.................................................63.
Study 3
Figure 1. The structure of the networks of influencers and commenters to their posts
on Instagram and TikTok………………………………………………………………………..87.
Figure 2. Weighted unipartite networks of Instagram micro-influencers (n=104)
and commenters to their posts (n=55,622) and TikTok micro-influencers (n=100)
and commenters to their posts (n=68,673)……………………… ……………………………...88.
Figure 3. Weighted unipartite networks of Instagram (n=104) and TikTok (n=100)
micro-influencers who commented on each other’s posts…………………………………........89.
v
Figure 4. Instagram communities by region. …………………………………………………...90.
Supplementary Figure 1. Distribution of the most popular content Instagram (N=104)
and TikTok (N=100) micro-influencers posted about in 2021-2022 along with
e- cigarettes……………………...................................................................................................94.
vi
ABSTRACT
Background:
Adolescent exposure to e-cigarette marketing on social media is a risk factor associated with
the use or intention to use e-cigarettes among youth. This project explored insufficiently
regulated marketing of micro-influencers, their network connections on Instagram and
TikTok, and the effects of content the micro-influencers post about on youths’ perceptions of
e-cigarettes.
Methods:
We used two randomized survey-based experiments: online (N=664 adolescents, Mean age =
15) and in classroom (N = 3,221 adolescents, Mean age = 17) to assess the effects of ecigarette-related content influencers post about (ie., e-cigarettes alongside different lifestyle
contexts versus e-cigarettes alone) on perceptions of influencer credibility, perceived harm and
appeal of, and susceptibility to use, e-cigarettes among California adolescents. We constructed
directed unipartite networks among Instagram (N = 104) and TikTok (N = 100) microinfluencers and users on Instagram (N = 55,622) and TikTok (N = 68,673) who commented on
influencers’ posts (including micro-influencers who commented on each other’s posts). We
calculated individual node measures: in/out degree and betweenness centrality, and network
measures: density, reciprocity, transitivity, and assortativity for these networks.
Results:
Among participants who perceived influencers as credible, adolescents who saw posts of microinfluencers promoting e-cigarettes alongside healthy lifestyle (Study 1) or alongside marijuana
(Study 2) reported higher appeal, lower harm perceptions of, or susceptibility to use, e-cigarettes
(compared to controls). Micro-influencers who promoted e-cigarettes alongside lifestyle content
vii
had higher betweenness compared to the micro-influencers who promoted e-cigarettes alone
(Study 3).
Conclusions:
The findings emphasize the need for strengthening influencer marketing regulation.
1
INTRODUCTION
Micro-influencers: a new form of e-cigarette advertising
E-cigarette use among adolescents remains a concern. Despite an over 15% decline in ecigarette use among high school students in 2023, over 2 million (8%) of high school and
middle school students in the United States (U.S.) reported past 30-day use of e-cigarette
products1,2 Among current users, 25% reported using e-cigarettes frequently. E-cigarette use
is harmful to adolescents’ brain development, can lead to addiction, and are linked to depression
and other mental health problems.3–5 Numerous risk factors for e-cigarette use were described
in scientific literature, including personal and sociodemographic characteristics (e.g., age, sex,
race, ethnicity); substance use history; social environment (e-cigarette use by friends and
family members, parental educational level, and family economic status). 3–8 An ongoing
assessment of the risk factors that may be contributing to e-cigarette uptake among adolescents
is important for protection of their health.
Exposure to visual posts featuring e-cigarette products on social media, including marketing
e-cigarette-related content, has also been identified as one of the risk factors associated with
e-cigarette use,
9,10 intentions to use e-cigarettes, more positive attitudes, greater norm
perceptions,
11 and lower perceived risks12 of e-cigarette use among U.S. adolescents. Despite
existing restriction policies on advertising tobacco products (e.g., prohibition of paid ads or
sponsored content),
13 social media platforms popular among adolescents (e.g., TikTok,
Instagram)14 host promotional e-cigarette-related content. Only a fraction of this content gets
removed by the platforms’ algorithm, indicating that better enforcement of the existing
restrictions in the community guidelines (e.g., tobacco-sponsored promotional content posted
by influencers is not allowed on Instagram and TikTok; tobacco-related content depicted by
young-looking individuals is not allowed on TikTok) is necessary.15,16
2
U.S. federal authorities took actions against youth-oriented direct advertising of e-cigarette
products by large e-cigarette brands. For example, e-cigarette manufacturer JUUL used young
models and cartoon characters in its social media advertising in 2015-2018, which was considered
by federal legislation an intentional adolescent targeting that contributed to soaring levels of
vaping among adolescents.17–19 In response to the US Food and Drug Administration (FDA)
warnings in 2018 about selling and marketing of e-cigarettes to kids, JUUL (along with several
other major e-cigarette brands) voluntarily suspended their social media youth-oriented
marketing.20 However, it did not eradicate e-cigarette marketing on social media that could reach
youth. E-cigarette brands started utilizing indirect marketing, i.e., using influencers to promote
tobacco products (when these influencers post e-cigarette content from their own social media
accounts). Over the last five to 10 years, such marketing has been growing on social media
platforms (e.g., Instagram, TikTok) popular among Generation Z (i.e., those 11 to 26 years of age
in 2024).21–25 Influencer content is often perceived by consumers as more authentic and relatable
compared to direct advertising from brands or endorses (e.g., brands’ spokespeople) used in
traditional advertizing.
26–28 This phenomenon could be explained by the Attribution theories,
e.g., Correspondent Inference Theory.
29 It posits that consumers could be more strongly
persuaded by a spokesperson (e.g., an influencer who promotes e-cigarette product) if
consumers believe that this spokesperson is intrinsically (genuinely) motivated to promote this
product (as opposed to extrinsic motivation to promote a product solely for compensation
without being genuinely interest in the product, which decreases influencers’ perceived
credibility). While influencers do get sponsored by brands, they are often chosen based on a
certain domain of expertise (e.g., e-cigarette brands may partner with the influencers who have
3
been posting engaging content about e-cigarettes on social media and who are knowledgeable
and interested in e-cigarette products).30
Social media analysts categorize influencers (who also often identify themselves on social
media as models, bloggers, or brand ambassadors21–23) by the number of followers as celebrity or
macro-influencers (~100,000 to 1 million+ followers), micro-influencers (~1,000 to ~100,000
influencers) or nano-influencers (~1,000 or fewer followers).31 While depiction of e-cigarettes
and other tobacco products by celebrity influencers or macro-influencers in social media has been
documented in scientific literature and generated media attention news, such instances are not
ubiquitous.32,33 Hiring celebrities or experts like macro-influencers could be cost-prohibitive for
tobacco brands. Showing e-cigarettes could also create public backlash and be damaging for
celebrities’ reputations. Lately, tobacco brands have increased their reliance on influencers with
smaller audiences, especially micro- or nano-influencers, to promote their products or services.22,26
Micro-influencers (unlike celebrities) are typically non-famous general social media users
known among their niche followers and knowledgeable on specific topics they post about.28
Micro-influencers (versus celebrity influencers) have the added benefit of being cost-effective for
brands. As such, brands, including e-cigarette companies, often offer paid partnerships to microinfluencers to market e-cigarette products.
34 Market research showed that over 50% of surveyed
Gen Z respondents (13-26 year of age) considered social media influencing a reputable career
choice.35 The respondents indicated that they would be willing to quit their day job if influencer
job would pay enough for their lifestyle.35 Taken together, not only micro-influencer marketing
has become a new form of adverting favored by brands and having potential to be engaging
for youth, but it also provides an opportunity for youth to create their own content and
distribute it among their niche followers. In other words, becoming a micro- or a nano-
4
influencer is not a pipe dream (versus becoming a celebrity), but a real possibility for anyone.
Financial compensation and opportunities to create potentially engaging content on social media
can motivate young social media content creators to collaborate with a variety of brands, including
tobacco brands (or other substance brands like marijuana), resulting in the promotion of harmful
content on social media.
State of science and knowledge gap around micro-influencer marketing
Influencer marketing and its effect on youths’ perceptions of e-cigarettes has not been
extensively studied. Prior research assessed compliance of e-cigarette-related social media
posts featuring micro-influencers with federal marketing regulations: i.e., the presence of
nicotine warning labels, tobacco brand sponsorship disclosures, and access restriction to these
posts for youth. Vassey et al22 showed that influencers’ e-cigarette-related Instagram posts
frequently have non-compliant or not fully compliant disclosures of partnerships with tobacco
brands (when influencers disclose their partnership in some e-cigarette-related posts, but not
in all e-cigarette-related posts) and/or lack the required nicotine warning labels and youths’
age restrictions, in violation of the U.S. Food and Drug Administration (FDA) and the Federal
Trade Commission (FTC) marketing restrictions.36,37 Prior research also assessed the effects
of the warning labels or sponsorship disclosures in images posted by or featuring non-celebrity
influencers on e-cigarette perceptions among youth.38,39 Vogel et al investigated the effects of
exposure to sponsorship disclosures in e-cigarette influencer marketing images on Instagram
on young adult harm perceptions of e-cigarettes.39 The study found that viewing and
recognizing clear sponsorship disclosures (#sponsored or #paidad) resulted in significantly lower
intentions to engage with the Instagram post. Klein et al evaluated the effect of labeling
strategies, i.e., “ad” or “sponsored” in e-cigarette-related influencer images using eye-tracking
5
technology.38 The study found that the overall visual attention to any labeled post was lower
compared to unlabeled posts. Results were in line with the Correspondent Inference Theory
reflecting a trend toward overall skepticism of e-cigarette influencer posts that are perceived as
advertisements.39 Vassey et al 2022 survey-based experiment among high school students in
California showed that participants who viewed influencer e-cigarette-related Instagram posts
with a tobacco-free warning label were at greater odds of susceptibility to e-cigarette use
compared to adolescents who viewed posts with the FDA nicotine warning label.40 Vassey at
el and Vogel et al also found that adolescents and young adults perceived influencers as
credible: (trustworthy, knowledgeable, and honest) if they viewed e-cigarette-related promotional
posts with the FDA nicotine warning label.40,41 The results indicate the need for increasing
enforcement of the FDA nicotine warning labels in e-cigarette influencer marketing on social
media. Vassey et al also characterized a network of micro-influencers on Instagram where microinfluencers from around the world shared partnerships with over 600 e-cigarettes brands to
promote their products.22
In this dissertation project, we used experimental and social network studies to further
characterize micro-influencers and their marketing tactics. The randomized survey-based
experiments (Study 1 and Study 2) allowed us to assess the effects of micro-influencer
marketing tactics on adolescent perceptions of e-cigarettes these micro-influencers promote on
Instagram and TikTok. The social network analysis (Study 3) helped us understand the level
of interaction among micro-influencers, identify central micro-influencers, and estimate
tendencies of connection or tie formation among micro-influencers on Instagram and TikTok.
(Connection of the aims are shown in Figure 5). Each study is described in detail further in
6
subsequent sections.
Effects of micro-influencer marketing on adolescents’ perceptions of e-cigarettes
Beyond inconsistent use of required warning labels and tobacco sponsorship
disclosures, influencers use a variety of marketing tactics to generate higher attention to their
posts. Research has shown that influencers who promote e-cigarettes on social media often
depict them alongside youth-oriented, or lifestyle, posts featuring fashion, video games, nature,
and healthy activities (e.g., exercising), to make the content more appealing to social media
users.
22,42–45 The Social Learning Theory states that people imitate behavior performed by a
role model (e.g., social media influencer) if they perceive these role models like themselves,
attractive, likable, happy, or popular.
The observation of role models becomes particularly important during adolescence when
young individuals strive for independence and actively seek new sources of influence within their
social and media environments.46 Hence if influencers promote harmful content such as e-
7
cigarettes in youth-appealing context, there is a risk that youth might try to imitate the
influencers’ e-cigarette use behavior. Marketing tactics or features used by influencers to
increase their allure to audiences have been captured by the Content Appealing to Youth (CAY)
index, a measure of media content found to be appealing to youth developed in prior research.45
The CAY index captures over 40 content features, categorized under six broad dimensions:
production value, character appeals, youth-oriented themes, product appeals, rewarding appeals,
and miscellaneous content.45
We observed the following CAY content elements in e-cigarette-related microinfluencer marketing on Instagram and TikTok: character appeal e.g., posts featuring younglooking influencers, (Figure 1); theme appeal, e.g., use of generally positively perceived
lifestyle context (e.g., healthy lifestyle) alongside e-cigarette
marketing (Figure 2); production value, e.g., the use of bright
colors on e-cigarette devices (Figure 3). Such marketing could
engage youth cognitively, increase appeal and diminish
adolescents’ harm
perceptions of ecigarettes,
contribute to normalization of e-cigarette use, and
expose e-cigarette never-users to e-cigarette
content. We also observed micro-influencers
promoting e-cigarettes next to marijuana or hemp products (including psychoactive Delta-8 THC,
Figure 4), which could contribute to dual product use among youth (Vassey et al., Tobacco
8
Control blog, in print). The effects of e-cigarette
promotion by micro-influencers in these contexts on
perceptions and use patterns of e-cigarettes among youth
has not been studied. This project addressed the literature
gap.
We assessed the effects of specific marketing
features utilized in e-cigarette-related microinfluencer marketing in Instagram and TikTok posts,
i.e., videos (Study 1, Aim 1) and images (Study 2, Aim 2) on perceptions of influencer
credibility, appeal, harm perceptions of, and susceptibility to use, e-cigarettes among
California adolescents using randomized survey-based experiments.47 We stated the following
hypotheses for Study 1 and Study 2:
• Adolescents who viewed posts of micro-influencers promoting e-cigarettes
alongside lifestyle activities (e.g., exercising, modeling, playing video games) or
alongside marijuana products would be more likely to have lower harm perceptions
and higher perceived appeal of e-cigarettes and report susceptibility e-cigarette use
compared to adolescents who viewed posts of micro-influencers promoting ecigarettes alone (without any lifestyle or marijuana use contexts).
• These effects would also be observed if adolescents had positive perceptions of the
micro-influencers, i.e., perceive them as credible (honest, trustworthy, informed).
The hypotheses fit the Prototype Willingness Model (PWM), which posits that
adolescents are willing to engage in a health risky behavior under certain social circumstances,
e.g., while comparing themselves with the prototype performing a certain risky behavior.
10,48,49
9
The PWM might help explain which circumstances can motivate adolescents to get involved into
risky health behavior such as vaping even if adolescent vaping perception overall is negative.50
The PWM encompasses two types of decision making (i.e., dual processes) related to health
behavior: one involves analytic processing and a reasoned path, and the other one is a social
reaction path that explains the unintended and unplanned decision-making, specifically behavior
that can put adolescent health at risk. The latter is often a reaction to risk-conducive social
situations. This social reaction path incorporates two constructs: risk prototypes, which are images
of people who engage in a risky behavior (e.g., vaping) and behavioral willingness – openness to
engaging in risky behavior. The more favorable adolescents’ perceptions of the prototype images,
the more open they are to engage in a risky behavior and accept social consequences associate
with such behavior. Short form media (images or short video clips) and mostly entertaining content
that influencers use to portray e-cigarette on Instagram and TikTok likely activate the social
reaction path (versus a reasoned path) of imagery processing. If adolescents have favorable
perception of influencers who present e-cigarette use in a positive lifestyle context (e.g.,
exercising), they may relate to these influencers, be more open to being seeing in social
circumstances around such people and be willing to tolerate or accept their health risky behavior.
Connections and influence of micro-influencers on social media
While the experimental studies in this project looked at the effects of exposure to e-cigarette
content with e-cigarette outcomes among adolescent populations, the social network framework
shifts the focus from studying individual traits to analyzing interactions, relationships, and
communications.51 Assessing network structures and positions of individual influencers in
these networks could help understand the level of interaction (e.g., via content engagement
such as providing comments to influencers posts) and e-cigarette-related information sharing
10
among influencers as well as among influencers and their audiences. These interactions could
be assessed using the Two-Step Flow of Communication Model.
52 However, social media have
changed the traditional way of interpreting this framework. In its traditional interpretation, the
Two-Step Flow of Communication Modelsuggests that the flow of information and influence from
the mass media to their audiences involves two steps: from the media to certain individuals (i.e.,
the opinion leaders) and from these opinion leaders to the public. In other words, opinion leaders
mediate information flow from mass media to audiences. Due to decline of the conventional
communication channels, such as TV, newspapers, or radio and the rise of the more fragmented
media marketplace (i.e., internet and social media), opinion leaders have become original sources
of information that is directly communicated to the public. Opinion leaders are no longer limited
to only celebrities (as it used to be in the past) but include a wider range of influencers of different
caliber (from mega- to nano-influencers). In relation to this project, influencers create their own
content on social media, engage with each other’s content and directly interact with audiences who
consume this content. This project allowed us to assess connections among different microinfluencers and their audiences in the Instagram and TikTok networks, (Study 3, Aim 3).
Research on tendencies of influencers who promote e-cigarettes to form ties with each other
on social media platforms have been limited. Vassey et al discovered a dense network of
international micro-influencers and over 600 e-cigarette brands the influencers partnered with
to promote e-cigarette products.22 The study found that different e-cigarette brands often partner
with the same micro-influencers from all over the world, primarily from the U.S., Indonesia,
Malaysia, Germany, France, and Italy. Since social media space lacks borders, it increases the
possibility of content sharing and tie formation among influencers from different regions, which
is problematic from the tobacco regulatory science standpoint. This phenomenon was described as
11
cross-border promotion in the World Health Organization (WHO) Framework Convention on
Tobacco Control.53 Tobacco Advertising, Promotion, and Sponsorship (TAPS) is cross-border
whenever the content created, uploaded, or broadcast in one country may be consumed or shared
in another, thereby crossing geographical borders. In other words, if influencers from different
countries engage (e.g., comment on) each other’s social media posts, they potentially expose their
followers (including adolescents) to broader e-cigarette promotional materials (posted by
influencers from different countries).
53
Study 3 of this dissertation project further explored the network of micro-influencers by
assessing ties among micro-influencers who engage with each other’s content and ties among
micro-influencers and other users who engage with micro-influencers’ content on two
platforms widely popular among adolescents: i.e., Instagram and TikTok.
• We explored and compared Instagram and TikTok networks’ structural
characteristics: i.e., density, transitivity, reciprocity, and homophily, and assessed
whether influencers’ positions in these networks differ across platforms (Instagram
versus TikTok), (exploratory hypotheses).
We identified the most central micro-influencers on these two platforms and assessed how
geographic regions the influencers were from affected influencers’ positions in these networks.
Since Vassey et al found that micro-influencers from Asia and U.S. collaborated with a higher
number of e-cigarette brands and had higher numbers of followers compared to European
influencers,22 we hypothesized that:
• Asian and North American (including U.S.-based) micro-influencers will be more
central and influential in the network compared to European micro-influencers
within Instagram and within TikTok.
12
We also assessed tendencies of micro-influencers to form ties based on geographic regions
and linguistic commonalities. Based on Waldo Tobler’s First Law of Geography, which states
that “everything is related to everything else, but near things are more related than distant things,”54
we hypothesized that:
• Regional homophily (i.e., tendency of influencers from the same geographic regions to
form ties among each other) will be strong and driven by linguistic commonalities,
but heterophily (connection with dissimilar nodes based on the geographic attribute,
i.e., ties among micro-influencers from different regions) will be present. That is
because prior research by Vassey et al discovered the interconnected network of ecigarette brands and international micro-influencers on Instagram.22
Vassey et al also showed that micro-influencers who promote e-cigarettes often collaborate
with multiple industries (e.g., fashion, beauty products, healthy lifestyle) besides e-cigarette
brands.22 As such, we hypothesized the following:
• Influencers who post about e-cigarettes and other content will be more central and
influential in the network compared to influencers who post exclusively about ecigarettes on Instagram and TikTok. We expected this pattern because microinfluencers who promote e-cigarettes and other content may be forming ties with a
larger number of influencers since they share a wider range of interests (e-cigarettes
along with other topics vs e-cigarettes only). Such pattern increases the risk that
influencers who post on a variety of topics besides e-cigarettes could reach users who
are not interested in tobacco-related content and expose these users to harmful imagery
of e-cigarettes.
13
The overall goal of this dissertation research was to inform tobacco control authorities about
the impact of micro-influencers who promote e-cigarettes on social media popular among
youth. Regulating specific influencer marketing features (e.g., prohibiting depiction of healthy
lifestyle context next to e-cigarette products) could be challenging because it could fall under
the umbrella of free speech and be protected by the First Amendment of the U.S. Constitution.55
(However, challenging does not mean impossible, since we demonstrated that such content
could make influencers and e-cigarettes look appealing to youth). In any case, the findings
emphasized the need for regulators - at the very least – to better enforce legislation of
influencer marketing that is already in place. This includes restricting access to influencer ecigarette marketing content to youth under tobacco-purchasing age (younger than 21 years of age)
and following the FDA/FTC guidelines for sponsorship disclosures and nicotine warning labels,
since these restrictions get violated on social media.22,56 To improve enforcement, for example,
influencers and e-cigarette brands could get penalized for marketing restriction violations.
Social media platforms could also collaborate with legislation bodies on more effectively
enforcing such policies. New legislation, e.g., requiring e-cigarette brands not to work with
young-looking influencers under a certain age could also be considered.57 The findings also
emphasized the need for collaboration among the international public health agencies to
enforce the above-mentioned legislation globally and to prevent youth exposure to the harmful
content on social media.53,57 Finally, the findings also suggest the need to develop countermarketing campaigns using micro-influencers to create effective anti-tobacco messages on
social media. Such steps are important for protection of adolescents’ health.
