Article Text

Download PDFPDF

Public reactions to e-cigarette regulations on Twitter: a text mining analysis
  1. Allison J Lazard1,
  2. Gary B Wilcox2,3,
  3. Hannah M Tuttle4,
  4. Elizabeth M Glowacki3,5,
  5. Jessica Pikowski1
  1. 1 School of Media and Journalism, University of North Carolina at Chapel Hill, Chapel Hill, USA
  2. 2 Stan Richards School of Advertising and Public Relations, University of Texas at Austin, Austin, USA
  3. 3 Center for Health Communication, University of Texas at Austin, Austin, USA
  4. 4 Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, USA
  5. 5 Department of Communication Studies, University of Texas at Austin, Austin, USA
  1. Correspondence to Dr Allison J Lazard, School of Media and Journalism, University of North Carolina at Chapel Hill, Carroll Hall 384, Chapel Hill, NC 27599, USA; lazard{at}unc.edu

Abstract

Background In May 2016, the Food and Drug Administration (FDA) issued a final rule that deemed e-cigarettes to be within their regulatory authority as a tobacco product. News and opinions about the regulation were shared on social media platforms, such as Twitter, which can play an important role in shaping the public’s attitudes. We analysed information shared on Twitter for insights into initial public reactions.

Methods A text mining approach was used to uncover important topics among reactions to the e-cigarette regulations on Twitter. SAS Text Miner V.12.1 software was used for descriptive text mining to uncover the primary topics from tweets collected from May 1 to May 17 2016 using NUVI software to gather the data.

Results A total of nine topics were generated. These topics reveal initial reactions to whether the FDA’s e-cigarette regulations will benefit or harm public health, how the regulations will impact the emerging e-cigarette market and efforts to share the news. The topics were dominated by negative or mixed reactions.

Conclusions In the days following the FDA’s announcement of the new deeming regulations, the public reaction on Twitter was largely negative. Public health advocates should consider using social media outlets to better communicate the policy’s intentions, reach and potential impact for public good to create a more balanced conversation.

  • Electronic nicotine delivery devices
  • Public opinion
  • Media

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Use of electronic cigarettes (e-cigarettes) has dramatically risen in the USA.1 From 2010 to 2013, awareness of e-cigarettes rose nearly 40% and the percentage of adults who used an e-cigarette, at least once, more than doubled (from 3.3% to 8.5%)2 and use among high school students has skyrocketed from 1.5% in 2011 to 16% in 2015.3 Indicative of increased use, e-cigarette sales in the USA have more than doubled between 2012 and 2013, and are expected to continue to grow at a rapid pace.4

The popularity of e-cigarettes combined with the uncertainty of their long-term health impact has produced an active debate regarding e-cigarette use and regulation. Some fear that e-cigarettes may prolong the tobacco epidemic or serve as a gateway to traditional, combustible cigarettes, especially for youth.5 Others believe that e-cigarettes aid in smoking cessation or reduction, but evidence of their effectiveness in this regard is inconclusive to date.6 This widely contested public health debate—whether e-cigarettes (and other non-combustible products) are instrumental in efforts to phase out use of traditional cigarettes—continues7 and is further complicated by the variety of e-cigarettes on the market—some of which do not contain nicotine.

Until recently, no federal regulations on e-cigarettes existed, although various state-level regulations prohibited e-cigarette companies from selling their products to minors throughout the USA.8 However, in May 2016, the US Food and Drug Administration (FDA) issued a new rule that extends their regulatory authority to include all tobacco products, including e-cigarettes, cigars, pipe tobacco and hookah.9 The new regulation, which deems that e-cigarettes—categorised (rightly or wrongly) as tobacco products—are subject to regulatory authority, will allow the FDA to monitor the manufacturing, distribution and marketing of these tobacco products, in hopes to protect youth from the dangers of tobacco by prohibiting sales to minors. Understanding how the public perceives this policy is critical for regulatory efforts10–12 and much can be gleaned about initial public responses through social media.

