Smoking Selfies: Using Instagram to Explore Young Women’s Smoking Behaviors

Introduction

Cigarette use continues to be the number one preventable disease in the world and in the United States (Bauer, Briss, Goodman, & Bowman, 2014; Centers for Disease Control and Prevention, 2012). Young adults are particularly vulnerable to tobacco use due to targeted tobacco marketing campaigns (Cortese, Lewis, & Ling, 2009; Ling & Glantz, 2002), with approximately 20% of young adults smoking cigarettes, with 18 to 25 having the highest prevalence at 41% (Richardson, Williams, Rath, Villanti, & Vallone, 2014). Although the smoking trends for young adult men show higher prevalence when compared to young adult women (Nelson et al., 2008), it is not as if young adult women have little risk of tobacco use (Li, Holahan, & Holahan, 2015) which can lead to tobacco addiction and multiple tobacco use (Lee, Hebert, Nonnemaker, & Kim, 2014; Ling & Glantz, 2002). Anti-tobacco marketing to youth such as those from the Truth Campaign, as well as anti-tobacco advertising and interventions in bars and clubs targeting young adults, helps in reducing or preventing smoking (Agaku, King, & Dube, 2014; Ling, Holmes, Jordan, Lisha, & Bibbins-Domingo, 2017). As Laestadius, Wahl, and Cho (2016) demonstrate in their research using Instagram hashtags (e.g., #ecig and #vape), the tobacco industry is already circumventing advertising bans on Instagram with corporate users comprising half of their sample of e-cigarette or vape posts, and none of the posts they found were critical of vaping tobacco use. Are we witnessing the tobacco industry using the shifting landscape of social media to advertise by influencing the creation and perpetuation of the “smoking selfie?” Who is mostly sharing these selfies, and what are in these images that are shared? What social groups are most at risk by seeing a social reconstruction of smoking re-normalization by means of a rapidly growing and minimally monitored technology that captivates a demographic most vulnerable to smoking initiation?

In this article, we provide early data on tobacco use in selfies through Instagram, and provide examples of how the selfies Instagram users share normalize tobacco use, thereby putting at risk the smoking rate declines seen over the past decades. We conclude that time is of the essence for public health professionals to begin using media technologies in their anti-tobacco counter-marketing strategies to stave off an uptick in use because, as we argue, the tobacco industry is already perfecting their marketing in social media.

Instagram represents one of the fastest growing social networks; at the time the Instagram sample was extracted, the platform had more than 400 million daily users, 90% of whom are under age 35 (Instagram, 2016; Smith, 2018); in 2014, Pew reported that 53% of young adults aged 18 to 29 used Instagram, compared to just 37% who used Twitter (Duggan, Ellison, Lampe, Lenhart, & Madden, 2015). A huge draw to the platform is that Instagram is visually based, and users can manipulate their photographs through a variety of filters. Portraits of oneself (i.e., “selfies”) are highly common on Instagram (Hu, Manikonda, & Kambhampati, 2014).

Selfies may seem insignificant or fleeting to the casual observer, but it is an integral part of the self-identity construction, normative gender performance expectations, and the subculture of young adult women. Although nearly anyone can take—and probably has taken—a self-photo, or “selfie” as it is commonly called, the dearth of research on selfies suggests that these are gendered performances done by young women (Gorey, 2014; Tifentale & Manovich, 2014). In one such study that analyzed thousands of selfies from users in five cities across the world (Bangkok, Berlin, Moscow, New York, and Sao Paulo), women were more likely to take selfies than men (range 55.2% women in Bangkok to 82% in Moscow, with New York 61.6% women), and young women in particular (23.7 estimated median age, with women having the younger average age across all cities). Normative cultural expectations of young women center on physical appearances (Strahan, Wilson, Cressman, & Buote, 2006) and help explain why nearly half of all selfies fall into the “appearance” category (Deeb-Swihart, Polack, Gilbert, & Essa, 2017).