14
Chapter 1. Study 1
The impact of healthy lifestyle imagery alongside e-cigarettes in social media influencer
marketing on perceptions of e-cigarettes among adolescents: a survey-based experiment
Julia Vassey,1 Erin A. Vogel,2 Junhan Cho,1 Dayoung Bae,1 Jennifer B. Unger.
1
1
Department of Population and Public Health Sciences, University of Southern California, Los
Angeles, CA
2
TSET Health Promotion Research Center, University of Oklahoma Health Sciences Center,
Oklahoma City, OK
Declarations of Interest: All authors declare no competing interests.
Funding/Support: This study was funded by the Tobacco-Related Disease Research Program,
TRDRP, T33DT6620, the National Institute of Health, NIH, R01CA260459 and the NCI & FDA
Center for Tobacco Products (CTP) Award (NCI/FDA Grant #U54CA180905)
Role of Funder/sponsor: The funders had no role in study design; collection, analysis, and
interpretation of data; writing the report; and the decision to submit the report for publication. The
content is solely the responsibility of the authors and does not necessarily represent the official
views of the funders.
Data availability statement: Data will be made available upon request.
Running Head: E-cigarette promotion and adolescents’ perceptions of e-cigarettes.
Word Count: 3,345
Number of Tables: 3
Number of Figures: 2
Manuscript status: in preparation
15
INTRODUCTION
Promotional tobacco-related content posted by influencers on behalf of tobacco brands is
present on social media platforms (e.g., Instagram, TikTok) popular among Generation Z (i.e.,
those 11 to 26 years of age in 2024).21–24 These promotions exist despite the community guidelines
established by Instagram and TikTok, restricting paid advertisements or promotions of tobacco
products.13,58,59 This is troubling as research has shown that consumers perceive influencers as
trusted sources of information.
28 Moreover, market research showed that over 50% of surveyed
Gen Z respondents (13-26 year of age) considered social media influencing a reputable career
choice.35 As such, brands have increased their reliance on influencers, especially microinfluencers – users with about 1,000 – 100,000 followers – to promote their products or services.26
Micro-influencers (versus celebrity influencers) in particular have the added benefit of being costeffective for brands, and their content is often perceived by consumers as authentic and relatable.26–
28 Brands, including e-cigarette companies, often offer paid partnerships to micro-influencers to
market e-cigarette products.34 Financial compensation can motivate young social media content
creators to collaborate with tobacco brands, resulting in the promotion of harmful content on social
media.
Research has shown that influencers often promote e-cigarettes on social media alongside
youth-oriented, or lifestyle, posts featuring fashion, video games, nature, and healthy activities
(e.g., exercising), to make the content more appealing to social media users, which could contribute
to normalization of substance use.22,42–45 However, despite its growing presence on social media,
the effects of e-cigarette promotion by micro-influencers in these lifestyle contexts on perceptions
and use patterns of e-cigarettes among adolescents remain unknown.
16
This study assessed the impact of posts from social media influencers promoting ecigarettes alongside healthy lifestyle activities on harm perceptions of e-cigarettes, perceived
appeal of, and susceptibility to use, e-cigarettes among adolescents. The Prototype Willingness
Model suggests that people are more willing to engage in a risky health behavior (e.g., e-cigarette
use) when they have a positive view of the prototypical person (e.g., an influencer) who regularly
engages in the behavior.
48,60 As such, we hypothesized that adolescents who viewed videos of
influencers promoting e-cigarettes alongside healthy lifestyle activities (experimental condition)
would have lower harm perceptions of e-cigarettes, higher perceived appeal of, and susceptibility
to use, e-cigarettes, compared to adolescents who viewed videos of influencers promoting ecigarettes alone, without healthy lifestyle context (control condition), adjusting for perceptions of
influencer credibility (H1). We also hypothesized that adolescents who perceived influencers as
credible in the experimental condition, compared to those in the control condition, would have
lower harm perceptions of e-cigarettes, higher perceived appeal of, and susceptibility to use, ecigarettes (H2).
MATERIALS AND METHODS
Participants and Procedures
In January 2024, adolescents (13-17 years of age) living in California were recruited by
Qualtrics (a research panel agency that has been used in prior research to survey adolescents about
their substance use and other behaviors61–63) to complete an online randomized survey-based
experiment on tobacco-related attitudes and behaviors. The participants (N=664) completed a
15-minute online Qualtrics-programmed survey that consisted of text-based questions and 10-
second-long videos. Researchers obtained participants’ written informed assent and parental
consent prior to data collection (the participants access the survey via a URL link). Each
17
adolescent received compensation by the panel ($11) after providing a quality completion of
this survey. To ensure data quality, the Qualtrics team that administers the survey performed
fraud detection to check for duplicate IP addresses, bots straight-lining, speeding, gibberish,
or failure to answer quality verification (select the correct multiple-choice answer about the
content of the videos the participants watched) or age verification questions (provide the same
age at the beginning and the end of the survey). (Sample sizes of all eligible, included and
excluded participants are reported in Supplementary Figure 1). All low-quality responses were
replaced with the high-quality responses. The study was approved by the University of
Southern California Institutional Review Board (UP-21-00352).
Experimental stimuli development
The experimental videos were created from content (Instagram and TikTok posts) of the
10 influencers with public accounts identified in prior research22 who had disclosed that they were
sponsored by e-cigarette companies. These influencers were known to post e-cigarette content,
and, in some instances, post e-cigarette use alongside a healthy activity.
To create homogenous stimuli (e.g., to ensure that no influencer is perceived drastically
differently from the others by age and gender characteristics), 10 undergraduate students
reviewed the stimuli prior to the experiment, answering the questions about perceived gender
(“How likely that the person featured in this promotional image is Man/Woman/Transgender?”)
and perceived age of the influencers (“How likely that the person featured in this promotional
post is: a) younger than 21; b) 21-30; c) 31-40; d) older than 40?”). Posts with the highest
rating agreement and highest rating scores (8 or above) were selected for the experiments (five
male and five female profiles of the perceived age between “younger than 21” and “21-30”, out of
12 profiles).
18
Experimental design
Each participant viewed 10 randomly assigned posts (10-second-long videos featuring 10
different influencers) in two experimental conditions: 1) influencers promoting e-cigarettes
alongside a healthy activity (experimental group), or 2) influencers promoting e-cigarettes with
the neutral profiles of those influencers (control). To ensure that the content of the stimuli only
differed by the characteristics of interest, the first 5 seconds of each of the 10 videos featured
e-cigarette promotion by 10 different influencers and remained unchanged across the two
conditions. The conditions differed by the last five seconds of the videos (i.e., influencers
engaged in healthy lifestyle activities in the experimental group versus neutral profile of
influencers in the control group. (Figure 1 shows examples of the experimental conditions for
one influencer profile).
As part of the experiment, participants assessed perceptions of influencer credibility after
watching each video. Harm and appeal perceptions of, and susceptibility to use, e-cigarettes were
assessed after participants watched all the videos.
Measures of harm perceptions, appeal perceptions and susceptibility to e-cigarette use
Harm perceptions of e-cigarettes were measured using the validated two-item scale
(Table 1) adapted from the Population Assessment of Tobacco and Health (PATH).64
Responses to questions (referring to the e-cigarette products the participants saw in the videos):
Do you think using e-cigarettes for vaping nicotine is harmful to your health? and Do you think
people harm themselves when they use e-cigarettes for vaping nicotine? were measured on a 0
(not at all harmful) to 100 (very harmful).65 Since the items were closely related (α=0.87), they
were combined (by summing all the non-missing values of the items) into one variable. The
outcome variable was recoded to the 0-10 scale (since the distribution of this count variable
19
was a better fit for the Poisson model). To keep regression coefficients for all the outcomes in
the same direction, the harm perception variable was also reversed-coded with the higher score
representing lower harm perceptions of e-cigarettes.
Attitudes or perceptions of appeal of e-cigarettes were measured using the validated
three-item scale (Table 1)66 assessed on the 7-point semantic differential scale with the word
pairs anchored at each end: Using e-cigarette is: not cool/cool, unattractive/attractive,
boring/fun. The items were combined (by summing all the non-missing values of the items)
into one variable (α=0.93). The higher score represented higher appeal.
Susceptibility to e-cigarette use (among never-users of e-cigarettes) was measured
using the validated three-item scale (Table 1) adapted from the PATH study64 and combined
into one variable (α=0.93). Consistent with prior research,67 the measure was collapsed into
dichotomized responses related to interest in trying e-cigarettes: responding “definitely not”
to all items vs. responding “probably not,” “probably yes”, “definitely yes” to one or more
items.
Perceptions of influencer credibility
Perceptions of influencer credibility representing three personality traits (honesty,
trustworthiness, knowledge, Table 1) were assessed with a 0-100 scale. This scale has been
validated in prior research.
41,68 Scores for each of the 10 influencers the participants viewed
were measured per each trait (Supplementary Table 1). The three traits (honesty,
trustworthiness, knowledge) were then combined into one variable (α=0.87). The mean
“perceptions of influencer credibility” score (i.e., the mean of the non-missing scores for the
combined three traits) was then computed for each of the 10 videos that each participant watched.
Statistical analysis
20
To test H1, we examined the main effects of experimental condition (i.e., exposure to
videos of influencers promoting e-cigarettes alongside healthy lifestyle activities) on harm and
appeal perceptions of e-cigarettes (assessed after all the videos), adjusting for perceptions of
influencer credibility (assessed after each of the 10 videos featuring influencers). We used the
Poisson generalized linear mixed effects model (since the harm and appeal perception
outcomes followed the Poisson distribution) and tested it among all the participants who
completed the experiment (N=644). The main effect of the experimental condition and perceptions
of influencer credibility on susceptibility to use e-cigarettes were tested using the binary logistic
mixed effects regression model among never-users of e-cigarettes (n=514). Repeated measures
in experimental conditions (i.e., perceptions of influencer credibility) were modeled as random
effects (random intercept and fixed slope). This approach was chosen as a result of research
suggesting that a substantial bias could exist when stimulus repetitions go unanalyzed as a random
effect.69 The models were adjusted for fixed effects of sociodemographic and other covariates
described below. The choice of covariates was guided by prior research since these covariates were
expected to be associated with the outcome variables.41 Eight participants in the experimental
group and seven participants in the control group did not provide responses to the influencer
perception question. Listwise deletion was used to exclude these observations from the analysis.
Interaction effects models
To test the potential moderating effects (H2) of perceptions of influencer credibility on
associations of experimental condition (i.e., exposure to videos of influencers promoting ecigarettes alongside healthy lifestyle activities) with harm and appeal perceptions of, and
susceptibility to use, e-cigarettes, the interaction term of the experimental condition X perceptions
of influencer credibility was added to the subsequent interaction effects models. To interpret the
21
significant interaction effects of H2, post hoc analyses were performed to assess the effects of
experimental condition on harm and appeal perceptions of e-cigarettes, stratified by the level of
influencer credibility perceptions (i.e., higher influencer credibility perception score [the
credibility perception score at or above the median value] vs. lower influencer credibility
perception score).
All multivariable models were adjusted for socio-demographic characteristics, substance
use in the past 30 days, frequency of social media use and exposure to influencer e-cigarette
content on social media in the past 30 days (Table 2). Adjustment for self-reported exposure to ecigarette content posted by influencers accounted for participants’ familiarity with this type of
content on social media. All the covariates were assessed before the exposure to the experimental
stimuli.
Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were reported with
statistical significance set at p < 0.05 (two-tailed) after applying Benjamini-Hochberg multiple
testing corrections to control for the false discovery rate at 0.05. All statistical analyses were
conducted using R software (version 4.2.2). (Sample sizes for all the experimental conditions are
reported in Supplementary Figure 1).
RESULTS
Adolescents who completed the experiment (N = 664) were Mage=15 [SD=1.4]; 53%
female, 39% Hispanic. 10% reported using e-cigarettes in the past 30 days, and 77% reported never
using e-cigarettes. 51% reported at least weekly exposure to e-cigarette-related content on
Instagram and/or TikTok and 27% reported past 30-day exposure to e-cigarette-related social
media content posted by influencers (Table 2).
Main effects models
22
Participants (N = 664) in the experimental condition had higher odds of lower harm
perceptions (AOR = 1.08; CI = 1.04 –1.12) and higher perceived appeal of e-cigarettes (AOR =
1.04; CI = 1.01 –1.07) than participants in the control condition. The main effect of the
experimental conditions on susceptibility to e-cigarette use assessed among never-users of ecigarettes (N=514) was not significant (AOR = 1.19; CI = 0.94 – 1.51, Table 3).
Interaction effects models
We found a significant interaction effect of the experimental condition X perceived
influencer credibility on harm perceptions and perceived appeal of e-cigarettes (Figure 2). Among
the participants (n=305) who perceived influencers as credible (at or above the median of the
influencer credibility perception score), 74% of those in the experimental condition (who viewed
videos of e-cigarette promotion alongside healthy lifestyle) vs 57% in the control condition (ecigarette promotion alone) had lower harm perceptions of e-cigarettes (p-value < .001). Similarly,
among the participants who perceived influencers as credible, 62% of those in the experimental
condition vs 51% in the control condition had higher perceived appeal of e-cigarettes (p-value <
.001).
The post hoc analyses of associations between experimental condition and each outcome,
stratified by the perception of influencer credibility score, were presented in Table 3. Among the
participants who perceived influencers as credible, those in the treatment condition had higher
odds of lower harm perceptions (AOR = 1.11; 95% CI = 1.05 – 1.16) and higher perceived appeal
of e-cigarettes (AOR = 1.07; 95% CI = 1.02 – 1.11) compared to those in the control condition.
Among participants who perceived influencers as non-credible (n=343), there was no significant
difference in harm perceptions (AOR = 1.11; 95% CI = 1.01 – 1.22) and perceived appeal (AOR
= 1.03; 95% CI = 0.96 – 1.11) of e-cigarettes between treatment and control conditions (Table 3).
23
DISCUSSION
An influencer marketing tactic such as depiction of e-cigarette promotion alongside healthy
lifestyle activities in Instagram or TikTok videos may contribute to decreased harm perceptions
and increased appeal of e-cigarettes among adolescents. This randomized survey-based experiment
showed that participants who viewed influencers promoting e-cigarettes paired with a healthy
lifestyle were more likely to report lower harm perceptions and higher perceived appeal of ecigarettes if they perceived influencers as credible. The findings are in line with theoretical
research, i.e., the Prototype Willingness Model48,60 suggesting that people are more willing to
engage in a health risky behavior to the extent that they have a positive view of the prototypical
person who performs that behavior. Such prototypical person could be a peer, a celebrity, or an
influencer. The findings are also in line with the results from a similar online experiment conducted
by Vassey et al. among California’s young adults recruited by the YouGov marketing research
panel (manuscript in review).70 The study showed that among young adults who perceived microinfluencers as credible, those who viewed videos of influencers promoting e-cigarettes paired with
a healthy lifestyle were more likely to report lower harm perceptions of e-cigarettes and
susceptibility to e-cigarette use. (The effect for susceptibility in this study was borderline
significant. It is possible that with the sample size of 514 never-users of e-cigarettes among whom
susceptibility to e-cigarette use was assessed we were slightly underpowered to find a significant
effect for this outcome).
Prior research has focused on the effects of self-reported, or experimentally assessed,
exposure to e-cigarette marketing, finding that youth exposed to such content were more likely to
report e-cigarette initiation, willingness, and intention to use e-cigarettes, more positive attitudes
toward e-cigarettes, and greater perceptions of e-cigarette use as normative.11,71 Few studies have
24
analyzed the effect of a specific source of marketing or advertising (e.g., direct marketing from
tobacco brands or indirect marketing from celebrities,32 or influencers)41 on youth perceptions and
use patterns of e-cigarettes. Influencers use a variety of marketing tactics to make their social
media posts more appealing. In the case of e-cigarettes, research has shown that influencers often
promote e-cigarettes alongside youth-oriented or lifestyle contexts such as fashion, gaming,
nature, and health-related activities (e.g., exercising) in an attempt to make this content more
appealing to social media users.22,42–45 Similar marketing tactics were used by tobacco brands in
the past to advertise combustible cigarettes next to sports-related imagery.72 While influencers may
advertise unhealthy products to pursue commercial interests, consumers of their content (including
youth) may perceive these influencers as trusted, credible, and authentic sources,46 especially when
influencers promote e-cigarettes as part of their daily lifestyle (e.g., working out, playing video
games). Adding to the existing scientific literature, this study demonstrated that these marketing
tactics could contribute to normalization of e-cigarette use (lower harm perceptions and higher
appeal of e-cigarettes), if micro-influencers are perceived as credible, but not if micro-influencers
are perceived as non-credible. Results support public health efforts to harness influencers for antitobacco messaging and in health literacy campaigns about harmful effects of e-cigarette use. Since
influencers can be perceived as trusted sources of information,26,27 public health authorities may
harness influencers for anti-tobacco messaging and in health literacy campaigns about the harmful
effects of e-cigarette use46,73 or disseminating other positive public health messages.46 Consistent
with PWM and this study’s findings, pro-e-cigarette influencers used in intervention campaigns
could be portrayed in a more negative light, while anti-e-cigarette influencers could be portrayed
in a more positive light.
25
With the proliferation of e-cigarette-related influencer-driven content on image and video-based
social media platforms popular among youth,14 more research is needed to analyze the effects of
different contexts used in e-cigarette marketing by influencers with different number of followers
on youth perceptions and use patterns of e-cigarettes.
Influencer e-cigarette-related marketing on social media is not sufficiently regulated by
platforms’ community guidelines or federal legislation. Although most widely-used and wellknown social media platforms (e.g., Facebook, Instagram, YouTube, Twitter, TikTok, Reddit,
Snapchat) prohibit paid advertisements for tobacco products,41,59 few (Facebook, Instagram and
TikTok) explicitly prohibit influencers from promoting nicotine/tobacco products.11 Moreover,
influencers still promote e-cigarettes on the platforms that have restrictions in place. According to
the U.S. Food and Drug Administration (FDA) and the Federal Trade Commission (FTC)
guidelines, if influencers have any material connection with a tobacco brand – meaning that they
have been paid or given something of value to tout the product - such relationships need to be
disclosed in their social media posts.74 However, many influencers’ posts have ambiguous or
absent disclosures of their partnerships with tobacco brands.22 Influencers who promote products
on behalf of tobacco brands are also required to restrict access to e-cigarette marketing content
to only individuals who are at or above the federal minimum age (i.e., 21) of sale of tobacco
products. However, research has demonstrated that micro-influencers have not been consistently
using youth age restrictions to their social media accounts.22 Enforcement of the existing
influencer marketing regulations needs improvement.
Regulating influencer marketing features (e.g., prohibiting depiction of healthy lifestyle
context next to e-cigarette products) could be challenging because specific framing of content
could be protected by the First Amendment of the U.S. Constitution.55 However, federal
26
agencies and social media users (including youth) should be informed about such marketing
tactics, especially if such content is youth-appealing. Regulating attempts could still be made
since we demonstrated that such content could make influencers and e-cigarettes look
appealing to youth. Authorities could also consider prohibiting e-cigarette brands from hiring
influencers under a certain age,56,75 ban influencer advertising of e-cigarettes altogether, or
penalize brands and influencers for non-compliance with the existing federal marketing
regulations such as restricting access to e-cigarette marketing content to youth or following the
FTC guidelines for sponsorship disclosures. Additionally, further improvement in the
enforcement of social media community guidelines is necessary.15,16
Limitations
Since the participants of the experiment were based in California, the results from the
studies may not be generalizable to all U.S. adolescents. Concerns about generalizability are
mitigated by the fact that ethnic diversity of the samples represent the projected future ethnic
diversity of the U.S.76 Despite the randomized experimental design, self-reported survey
outcomes might be prone to social desirability bias (which could not rule out a possibility that to
e.g., e-cigarette use history was underreported or higher perceptions of e-cigarettes was
overreported). The measures were also assessed only at one point in time (vs multiple times
overtime). Since Qualtrics online panels recruited adolescents’ parents (versus recruiting
adolescents directly) to be in compliant with the IRB protocol, we cannot be completely certain
that the adolescents completed the survey themselves as opposed to parents completing it on
their behalf. Concerns are mitigating by inclusion of the prompt at the start of the survey
emphasizing the importance of obtaining responses from the adolescents directly to ensure the
validity of data and by inclusion the verification check questions. (Participants were asked
27
about their age twice: at the beginning and at the end of the survey. Those who provided
different age in responses to these two questions, were excluded from the analytic sample).
The measures in this study were assessed at one point in time, reducing generalizability of the
findings. While the experimental stimuli were created from actual Instagram and TikTok posts
featuring micro-influencers promoting e-cigarette products, the survey did not use an actual
Instagram or TikTok interface to simulate the real-life social media user experience. While we
assessed intentions to use e-cigarettes (susceptibility) among adolescents, it was not possible to
assess the actual e-cigarette use behavior in these experimental settings: i.e., the association
between the experimental exposure and initiation of e-cigarette use.
Conclusions
The study demonstrated that an influencer marketing tactic such as e-cigarette promotion
alongside healthy lifestyle activities in Instagram or TikTok videos may contribute to lower harm
perceptions and higher appeal of e-cigarettes among adolescents, particularly among those who
perceived influencers as credible. Future research may assess the effects of broader context and
themes (e.g., marijuana use, fashion, gaming) often present alongside e-cigarette promotion in
influencer marketing. Future research should also continue providing surveillance data to show
changes in e-cigarette influencer marketing over time, and investigating if tobacco companies
use influencers to introduce new or promote existing e-cigarette products in violation of
premarket authorization requirements (PMTA). Our findings indicate the need for further
regulation of influencer marketing. Results also support public health efforts to harness
influencers for anti-tobacco messaging and in health literacy campaigns about harmful effects of
e-cigarette use.
28
Contributorship Statement Concept and design – JV. Acquisition, analysis, or interpretation of
data—all authors. Draft of the manuscript - JV. Critical revision of the manuscript for important
intellectual content—all authors. Statistical analysis – JV. Obtaining funding – JV, JBU.
Administrative, technical, or material support – JV, JBU. Supervision – JBU. All co-authors
approved the final version.
Ethics approval The University of Southern California Institutional Review Board (UP-21-
00352) approved all study procedures.
29
Table 1. Experiment outcome measures
Table 3. Outcome measures
The outcome assessed after participants viewed each unique influencer profile
Perceptions of influencer
credibility 41,77
Based on the video you just watched, how would
you rate that influencer on each of the items
below?
dishonest–honest
untrustworthy–trustworthy
uninformed–informed
0 (not at all) -100 (a
lot), with the word
pairs anchored at
each end of the 0-
100 scale
Outcomes assessed after participants viewed ALL the posts
Harm perceptions64 of ecigarettes
Based on the videos you just watched, do you think
using e-cigarettes for vaping nicotine is harmful to
your health?
Based on the videos you just watched, do you think
people harm themselves when they use e-cigarettes
for vaping nicotine?
0 (Not harmful at
all) – 100 (Very
harmful)
0 (Not at all) – 100
(A lot)
Attitudes66 or
Perceptions of appeal
Based on the videos you just watched, using ecigarette is:
- not cool/cool,
- unattractive/attractive,
- boring/fun
7-point semantic
differential scale
with the word pairs
anchored at each
end
Susceptibility to ecigarette use,64
(assessed only among
never-users of ecigarettes)
Based on the videos you just watched:
1) Would you try e-cigarettes for vaping nicotine if
one of your best friends offered them to you?
2) If you had an opportunity to use e-cigarettes for
vaping nicotine, would you use them?
3) Do you think you would use e-cigarettes for
vaping nicotine in the next 6 months?
4) Are you curious about using e-cigarettes for
vaping nicotine?
Definitely Not/
Probably Not/
Probably Yes/
Definitely Yes
30
Table 2. Participant characteristics for adolescents (N = 664)
Participant characteristics % (n) / M (SD)
Age 15 (1.4)
Sex assigned at birth
Female 53% (355)
Sexual orientation
Heterosexual 87% (576)
Race
Asian 13% (85)
White 47% (315)
Black 9% (59)
Othera 13% (88)
Multiple race 18% (117)
Ethnicity
Hispanic/Latinax 39% (259)
Substance use
Never-users of e-cigarettes 77% (514)
Past 30-day use of e-cigarettesb (%/n yes) 10% (68)
Past 30-day use of marijuanac (%/n yes) 12% (77)
Past 30-day use of cigarettesd (%/n yes) 7% (45)
Past 30-day use of alcohole (%/n yes) 12% (81)
Social media use
Instagram and/or TikTok use
(daily or multiple times per day)
78% (516)
Exposure to e-cigarette contentf on social media
Exposure to e-cigarette posts on Instagram and/or TikTok
(weekly, daily, or multiple times per day)
51% (336)
Exposure to e-cigarette content on social media posted by
influencers (including micro-influencers)
27% (178)
n in the header stands for the sample size; M (SD) stands for Mean and Standard Deviation.
Percent may not add up to 100%, because of missing value or may be above 100%, because of rounding or if categories are not
mutually exclusive (race).
a “Other” includes Native Hawaiian or Pacific Islander, American Indian or Alaska Native, a category defined as “Other” and
undisclosed. b Participants were asked the following question: “Have you ever used any electronic cigarette with nicotine (for example, e-cigs,
vaporizer, JUUL, Puff Bar, Elf Bar, or similar products)
in the past 30 days?”
c Participants were asked the following question: “Have you ever used any marijuana/cannabis products (for example, smoked
weed or bud, used marijuana or THC edibles, gummies, lozenges, used electronic device to vape marijuana or hash oil: weed
pen, dry flower or similar products) in the past 30 days?” d Participants were asked the following question: “Have you ever smoked a cigarette (for example, Marlboro, Camel, Newport,
etc.) in the past 30 days?” e Participants were asked the following question: “Have you used alcohol (for example, beer, wine, liquor) in the past 30 days?”
f E-cigarette posts may include a mix of pro- and anti-e-cigarette content.
31
Table 3. Multivariableab analyses to assess the effects of healthy lifestyle activities
performed by micro-influencers in e-cigarette promotional videos on harm perceptions and
susceptibility to e-cigarette use among adolescents.
Regressors
Outcomes
Harm perceptions of
e-cigarettesc
(N=664)
Perceptions of ecigarette appeal
(N=664)
Susceptibility to ecigarette use
(N=514)
AOR (95%CI)
(AOR > 1 represents lower
harm perceptions of ecigarettes)
AOR (95%CI)
(AOR > 1 represents
perceptions of e-cigarettes as
appealing: attractive, cool, or
fun)
AOR (95%CI)
(AOR > 1 represents
susceptibility to e-cigarette use)
Main effects models
Presence of healthy
lifestyle activities 1.03** (1.01; 1.05) 1.04** (1.02; 1.06) 1.10 (0.97; 1.25)
Perceptions of
influencer credibilityd 3.84*** (3.67;4.01) 5.88*** (5.63;6.14) 1.56** (1.19;2.04)
Interaction effects
models
Experimental
condition ×
perceptions of
influencer credibility
1.04** (1.02-1.07) 1.07*** (1.05; 1.09) 1.10 (0.97; 1.25)
Post-hoc multigroup
analyses
Effects of healthy
lifestyle presence
assessed among
respondents with
higher influencer
credibility perception
scoree
1.08*** (1.04;1.11) 1.08*** (1.06; 1.11) NA
Effects of healthy
lifestyle presence
assessed among
respondents with
lower influencer
0.99 (0.95;1.02) 0.98 (0.95; 1.01) NA
32
credibility perception
scoree
*** p-value < .001, ** p-value < 0.01, * p-value < 0.05, ’p-value < 0.1. Adjusted p-values are reported in Table 3. BenjaminiHochberg multiple testing corrections were applied to control the false discovery rate at 0.05 (based on 2-tailed corrected P
value).
Control condition: videos featuring influencers showing promoting e-cigarettes alone. a Poisson generalized linear mixed effects models to assess harm perceptions of e-cigarettes were adjusted for age, sex assigned
at birth, past 30-day e-cigarette, cigarette, marijuana, and alcohol use status, past 30-day exposure to e-cigarette-related social
media content posted by influencers, and frequency of social media use (no, less than monthly, monthly, weekly, daily). b Binomial generalized linear mixed effects models to assess susceptibility to e-cigarette use among never users of e-cigarettes
were adjusted for age, sex assigned at birth, past 30-day cigarette, marijuana, and alcohol use status, past 30-day exposure to ecigarette-related social media content posted by influencers, and frequency of social media use (no, less than monthly, monthly,
weekly, daily).
c Harm perceptions responses measured on the 0-100 scale were reversed-coded on the 0 to 10 scale: 0 (very harmful) – 10
(not harmful at all). AOR above 1 represents lower harm perceptions of e-cigarettes. d Susceptibility was collapsed into dichotomized responses related to interest in trying e-cigarettes: responding “definitely
not” to all the items vs. responding “probably not,” “probably yes”, “definitely yes” to one or more items.
e
To test the main and interaction effects, the continuous scale (0-100) of the perception of influencer credibility score was
standardized in the models (Mean = 0; SD = 1).
f
To conduct the post-hoc multigroup analyses, the perception of influencer credibility score was dichotomized using the median
split (Median=27): the higher score was set at or above 27; the lower score was set below 27.
Sample size: participants in the experimental group with higher (n = 157) and lower (n=167) influencer credibility perceptions.
Participants in the control group with higher (n = 148) and lower (n =177) influencer credibility perceptions. 8 participants in the
experimental group and 7 participants in the control group did not provide responses to the influencer perception question.
33
Figure 1. Experimental conditions. Healthy lifestyle activity featured in a social media video
clip alongside e-cigarette promotion by a micro-influencer used in the experimental stimuli
among adolescents. A screenshot of the video in the treatment condition (left) features an
influencer promoting a disposable vaping device (first five second of the 10-second clip)
followed by a tennis game (last 5 seconds of the 10-second clip). A screenshot of the video in the
control condition (right) features the same influencer promoting a disposable vaping device (first
five second of the 10-second clip) followed by a neutral profile of that influencer (last 5 seconds
of the 10-second clip).
34
Figure 2. Interaction plot. Proportion of participants who reported lower harm perceptions of
e-cigarettes (dichotomized using the median split of the harm perception score) and higher
perceived appeal to e-cigarettes (dichotomized using the median split of the appeal score) in the
experimental and control groups with higher or lower influencer credibility perceptions
(dichotomized using the median split). Differences in proportions between experimental and
control groups among participants who reported higher or lower influencer credibility perception
score in each of the 10 videos were compared using 2-sample test for equality of proportions
with continuity correction.
74%
34%
57%
35%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Higher influencer credibility perception Lower influencer credibility perception
Proportion of participants with lower harm
perceptions of e-cigarettes
Treatment Control
62%
22%
51%
23%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Higher influencer credibility perception Lower influencer credibility perception
Proportion of participants with higher perceived
appeal of e-cigarettes
Treatment Control
p-value < .001
p-value < .001
35
Sample size: participants in the experimental group with higher (n = 157) and lower (n=167)
influencer credibility perceptions. Participants in the control group with higher (n = 148) and
lower (n =177) influencer credibility perceptions. 8 participants in the experimental group and 7
participants in the control group did not provide responses to the influencer perception question.
36
Supplementary figure 1. Profile of a Randomized Clinical Trial
37
Supplementary table 1. Influencer credibility perception scores per each influencer
featured in the stimuli in the treatment and control condition.
Influencer
1
female
2
female
3
female
4
male
5
male
6
male
7
female
8
female
9
male
10
male
TREATMENT M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
Honest 24
(27.2)
19.5
(27.4)
27
(27.8)
25
(27.6)
22.5
(27.1)
23
(27.6)
27
(26.9)
25.5
(26.8)
27
(26.7)
22.5
(26.7)
Trustworthy 21
(26.5)
19.5
(25.5)
23
(26.5)
23
(25.1)
21
(25.5)
21.5
(25.5)
22
(26)
27
(25.3)
22
(24.9)
20
(25.2)
Informed 24
(28.1)
19
(26.8)
21
(25.2)
24
(26.7)
20
(26)
23
(25.7)
22
(27.6)
24
(25.7)
23
(24.8)
20
(24.9)
CONTROL M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
M
(SD)
Honest 22
(26.3)
28
(29.1)
30.5
(29.1)
31
(29.7)
31
(29)
28
(28.1)
28
(28.3)
31
(29)
32
(27.6)
25.5
(27.6)
Trustworthy 19
(25.7)
26
(27.7)
28
(28.6)
30
(27.9)
25
(27.1)
24
(27.6)
24
(25.8)
29.5
(29)
25
(26)
22
(26)
Informed 18
(27.7)
25
(29.1)
27
(29.1)
30
(29.5)
28
(30)
26.5
(28.7)
22
(27.7)
24
(29.9)
22
(28.9)
20
(27.2)
a b c d e f g h i j
M (SD) stands for Median and Standard Deviation.
The credibility perception scores per each trait (honest, trustworthy, informed) were measured using Kruskal-Wallis test and
Wilcoxon Test as a post hoc test for pairwise comparisons on a video level separately in the treatment and control
conditions.
Superscripts (a through j) indicate statistically significant (p-value < 0.05) pairwise differences in influencer perception scores
for each video per a respective personality trait in each condition. There were no statistically significant differences in influencer
perceptions among the 10 videos in each condition.
38
Chapter 2. Study 2
The impact of social media influencer marketing on perceptions of e-cigarettes among
adolescents: a survey-based experiment
Julia Vassey,
1 Erin A. Vogel,
2 Junhan Cho,1 Dayoung Bae,1 Jessica JL BarringtonTrimis,1
Jennifer B. Unger.
1
1
Department of Population and Public Health Sciences, University of Southern California, Los
Angeles, CA
2
TSET Health Promotion Research Center, University of Oklahoma Health Sciences Center,
Oklahoma City, OK
Declarations of Interest: All authors declare no competing interests.
Funding/Support: This study was funded by the Tobacco-Related Disease Research Program,
TRDRP, T33DT6620 and the NCI & FDA Center for Tobacco Products (CTP) Award
(NCI/FDA Grant #U54CA180905)
Role of Funder/sponsor: The funders had no role in study design; collection, analysis, and
interpretation of data; writing the report; and the decision to submit the report for publication.
The content is solely the responsibility of the authors and does not necessarily represent the
official views of the funders.
Data availability statement: Data will be made available upon request.
Running Head: E-cigarette promotion and adolescents’ perceptions of e-cigarettes.
Word Count: 4,127
Number of Tables: 3
Number of Figures: 2
Manuscript status: in preparation
39
INTRODUCTION
E-cigarette use among adolescents remains a global public health concern. Despite a
decline in e-cigarette use among high school students in 2023, over 2 million (8%) of high
school and middle school students in the United States (U.S.) reported past 30-day use of ecigarette products.1 Among current users, 25% reported using e-cigarettes daily. For
comparison, in the United Kingdom (UK),78 over 20% of adolescents tried e-cigarettes in 2023,
up from 16% in 2022. E-cigarette use has also been growing among adolescents in Southeast
Asia.79
Exposure to e-cigarette-related marketing content on social media increases risk of ecigarette use9,10,71 and use intentions,11 positive attitudes toward e-cigarettes, greater perceptions
of e-cigarette use as normative and popular11 and lower perceived risks of e-cigarette use among
adolescents.11,12 Particularly vulnerable groups of adolescents include e-cigarette never-users80 and
social media users.9 Despite policies restricting tobacco advertising,13,58,59 social media
platforms popular among youth (e.g., Instagram, TikTok)81 host promotional e-cigarette content
posted directly by tobacco brands or by influencers21–23 (i.e., models, bloggers, or brand
ambassadors who partner with tobacco brands to post on their behalf). Brands often use microinfluencers or nano-influencers (i.e., influencers with fewer than 100,000 followers31 who occupy
specific audience niches21) in their promotional content.26 These influencers may be perceived as
more trusted sources compared to direct advertising from brands because of greater perceived
authenticity and relatability of the content.
26–28 We recently identified engaging Instagram microinfluencers (n=55) from around the world (mostly U.S., UK, France, Germany, Indonesia and
Malaysia) who collaborated with over 600 e-cigarette brands.22 We also found that prevalence
of e-cigarette-related products promoted by micro-influencers has been growing on TikTok over
40
the last several years.25 E-cigarette-related marketing content in prior research has often been
defined broadly as “advertising,” with few studies analyzing the effect of a specific source of
marketing (e.g., tobacco brand, celebrity,32 or influencer)41 on youth perceptions and use patterns
of e-cigarettes. With proliferation of e-cigarette-related micro-influencer-driven content, more
research is needed to analyze the effect of content posted by these influencers on adolescent
perceptions and use patterns of e-cigarettes.
Micro-influencers promoting e-cigarettes use marketing tactics such as bright colors or
youth-oriented design of e-cigarette products,82 stylization of influencers to look young and
fashionable,44 and promotion of e-cigarettes in positive lifestyle context (e.g., nature or active and
healthy lifestyle).22,42,45,83 This could normalize e-cigarette use. Indicate findings from prior
studies. We found evidence that some micro-influencers who promote e-cigarettes on social media
also promote marijuana or hemp products (including CBD and THC).22,84 Promoting marijuana
and tobacco products next to each other may create youth misperceptions about which product
(e-cigarette or marijuana) is being promoted in a post because of the high resemblance of
disposable and cartridge-based nicotine vape products and marijuana vape products. Such
marketing can contribute to normalization of substance use, lower harm perceptions of ecigarettes (if marijuana is perceived as a safer substance than e-cigarettes as indicated by prior
research and our preliminary findings from focus groups among California high school
graduates), increase the risk of nicotine and/or marijuana use initiation or dual product use
among youth. This study assessed the effects of positive lifestyle contexts such as healthy lifestyle
activities, fashion or entertainment, and the effect of marijuana products shown alongside ecigarettes in promotional Instagram images featuring micro-influencers on harm perceptions of,
and susceptibility to use, e-cigarettes among adolescents. The Prototype Willingness Model
41
suggests that people are more willing to engage in a risky behavior when they have a positive
view of the prototypical person performing that behavior.
48,60 As such, we hypothesized that
adolescents would have lower harm perceptions of e-cigarettes and susceptibility to e-cigarette use
if they viewed the images where influencers promote e-cigarettes alongside any lifestyle context
or marijuana products versus images where influencers promote e-cigarettes alone, with no
lifestyle contexts, (control), adjusting for perceptions of influencer credibility (H1).
We also hypothesized that adolescents who viewed the images where influencers promote
e-cigarettes alongside any lifestyle contexts or alongside marijuana would have lower harm
perceptions of e-cigarettes and higher odds of susceptibility to e-cigarette use compared to the ecigarette only control condition if they perceived influencers as credible, (H2).
MATERIALS AND METHODS
Participants and Procedures
Participants (on average, 17 years of age) were part of the ongoing ADVANCE cohort
(Wave 7) run semi-annually by the University of Southern California (USC) Tobacco Center of
Regulatory Science (TCORS).85 Adolescents (N = 3,221) from 15 schools in the
socioeconomically and racially diverse Los Angeles, California metropolitan area were in 12th
grade when they took the survey between September 2023-January 2024. They completed the 30-
minute online survey that included the randomized experiment in class, programmed in REDCap,
on their classroom Chromebooks. Researchers obtained participants’ written informed assent and
parental consent prior to data collection. The study was approved by the University of Southern
California Institutional Review Board (UP-23-00810).
Sample for the experimental stimuli
42
The experimental videos were created from content (Instagram posts) of the 10 influencers
with public accounts identified in prior research22 who had disclosed that they were sponsored by
e-cigarette companies. These influencers were known to post e-cigarette content, and, in some
instances, post e-cigarette use alongside lifestyle activities or alongside marijuana products.
Experimental stimuli development
Each participant viewed two paired images per influencer profile (a total of six mages
featuring three out of four different influencers) in their respective condition. To ensure that
the content of the stimuli only differed by the characteristics of interest, the e-cigarette image
per each influencer (image 1) remained unchanged across conditions (except for the nonsubstance use condition, which featured two profile images without e-cigarettes). The
conditions differed by the second image (image 2) per their respective condition (Figure 1
shows examples of the experimental conditions for one influencer profile). To create
homogenous stimuli (e.g., to ensure that no influencer is perceived drastically differently from
the others by age and gender characteristics), 10 undergraduate students reviewed the stimuli,
answering the questions about perceived gender (“How likely that the person featured in this
promotional image is Man/Woman/Transgender?”) and perceived age of the influencers (How
likely that the person featured in this promotional post is: a) younger than 21; b) 21-30; c) 31-
40; d) older than 40?”). Posts with the highest rating agreement and highest rating scores (8 or
above, Table 1) were selected for the experiments. (Two male and two female profiles with the
perceived age of “between 21-30” were selected).
Experimental design
Each participant was randomly assigned to view three unique influencer profiles in one
of five versions of the influencers’ content representing context conditions: an influencer
43
promoting e-cigarette products along with marijuana products, (condition 1); an influencer
promoting e-cigarette products alongside fashion (e.g., modeling) or entertainment (e.g., DJing, gaming), (condition 2), an influencer promoting e-cigarette alongside healthy lifestyle
activity (e.g., playing tennis), (condition 3), an influencer promoting e-cigarette products
paired with a neutral profile post of this influencer (condition 4, control) or a non-substance
use condition: two profile posts featuring an influencer without e-cigarettes (condition 5,
placebo treatment). Each image had a social media username (not the actual name) of the
influencer. (Paired images had the same username since they feature the same influencer,
Figure 1).
Measures of harm perceptions of and susceptibility to e-cigarette use
Harm perceptions of e-cigarettes were measured using the validated two-item scale
(Table 1) adapted from the Population Assessment of Tobacco and Health (PATH).64
Responses to questions: Do you think using e-cigarettes for vaping nicotine is harmful to your
health? and Do you think people harm themselves when they use e-cigarettes for vaping nicotine?
were measured on a 0 (not at all harmful) to 100 (very harmful). Since the items were closely
intercorrelated (α=0.87), they were combined (by summing all the non-missing values of the
items) into one variable. The outcome variable was recoded to the 0-10 scale (since the
distribution of this count variable was a better fit for the Poisson model). The variable was also
reversed-coded with the higher score representing lower harm perceptions of e-cigarettes. For
easier interpretation of the interaction plots (Figures 2a and Figure 2b), the reversed-coded 0-
10 harm perception score in these figures was dichotomized by the median split (i.e., low harm
perceptions [harm perceptions of e-cigarettes at or above the median value] vs. high harm
perceptions).
44
Susceptibility to e-cigarette use (among never-users of e-cigarettes) was measured
using the validated three-item scale (Table 1) adapted from PATH64 and combined into one
variable (α=0.93). Consistent with prior research,67 the measure was collapsed into
dichotomized responses related to interest in trying e-cigarettes: responding “definitely not”
to all items vs. responding “probably not,” “probably yes”, “definitely yes” to one or more
items.