Sociologists and social psychologists often note the importance of this type of presentation of self (Goffman, 1959), as the clothes we wear and the roles we perform help establish and perpetuate self-identity. Selfies are an even faster way in which individuals can manage these performances because they can receive almost instantaneous validation (or correction) in their gendered performances. Although selfie images are consumed quickly and individually, their impacts on the consumer are cumulative like advertising (Kilbourne, 2016) and other print media (Cortese & Ling, 2011), with a recent study of young women finding that even the most fleeting perusal of selfies had a negative impact on their self-image (Eckler, Kalyango, & Paasch, 2017). Selfies can attract followers and increase online popularity (Marwick, 2015), making users “mircocelebrities” (see Senft, 2013) through their consistent online self-presentations. In short, selfies are impactful to both the selfie-taker and the observer consuming the imagery.

We caution the reader not to interpret selfies as exclusively a young woman’s domain. The selfie is a nuanced cultural phenomenon used by all genders, documenting and exhibiting experiences from the banal to the advert of all users—from the common-folk to the celebrity (Abidin, 2016; Marwick, 2015). However, it is important to note that our research findings comport with earlier research on stealth tobacco advertising in bars and clubs that use popular people as “influencers” to use and thereby encourage product use of others. Although the tobacco industry historically built advertising campaigns on the use of female gender-normative performances in tobacco print advertising in the 1970s and 1980s (Anderson, Glantz, & Ling, 2005), in the 1990s and 2000s, the tobacco industry took their advertising to a stealthier method. The tobacco industry spent years psychographic profiling young men and women to understand their lifestyles, desires, aspirations, and gender expressions to market their tobacco products beyond traditional print advertising (Cortese et al., 2009; Cortese & Ling, 2011; Katz & Lavack, 2002; Ling & Glantz, 2002; Sepe, Ling, & Glantz, 2002). This article reifies another way the tobacco industry uses the gendered performances of young women—unknowingly or not—as tools for and targets of their stealth tobacco advertising methods, and that by doing so, it makes young adult women who use social media particularly vulnerable to these “crafty” and “cool” tobacco marketing efforts.

The recent rapid evolution of the media landscape and the widespread use of social media fundamentally altered the ways in which audiences receive health-related information (Cappella, Kim, & Albarracín, 2015; Himelboim, Smith, & Shneiderman, 2013; Napoli, 2011). Not only are people simply exposed to such information, but they also search for, create, share, and respond to content across multiple media platforms (Cappella et al., 2015). Researchers increasingly turn to the rich datasets offered by social media (e.g., Twitter) to conduct public health surveillance, tracking such topics as disease outbreaks (Culotta, 2010; Lampos & Cristianini, 2010), depression (De Choudhury, Gamon, Counts, & Horvitz, 2013), insomnia (Jamison-Powell, Linehan, Daley, Garbett, & Lawson, 2012), and overall county health (Schwartz et al., 2013). This evolving communications ecosystem has implications for substance use trends, including tobacco use; indeed, the tobacco control community has begun to recognize the importance of social media for influencing health behaviors (Allem et al., 2016; Hefler, Freeman, & Chapman, 2013).

Evidence has shown that substance use frequently is portrayed and discussed on social media, often in positive terms, and that young people easily may access and engage with these posts (Egan & Moreno, 2011; Morgan, Snelson, & Elison-Bowers, 2010; Myslín, Zhu, Chapman, & Conway, 2013; Seidenberg, Rodgers, Rees, & Connolly, 2012). This finding is meaningful when considering tobacco use because, according to social cognitive theory (Bandura, 2001; Stefanone, Lackaff, & Rosen, 2010), people—especially young people—are strongly influenced by peers or others in social media and reality shows who “look” like they could be peers (Stefanone et al., 2010). Actual or perceived behaviors of peers and/or other valued social referents (such as celebrities) have been established as key risk factors for tobacco and other substance use among adolescents and young adults (Bandura, 1971; Stefanone et al., 2010; Steinberg & Monahan, 2007; Turner, Mermelstein, & Flay, 2004), and the advertising methods of microcelebrities (Abidin, 2016; Marwick, 2015) may be an additional risk factor in the rapidly growing social media advertising landscape. Repeated exposure to images modeling tobacco use behavior by members of one’s social network may increase the perceived prevalence of tobacco use among one’s peers, serving to normalize the behavior and potentially resulting in imitation (Cialdini & Trost, 1998; Rivis & Sheeran, 2003).