Perceptions of influencer credibility
Perceptions of influencer credibility (honesty and trustworthiness, Table 1) were
assessed with a 0-100 scale. This scale has been validated in prior research.41,68 Scores for each
of the three influencers whose images the participants viewed were measured and presented in
Supplementary Table 1.
Statistical analysis
Main effects models
To test H1, we examined the main effects of experimental conditions (i.e., exposure to images of
influencers promoting e-cigarettes alongside health, fashion or marijuana or not promoting ecigarettes) on harm perceptions of e-cigarettes (assessed after all the six images), adjusting for
perceptions of influencer credibility (assessed after each pair of three images featuring
influencers). We used the Poisson generalized linear mixed effects model (since the harm
perception outcome followed the Poisson distribution) and tested it among all the participants
who completed the experiment (N=3,221). The main effect of the experimental conditions and
perceptions of influencer credibility on susceptibility to use e-cigarettes were tested using the
binary logistic mixed effects regression model among never-users of e-cigarettes (n=2,408).
Repeated measures in experimental conditions (i.e., perceptions of influencer credibility) were
45
modeled as random effects. This approach was chosen as a result of research suggesting that a
substantial bias could exist when stimulus repetitions go unanalyzed as a random effect.69 (The
influencer perception score was modeled as a random intercept and fixed slope. We also included
schools as a random intercept to account for clustering by 15 schools and 1 cluster [n=117] with
missing school ID retained in the dataset). The models were adjusted for fixed effects of
sociodemographic and other covariates described below.
Interaction effects models
To test the potential moderating effects of perceptions of influencer credibility on
associations of experimental conditions with harm perceptions of e-cigarettes and susceptibility to
e-cigarette use outcomes, the interaction term of experimental condition X perceptions of
influencer credibility was added to the subsequent interaction effects models. To interpret the
significant interaction effects of H2, post hoc analyses were performed to assess the effects of
experimental conditions on harm perceptions of e-cigarettes and susceptibility to e-cigarette use,
stratified by the level of influencer credibility perceptions (i.e., high [perception of influencer
credibility score at or above the mean] vs. low).
Perceptions of influencer credibility
The effect of the experimental conditions on perceptions of influencer credibility was
assessed using the Analysis of variance model (ANOVA). The mean credibility perception
scores were measured using one-way ANOVA and Tukey’s Honest Significant Difference
(HSD) post hoc test for pairwise comparisons among influencers in each condition.
All multivariable models were adjusted for socio-demographic characteristics, substance
use in the past 30 days, frequency (weekly or more frequent vs less frequent or none) of exposure
to e-cigarette content on Instagram or TikTok and exposure to influencer e-cigarette content on
46
social media in the past 30 days assessed prior to the exposure to the experimental stimuli (Table
2). Adjustment for self-reported exposure to e-cigarette content posted by influencers accounted
for participants’ familiarity with this type of content on social media.
Adjusted odds ratios (AORs) with 95 % confidence intervals (CIs) were reported with
statistical significance set at p < 0.05 (two-tailed) after applying Benjamini-Hochberg multiple
testing corrections to control for the false discovery rate at 0.05. All statistical analyses were
conducted using R software (version 4.2.2). (Sample sizes for all the experimental conditions are
reported in Supplementary figure 2).
RESULTS
Adolescents who completed the experiment (N = 3,221) were Mage=17 [SD=0.6]; 53%
female, 46% Hispanic. 5% reported using e-cigarettes with nicotine, 5% reported using an
electronic device to vape marijuana in the past 30 days, and 11% reported marijuana use including
vaping, smoking, or using any THC, CBD, or hemp products. 18% reported at least weekly
exposure to e-cigarette-related content on Instagram and/or TikTok and 6% reported exposure to
e-cigarette-related social media content posted by micro-influencers (Table 2). Among all
adolescents who completed the experiment, 47% had low harm perceptions of e-cigarettes
(dichotomized as low vs high harm perceptions using the median split of the score for harm
perceptions of e-cigarettes); 73% were never users of e-cigarettes, and among those neverusers, 19% were susceptible to e-cigarette use.
Main effects models
Participants (N = 3,221) had higher odds of lower harm perceptions of e-cigarettes if they
viewed images of influencers promoting e-cigarettes alongside fashion or entertainment activities
(AOR = 1.05; CI = 1.01 –1.09) or neutral profiles of influencers with no substance use (AOR =
47
1.07; CI = 1.03 –1.11) compared to controls who viewed images of influencers promoting ecigarettes only. Participants had lower odds of lower harm perceptions (AOR = 0.95; CI = 0.92 –
0.99) and susceptibility to e-cigarette use (AOR = 0.67; CI = 0.53 –0.85 [with susceptibility
assessed among never-users of e-cigarettes, N = 2,408]) if they viewed images of influencers
promoting e-cigarettes alongside healthy lifestyle activities compared to controls, (Table 3).
Interaction effects models
We found significant interaction effects of the experimental condition X perceived
influencer credibility on harm perceptions of, and susceptibility to use, e-cigarettes (Figure 2a and
2b). Among the participants (n=1,600) who perceived influencers as non-credible (below the mean
of the influencer credibility perception score), 51% of those who viewed images of e-cigarette
promotion alongside fashion or entertainment vs 39% who viewed images in the control condition
(e-cigarettes alone) had lower harm perceptions of e-cigarettes (p-value < .001). Similarly, among
participants who perceived influencers as non-credible, 55% of those who viewed images in the
no substance use condition vs 39% in the control condition (e-cigarettes alone) had lower harm
perceptions of e-cigarettes (p-value < .001). Among never-users of e-cigarettes who perceived
influencers as credible (n=1,741), 41% of those who viewed e-cigarette promotion alongside
marijuana vs 26% who viewed images in the control condition were susceptible to e-cigarette use
(p-value < .001).
The post hoc analyses of associations between experimental conditions and each outcome,
stratified by the perception of influencer credibility score, are presented in Table 3. Among
participants who perceived influencers as non-credible, those who viewed images in the no
substance use condition had higher odds of lower harm perceptions of e-cigarettes (AOR = 1.08;
95% CI = 1.04 – 1.12) compared to those in the control condition.
48
Among never-users of e-cigarettes who perceived influencers as credible, participants who
viewed images of e-cigarette promotion alongside marijuana had slightly higher odds of
susceptibility to use e-cigarettes (AOR = 1.32; 95% CI = 1.03 – 1.71, unadjusted p-value = 0.03,
adjusted p-value = 0.06) compared to the participants in the control condition (Table 3). Among
never-users of e-cigarettes who perceived influencers as non-credible, participants who viewed
images of e-cigarette promotion alongside marijuana had lower odds of susceptibility to use ecigarettes (AOR = 0.65; 95% CI = 0.47 – 0.91) compared to the participants in the control
condition.
Perceptions of influencer credibility
Participants in the no substance use condition had the highest influencer credibility
perception score (Mean = 48.4) compared to all the other conditions with e-cigarette presence
(Figure 3), while participants who viewed images of influencers promoting e-cigarettes alongside
marijuana had the lowest influencer credibility perception score (Mean = 34.1, p-value < .001,
Supplementary figure 1).
DISCUSSION
An influencer marketing tactic such as promotion of e-cigarettes alongside marijuana
products or alongside fashion or entertainment activities in Instagram images may contribute to
lower harm perceptions of, and susceptibility to use, e-cigarettes among adolescents. Participants
of this randomized survey-based experiment who viewed influencers promoting e-cigarette use
alongside marijuana products (compared to influencers who promoted e-cigarettes only) were
more likely to report susceptibility to e-cigarette use if they perceived influencers as credible.
However, if the participants perceived influencers as non-credible, those who viewed images of ecigarettes alongside marijuana were less likely to report susceptibility to e-cigarette use. The
49
findings are in line with the Prototype Willingness Model48,60 suggesting that people are more
willing to engage in a health risky behavior to the extent that they have a positive view of the
prototypical person who performs that behavior. Such prototypical person could be a peer, a
celebrity, or an influencer. The finding is concerning considering evidence of co-promotion of ecigarettes and marijuana or hemp products (including CBD and psychoactive Delta-8 THC) by
influencers on social media. Promoting marijuana and tobacco products next to each other may
create youth misperceptions about which product (e-cigarette or marijuana) is being promoted
in a post because of the high resemblance of disposable and cartridge-based nicotine vape
products and marijuana vape products, which could contribute to increased risk of co-use of both
products (e-cigarette and marijuana). Besides promoting marijuana vape products, microinfluencers also promote edible marijuana products next to e-cigarettes, which is also concerning
(Tobacco Control blog, in print). While the prevalence of e-cigarette use (5%) and vaping
marijuana (5%) among participants of this experiment was the same and relatively low, a higher
proportion of participants (11%) reported another form of marijuana use besides vaping, such as
smoking or eating THC, CBD, or hemp products.
Participants who viewed images of influencers promoting e-cigarettes alongside fashion or
entertainment or influencers showing no substance use (compared to influencers who promoted ecigarettes only) had higher odds of lower harm perceptions of e-cigarettes. E-cigarettes were also
perceived as less harmful in the no substance use condition compared to the e-cigarette only
condition if adolescents perceived influencers as non-credible. The finding may imply that
adolescents might not wish to emulate e-cigarette use behavior when they see influencers
promoting e-cigarettes with no lifestyle context. Adolescents may believe that these influencers
engage in a harmful behavior for the sole purpose of promoting e-cigarettes, which increases
50
negative perceptions of these influencers. However, if e-cigarette promotion occurs as part of
influencers’ daily life: e.g., entertainment activities (playing DJ turn tables, gaming) or modeling
or if influencers do not show substance use at all, e-cigarettes might be perceived as less harmful.
These findings are consistent with preliminary focus group results among California 18-year-old
high-school graduates were participants considered influencers less credible if they perceived them
solely as promoters of e-cigarettes (as opposed to influencers who show e-cigarette use on social
media as part of their lifestyle, (Vassey et al., manuscript in preparation).
Participants who viewed images of influencers promoting e-cigarette use alongside healthy
lifestyle activities (e.g., working out in a gym) had higher harm perceptions of, and lower odds of
susceptibility to use, e-cigarettes. This finding seems inconsistent with the preliminary results from
two of our online experiments among adolescents (Study 1 of the dissertation) and among young
adults (Vassey et al, manuscript in review). In these experiments, participants who viewed videos
of influencers promoting e-cigarette use alongside healthy lifestyle activities had lower harm
perceptions, higher perceived appeal of, or higher odds of susceptibility to use, e-cigarettes if they
perceived influencers positively. It is possible that shorter duration of exposure (images in the
current study versus videos in the referred online experiments) contributed to differences in
experimental findings. Differences could also be explained by the fact that in the current study
participants completed the survey in their classrooms, while the other two experiments took place
online. (While online panels were anonymous, classroom surveys were not. It is possible that when
a survey is taken in classroom students may be stronger prone to social desirability bias while
being supervised by teachers or staff). It is also possible that demographic differences contributed
to different results. For example, the population of this study had higher prevalence (38%) of Asian
participants compared to 13% among adolescents and young adults in the two online experiments.
51
A study among a state-representative sample of 58,689 Utah adolescents showed that Asian youth
(8 to 12th graders) had lower rates of e-cigarette use compared to White youth.
86 So was this case
in this study with 1% of Asian versus 8% of White students reporting using e-cigarettes in the past
30 days. Thus, Asian participants potentially could be less influenced by social media stimuli
featuring e-cigarettes compared to other racial/ethnic groups. Demographic differences in the
survey populations could also explain that only 6% of participants in the currentstudy self-reported
exposure to influencer e-cigarette marketing versus 26% and 18% respectively among adolescents
and young adults in the online experiments. Familiarity with Instagram and TikTok interface and
tobacco-related content may increase persuasiveness of promotional messages. Taken together, it
is possible that the participants of the current study, less familiar with tobacco-related content on
social media, experienced cognitive dissonance while seeing healthy activities shown next to ecigarette promotion and perceived such imagery more negatively relative to e-cigarettes promoted
alone, which could explain the opposite effect of this experimental condition to what was observed
in the two online experiments among adolescents (Study 1) and young adults (Vassey et al, in
review).
Influencer e-cigarette-related marketing on social media is underregulated. Although most
widely-used and well-known social media platforms (e.g., Facebook, Instagram, YouTube,
Twitter, TikTok, Reddit, Snapchat) prohibit paid advertisements for tobacco products,41,59 few
(Facebook, Instagram and TikTok)11 explicitly prohibit influencers from promoting
nicotine/tobacco products. And despite these policies, influencers still promote e-cigarettes on the
platforms that have the restrictions (indicating that community guidelines are not very well
followed and algorithms for removal of such content are not very well developed or functional).
While influencers may advertise unhealthy products to pursue commercial interests, consumers of
52
their content (including youth) may perceive these influencers as trusted and authentic sources,46
especially when influencers promote e-cigarettes as part of their daily lifestyle (e.g., working out,
playing video games) or along with marijuana, which could contribute to normalization of different
substance use. While regulating influencer marketing features (e.g., prohibiting depiction of
healthy lifestyle context next to e-cigarette products) could be challenging because of the First
Amendment of the U.S. Constitution,55 regulating promotion of e-cigarettes and marijuana
products could be considered. Not only co-promotion of e-cigarette and marijuana could
contribute to normalization of substance use, but it could also increase the risk of nicotine and/or
marijuana use initiation or dual product use among youth. Moreover, the marijuana industry
has started adopting marketing strategies of the tobacco industry by utilizing influencer
marketing in marijuana product promotion on social media, which is exacerbated by the lack
of legislation (e.g., in the U.S.) related to marijuana marketing restrictions.87 Authorities could
consider prohibiting e-cigarette brands from co-promotion of e-cigarettes and marijuana (even
though the FDA does not regulate marijuana marketing) or penalize brands and influencers for
non-compliance with federal marketing regulations such as restricting access to e-cigarette
marketing content to youth or following the Federal Trade Commission (FTC) guidelines for
sponsorship disclosures.
36,74 Further improvement in the enforcement of social media
community guidelines is necessary, too.15,16 At the same time, since influencers can be perceived
as trusted sources of information, public health authorities may harness influencers for antitobacco messaging and in health literacy campaigns about harmful effects of e-cigarette use46,73 or
disseminating other positive public health messages.46 The findings can also be disseminated via
media literacy curriculums in schools.
Limitations
53
Since the participants in both experiments were based in California, the results from the
studies may not be generalizable to all U.S. adolescents. Concerns about generalizability are
mitigated by the fact that ethnic diversity of the samples represent the projected future ethnic
diversity of the U.S.76 Despite the randomized experimental design, self-reported survey
outcomes might be prone to social desirability bias, especially when they are not anonymous and
taken in classroom as was the case in this study. The measures were also assessed only at one point
in time (vs multiple times overtime). Participants were only exposed to three paired images of
influencer profiles. Such brief duration of exposure might have affected the results. While the
experimental stimuli were created from actual Instagram posts featuring micro-influencers
promoting e-cigarette products, the survey did not use an actual Instagram interface to simulate
a real-life social media use experience.
Conclusions
The study demonstrated that an influencer marketing tactic such as e-cigarette promotion
alongside fashion or entertainment or next to marijuana products may contribute to lower harm
perceptions and susceptibility to e-cigarette use among adolescents, particularly those who
perceived influencers as credible. Future research should continue providing surveillance data to
show changes in e-cigarette as well as marijuana influencer marketing over time, and
investigating if tobacco companies use influencers to introduce new or promote existing ecigarette products in violation of premarket authorization requirements (PMTA). Future
research should also examine perceptions of influencer credibility with and without FTCrequired e-cigarette sponsorship disclosures, since motivations of influencers who promote ecigarettes without such disclosures could be perceived as more genuine and authentic versus
purely promotional.88 Our findings indicate the need for further regulation of influencer
54
marketing. Results also support public health efforts to harness influencers for anti-substance use
messaging and in health literacy campaigns about harmful effects of e-cigarette use, marijuana
use, or dual substance use for adolescent health.
Contributorship Statement Concept and design – JV. Acquisition, analysis, or interpretation of
data—all authors. Draft of the manuscript - JV. Critical revision of the manuscript for important
intellectual content—all authors. Statistical analysis – JV. Obtaining funding – JV, JBU.
Administrative, technical, or material support – JV, JBU. Supervision – JBU. All co-authors
approved the final version.
Ethics approval The University of Southern California Institutional Review Board (UP-23-
00810) approved all study procedures.
55
Table 1. Experiment outcome measures
Table 3. Outcome measures
Outcomes assessed after participants viewed ALL the posts
Harm perceptions64 of ecigarettes
Do you think using e-cigarettes for vaping nicotine
is harmful to your health?
Do you think people harm themselves when they
use e-cigarettes for vaping nicotine?
0 (Not harmful at
all) – 100 (Very
harmful)
0 (Not at all) – 100
(A lot)
Susceptibility to ecigarette use,64
(assessed only among
never-users of ecigarettes)
1) Would you try e-cigarettes for vaping nicotine if
one of your best friends offered them to you?
2) If you had an opportunity to use e-cigarettes for
vaping nicotine, would you use them?
3) Do you think you would use e-cigarettes for
vaping nicotine in the next 6 months?
4) Are you curious about using e-cigarettes for
vaping nicotine?
Definitely Not/
Probably Not/
Probably Yes/
Definitely Yes
The outcome assessed after participants viewed each unique influencer profile
Perceptions of influencer
credibility 41,77
Based on the video you just watched, how would
you rate that influencer on the items below?
dishonest–honest
untrustworthy–trustworthy
0 (not at all) -100 (a
lot), with the word
pairs anchored at
each end of the 0-
100 scale
56
Table 2. Participant characteristics for adolescents (N = 3,221)
Participant characteristics % (n) / M (SD)
Age 17 (0.6)
Sex assigned at birth
Female 53% (1,717)
Sexual orientation
Heterosexual 68% (2,196)
Race
Asian 38% (1,225)
White 16% (517)
Othera 9% (304)
Multiple race 14% (457)
Unreported 22% (718)
Ethnicity
Hispanic/Latinax 46% (1,488)
Substance use
Past 30-day use of e-cigarettesb (%/n yes) 5% (151)
Past 30-day vape of marijuanac (%/n yes) 5% (159)
Past 30-day use of all marijuana productsd (%/n yes) 11% (344)
Past 30-day use of other tobacco productse (%/n yes) 1% (44)
Past 30-day use of alcoholf (%/n yes) 11% (370)
Social media use
Instagram and/or TikTok use
(daily or multiple times per day)
80% (2,585)
Exposure to e-cigarette contentg on social media
Exposure to e-cigarette posts on Instagram and/or TikTok
(weekly, daily, or multiple times per day)
18% (579)
Exposure to e-cigarette content on social media posted by
micro-influencers
6% (191)
n in the header stands for the sample size; M (SD) stands for Mean and Standard Deviation.
Percent may not add up to exactly 100%, because of missing value or rounding. a “Other” includes Native Hawaiian or Pacific Islander, American Indian or Alaska Native, Black, and a category defined as
“Other.” b Participants were asked the following question: “Have you used any electronic cigarette for vaping nicotine (E-cigs, vaporizer,
JUUL, Puff Bar)in the past 30 days?” c Participants were asked the following question: “Have you used electronic device to vape THC or hash oil (liquid pot,
marijuana oil, weed pen, Kandypens, Vessel, Linx, PAX) in the past 30 days?” d
Marijuana products include: smoking marijuana (pot, weed, hash, reefer, bud, or grass); using electronic device to vape THC or
hash oil (liquid pot, marijuana oil, weed pen, Kandypens, Vessel, Linx, PAX); using marijuana or THC foods or drinks (edibles,
pot brownies, cookies, cakes, butter, oil); using any CBD or hemp products (gummies, butter, oil, food products). e Other tobacco products include combustible cigarettes, IQOS or other heated tobacco devices, oral nicotine products, big cigars,
little cigars and cigarillos, and hookah. f Participants were asked the following question: “Have you used one full drink of alcohol (can of beer, glass of wine, wine
cooler, or shot of liquor) in the past 30 days?” g E-cigarette posts may include a mix of pro- and anti-e-cigarette content.
57
Table 3. Multivariableab analyses to assess the effects of experimental conditions in ecigarette promotional images featuring influencers on harm perceptions and susceptibility
to e-cigarette use among adolescents.
Regressors
Outcomes
Harm perceptions of ecigarettesc
(N=3,221)
Susceptibilityd to ecigarette use
(N=2,408)
AOR (95%CI)
(AOR > 1 represents lower harm
perceptions of e-cigarettes)
AOR (95%CI)
(AOR > 1 represents susceptibility
to e-cigarette use)
Main effects models
E-cigarette and marijuana 1.01 (0.97; 1.05) 1.08 (0.94; 1.26)
E-cigarette and fashion/entertainment 1.09*** (1.04; 1.13) 0.99 (0.85; 1.15)
E-cigarette and healthy lifestyle activities 0.99 (0.95; 1.03) 0.80** (0.67; 0.92)
No-substance use presence 1.11*** (1.07; 1.16) 0.96 (0.83; 1.11)
Perceptions of influencer credibilitye 2.43*** (2.28;2.59) 0.44*** (0.34;0.55)
Interaction effects models
E-cigarette and marijuana × perceptions
of influencer credibility 1.06* (1.02-1.11) 1.41*** (1.21-1.65)
E-cigarette and
fashion/technology/gaming ×
perceptions of influencer credibility
0.95* (0.91; 0.99) 0.87 (0.45; 1.64)
E-cigarette and healthy lifestyle
activities × perceptions of influencer
credibility
0.94’ (0.91; 1.00) 1.10 (0.60; 2.03)
No-substance use presence ×
perceptions of influencer credibility 0.92*** (0.89; 0.96) 0.83* (0.72; 0.96)
Post-hoc multigroup analyses
Effects of the experimental conditions
assessed among respondents with higher
influencer credibility perception scoref
E-cigarette and marijuana 1.05 (0.99;1.12) 1.57*** (1.28-1.92)
E-cigarette and
fashion/technology/gaming 1.04 (0.98;1.10) NA
58
E-cigarette and healthy lifestyle
activities 0.94 (0.89;1.01) NA
No-substance use presence 1.03 (0.98;1.08) 0.83’ (0.69; 0.99)
Effects of the experimental conditions
assessed among respondents with lower
influencer credibility perception scoref
E-cigarette and marijuana 0.99 (0.93;1.05) 0.73* (0.59; 0.91)
E-cigarette and
fashion/technology/gaming 1.14*** (1.07;1.21) NA
E-cigarette and healthy lifestyle
activities 1.04 (0.98;1.11) NA
No-substance use presence 1.23*** (1.15; 1.31) 1.18 (0.93; 1.49)
*** p-value < .001, ** p-value < 0.01, * p-value < 0.05, ’p-value < 0.1. Adjusted p-values are reported in Table 3. BenjaminiHochberg multiple testing corrections were applied to control the false discovery rate at 0.05 (based on 2-tailed corrected P
value).
Control condition: videos featuring influencers showing promoting e-cigarettes alone. a Poisson generalized linear mixed effects models to assess harm perceptions of e-cigarettes were adjusted for age, sex assigned
at birth, past 30-day e-cigarette, cigarette, marijuana, and alcohol use status, past 30-day exposure to e-cigarette-related social
media content posted by influencers, and frequency of social media use (no, less than monthly, monthly, weekly, daily). b Binomial generalized linear mixed effects models to assess susceptibility to e-cigarette use among never users of e-cigarettes
were adjusted for age, sex assigned at birth, past 30-day cigarette, marijuana, and alcohol use status, past 30-day exposure to ecigarette-related social media content posted by influencers, and frequency of social media use (no, less than monthly, monthly,
weekly, daily).
c Harm perceptions responses measured on the 0-100 scale were reversed-coded on the 0 to 10 scale: 0 (very harmful) – 10
(not harmful at all). AOR above 1 represents lower harm perceptions of e-cigarettes. d Susceptibility was collapsed into dichotomized responses related to interest in trying e-cigarettes: responding “definitely
not” to all the items vs. responding “probably not,” “probably yes”, “definitely yes” to one or more items.
e
To test the main and interaction effects, the continuous scale (0-100) of the perception of influencer credibility score was
standardized in the models (Mean = 0; SD = 1).
f
To conduct the post-hoc multigroup analyses, the perception of influencer credibility score was dichotomized using the mean
split (Mean=40.7) to assess harm perceptions of e-cigarettes: the higher score was set at or above 40.7; the lower score was set
below 40.7.
The score was dichotomized using the mean split of 39.9 to assess susceptibility to e-cigarette use: the higher score was set at or
above 39.9; the lower score was set below 39.9.
Sample sizes to assess harm perceptions of e-cigarettes: participants with higher influencer credibility perceptions: e-cigarette
alone (n = 424); e-cigarette + fashion/entertainment (n = 441); e-cigarette + health (n = 489); e-cigarette + marijuana (n = 380);
no substance (n = 556). Participants with lower influencer credibility perceptions: e-cigarette alone (n = 393); e-cigarette +
fashion/entertainment (n = 414); e-cigarette + health (n = 485); e-cigarette + marijuana (n = 447); no substance (n = 402).
Sample size to assess susceptibility to e-cigarette use among never-users of e-cigarettes: participants with higher influencer
credibility perceptions: e-cigarette alone (n = 424); e-cigarette + fashion/entertainment (n = 441); e-cigarette + health (n = 489);
e-cigarette + marijuana (n = 380); no substance (n = 556). Participants with lower influencer credibility perceptions: e-cigarette
alone (n = 393); e-cigarette + fashion/entertainment (n = 414); e-cigarette + health (n = 485); e-cigarette + marijuana (n = 447);
no substance (n = 402).
59
Figure 1. Experimental conditions. Experimental conditions featured in social media images
used in the experimental stimuli among adolescents. Condition 1: an influencer promoting ecigarette products paired with promoting marijuana products; condition 2: an influencer
promoting e-cigarette products paired with fashion or entertainment; condition 3: an
influencer promoting e-cigarette products along with healthy lifestyle; condition 4 (placebo
treatment): profile posts featuring an influencer without e-cigarettes; condition 5 (control):
an influencer promoting e-cigarette products alone paired with a neutral profile post of this
influencer. Each participant viewed two paired images per influencer profile (a total of six
mages) in their respective condition.
60
Figure 2a. Interaction plots. Proportion of participants who reported low harm perceptions of
e-cigarettes in the experimental conditions with higher or lower influencer credibility perceptions
(dichotomized using the mean split). In this figure the 1-10 score for harm perceptions of ecigarettes was dichotomized by the median split (i.e., low harm perceptions [harm perceptions
of e-cigarettes at or above the median value] vs. high harm perceptions).
Pairwise differences in proportions between experimental conditions and the control condition
among participants who reported higher or lower influencer credibility perception score (Figure
2a and 2b) were compared using 2-sample test for equality of proportions with continuity
correction.
Sample size: participants with higher influencer credibility perceptions: e-cigarette alone (n =
424);
e-cigarette + fashion/entertainment (n = 441); e-cigarette + health (n = 489); e-cigarette +
marijuana (n = 380); no substance (n = 556). Participants with lower influencer credibility
perceptions: e-cigarette alone (n = 393); e-cigarette + fashion/entertainment (n = 414); ecigarette + health (n = 485); e-cigarette + marijuana (n = 447); no substance (n = 402).
61
Figure 2b. Interaction plots. Proportion of participants who reported susceptibility to ecigarette use in the experimental conditions with higher or lower influencer credibility
perceptions (dichotomized using the mean split).
Sample size: participants with higher influencer credibility perceptions: e-cigarette alone (n =
322);
e-cigarette + fashion/entertainment (n = 317); e-cigarette + health (n = 370); e-cigarette +
marijuana (n = 287); no substance (n = 445). Participants with lower influencer credibility
perceptions: e-cigarette alone (n = 302); e-cigarette + fashion/entertainment (n = 282); ecigarette + health (n = 356); e-cigarette + marijuana (n = 350); no substance (n = 310).
62
Supplementary figure 1. Influencer perception score by experimental condition. The mean
influencer credibility perception score assessed among the participants who completed the
experiment (N = 3,221) by the experimental condition. The mean credibility perception scores
were measured using one-way ANOVA and Tukey’s Honest Significant Difference (HSD)
post hoc test for pairwise comparisons among influencers in each condition. *** p-value <
.001, ** p-value < 0.01, * p-value < 0.05, ’p-value < 0.1.
34.1
39.8 38.6 39
48.4
0
10
20
30
40
50
60
70
80
90
100
E-cigarette +
Marijuana
E-cigarette +
Fashion or
Entertainment
E-cigarette +
Health
E-cigarette alone Non-substance
Influencer crediblity perception score
***
***
***
***
***
63
Supplementary figure 2. Profile of a Randomized Clinical Trial
64
Supplementary table 1. Influencer credibility perception scores per each
influencer featured in each experimental condition.
Influencer
1 2 3
Condition M (SD) M (SD) M (SD)
Trait: honest
E-cigarette + marijuana 38 (24.3) c 35 (23.7) 35.1 (24.2)
E-cigarette + fashion/entertainment 43.9 (23.6) 41.8 (22.8) 40.8 (23.4)
E-cigarette + health 42.6 (22.4) 40.7 (23.1) 40.3 (22.8)
E-cigarette + alone 43.8 (23) c 41.4 (23.2) 39.4 (23.4)
No substance 48 (23.7) 49.4 (23.1) 50.8 (24.2)
a b c
Trait: trustworthy
E-cigarette + marijuana 33.8 (23.2) 31.8 (22.9) 32.3 (23.2)
E-cigarette + fashion/entertainment 39.5 (22.1) 39.3 (22) 39.4 (23.5)
E-cigarette + health 39.1 (21.3) 37 (22) 37.6 (22.5)
E-cigarette + alone 39.8 (21.6) 37.7 (22.3) 37 (22.8)
No substance 45.4 (22.7) c 47.2 (23.2) 49.5 (24.9)
a b c
M (SD) stands for Mean and Standard Deviation.
The mean credibility perception scores per each trait (honest, trustworthy) were measured using one-way ANOVA and
Tukey’s Honest Significant Difference (HSD) post hoc test for pairwise comparisons among influencers in each condition.
Superscripts (a through c) indicate statistically significant (p-value < 0.05) pairwise differences in influencer perception
scores for each image per a respective personality trait in each condition. (For example, the honesty score for influencer 1 in the
e-cigarette + marijuana condition is statistically significantly different from the honesty scores of the influencer 3 represented by
a superscript c).
65
Chapter 3. Study 3
Worldwide connections of micro-influencers who promote e-cigarettes on Instagram and
TikTok: a social network analysis.
Julia Vassey1
, Herbert Ho-Chun2,3, Tom Valente1
, Jennifer B. Unger1
1. Department of Population and Public Health Sciences, University of Southern California, Los
Angeles, CA
2. Annenberg School of Communication and Journalism, University of Southern California, Los
Angeles, CA
3. Information Sciences Institute, Viterbi School of Engineering, University of Southern
California, Los Angeles, CA
Declarations of Interest: All authors declare no competing interests.
Funding/Support: This study was funded by the NCI & FDA Center for Tobacco Products
(CTP) Award (NCI/FDA Grant #U54CA180905) and the National Institute of Health, NIH,
R01CA260459
Role of Funder/sponsor: The funders had no role in study design; collection, analysis, and
interpretation of data; writing the report; and the decision to submit the report for publication.
The content is solely the responsibility of the authors and does not necessarily represent the
official views of the funders.
Data availability statement: Data will be made available upon request.
Running Head: Connections of micro-influencers who promote e-cigarettes on social media.
Word Count: 5,091
Number of Tables: 1
Number of Figures: 4
Manuscript status: in preparation
66
INTRODUCTION
Despite existing policies restricting paid advertisements or promotions of tobacco
products,13,58,59 social media platforms (e.g., Instagram, TikTok) popular among Generation Z (11
to 26 years of age)24 host promotional e-cigarette content posted by tobacco brands and influencers
– models, bloggers, brand ambassadors who post on behalf of these brands.21–23 Brands have been
increasingly relying on micro-influencers – users with about 1,000 – 100,000 followers – to
promote their products or services,
26 since working with micro-influencers is more cost-effective
compared to celebrities. Content posted by micro-influencers is also perceived by consumers as
more authentic, trustworthy and relatable compared to direct advertising from brands.
27,89
Instagram has been the most popular platform for micro-influencer marketing, with TikTok
gaining popularity among brands as well.90 Influencers, including those who promote e-cigarettes
on social media, often share their content on both platforms, cross-promoting harmful content on
both platforms widely used by youth.23 This is concerning since exposure to tobacco-related
content on Instagram and TikTok has been associated with adolescent and young adult e-cigarette
use.9-12
While experimental or observational studies analyze association of exposure to e-cigarette
content with e-cigarette outcomes among certain populations, the social network framework shifts
the focus from studying individual traits to analyzing interactions, relationships, and
communications.51 Analyzing network structures and positions of individual influencers in
social media networks could help researchers understand the level of interaction (e.g., via
content engagement such as comments) and e-cigarette-related information sharing among
influencers as well as among influencers and their audiences. These interactions fit the
framework of the Two-Step Flow of Communication Model.
52 In its traditional interpretation,
67
the Two-Step Flow of Communication Model suggests that the flow of information and influence
from the mass media to their audiences involves two steps: from the media to certain individuals
(i.e., the opinion leaders) and from these opinion leaders to the public. In other words, opinion
leaders mediate information flow from mass media to audiences. Due to decline of conventional
communication channels, such as TV, newspapers, or radio and the rise of the more fragmented
media marketplace (i.e., internet and social media), opinion leaders have become original sources
of information that is directly communicated to the public. Opinion leaders are no longer only
celebrities or notorious experts (as it used to be in the past), but now include a wider range of
influencers of different caliber (from mega- to micro- and nano-influencers who are considered
experts on a specific topic and are popular among their niche audience).21,22 These non-celebrity
influencers create their own content on social media, engage with each other’s content and directly
interact with audiences who consume this content.
Research on influencer networks related to tobacco products on social media has been
limited. Vassey et al discovered a dense network of international micro-influencers and over
600 e-cigarette brands the influencers partnered with to promote e-cigarette products on
Instagram.
22 Gu et al constructed a network of well-known tobacco brands who posted about IQOS
heated tobacco products on Instagram.
91 Zhou et al analyzed the follower networks of 33
influencers who posted about e-cigarettes on Twitter and found that the influencers’ followers who
posted about e-cigarettes were more interconnected compared to the followers who did not post
about e-cigarettes.92 Social networks of micro-influencers who promote e-cigarettes on TikTok
have not been studied in scientific literature. Comparing tobacco-related influencer networks on
different platforms has not been studied either. Considering that influencers who promote ecigarettes on Instagram and TikTok often cross-promote their content on both platforms, we
68
explored and compared Instagram versus TikTok networks’ structural characteristics: i.e.,
interconnectedness, clustering; and individual characteristics: i.e., influencers’ central positions
in these networks, (exploratory hypotheses). (The measures are described in detail in the
Methods section).
We also compared differences in individual characteristics within Instagram and within
TikTok by geographic region (i.e., North vs South America vs Asia vs Europe). Prior research
found that micro-influencers from Asia and U.S. collaborated with a higher number of ecigarette brands and had higher numbers of followers compared to European influencers.22 As
such, we hypothesized that Asian and North American (including U.S.-based) microinfluencers will be more central and influential in the network compared to European microinfluencers within Instagram and within TikTok (H1). We also assessed tendency of tie
(connection) formation among micro-influencers from different regions. Regional homophily (i.e.,
tendency of influencers from the same regions to form ties among each other as opposed to forming
ties with influencers from different geographic regions) are explained by certain regional
commonalities or interests, including linguistic commonalities.
54 As Waldo Tobler’s First Law of
Geography states, “everything is related to everything else, but near things are more related than
distant things.”54 However, international borders on social media are absent, which increases the
possibility of tie formation among influencers from different regions. This is a concern from the
tobacco regulatory science standpoint. E-cigarette content posted by international influencers can
reach other countries’ audiences if influencers from different countries engage (comment on or
‘like’) with each other’s social media posts, potentially exposing their followers (that may include
adolescents) to broader e-cigarette promotional materials (i.e., e-cigarette posts from different
countries). Moreover, e-cigarette products promoted by international influencers may also be
69
available for online purchase and shipment to different countries (e.g., the U.S.), and such products
could be discovered and ordered by youth (if age verification checks on such purchases and
shipments are absent). Prior research found that different e-cigarette brands often partner with
the same influencers from all over the world, primarily from the U.S., Indonesia, Malaysia,
Germany, France, and Italy.22 As such, we hypothesized that the network’s regional homophily
(ties among micro-influencers from the same regions) will be strong and driven by linguistic
homophily (ties among users who share the same language in comments to micro-influencers’
posts); however, heterophily (connection with dissimilar nodes based on the geographic attribute,
i.e., ties among micro-influencers from different regions) will also be present, (H2). Prior
research also demonstrated that influencers often collaborate with multiple industries (e.g.,
fashion, gaming, healthy lifestyle) besides e-cigarette brands.22 These influencers could potentially
expose their non-e-cigarette-focused audience (including non-users of e-cigarettes) to e-cigarette
content, which is problematic. As such, we hypothesized that influencers who post about ecigarettes and other content will be more central and influential in the network compared to
influencers who post exclusively about e-cigarettes on Instagram and TikTok (H3). We
expected this pattern because micro-influencers who promote e-cigarettes and other content may
be forming ties with a larger number of influencers since they share a wider range of interests
(e-cigarettes along with other topics or contexts [e.g., fashion, gaming, healthy lifestyle] vs ecigarettes only).
METHODS
Data collection
Using media intelligence platform Meltwater93 and influencer profile collection method
adapted from prior research,22 we tracked influencers with public profiles on Instagram (N=104)
70
and TikTok (N=100) who posted promotional e-cigarette content in 2012-2022 on Instagram and
TikTok, had over 1,000 followers on either of the platforms, and high user engagement rates of
1%–25% per post (calculated as six-month average of the ratio of “likes” and comments to
followers).22 Verification of the profile characteristics of each selected influencer by two
University of Southern California (USC) graduate students and selection of the profiles with high
engagement rates minimized the risk that any of the influencers were bots. Then, we collected
publicly available usernames of Instagram (N=55,622) and TikTok (N=68,673) users who
commented on the tracked influencers’ e-cigarette-related posts over 2021-2022 to construct
networks of influencers and commenters.
In addition, two USC graduate student coders manually reviewed images and videos on
Instagram and TikTok posted in 2021-2022 (N = 15,480) by these influencers. The coders
documented if the influencers posted solely about e-cigarettes or e-cigarettes alongside other
contexts (e.g., fashion, gaming, healthy lifestyle). The coders also collected these influencers’
geolocation (country) from Meltwater that had these metadata.93 The study was approved by the
University of Southern California Institutional Review Board. Usernames of influencers or
commenters were not shown in the network graphs to keep them anonymous.
Network structure
We constructed directed networks (i.e., where ties among individuals may point in one
direction or both directions) among Instagram and TikTok influencers and users who commented
on the influencers’ posts. The structure of the constructed network is schematically shown in
Figure 1. Nodes (circles) represent influencers and commenters to their posts. Edges also known
as ties (lines) represent comments (a tie is formed if a user commented on an influencer’s post).
Edge weights represent multiple comments from one user to an influencer (shown by thicker lines
71
from nodes c-f to Profile 2). This network type can be described as star typology, where most of
the ties exist between two distinct types of nodes: influencers and commenters (and the number of
commenters by far exceeds the number of influencers), however, since some of the influencers are
among the commenters, ties also exist among influencers. This structure allowed us to construct
unipartite networks (vs bipartite where ties strictly exist between distinct types of nodes).
Measures
Individual node measures
We calculated centrality measures (indegree, outdegree and betweenness) for individual
nodes of the Instagram and TikTok networks that provide an opportunity to assess which nodes
in the network are most central and influential.
94 Indegree, i.e. the number of incoming ties (the
number of users who provided comments to an influencer posts) reflects popularity of a node; and
outdegree, i.e., the number of outgoing ties a node had to other nodes (the number of users an
influencer provided comments to) reflects a node’s influence in terms of outward information
distribution. Betweenness indicates the number of times a node is located on the shortest paths
between the pair of nodes. Betweenness assesses which nodes connect, i.e., serve as bridges
between the other nodes, facilitating interactions across the network, and thus have more influence
in the network. Nodes with high betweenness centrality are important for information flow.95
Network measures
We used density, mutuality, transitivity and assortativity to describe and compare
Instagram and TikTok networks’ structure. Density (the proportion of ties in the network out of all
possible ties) assesses whether the network is sparse or interconnected (dense). It can range from
0 (no ties between any nodes) to 1 (a graph where each pair of nodes is connected). Mutuality
assesses the number of reciprocal ties and range from 0 (no reciprocal ties) to 1 (all ties are mutual).
72
Transitivity (triads or the ties between three nodes in a network) assesses how effective or cohesive
the network is (nodes form triads in transitive networks that resemble commonly-observed human
interactions, i.e., the friends of my friends are my friends,
94 and reveal potential clusters within
networks, which makes it easier to understand social dynamics and tie formation). Transitivity can
range from 0 (no triads) to 1 (a network that consists of triads). Assortativity assesses the network’s
homophily (the tendency of nodes to connect with similar nodes based on an attribute, e.g., nodes
from the same region) or heterophily (the tendency of nodes to connect with dissimilar nodes, e.g.,
nodes from different regions). Assortativity can range from 1 (strong homophily) to -1 (strong
heterophily), with 0 representing no homophilic or heterophilic mixing.
In this study, since geolocation (region) information was only available for micro-influencers, but
not the rest of the commenters who were not among the micro-influencers, assortativity based on
geolocation was assessed for the micro-influencer networks on Instagram (104 nodes and 797
edges) and TikTok (100 nodes and 23 edges). We also calculated assortativity based on the
language attribute (i.e., the language used in comments to micro-influencers’ posts). FastText R
package was used for language detection in comments. Undetected languages in comments that
consisted of only slang or only emoji were coded as missing; multiple imputation method for multilevel categorical variables (R package mice) was also used as an alternative way to handle missing
values in the language attribute. Language assortativity was computed using two versions of the
language attribute: with coded and imputed missing values. The results on the coded version,
which were very similar to the imputed version, were reported.
To test H2, the assortativity was computed using assortativity function of the igraph
package in R software (version 4.2.2). The igraph was also used to compute all the other network
and individual measures and produce data visualization (Figures 2-4).
73
Statistical analysis
To test the exploratory hypotheses, H1 and H3, we applied Kruskal-Wallis and Dunn
tests with Bonferroni correction for multi-group differences to compare the mean ranks in
in/out degree and betweenness centrality. We compared these differences between individual
nodes on Instagram versus TikTok overall, by region micro-influencers posted from (North
America vs Asia vs Europe vs South America within Instagram and TikTok and between each
platform per the respective region [e,g., North American influencers on Instagram versus North
American influencers on TikTok]), and by content micro-influencers posted about (exclusively
about e-cigarettes versus e-cigarettes along with other content within Instagram and TikTok
and between each platform per the respective type of content [e.g., influencers who posted
exclusively about e-cigarettes on Instagram versus TikTok]). Kruskal-Wallis and Dunn tests
are non-parametric tests used to determine statistically significant differences between the
distributions of two or more independent groups by comparing ranks within the groups. We used
median values for comparisons. When median values were equal, but the tests showed statistically
significant differences between the groups (which is common for highly skewed data), we
provided mean values and range.