Social media platforms also play a significant role in tobacco marketing; such marketing falls largely under the radar of the tobacco control community and outside the restrictions stipulated by the 1998 Master Settlement Agreement (Huang, Kornfield, & Emery, 2016; Huang, Kornfield, Szczypka, & Emery, 2014; Kostygina, Huang, & Emery, 2016; Kostygina, Tran, Shi, Kim, & Emery, 2016; National Association of Attorneys General, 1998). Advertising for tobacco products on social media offers high brand visibility (Richardson, Ganz, & Vallone, 2013), reaches a primarily young audience, and effectively circumvents the restrictive media marketing policies for tobacco. Users promote tobacco products organically on social media by incorporating them into narratives about their lives, while brands leverage consumer-generated content to extend their marketing reach when users post images in which those brands are visible (Carah & Shaul, 2016). Furthermore, celebrity presence on Instagram is quite high, and the extant studies examining celebrity substance use–related Instagram posts to date suggest that these communications have potential to reshape social norms about substance use behavior (Addiction-Treatment.com, 2014; Richardson et al., 2013; Richardson & Vallone, 2014). Tobacco control seized upon this opportunity in the Truth Initiative’s (2015) “unpaid spokesperson” television advertisement which featured images of young celebrities smoking and tagged them as unwitting marketers for the tobacco industry (Ad Age, 2014). The “unpaid spokesperson” ad first aired during the MTV Music Awards program and was part of Truth®’s “Finish It” campaign (Truth Initiative, 2015).

Given Instagram’s popularity among young adults and celebrities—high-profile and microcelebrities alike—(Khamis, Ang, & Welling, 2017; Laestadius et al., 2016; Marwick, 2015) and the platform’s potential to influence social norms related to tobacco, characterizing the extent and nature of consumer-generated image marketing of tobacco products on Instagram becomes an important research question. In this study, we undertook a content analysis of smoking imagery on Instagram and analyzed how such tobacco-related imagery is shared among users.

Methods

Results

Content coding of the 5,721 Instagram photographs relevant to tobacco smoking or vaping behavior revealed that combustible tobacco cigarettes were the most popular images displayed, with the most prevalent type of image posted being one of a young White woman smoking a cigarette in a conventional fashion.

Coding of demographics among the images is presented in Table 1. By far the highest proportion of smoking-related images was coded for females at 67.4% compared to 29.7% showing males, and 2.7% in which both genders were visible. A majority of the coded photographs were found to be selfies showing a single individual at 61.4%, while 4.1% were group selfies containing two or more people and 34.5% did not include any people in the image frame. The most prevalent age group was college-aged at 87.5%, followed by high school age or younger at 8.9%, and the least prevalent was the “older than 30” age group at 3.6%. More than half of the images were judged to depict people of White ethnic background at 85.7%, with 4.6% Black and 9.6% “all other” ethnicities.

OPEN IN VIEWER

Table 1.

Demographics Among Smoking-Related Photographs on Instagram.

Characteristic	No. of photos	%
Age
College	2,566	87.5
High school	261	8.9
Older than 30	106	3.6
Gender
Female	2,383	67.4
Male	1,056	29.7
Both visible	97	2.7
Ethnicity
White	2,699	85.7
Black	162	4.6
All others	338	9.7
Number
Individual selfie	3,514	61.4
Group selfie	235	4.1

Content coding also indicated high presence of smoking artifacts and branding in the images. As shown in Table 2, of the 5,721 photographs coded, 2,299 (40.2%) contained artifacts likely displaying a brand name or logo. Artifacts featured in the photographs were most likely to be cigarettes (47.2%), followed by e-cigarettes, other vaping devices, or vaping liquid at 32.1%, cigarillos at 3.6%, and hookah (2.4%).

OPEN IN VIEWER

Table 2.

Tobacco Artifacts and Brands Present in Instagram Images.