Since Instagram and TikTok micro-influencers had similar characteristics (promotional
content they posted about and the average number of followers) and since individual measures
(degree and betweenness centrality) were on the same scale and referred to homogeneous units
(e.g., micro-influencers on Instagram and TikTok from the same geographic regions), we
compared them directly without normalization. Normalized values on highly skewed data have
excessive zero values, which are difficult for interpretation. (However, as a sensitivity analysis,
74
we compared differences between influencer-only Instagram and TikTok networks using
normalized centrality measures).
RESULTS
The micro-influencers had on average 44,074 followers on Instagram and 89,497 on
TikTok.
The Instagram network included 104 influencers and 55,622 commenters to their posts. 100 out of
104 influencers were also among the commenters. The TikTok network included 100 influencers
and 68,673 commenters to their posts. Only 7 out of 100 influencers were also among the
commenters.
Individual node measures on Instagram and TikTok
Exploratory hypothesis. Centrality measures on Instagram versus TikTok. Microinfluencers on Instagram had 2.5 times as high median indegree or the number of incoming ties
(Median = 286, p-value = .008) as micro-influencers on TikTok, higher median outdegree or the
number of outgoing ties (Median = 5) than micro-influencers on TikTok (Median = 0, p-value <
.001) and much higher median betweenness centrality (Median = 32,958) than micro-influencers
on TikTok (Median = 0, p-value < .001) (Table 1).
Exploratory hypothesis. Centrality measures on Instagram versus TikTok by region. The
distributions of centrality measures (especially for betweenness centrality) in Instagram (Median
betweenness = 0; Mean = 183; range = 0 – 650,370) and TikTok (Median betweenness = 0; Mean
= 1.9; range = 0 – 49,960) networks were very skewed (most nodes had very few ties, while few
nodes had a very large number of ties, Table 1). Among the top 10 Instagram micro-influencers
with the highest indegree centrality (range: 1,689-5,364), five were from Indonesia, two were from
Canada, one from the U.S., and two were from Italy. Among the top 10 Instagram micro-
75
influencers with the highest betweenness (range: 416,984 – 768,662), four influencers were from
Indonesia, four were from the U.S., and two were from Canada. Among the top 10 TikTok microinfluencers with the highest indegree centrality (range: 2,050-10,011), two were from Spain, two
were from Indonesia, two from the Philippines, and one from each of these countries: U.S, Brazil,
France, and Malaysia. On TikTok, only seven micro-influencers had betweenness above zero
(range: 6,017 – 49,960). Five of these influencers were from the U.S., one was from Germany and
one from Spain.
Among North American, primarily U.S. and Canadian micro-influencers, those on
Instagram had 3.5 times as high median indegree (Median = 323, p-value = .01) as microinfluencers on TikTok. Among European micro-influencers, those on Instagram had 16 times as
high median indegree (Median = 267, p-value = .006) as micro-influencers on TikTok. Instagram
micro-influencers in each region: i.e., North America, Asia, South America, and Europe had higher
outdegree and betweenness centrality than TikTok micro-influencers in each of these regions,
respectively (Table 1).
Exploratory hypothesis. Centrality measures on Instagram versus TikTok by content.
Among micro-influencers who promoted e-cigarettes exclusively, those on Instagram had 18 times
as high indegree (Median = 92, p-value = < .001) as TikTok micro-influencers. However, among
micro-influencers who promoted e-cigarettes along with other themes (e.g., health, marijuana,
gaming, Supplementary figure 2), those on TikTok had higher indegree than Instagram microinfluencers, but the difference was not statistically significant (Table 1). Instagram microinfluencers who promoted e-cigarettes along with other themes had higher outdegree and
betweenness centrality compared to TikTok micro-influencers (p-value < .001), (Table 1).
76
Results from the sensitivity analysis that compared differences between individual
measures on influencer-only Instagram and TikTok networks using normalized centrality
measures (Supplementary table 1) were also consistent with the results using non-normalized
values.
H1. Centrality measures within Instagram and within TikTok.
On Instagram, on average, users who were micro-influencers provided 5 times as many
comments to other micro-influencers’ posts (Median outdegree = 5, p-value < .001) as users who
were not micro-influencers. On TikTok, on average, users who were not among micro-influencers
provided more comments (Median = 1.1) than users who were micro-influencers (Median = 0, pvalue < .001). However, commenters with the highest outdegree on TikTok were among microinfluencers.
H1. Centrality measures by region within Instagram and within TikTok. Among
Instagram micro-influencers, Asian had the highest indegree (Median = 395), followed by the
North American (Median = 323), European (Median = 267), and South American microinfluencers (Median = 156), but the differences among the groups were not statistically significant.
(The sample size for South American influencers was too small [n=2], and the results for this group
should be interpreted with caution).
The European influencers had the lowest outdegree. North American micro-influencers
had 2.25 times as high outdegree (Median = 4.5), Asian micro-influencers had 4.5 times as high
outdegree (Median = 9) and South American micro-influencers had 4.75 times as high outdegree
(Median = 9.5) as the European influencers (Median = 2) (p-value < .001).
Asian micro-influencers had the highest betweenness: almost twice as high (Median =
81,841) as North American micro-influencers, 1.5 times as high as the South American micro-
77
influencers, and over 7 times as high as the European micro-influencers. Pairwise differences
between the groups were statistically significant (p-value < .001), (Table 1).
Among TikTok micro-influencers, those from North America had higher outdegree (Mean
= 0.9, range = 0-7) and higher betweenness (Mean = 3,010, range = 0-29,292) compared to microinfluencers from Asia (p-value < .001).
H3. Centrality measures by content within Instagram and within TikTok.
Among Instagram micro-influencers, those who promoted e-cigarettes alongside lifestyle
or marijuana content had 1.3 as high outdegree (Median = 5) and 4 times as high betweenness
(Median = 69,404) as micro-influencers who promoted e-cigarettes exclusively, (p-value <
.001).
Among TikTok micro-influencers, those who promoted e-cigarettes alongside lifestyle or
marijuana content had higher outdegree (Mean = 0.26, range = 0-7) and higher betweenness (Mean
= 1,500, range = 0-49,960) compared to micro-influencers who promoted e-cigarettes
exclusively, (p-value < .001).
Network measures on Instagram and TikTok
Exploratory hypothesis. Instagram vs TikTok network. Both Instagram and TikTok
networks were sparse, i.e., had low density. They represent scale-free networks, which are
common on social media, where most nodes have very few ties, while few nodes have a very large
number of ties.96 However, the Instagram network was much denser compared to the TikTok
network (Figure 2). The Instagram network had 1.48 times as high density, 281 times as high
transitivity, and 85 times as high reciprocity as the TikTok network (Table 1).
H2. Tendency of tie formation on Instagram versus TikTok by region and language.
78
The Instagram network was characterized by homophilous regional clustering (i.e.,
tendency for connections among micro-influencers from the same regions); however, heterophily
(the presence of nodes from different regions in the same clusters) was also observed (Table 1,
Figure 2, Figure 3, and Figure 4). The TikTok network of micro-influencers consisted primarily
of isolates with several international nodes forming ties. The largest cluster on TikTok consisted
of influencers from different regions but was dominated by the influencers from North America
(i.e., U.S.), (Figure 3). The networks’ assortativity was moderate (0.34 for Instagram and 0.56 for
TikTok).
Most of the comments were made in English, Indonesian, Malaysian, French, Italian,
Spanish, and Portuguese (sample sizes of comments per each language are provided in the
footnotes for Table 1 and Supplementary table 1). The language assortativity (homophily) of the
Instagram (0.13) and TikTok full networks (0.07) was much weaker than regional assortativity
(homophily). Language assortativity of the micro-influencer Instagram network (-0.04) and
TikTok network (-0.21) represented dissortativity (i.e., tendency for connections among microinfluencers who used different languages), especially on TikTok.
DISCUSSION
This study explored and compared the networks of micro-influencers who promote ecigarettes on Instagram and TikTok. Overall, inconsistent with the Two-Step Flow of
Communication Model in its traditional interpretation where opinion leaders (usually celebrities)
mediate information flow from media sources to the public, the study showed that microinfluencers (non-celebrity influencers) create their own content (e.g., related to e-cigarettes) on
Instagram and TikTok, engage with each other’s content and directly interact with audiences who
consume this content. We found that the Instagram network was denser (more interconnected) and
79
had higher transitivity, i.e., considered more effective in terms of tie formation compared to the
TikTok network. On average, micro-influencers on Instagram also had higher in/out degree and
betweenness centrality. In addition, more comments on Instagram were provided by users who
were also micro-influencers than by users who were not among micro-influencers. This is
consistent with overall marketing trends on social media: i.e., Instagram has been the most popular
platform for micro-influencer marketing. However, lately TikTok has been gaining popularity
among brands as well,90 especially considering that micro-influencers cross-share their content on
Instagram and TikTok. Since, on average, more comments were provided by users who were not
among micro-influencers than by users who were micro-influencers on TikTok, this could indicate
that e-cigarette content on TikTok generate more general-user engagement compared to Instagram.
However, this finding should be interpreted with caution, since we did not have information on
whether commenters who were not among influencers posted any e-cigarette promotional content
or if they were just general TikTok users not engaged in any e-cigarette promotional activities.
Future studies should apply qualitative and machine learning methods to explore the characteristics
of general users (beyond micro-influencers) who engage with micro-influencer e-cigarette-related
content. (Although, regardless of the data collection and analysis method, metadata on general
users will only be available for public, but not for private accounts, which would require data
access authorization from these users).
The study results were consistent with and added more knowledge to prior research that
showed interaction among international influencers who promoted e-cigarettes on Instagram.22
In support of our hypothesis based on Waldo Tobler’s First Law of Geography, the Instagram
network was characterized by the presence of homophilous regional ties (tendency of nodes from
the same geographic regions to form ties); however, heterophily (tendency of nodes from different
80
geographic regions to form ties) was also present. The TikTok network consisted primarily of
isolated nodes. However, two clusters observed in the network consisted of international nodes
(influencers), indicating heterophily. Linguistic homophily of both Instagram and TikTok
networks was weaker compared to the regional homophily, which was unexpected. Moreover,
when the networks were restricted to micro-influencers only, we observed the tendency for
linguistic heterophily, especially on TikTok. Multiple languages were present in comments,
including English, Indonesian, Malaysian, Spanish, French and Portuguese. The finding suggests
that engagement with promotional Instagram and TikTok posts may occur in the absence of
linguistic commonalities, since context of this promotional content can be inferred from visual
cues. Moreover, comments in such posts are limited to very brief reactions or descriptions of the
marketed products, including emoji, product names, and hashtags, easily understood in any
language. It is also possible that comments get translated into a native language of a social media
user on their mobile devices. The findings from this study add evidence of cross-border promotion
in e-cigarette influencer marketing, which is concerning from the tobacco regulatory science
standpoint. The problem of cross-border promotion is described in the World Health Organization
(WHO) Framework Convention on Tobacco Control.53 Tobacco Advertising, Promotion, and
Sponsorship (TAPS) is cross-border whenever the content created, uploaded, or broadcast in one
country may be consumed or shared in another, thereby crossing geographical borders.53 In other
words, since influencers from different countries engage (comment on) each other’s social media
posts, they potentially expose their followers (that could include adolescents) to broader e-cigarette
promotional materials. Based prior research, micro-influencers who promoted e-cigarettes on
Instagram had followers among adolescents and most of these micro-influencers did not have agegating (access restriction to e-cigarette content for those younger than 18 or 21 years of age).
22
81
Moreover, influencers who post about e-cigarettes are from countries with different levels of
e-cigarette marketing regulations. For example, Asian: i.e., Indonesian micro-influencers,
who, in support of our hypothesis, were among the most central and influential in the networks,
actively promote e-cigarettes on social media, and Indonesia has no regulations of e-cigarette
marketing.97 If the U.S. and Indonesian influencers engage with each other’s content (which is
possible since we observed the presence of international micro-influencers in the same
clusters), it increases the risk for the U.S. users, including youth, to encounter e-cigarette
content promoted by Indonesian influencers. Collaboration among international public health
agencies might be needed to enforce regulation of influencer marketing on social media as well
as to develop global educational campaigns to inform youth about risks of exposure to
influencer marketing related to harmful substances on social media.
Micro-influencers have been also increasingly using audio-visual effects, including
stylization of influencers to look young,44 promotion of e-cigarettes in youth-oriented or
lifestyle contexts (e.g., nature, fashion, healthy lifestyle, gaming, technology) or next to other
substances (e.g., marijuana), and the use of youth-appealing design of e-cigarette devices
(bright colors,23 design resembling toys, candy, or other consumables).
82 In support of our
hypothesis, this study showed that influencers who promote e-cigarettes alongside lifestyle
content (e.g., health, gaming, fashion) or other substance use (e.g., marijuana) versus promoting
e-cigarettes alone had higher median number of ties and higher median betweenness, indicating
that they are more active and influential in the network. Such pattern increases the risk that
influencers who post on a variety of topics besides e-cigarettes could reach users who are not
interested in tobacco-related content and expose these users to harmful imagery of e-cigarettes.
Regulating influencer marketing features (e.g., prohibiting depiction of lifestyle context[e.g.,
82
health-related activities, gaming, fashion] next to e-cigarette products) could be challenging
because it could fall under the umbrella of free speech and be protected by the First
Amendment of the U.S. Constitution.55 However, federal agencies could consider prohibiting
e-cigarette brands from hiring influencers under a certain age (the FDA supports e-cigarette
brands’ intentions to use older models or influencers in their marketing materials but does not
require it),
75,98 ban influencer advertising of e-cigarettes altogether, or penalize brands and
influencers for non-compliance with federal marketing regulations, e.g., not restricting access
to e-cigarette marketing content to youth.57 Additionally, improvement in the enforcement of
social media community guidelines is necessary, since only a fraction of tobacco-related
content prohibited by platforms’ community guidelines (e.g., on Instagram and TikTok) gets
removed by the platforms’ algorithms.
15,16 Although it can be challenging to fully remove all ecigarette content online, the presence of branded posts (where e-cigarette brand sponsorship
disclosures are indicated explicitly and thus should be relatively easy for social media platforms
to flag), suggests insufficient enforcement.99 Using machine learning approaches in social media
algorithms to detect any type of e-cigarette content could also facilitate its prompt removal.25
Limitations. Since accessing the Instagram Application Programming Interface directly became
impossible for non-commercial parties after 2016, we used a commercial social media listening
platform Meltwater for data collection. Although Meltwater, based on their agreement with
Instagram, provides access to a rich dataset of over 10 million influencers with public profiles who
post on a variety of topics, the sample from which we selected influencers who promote ecigarettes might have been not random and should be considered a convenience sample. We did
not have geolocation information for commenters who were not among the influencers and had
limited metadata on these users overall. For example, it was not possible to infer whether
83
commenters who were not among influencers posted any e-cigarette promotional content. This
limits our ability to conclude what other types of users (beyond micro-influencers themselves)
engage with micro-influencer e-cigarette-related content. The language was not identified in a
large proportion of comments that contained only emoji or slang, which could have bias the
assortatitivty measure. Concerns are mitigated by the fact that the assortativity coefficients
computed on the language attribute, where the undetected languages were coded as missing and
where they were imputed, were similar, which indicates robustness of the finding. Data included
posts and comments from 2021-2022. More recent data (2023-2024) was not collected due to
difficulties accessing the Instagram API to collect usernames of commenters.
Conclusion This study explored and compared the connections among micro-influencers who
promote e-cigarettes on Instagram and TikTok, social media platforms highly popular among
youth. International borders on social media are absent, which is a concern from the tobacco
regulatory science standpoint. Micro-influencers who promote e-cigarette content on these
platforms interact not only with micro-influencers from the same geographic regions, but also
with micro-influencers from different geographic regions, potentially exposing their followers
(including adolescents) to broader e-cigarette promotional materials. This finding adds evidence
of cross-border promotion in e-cigarette influencer marketing. The problem of cross-border
promotion is described in the World Health Organization (WHO) Framework Convention on
Tobacco Control.53 Tobacco Advertising, Promotion, and Sponsorship (TAPS) is cross-border
whenever the content created, uploaded, or broadcast in one country may be consumed or shared
in another, thereby crossing geographical borders. Influencers who post about different content
besides e-cigarettes are more active and influential in the networks compared to influencers who
post only about e-cigarettes. Such pattern increases the risk that influencers who post on a variety
84
of topics besides e-cigarettes could reach users who are not interested in tobacco-related content
and expose these users to harmful imagery of e-cigarettes. Overall, the study’s findings emphasize
the need for stronger influencer marketing regulation on social media.
Contributorship Statement Concept and design – JV. Acquisition, analysis, or interpretation of
data—all authors. Draft of the manuscript - JV. Critical revision of the manuscript for important
intellectual content—all authors. Statistical analysis – JV. Obtaining funding – JV, JBU.
Administrative, technical, or material support – JV, JBU. Supervision – JBU. All co-authors
approved the final version.
Ethics approval The University of Southern California Institutional Review Board (UP-21-
00352) approved all study procedures.
85
Table 1. Individual node centrality measures and network measures of directed networks of 104
influencers and 55,622 commenters on Instagram and 100 influencers and 68,673 commenters on
TikTok.
INSTAGRAM TIKTOK
Nodes 55,626 68,766
Influencers 104 100
Commenters 55,522 (excluding commenters among influencers) 68,666 (excluding commenters among influencers)
Edges
(including
multiple ties)
328,893 88,819
Unique edges
(excluding
multiple ties)
67,196 69,495
Network measures
Density 0.00002171679 0.00001469645
Reciprocity 0.006 0.00007
Transitivity 0.0008128361 0.00000288988
Assortativity1
(by region)
0.34 0.56
Assortativity2
(by language)
0.13 0.07
Individual measures
Centrality N (%) nodes Median Mean SD Range
(minmax)
N, nodes (%) Median Mean SD Range
(minmax)
p-value
Full network
Indegree 55,626 (100%) 1 1.21 48.59 0 –
5,364
68,766
(100%)
0 1.01 58.4 0 –
10,011
< .001
Outdegree 55,626 (100%) 1 1.21 1.07 0 – 35 68,766
(100%)
1 1.01 0.14 0 – 10 .01
Betweenness 55,626 (100%) 0 183 7,711 0 –
650,370
68,766
(100%)
0 1.9 248 0 –
49,960
< .001
Commenters
(outdegree)
55,522 (99.8%) 1 1.2 0.98 1- 34 68,666
(99.8%)
1.01 1.01 0.13 1-10 < .001
Influencers
Indegree 104 (0.2%) 286 646 924 3 –
5,364
100 (0.1%) 115 695 1,372 1 –
10,011
.008
Outdegree 104 (0.2%) 5 7.66 7.57 0 – 35 100 (0.1%) 0 0.23 1 0 -7 < .001
Betweenness 104 (0.2%) 32,958 97,965 149,786 0 -
650,370
100 (0.1%) 0 1,320 6,401 0 -
49,960
< .001
Influencer
network
N (%) nodes,
(out of 104
influencers)
Median Mean SD Range N (%) nodes,
(out of 100
influencers)
Median Mean SD Range
Indegree
(by region)
N. Americaa 36 (35%) 323 656 992 10 -
5,364
20 (20%) 91 275 496 1-2,190 .01
Asiab 44 (42%) 395 685 904 16 –
4,511
50 (50%) 262 737 1,082 1-5,470 0.6
S. Americac 2 (2%) 156 156 159 43-268 4 (4%) 1,384 1,789 1,792 81-4,307 0.35
Europed 22 (21%) 267 597 923 3-3,414 25 (25%) 17 796 2,098 1-10,001 .006
Outdegree
(by region)
N. Americaa 36 (35%) 4.5 7.75 8.55 0 – 35 20 (20%) 0b 0.9 2 0-7 < .001
Asiab 44 (42%) 9 10 7.11 0 – 26 50 (50%) 0 0 0 0 < .001
S. Americac 2 (2%) 9.5 9.5 6.36 5-14 4 (4%) 0 0 0 0 0.16
Europed 22 (21%) 2abc 3.09 4.68 0-18 25 (25%) 0 0.2 0.71 0-3 < .001
Betweenness
(by region)
86
N. Americaa 36 (35%) 46,471d 177,354 236,577 0 –
768,662
20 (20%) 0b 3,010 7,523 0-29,292 < .001
Asiab 44 (42%) 81,841acd 137,740 143,551 0 –
546,025
50 (50%) 0 0 0 0 < .001
S. Americac 2 (2%) 52,677abd 52,677 69,840 3,292 -
102,062
4 (4%) 0 0 0 0 0.16
Europed 22 (21%) 11,102 73,199 112,235 0 –
329,549
25 (25%) 0 2,873 10,739 0-49,960 < .001
Indegree
(by content)
e-cigarette
onlya
25 (24%) 289 551 637 9 –
2,628
10 (10%) 5 16 25 1 – 78 < .001
e-cigarette &
other contentb
78 (76%) 286 684 1,004 3 –
5,364
88 (88%) 149 785 1,440 1-10,011 0.1
Outdegree
(by content)
e-cigarette
onlya
25 (24%) 4b 6 6.8 0 – 26 10 (10%) 0b 0 0 0 < .001
e-cigarette &
other contentb
78 (76%) 5 8.2 7.8 0 – 35 88 (88%) 0 0.26 1.07 0 – 7 < .001
Betweenness
(by content)
e-cigarette
onlya
25 (24%) 16,832b 108,738 177,996 0 –
686,014
10 (10%) 0b 0 0 0 < .001
e-cigarette &
other contentb
78 (76%) 69,404 146,441 179,315 0 –
768,662
88 (88%) 0 1,500 6,809 0 –
49,960
< .001
1
Assortativity by region was computed on the network of micro-influencers (104 nodes and 797 edges on Instagram; and 100
nodes and 23 edges on TikTok), excluding commenters who were not micro-influencers, since their regional and content
attributes were not available. 2
Assortativity by language. Language of comments on Instagram included English (n=15,218 [27%]), Indonesian (n=7,187
[13%]),
Malaysian (n=1,160 [2%]), Italian (n = 2,356 [4%]), Other (n = 4,566 [8%]), emoji or unidentified (n=25,139 [45%]).
Language of comments on TikTok included English (n=10,013 [15%]), Indonesian (n=11,536 [17%]), Malaysian (n=2,856
[4%]), Spanish (n = 4,454 [6%]), French (n=2,387[3%]), Other (n = 9,337 [14%]), emoji or unidentified (n=28,183 [41%]).
Kruskal-Wallis non-parametric test was conducted to assess differences in sample distributions (which often, but not exclusively,
are due to differences in median values) between individual measures on Instagram and TikTok and within Instagram and
TikTok. Dunn Test with Bonferroni correction was conducted to assess multi-group differences by region. P-values indicate
statistically significant differences in individual measures between Instagram and TikTok. Superscripts indicate statistically
significant differences ( < .05) by region and content influencers post about within each of the two platforms. (For example, a
superscript d next to a median value indicates that the distribution of values for this specific region is statistically significantly
different from the distribution of values for the region represented by this letter: i.e., from Europe). Individual measures are not
normalized.
87
Figure 1. The structure of the networks of influencers and commenters to their posts on
Instagram and TikTok
88
Figure 2. Weighted unipartite networks of Instagram micro-influencers (n=104) and
commenters to their posts (n=55,622) and TikTok micro-influencers (n=100) and
commenters to their posts (n=68,673)
89
Figure 3. Weighted unipartite networks of Instagram (n=104) and TikTok (n=100) microinfluencers who commented on each other’s posts.
90
Figure 4. Instagram communities by region.
91
Supplementary table 1. Normalized centrality measures and network measures assessed on the
directed network of 104 influencers on Instagram and 100 influencers on TikTok.
INSTAGRAM TIKTOK
Nodes 104 100
Edges (excluding
multiple ties)
797 23
Network measures
Density 0.07 0.002
Reciprocity 0.38 0.04
Transitivity 0.44 0.18
Assortativity (by
region)
0.34 0.56
Assortativity (by
language1
)
-0.004 -0.21
Individual measures
Centrality,
normalized
N, nodes
(%)
Median Mean SD Range N, nodes
(%)
Median Mean SD Range pvalue
Indegree overall
104 (100%)
0.05 0.07 0.07 0.34
100 (100%)
0 0 0.01 0.07 < .001
Outdegree overall 0.05 0.07 0.08 0.33 0 0 0.01 0.03 < .001
Betweenness
overall
0.01 0.03 0.04 0.2 0 0 0 0 < .001
Indegree
(by region)
N. Americaa 36 (35%) 0.04d 0.08 0.08 0.34 20 (20%) 0b 0.01 0.02 0.07 .005
Asiab 44 (42%) 0.09d 0.1 0.07 0.25 50 (50%) 0 0 0 0 0.2
S. Americac 2 (2%) 0.04 0.09 0.06 0.09 4 (4%) 0 0 0 0 0.35
Europed 22 (21%) 0.01 0.03 0.05 0.17 25 (25%) 0 0 0.01 0.03 .003
Outdegree
(by region)
N. Americaa 36 (35%) 0.04d 0.07 0.08 0.29 20 (20%) 0b 0.01 0.01 0.03 < .001
Asiab 44 (42%) 0.09d 0.1 0.08 0.33 50 (50%) 0 0 0 0.01 < .001
S. Americac 2 (2%) 0.09 0.04 0.05 0.07 4 (4%) 0 0 0.01 0.01 0.15
Europed 22 (21%) 0.02 0.02 0.02 0.07 25 (25%) 0 0 0 0.01 < .001
Betweenness (by
region)
N. Americaa 36 (35%) 0 0.04 0.06 0.2 20 (20%) 0b 0 0 0-0.003 < .001
Asiab 44 (42%) 0.01 0.02 0.03 0.13 50 (50%) 0 0 0 0 < .001
S. Americac 2 (2%) 0.05 0.05 0.07 0.1 4 (4%) 0 0 0 0 0.16
Europed 22 (21%) 0 0.02 0.04 0.13 25 (25%) 0 0 0 0 < .001
Indegree
(by content)
e-cigarette onlya 25 (24%) 0.03 0.06 0.07 0.25 10 (10%) 0 0 0 0 < .001
e-cigarette &
other contentb
78 (76%) 0.05 0.08 0.08 0.34 88 (88%) 0 0 0.01 0.07 .01
Outdegree
(by content)
e-cigarette onlya 25 (24%) 0.04 0.06 0.07 0.29 10 0 0 0 0 < .001
e-cigarette &
other contentb
78 (76%) 0.05 0.08 0.08 0.33 88 0 0 0.01 0.03 < .001
Betweenness (by
content)
e-cigarette onlya 25 (24%) 0 0.02 0.05 0.2 10 0 0 0 0 < .001
e-cigarette &
other contentb
78 (76%) 0.01 0.03 0.04 0.18 88 0 0 0 0 < .001
1
Language of micro-influencers on Instagram included English (n=24 [23%]), Indonesian (n=7 [7%]), Other (n = 5 [5%]), emoji
or unidentified (n=68 [65%]). Language of micro-influencers on TikTok included English (n=18 [18%]), Indonesian (n=18
92
[18%]), Malaysian (n=6 [6%]), Spanish (n = 4 [4%]), Portuguese (n=4 [4%]), Other (n = 8 [8%]), emoji or unidentified (n=42
[42%]).
Kruskal-Wallis non-parametric test was conducted to assess differences in sample distributions (which often, but not exclusively,
are due to differences in median values) between individual measures on Instagram and TikTok and within Instagram and
TikTok. P-values indicate statistically significant differences in individual measures between Instagram and TikTok. Superscripts
indicate statistically significant differences ( < .05) by region and content influencers post about within each of the two platforms.
(For example, a superscript d next to a median value indicates that the distribution of values for this specific region is statistically
significantly different from the distribution of values for the region represented by this letter: i.e., from Europe). Individual
measures are normalized.
93
Instagram TikTok
Supplementary Figure 1. Distribution of the most popular content Instagram (N=104) and
TikTok (N=100) micro-influencers posted about in 2021-2022 along with e-cigarettes.
Percentage indicates the proportion of influencers who post about the identified content.
16%
20%
20%
20%
28%
29%
41%
marijuana
gaming
health
food
nature
only e-cigarettes
clothing
1%
10%
16%
19%
34%
38%
59%
marijuana
only e-cigarettes
gaming
health
nature
food
clothing
94
DISCUSSION
The findings from the three studies of this dissertation demonstrated that e-cigaretterelated micro-influencer marketing (especially those marketing posts promoting healthy
lifestyles) is a viable new form of indirect marketing on social media, with active networking
among international micro-influencers who create e-cigarette promotional content that can
diminish harm perceptions, increase appeal of, and susceptibility to use, e-cigarettes among
adolescents. Not only does this new form of advertising have an impact on youth, but it also
provides an opportunity for youth to get involved in content creation and become microinfluencers. Brands, including e-cigarette companies, often offer paid partnerships to microinfluencers to market e-cigarette products.34 Market research showed that over 50% of surveyed
Gen Z respondents (13-26 year of age) considered social media influencing a reputable career
choice.35 The respondents indicated that they would be willing to quit their day job if influencer
job would pay enough for their lifestyle.35 Qualitative research (e.g., interviewers with microinfluencers) is needed to understand the role financial incentives from partnerships with
tobacco brands and from social media monetization (e.g., getting paid by engagement or by
clicks on a brand name tagged in a social media post) play in youths’ decision to become
content creators to promote e-cigarette products on social media.
While influencers may advertise unhealthy products to pursue commercial interests,
consumers of their content (including youth) may perceive these influencers as trusted, credible,
and authentic sources,46 especially when influencers promote e-cigarettes as part of their daily
lifestyle (e.g., working out, playing video games). Adding to the existing scientific literature, this
study demonstrated that these marketing tactics could contribute to normalization of e-cigarette
use (higher appeal and lower harm perceptions of, and susceptibility to use, e-cigarettes), if
95
influencers are perceived as credible, but not if influencers are perceived as non-credible. In other
words, the findings from this dissertation are theoretically consistent with the Prototype
Willingness Model,48,60 suggesting that people are more willing to engage in a health risk
behavior to the extent that they have a positive view of the prototypical person who performs that
behavior. Public health authorities may harness influencers for anti-tobacco messaging and in
health literacy campaigns about harmful effects of e-cigarette use46,73 or disseminating other
positive public health messages.46 Consistent with PWM, pro-e-cigarette influencers used in
intervention campaigns could be portrayed in a more negative light, while anti-e-cigarette
influencers could be shown in a more positive light.
More research is needed to assess if other features present in micro-influencer marketing
but not assessed in this dissertation (e.g., sound features such as narration, music, or Autonomous
Sensory Meridian Response (ASMR) audio-visual triggers100 associated with pleasant
sensation) have similar effects on e-cigarette perceptions by youth. Analyzing narration could
provide new insights about influencer marketing perceptions. Based on PWM, narration, which is
usually present in longer (at least 30-second long) Instagram, TikTok or YouTube videos, could
more likely involve analytic processing via a reasoned path that requires more cognitive efforts
than images or short 10-second videos used in this project that were likely processed by the survey
participants without much cognitive efforts, via a social reaction path. As such, influencers and
their marketing content could be perceived differently depending on the type (short or long form)
of e-cigarette marketing content.
The perceived age of influencers (e.g., younger than 21 versus older models) is another
feature that can be assessed in future research. The FDA stated in the marketing granted orders
to several e-cigarette companies (NJOY, RJ Reynolds and Logic Technology) that it supports
96
their intention to use older models (over the age of 25-45) since young models may increase
youth appeal to e-cigarette content these models promote.
56,101 Indeed, the preliminary findings
from our focus group research among California high school students demonstrated that
“young-looking” influencers (perceived as younger than 21) were considered by youth as more
relatable than “older-looking” influencers (perceived as 30 or older). However, this does not
mean that content produced by older-looking influencers has no effect on youth. Additional
research that assesses the effect of influencer age could contribute to the regulatory narrative
about the need for public health agencies to require that tobacco brands use models over a
specific age or set a standardized required age limit for the models (as opposed to only
supporting brands’ intentions to use older models), or to go even further and require that
tobacco companies do not partner with social media influencers of any age, since any
influencers could create youth-appealing e-cigarette-related content and enhance this appeal
by using a variety of marketing features.57
The experimental study design used in this project did not provide full assurance that the
message (visual content) was fully disentangled from the messenger (micro-influencer). Future
research should explore direct measures assessed via e.g., eye-tracking experiments where an eye
fixation on either influencers’ faces or e-cigarette devices are separately measured followed by
self-reported assessment of influencer credibility and e-cigarette perceptions, which could
potentially allow to better separate the effects of the messenger from the message.
The social network analysis findings of this project were, overall, inconsistent with the
Two-Step Flow of Communication Model in its traditional interpretation where opinion leaders
(usually celebrities) mediate information flow from media sources to the public. This study showed
that micro-influencers (non-celebrity influencers) create their own content (e.g., related to e-
97
cigarettes) on Instagram and TikTok, engage with each other’s content and directly interact with
audiences who consume this content. The findings were consistent with and added more
knowledge to prior research that discovered an interconnected network of international
influencers who partnered with the same e-cigarette brands to promote their products on
Instagram.22 This project also showed interaction among international influencers: mostly on
Instagram, but also on TikTok. Asian, American, and European micro-influencers were present
in the same social network clusters, which indicates interaction (engagement with each other’s
e-cigarette content on social media) of influencers from different countries with different levels
of e-cigarette marketing regulations. For example, Indonesian micro-influencers actively
promote e-cigarettes on social media but have no regulations of e-cigarette marketing.97 Since
the U.S. and Indonesian influencers engage with each other’s content (as this project has
shown), it increases the risk for the U.S. youth to encounter youth-appealing e-cigarette content
promoted by Indonesian influencers. The finding adds to the evidence of cross-border promotion
described in the World Health Organization (WHO) Framework Convention on Tobacco
Control.53 Tobacco Advertising, Promotion, and Sponsorship (TAPS) is cross-border whenever
the content created, uploaded, or broadcast in one country may be consumed or shared in another,
thereby crossing geographical borders.53 International collaboration among public health agencies
might be needed to enforce regulation of influencer marketing on social media.
The social network analysis also showed that micro-influencers who post about ecigarettes along with other content (e.g., healthy lifestyle, gaming, fashion) occupy more
central positions in the Instagram and TikTok networks than those who post about e-cigarettes
only. Such pattern increases the risk that influencers who post on a variety of topics besides e-
98
cigarettes could reach users who are not interested in tobacco-related content and expose these
users to harmful imagery of e-cigarettes.
Finally, the findings from this project emphasize the need for further influencer marketing
regulation. Influencer e-cigarette-related marketing on social media is not sufficiently regulated
by platforms’ community guidelines or federal legislation. Although the most widely-used and
well-known social media platforms (e.g., Facebook, Instagram, YouTube, Twitter, TikTok,
Reddit, Snapchat) prohibit paid advertisements for tobacco products,41,59 few (Facebook,
Instagram and TikTok) explicitly prohibit influencers from promoting nicotine/tobacco products.59
Moreover, influencers still promote e-cigarettes on the platforms that have restrictions in place.
Regulating influencer marketing features (e.g., prohibiting depiction of healthy lifestyle
context next to e-cigarette products) could be challenging because specific framing of content
could be protected by the First Amendment of the U.S. Constitution.55 However, federal
agencies and social media users (including youth) should be informed about such marketing
tactics, especially if such content is youth-appealing. Regulating attempts could still be made
since we demonstrated that such content could make influencers and e-cigarettes look
appealing to youth. Authorities could also consider prohibiting e-cigarette brands from hiring
influencers under a certain age,75,98 ban influencer advertising of e-cigarettes altogether, or
penalize brands and influencers for non-compliance with federal marketing regulations such
as restricting access to e-cigarette marketing content to youth or following the Federal Trade
Commission (FTC) guidelines for sponsorship disclosures. According to the U.S. Food and Drug
Administration (FDA) and the Federal Trade Commission (FTC) guidelines, if influencers have
any material connection with a tobacco brand – meaning that they have been paid or given
something of value to tout the product - such relationships need to be disclosed in their social
99
media posts.74 However, enforcement is inconsistent, and many influencers’ posts have ambiguous
or absent disclosures of their partnerships with tobacco brands.22 Influencers who promote
products on behalf of tobacco brands are also required to restrict access to e-cigarette marketing
content to only individuals who are at or above the federal minimum age (i.e., 21) of sale of tobacco
products. However, research has demonstrated that micro-influencers have not been consistently
using youth age restrictions to their social media accounts.22 Additionally, further improvement
in the enforcement of social media community guidelines is necessary.15,16 Elaborate on using
branded content feature, which is not working properly because sponsored content is still
present (Linnea et all) along with enforcing FTC disclosures.74,99
Conclusion
This project explored the emerging and insufficiently regulated marketing of microinfluencers, their network connections on social media platforms popular among youth, and
the effects of content the micro-influencers post about on adolescent perceptions of ecigarettes. Influencers who promote e-cigarettes on Instagram or TikTok often use covert
marketing tactics to make their content more appealing, including e-cigarette promotion in
positive lifestyle contexts (e.g., healthy lifestyle) or next to marijuana products. As this project
demonstrated, these tactics could diminish adolescents’ harm perceptions of e-cigarettes,
increase appeal and susceptibility to e-cigarette use among adolescents and contribute to
normalization of e-cigarette use. Micro-influencers from different geographic regions engage
with each other’ content potentially exposing their mutual audiences to a variety of e-cigarette
content. Micro-influencers who post about e-cigarettes along with other content (e.g., healthy
lifestyle, gaming, fashion) occupy more central positions in the Instagram and TikTok
networks than those who post about e-cigarettes only. Such pattern increases the risk that
100
influencers who post on a variety of topics besides e-cigarettes could reach users who are not
interested in tobacco-related content and expose these users to harmful imagery of e-cigarettes.
The findings from this project emphasize the need for strengthening influencer marketing
regulation, which is not sufficiently regulated by social media platforms’ community guidelines
or federal legislation.
101
References
1. Birdsey J. Tobacco Product Use Among U.S. Middle and High School Students — National
Youth Tobacco Survey, 2023. MMWR Morb Mortal Wkly Rep. 2023;72.
doi:10.15585/mmwr.mm7244a1
2. Cooper M. Notes from the Field: E-cigarette Use Among Middle and High School Students
— United States, 2022. MMWR Morb Mortal Wkly Rep. 2022;71.
doi:10.15585/mmwr.mm7140a3
3. Health CO on S and. Smoking and Tobacco Use; Electronic Cigarettes. Centers for Disease
Control and Prevention. Published November 2, 2023. Accessed March 26, 2024.
https://www.cdc.gov/tobacco/basic_information/e-cigarettes/about-e-cigarettes.html
4. Truth Initiative: inspiring lives free from smoking, vaping, and nicotine. Accessed February
26, 2024. https://truthinitiative.org/
5. CDC. Quick Facts on the Risks of E-cigarettes for Young People. Centers for Disease
Control and Prevention. Published January 16, 2024. Accessed March 23, 2024.
https://www.cdc.gov/tobacco/basic_information/e-cigarettes/Quick-Facts-on-the-Risks-of-Ecigarettes-for-Kids-Teens-and-Young-Adults.html
6. Liu PI, Lin MN, Ho PS, et al. Prediction and potential risk factors for electronic cigarette use
behaviors among adolescents: a pilot study in Chiayi, Taiwan. Front Public Health. 2023;11.
doi:10.3389/fpubh.2023.1140615
7. Valente TW, Piombo SE, Edwards KM, Waterman EA, Banyard VL. Social Network
Influences on Adolescent E-cigarette Use. Subst Use Misuse. 2023;58(6):780-786.
doi:10.1080/10826084.2023.2188429
8. Barrington-Trimis JL, Berhane K, Unger JB, et al. The E-cigarette Social Environment, Ecigarette Use, and Susceptibility to Cigarette Smoking. Journal of Adolescent Health.
2016;59(1):75-80. doi:10.1016/j.jadohealth.2016.03.019
9. Donaldson SI, Dormanesh A, Perez C, Majmundar A, Allem JP. Association Between
Exposure to Tobacco Content on Social Media and Tobacco Use: A Systematic Review and
Meta-analysis. JAMA Pediatr. 2022;176(9):878-885. doi:10.1001/jamapediatrics.2022.2223
10. Vassey J, Galimov A, Kennedy CJ, Vogel EA, Unger JB. Frequency of social media use and
exposure to tobacco or nicotine-related content in association with E-cigarette use among
youth: A cross-sectional and longitudinal survey analysis. Preventive Medicine Reports.
2022;30:102055. doi:10.1016/j.pmedr.2022.102055
11. Vogel EA, Ramo DE, Rubinstein ML, et al. Effects of Social Media on Adolescents’
Willingness and Intention to Use E-Cigarettes: An Experimental Investigation. Nicotine Tob
Res. doi:10.1093/ntr/ntaa003
102
12. Kim M, Popova L, Halpern-Felsher B, Ling PM. Effects of e-Cigarette Advertisements on
Adolescents’ Perceptions of Cigarettes. Health Communication. 2019;34(3):290-297.
doi:10.1080/10410236.2017.1407230
13. Branded Content Policies. Facebook Business Help Center. Accessed May 20, 2020.
https://www.facebook.com/business/help/221149188908254
14. Vogels E a, Gelles-Watnick R. Teens and social media: Key findings from Pew Research
Center surveys. Pew Research Center. Accessed October 1, 2023.
https://www.pewresearch.org/short-reads/2023/04/24/teens-and-social-media-key-findingsfrom-pew-research-center-surveys/
15. Vassey J, Donaldson SI, Dormanesh A, Allem JP. Themes in TikTok Videos Featuring Little
Cigars and Cigarillos: Content Analysis. Journal of Medical Internet Research.
2022;24(11):e42441. doi:10.2196/42441
16. Rutherford BN, Sun T, Lim CCW, et al. Changes in Viewer Engagement and Accessibility
of Popular Vaping Videos on TikTok: A 12-Month Prospective Study. International Journal
of Environmental Research and Public Health. 2022;19(3):1141.
doi:10.3390/ijerph19031141
17. Huang J, Duan Z, Kwok J, et al. Vaping versus JUULing: how the extraordinary growth and
marketing of JUUL transformed the US retail e-cigarette market. Tobacco Control.