Characteristic	No. of photos	%
Branding
Artifact likely has a brand name or logo	2,299	40.2
Tobacco artifact
Cigarette	2,803	47.2
E-cigarette or other vaping device	948	16
Vaping liquid	886	14.9
Tobacco package	357	6
Hookah	140	2.4
None	590	10.3

Table 3 presents the results of coding for presence of smoke and/or smoke play in the images. Smoke or vape clouds were visible in 60% of the photographs, and 16% depicted smoke play or vape play. A very small proportion of the coded sample of photographs was judged to be presenting an anti-tobacco message (1.9%).

OPEN IN VIEWER

Table 3.

Smoke Imagery in Instagram Photographs.

Characteristic	No. of photos	%
Smoke play
Smoking or vaping	3,430	60
Smoke play or vape play	894	16
Marijuana	584	10.2
Anti-smoking message	108	1.9

Discussion

In this exploratory study, the results indicate that smoking selfies are prevalent and widely shared across the Instagram platform, which is particularly popular among youth and young adult women. The prevailing sentiment in smoking-related Instagram posts appears to be pro-use and the most common form of smoking image is one of a young White woman smoking a combustible cigarette. Public health professionals should note this gendered social media experience, given that smoking rates for young adult males are higher than for females (Jamal, 2016), but that the smoking presence on Instagram skewed toward young adult women. Greenwood, Perrin, and Duggan (2016) have reported that teenage girls share more on visually oriented platforms like Instagram than boys do, and the same holds true for female adults (Lenhart, 2015). Therefore, even though posts are more common among young women, it does not necessarily reflect higher prevalence of tobacco use. However, when selfies demonstrate smoking behavior, they are easily seen and potentially shared among young adult social networks. Young White women in particular, who are establishing their social media personas, may be inadvertently encouraging and reinforcing pro-smoking norms, with potential to influence the uptake or maintenance of smoking behavior (Rivis & Sheeran, 2003). The prevalence and glamorization of such images may serve to counteract recent public health gains toward de-normalizing tobacco use, particularly in groups who may not be considered health disparate. Tobacco companies are likely exploiting this gendered selfie norms through under-the-radar advertising that is to the tobacco industry’s benefit only.

These findings imply that Instagram can be a powerful tool for the surveillance of tobacco use. Tobacco brands also appear to be commonly shared in user-generated Instagram posts. This finding has policy implications because brand sharing in the “organically sharing” features of popular image-based social media platforms essentially represents covert, unpaid advertising for the tobacco industry that it would have little interest in curbing. The present study is consistent with other research indicating that the tobacco companies remain a step ahead of tobacco control, subverting advertising restrictions by exploiting Web 2.0 for marketing purposes (Elkin, Thomson, & Wilson, 2010; Freeman, 2012), and offers support for including Instagram and other social media in future policy considerations. Furthermore, very few images were found to promote anti-tobacco sentiments and, given Instagram’s potential (and social media in general) to reach and engage with young smokers, anti-tobacco advocates and public health professionals may be inadvertently overlooking an important opportunity to spread counter-marketing messages.

This study has certain limitations and suggests directions for future research. First, our methodology relied upon Instagram text tags to capture relevant content during data collection. There is risk that some relevant tags may have been omitted, the inclusion of which might have altered the findings. Second, and related to the above, the coding of data was primarily of image-only, and textual content such as hashtags, comments, or emojis were used as supplementary material to clarify ambiguous or unclear imagery. Therefore, the original context of many users’ posts may have been lost (see Highfield & Leaver, 2016). Third, as described in the “Methods” section, we could neither replicate this study under the current API of Instagram nor include examples of photos because of copyright restrictions placed on user content by Instagram’s TOS. Gaining future access to Instagram data presents a challenge which tobacco control must surmount to remain relevant in today’s evolving communications environment. Public policy professionals must inform policymakers on the need to re-open social media to examination, given the ubiquity of the platform to shape and transform the Zeitgeist on a whole range of social norms, ideas, and belief systems, particularly with regard to tobacco and vape use among young adults.

This study offers preliminary evidence that pro-smoking imagery is widely viewed and shared across increasingly popular image-based media platforms such as Instagram. Smoking behavior may become re-normalized as the tobacco industry gains covert and unregulated advertising through this medium. Future research should confirm these results and further explore Instagram’s potential as a surveillance tool to inform behavioral and regulatory science.