2019;28(2):146-151. doi:10.1136/tobaccocontrol-2018-054382
18. Belluz J. The vape company Juul said it doesn’t target teens. Its early ads tell a different
story. Vox. Published January 25, 2019. Accessed April 29, 2024.
https://www.vox.com/2019/1/25/18194953/vape-juul-e-cigarette-marketing
19. Juul to pay N.C. $40M and stop targeting kids after igniting vaping “epidemic“ | Ars
Technica. Accessed April 29, 2024. https://arstechnica.com/science/2021/06/juul-to-pay-nc40m-and-stop-targeting-kids-after-igniting-vaping-epidemic/
20. FDA takes new steps to address epidemic of youth e-cigarette use, including a historic action
against more than 1,300 retailers and 5 major manufacturers for their roles perpetuating
youth access. FDA. Published March 24, 2020. Accessed April 29, 2024.
https://www.fda.gov/news-events/press-announcements/fda-takes-new-steps-addressepidemic-youth-e-cigarette-use-including-historic-action-against-more
21. What Is a Micro-Influencer? Why Do Brands Use Them? | Built In. Accessed November 4,
2023. https://builtin.com/marketing/micro-influencer
22. Vassey J, Valente T, Barker J, et al. E-cigarette brands and social media influencers on
Instagram: a social network analysis. Tobacco Control. Published online February 6, 2022.
doi:10.1136/tobaccocontrol-2021-057053
103
23. Vassey J, Allem JP, Barker J, et al. E-cigarette use and promotion by social media
influencers during videogame play on Twitch. Tobacco Control. Published online October 8,
2021. doi:10.1136/tobaccocontrol-2021-056828
24. Topic: Social media and Generation Z in the United States. Statista. Accessed March 1,
2024. https://www.statista.com/topics/10943/social-media-and-generation-z-in-the-unitedstates/
25. Vassey J, Kennedy CJ, Chang HCH, Smith AS, Unger JB. Scalable surveillance of ecigarette products on Instagram and TikTok using computer vision. Nicotine & Tobacco
Research. Published online November 8, 2023:ntad224. doi:10.1093/ntr/ntad224
26. The Future of Influencer Marketing: Trends to Watch Out for in 2023. Accessed November
4, 2023. https://www.linkedin.com/pulse/future-influencer-marketing-trends-watch-out2023-kosmc
27. 37% of Consumers Trust Social Media Influencers Over Brands. Accessed November 20,
2023. https://www.oracle.com/news/announcement/consumers-turn-to-social-mediainfluencers-2022-05-03/
28. Childers CC, Lemon LL, Hoy MG. #Sponsored #Ad: Agency Perspective on Influencer
Marketing Campaigns. Journal of Current Issues & Research in Advertising.
2019;40(3):258-274. doi:10.1080/10641734.2018.1521113
29. Jones EE, Davis KE. From Acts To Dispositions The Attribution Process In Person
Perception11Much of the research reported herein was supported by National Science
Foundation Grants 8857 and 21955 to the first author. In: Berkowitz L, ed. Advances in
Experimental Social Psychology. Vol 2. Academic Press; 1965:219-266. doi:10.1016/S0065-
2601(08)60107-0
30. YEC. Council Post: Five Steps For Finding The Right Influencer For Your Brand. Forbes.
Accessed May 17, 2024. https://www.forbes.com/sites/theyec/2022/06/28/five-steps-forfinding-the-right-influencer-for-your-brand/
31. 31 Influencer Marketing Stats to Know in 2023. Accessed March 19, 2023.
https://blog.hubspot.com/marketing/influencer-marketing-stats
32. Donaldson SI, Dormanesh A, Escobedo P, Majmundar A, Kirkpatrick M, Allem JP. The
impact of e-cigarette product place in music videos on susceptibility to use e-cigarettes
among young adults: An experimental investigation. Addictive Behaviors. 2022;130:107307.
doi:10.1016/j.addbeh.2022.107307
33. Dreyfuss E. Opinion | Our Kids Are Living in a Different Digital World. The New York
Times. https://www.nytimes.com/2024/01/12/opinion/children-nicotine-zyn-socialmedia.html. Published January 12, 2024. Accessed April 29, 2024.
34. 13 Best Vaping Affiliate Programs To Promote In 2024. Published January 30, 2024.
Accessed April 17, 2024. https://highpayingaffiliateprograms.com/vaping-affiliate-programs/
104
35. Liu J. More than half of Gen Zers think they “can easily make a career in influencing,” says
branding expert. CNBC. Published September 20, 2023. Accessed April 17, 2024.
https://www.cnbc.com/2023/09/20/more-than-half-of-gen-zers-think-they-can-easily-makea-career-in-influencing.html
36. The FTC Releases New Guidelines (And Its First Creator-Friendly Brochure) For Influencers
Who Post Sponsored Content. Tubefilter. Published November 7, 2019. Accessed February
28, 2024. https://www.tubefilter.com/2019/11/06/the-ftc-releases-new-guidelines-and-itsfirst-creator-friendly-brochure-for-influencers-who-post-sponsored-content/
37. Advertising and Promotion. FDA. Published January 28, 2022. Accessed April 30, 2024.
https://www.fda.gov/tobacco-products/products-guidance-regulations/advertising-andpromotion
38. Klein EG, Czaplicki L, Berman M, Emery S, Schillo B. Visual Attention to the Use of #ad
versus #sponsored on e-Cigarette Influencer Posts on Social Media: A Randomized
Experiment. Journal of Health Communication. 2020;0(0):1-6.
doi:10.1080/10810730.2020.1849464
39. Vogel EA, Guillory J, Ling PM. Sponsorship Disclosures and Perceptions of E-cigarette
Instagram Posts. Tobacco Regulatory Science. 2020;6(5):355-368. doi:10.18001/TRS.6.5.5
40. Vassey J, Vogel EA, Unger JB. US adolescents’ response to nicotine warning labels in
influencer e-cigarette marketing social media posts: a survey-based randomised betweensubject experiment. Tobacco Control. Published online May 15, 2024. doi:10.1136/tc-2023-
058534
41. Vogel EA, Unger JB, Vassey J, Barrington-Trimis JL. Effects of a nicotine warning label
and vaping cessation resources on young adults’ perceptions of pro-vaping instagram
influencer posts. Addictive Behaviors. 2024;149:107888. doi:10.1016/j.addbeh.2023.107888
42. Phua J, Lim D (Jay). Can warning labels mitigate effects of advertising message claims in
celebrity-endorsed Instagram-based electronic cigarette advertisements? Influence on social
media users’ E-cigarette attitudes and behavioral intentions. Journal of Marketing
Communications. 2023;29(5):455-475. doi:10.1080/13527266.2022.2037008
43. Willis E, Haught MJ, Morris II DL. Up in Vapor: Exploring the Health Messages of ECigarette Advertisements. Health Communication. 2017;32(3):372-380.
doi:10.1080/10410236.2016.1138388
44. Smith MJ, Hilton S. Youth’s exposure to and engagement with e-cigarette marketing on
social media: a UK focus group study. BMJ Open. 2023;13(8):e071270.
doi:10.1136/bmjopen-2022-071270
45. Padon AA, Maloney EK, Cappella JN. Youth-Targeted E-cigarette Marketing in the US. Tob
Regul Sci. 2017;3(1):95-101. doi:10.18001/TRS.3.1.9
105
46. Gell S, Engel E, Heiss R, Karsay K. Social media influencers and adolescents’ health: A
scoping review of the research field. Social Science & Medicine. Published online November
15, 2023:116387. doi:10.1016/j.socscimed.2023.116387
47. Online Research Panels & Samples for Surveys. Qualtrics. Accessed October 1, 2023.
https://www.qualtrics.com/research-services/online-sample/
48. Gibbons FX, Gerrard M. Predicting young adults’ health risk behavior. Journal of
Personality and Social Psychology. 1995;69(3):505-517. doi:https://doi.org/10.1037/0022-
3514.69.3.505
49. Gibbons FX, Gerrard M, Blanton H, Russell DW. Reasoned action and social reaction:
willingness and intention as independent predictors of health risk. J Pers Soc Psychol.
1998;74(5):1164-1180. doi:10.1037//0022-3514.74.5.1164
50. Gerrard M, Gibbons FX, Houlihan AE, Stock ML, Pomery EA. A dual-process approach to
health risk decision making: The prototype willingness model. Developmental Review.
2008;28(1):29-61. doi:10.1016/j.dr.2007.10.001
51. Himelboim I, Golan GJ. A Social Networks Approach to Viral Advertising: The Role of
Primary, Contextual, and Low Influencers. Social Media + Society.
2019;5(3):2056305119847516. doi:10.1177/2056305119847516
52. Two-Step Flow of Communication - an overview | ScienceDirect Topics. Accessed April 30,
2024. https://www.sciencedirect.com/topics/social-sciences/two-step-flow-ofcommunication
53. What is cross-border TAPS? https://portaluat.who.int/fctcapps/fctcapps/fctc/infohubs/crossborder-taps
54. Sigler T, Martinus K, Loginova J. Socio-spatial relations observed in the global city network
of firms. PLOS ONE. 2021;16(8):e0255461. doi:10.1371/journal.pone.0255461
55. Commercial Speech: Overview | U.S. Constitution Annotated | US Law | LII / Legal
Information Institute. Accessed November 6, 2023.
https://www.law.cornell.edu/constitution-conan/amendment-1/commercial-speech-overview
56. Premarket Tobacco Product Marketing Granted Orders. FDA. Published online March 28,
2024. Accessed April 30, 2024. https://www.fda.gov/tobacco-products/premarket-tobaccoproduct-applications/premarket-tobacco-product-marketing-granted-orders
57. Vassey J, Unger JB. Should Tobacco-Related Marketing on Social Media Have Stronger
Restrictions? Commentary. Substance Use & Misuse. 2023;58(12):1615-1619.
doi:10.1080/10826084.2023.2223287
58. TikTok. Accessed April 4, 2021. https://www.tiktok.com/community-guidelines?lang=en
106
59. Kong G, Laestadius L, Vassey J, et al. Tobacco promotion restriction policies on social
media. Tob Control. Published online November 3, 2022:tobaccocontrol-2022-057348.
doi:10.1136/tc-2022-057348
60. Howell JL, Ratliff KA. Investigating the role of implicit prototypes in the prototype
willingness model. J Behav Med. 2017;40(3):468-482. doi:10.1007/s10865-016-9814-y
61. Coyne SM, Weinstein E, Sheppard JA, et al. Analysis of Social Media Use, Mental Health,
and Gender Identity Among US Youths. JAMA Network Open. 2023;6(7):e2324389.
doi:10.1001/jamanetworkopen.2023.24389
62. Romer D, Jamieson KH. Conspiracy theories as barriers to controlling the spread of COVID19 in the U.S. Soc Sci Med. 2020;263:113356. doi:10.1016/j.socscimed.2020.113356
63. Friedman JA. Priorities for Preventive Action: Explaining Americans’ Divergent Reactions
to 100 Public Risks. American Journal of Political Science. 2019;63(1):181-196.
doi:10.1111/ajps.12400
64. PATH (Population Assessment of Tobacco and Health) Study - Home. Accessed October 1,
2023. https://pathstudyinfo.nih.gov/
65. Harm Perceptions of Tobacco/Nicotine Products and Child Exposure: Differences between
Non-Users, Cigarette-Exclusive, and Electronic Cigarette-Exclusive Users.
doi:10.1177/1179173X21998362
66. Duke JC, Allen JA, Eggers ME, Nonnemaker J, Farrelly MC. Exploring Differences in
Youth Perceptions of the Effectiveness of Electronic Cigarette Television Advertisements.
Nicotine & Tobacco Research. 2016;18(5):1382-1386. doi:10.1093/ntr/ntv264
67. Cheng HG, Lizhnyak PN, Knight NA, Vansickel AR, Largo EG. Youth susceptibility to
tobacco use: is it general or specific? BMC Public Health. 2021;21(1):1913.
doi:10.1186/s12889-021-11956-6
68. Grandpre J, Alvaro EM, Burgoon M, Miller CH, Hall JR. Adolescent reactance and antismoking campaigns: a theoretical approach. Health Commun. 2003;15(3):349-366.
doi:10.1207/S15327027HC1503_6
69. Reeves B, Yeykelis L, Cummings JJ. The Use of Media in Media Psychology. Media
Psychology. 2016;19(1):49-71. doi:10.1080/15213269.2015.1030083
70. About our panel | YouGov. Accessed May 17, 2024. https://today.yougov.com/about/panel
71. Camenga D, Gutierrez K, Kong G, Cavallo D, Simon P, Krishnan-Sarin S. E-cigarette
advertising exposure in e-cigarette naïve adolescents and subsequent e-cigarette use: A
longitudinal cohort study. Addictive Behaviors. 2018;81. doi:10.1016/j.addbeh.2018.02.008
107
72. Pollay RW, Dewhirst T. The dark side of marketing seemingly “Light” cigarettes: successful
images and failed fact. Tobacco Control. 2002;11(Supplement 1):i18-i31.
doi:10.1136/tc.11.suppl_1.i18
73. truthorange - YouTube. Accessed June 1, 2022.
https://www.youtube.com/user/truthorange/featured
74. FTC’s Endorsement Guides: What People Are Asking. Federal Trade Commission.
Published September 7, 2017. Accessed February 28, 2024. https://www.ftc.gov/businessguidance/resources/ftcs-endorsement-guides-what-people-are-asking
75. 21 CFR Part 1114 -- Premarket Tobacco Product Applications. Accessed August 31, 2022.
https://www.ecfr.gov/current/title-21/chapter-I/subchapter-K/part-1114
76. Bureau UC. Projections of the Size and Composition of the U.S. Population: 2014 to 2060.
Census.gov. Accessed February 22, 2024.
https://www.census.gov/library/publications/2015/demo/p25-1143.html
77. Grandpre J, Alvaro EM, Burgoon M, Miller CH, Hall JR. Adolescent Reactance and AntiSmoking Campaigns: A Theoretical Approach. Health Communication. 2003;15(3):349-366.
doi:10.1207/S15327027HC1503_6
78. ASH. Use of e-cigarettes among young people in Great Britain. ASH. Accessed November
26, 2023. https://ash.org.uk/resources/view/use-of-e-cigarettes-among-young-people-ingreat-britain
79. Jane Ling MY, Abdul Halim AFN, Ahmad D, Ahmad N, Safian N, Mohammed Nawi A.
Prevalence and Associated Factors of E-Cigarette Use among Adolescents in Southeast Asia:
A Systematic Review. Int J Environ Res Public Health. 2023;20(5):3883.
doi:10.3390/ijerph20053883
80. Padon AA, Lochbuehler K, Maloney EK, Cappella JN. A Randomized Trial of the Effect of
Youth Appealing E-Cigarette Advertising on Susceptibility to Use E-Cigarettes Among
Youth. Nicotine Tob Res. 2018;20(8):954-961. doi:10.1093/ntr/ntx155
81. U.S. teens on social media by platform 2026. Statista. Accessed November 4, 2023.
https://www.statista.com/statistics/1367662/teens-us-using-social-media-by-platform/
82. Tatum Z, Leventhal A, Wipfli HL. Playtime: vaping devices designed as cartoons and toys
may appeal to kids. Tobacco Control. Published online March 28, 2023. doi:10.1136/tc2022-057908
83. Willis E, Haught MJ, Morris Ii DL. Up in Vapor: Exploring the Health Messages of ECigarette Advertisements. Health Communication. 2017;32(3):372-380.
doi:10.1080/10410236.2016.1138388
84. J. Vassey, J. Chen-Sankey, and J. B. Unger. E-cigarette and cannabis influencer marketing
on social media despite regulations. Blog - Tobacco Control. Published June 3, 2024.
108
Accessed June 3, 2024. https://blogs.bmj.com/tc/2024/06/03/e-cigarette-and-cannabisinfluencer-marketing-on-social-media-despite-regulations/
85. ADVANCE Study – Epidemiology of Substance Use (EOS). Accessed October 7, 2022.
https://eosresearch.usc.edu/advance-study/
86. Cambron C. Racial/Ethnic Differences in Vaping Product Use among Youth: A State-Level
Analysis. International Journal of Environmental Research and Public Health.
2023;20(9):5729. doi:10.3390/ijerph20095729
87. Ouellette RR, Selino S, Kong G. Electronic Nicotine Delivery Systems and E-Liquid
Modifications to Vape Cannabis Depicted in Online Videos. JAMA Network Open.
2023;6(11):e2341075. doi:10.1001/jamanetworkopen.2023.41075
88. Jones EE, Davis KE. From Acts To Dispositions The Attribution Process In Person
Perception1. In: Berkowitz L, ed. Advances in Experimental Social Psychology. Vol 2.
Academic Press; 1965:219-266. doi:10.1016/S0065-2601(08)60107-0
89. Monthly Active Users by Social Media Platform (in millions). Accessed November 4, 2023.
https://datawrapper.dwcdn.net/7pJrd/1/
90. Zote J. Micro-Influencer Marketing Guide: Benefits and Steps. Influencer Marketing Hub.
Published December 29, 2023. Accessed April 30, 2024.
https://influencermarketinghub.com/micro-influencer-marketing-guide/
91. Gu J, Abroms LC, Broniatowski DA, Evans WD. An Investigation of Influential Users in the
Promotion and Marketing of Heated Tobacco Products on Instagram: A Social Network
Analysis. International Journal of Environmental Research and Public Health.
2022;19(3):1686. doi:10.3390/ijerph19031686
92. Zhou R, Xie Z, Tang Q, Li D. Social Network Analysis of e-Cigarette–Related Social Media
Influencers on Twitter/X: Observational Study. JMIR Formative Research.
2024;8(1):e53666. doi:10.2196/53666
93. Meltwater: Media, Social & Consumer Intelligence. Meltwater. Accessed May 16, 2024.
https://www.meltwater.com/en
94. Valente TW. Social Networks and Health: Models, Methods, and Applications. Oxford
University Press Accessed May 17, 2021.
https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780195301014.001.
0001/acprof-9780195301014
95. Freeman LC. Centrality in social networks conceptual clarification. Social Networks.
1978;1(3):215-239. doi:10.1016/0378-8733(78)90021-7
96. Barabási AL, Bonabeau E. Scale-Free Networks. Scientific American. 2003;288(5):60-69.
109
97. Vape store density and proximity to schools in Denpasar, Bali, Indonesia | Tobacco Control.
Accessed December 11, 2023. https://tobaccocontrol.bmj.com/content/early/2023/08/04/tc2023-058037
98. Graham DD. MARKETING GRANTED ORDERS. :18.
99. Laestadius L, Hoorn KV, Vassey J, et al. Tobacco, nicotine and counter-marketing
promotions using Instagram’s branded content tool. Tobacco Control. Published online
December 30, 2023. doi:10.1136/tc-2023-058301
100. Poerio GL, Blakey E, Hostler TJ, Veltri T. More than a feeling: Autonomous sensory
meridian response (ASMR) is characterized by reliable changes in affect and physiology.
PLOS ONE. 2018;13(6):e0196645. doi:10.1371/journal.pone.0196645
101. Alhabash S, Road W, Lansing E, et al. Effects of models’ age on effectiveness of social
media advertising of alcohol among underage minors. :6.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use
PDF
Adolescent social networks, smoking, and loneliness
PDF
Tobacco and marijuana surveillance using Twitter data
PDF
Anxiety symptoms and nicotine use among adolescents and young adults
PDF
Social network influences on depressive symptoms among Chinese adolescents
PDF
Examining the built and online environments to understand cannabis and e-cigarette availability, marketing, and product use in Southern California
PDF
TikToking and Instagramming: following high school teacher influencers' roles in supporting and informing teacher practices
PDF
Effects of flavorings in electronic cigarettes on the use and appeal of e-cigarettes among youth and adults
PDF
The role of depression symptoms on social information processing and tobacco use among adolescents
PDF
Contextualizing experiences and developmental stages of immigration and cultural stressors in Hispanic/Latinx adolescents
PDF
Influences of specific environmental domains on childhood obesity and related behaviors
PDF
Exploring the role of peer influence, linguistic acculturation, and social networks in substance use
PDF
Relationship formation and information sharing to promote risky health behavior on social media
PDF
Smoke-free housing policies and secondhand smoke exposure in low income multiunit housing in Los Angeles County
PDF
Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California
PDF
Social media and health: social support and social capital on pregnancy-related social networking sites
PDF
The threat of exploitation of "kidfluencers" & the role PR must play
PDF
Exploring mental health care utilization and academic persistence among college students: evaluating racial and stigma-related disparities
PDF
The role of social media in influencing consumer opinions and preferences in the fintech industry
PDF
Multilevel influences of care engagement and long-term survival among childhood, adolescent, and young adult cancer survivors
Asset Metadata
Creator
Vassey, Julia
(author)
Core Title
Evaluating social networks and impact of micro-influencers who promote e-cigarettes on social media
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Health Behavior Research
Degree Conferral Date
2024-08
Publication Date
07/11/2024
Defense Date
06/10/2024
Publisher
Los Angeles, California
(original),
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
e-cigarettes,ENDS,influencers,Marketing,OAI-PMH Harvest,social media,tobacco control
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Unger, Jennifer (
committee chair
), Chen-Sankey, Julia (
committee member
), Pickering, Trevor (
committee member
), Valente, Thomas (
committee member
)
Creator Email
juliavassey@gmail.com,vassey@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC113997L79
Unique identifier
UC113997L79
Identifier
etd-VasseyJuli-13187.pdf (filename)
Legacy Identifier
etd-VasseyJuli-13187
Document Type
Dissertation
Format
theses (aat)
Rights
Vassey, Julia
Internet Media Type
application/pdf
Type
texts
Source
20240712-usctheses-batch-1178
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
e-cigarettes
ENDS
influencers
social media
tobacco control