Social media platforms, such as Twitter, can play an important role in shaping the public’s attitudes towards novel health issues,13 14 and can be used by health communicators to gauge public knowledge of health issues and inform initial response strategies.15–18 Although Twitter was originally intended to serve as a platform for connecting people and facilitating communication within social networks, it has become a popular means for marketing and spreading information about health topics, such as the promotion, distribution and social acceptance of e-cigarettes.19–22 Using a combination of human coding and machine learning, Twitter has become an important source for surveillance of public opinion about e-cigarettes, among other topics.23–26 Potential e-cigarette consumers can easily access information and even shape the discussions around new products.27 28 Furthermore, the public may rely on information provided by social media outlets even more if there is uncertainty about the regulations, policies and long-term consequences associated with a new product, such as e-cigarettes.

In their search for information, current and future users may encounter a variety of content that is sponsored by tobacco companies and marketing firms, which includes tweets that promote their products and refute claims against them.21 29 However, social media can also be used to combat the spread of false information and ignite health advocacy campaigns.30 Many organisations have taken advantage of Twitter’s affordances by responding to questions and concerns raised by the public, such as Centers for Disease Control and Prevention (CDC)-hosted live Twitter chats responding to the public’s questions about Ebola16 and the Zika virus.31 To date, there is little evidence that individuals with an anti-electronic nicotine delivery systems stance or supporters of e-cigarette regulations have a large presence on social media discussions.32

Because of Twitter’s open platform, it is crucial that health organisations and governmental agencies are aware of the information that is being spread on social media.32 Thus, this study seeks to detect prevalent themes among the public conversation surrounding new regulations for e-cigarettes, via textual analytics methods, to understand initial reactions to the FDA’s regulations, who is driving the conversation33 and how this participation in this discussion fits within the current regulatory landscape.8

Methods

Text analytics and data acquisition

This study used textual analytics to identify topics and extract meanings contained in unstructured textual data. Twitter messages were captured during a period beginning 1 May 2016 and ending 17 May 2106. Tweets and retweets were collected using NUVI software in conjunction with Twitter’s Search application programming interface (API). Because our focus was on the FDA’s deeming rule announced on 5 May 2016, keywords included variations of the word ‘e-cigarette’, combined with the governing authority, the ‘FDA’,’ or issue of interest, ‘regulation’. These keyword combinations were selected to avoid biassed language used by different audiences, filter irrelevant information and increase the likelihood of capturing a greater signal-to-noise ratio for reactions to the FDA ruling.34 The following keywords or phrases were used to capture relevant Twitter messages:

‘electronic cigarettes’ FDA, ‘electronic cigarette’ FDA, ‘e-cig’ FDA, ecig FDA, ecigs FDA, #ecig FDA, ‘electronic cigarettes’ #FDA, ‘electronic cigarette’ #FDA, ‘e-cig’ #FDA, ecig #FDA, ecigs #FDA, #ecig #FDA, ‘electronic cigarettes’ Rules, ‘electronic cigarette’ Rules, ‘e-cig’ Rules, ecig Rules, ecigs Rules, #ecig Rules, ‘electronic cigarettes’ Ruling, ‘electronic cigarette’ Ruling, ‘e-cig’ Ruling, ecig Ruling, ecigs Ruling, #ecig Ruling, ‘electronic cigarettes’ regulation, ‘electronic cigarette’ regulation, ‘e-cig’ regulation, ecig regulation, ecigs regulation, #ecig regulation, ‘electronic cigarettes’ regulations, ‘electronic cigarette’ regulations, ‘e-cig’ regulations, ecig regulations, ecigs regulations, #ecig regulations

The keywords returned 6527 messages. The original tweets were extracted resulting in 4629 messages, with a spike in Twitter activity after the 5 May announcement. Notably less than 200 tweets and retweets were captured from 1 May to 4 May, indicating the sample captured was specific to the FDA’s deeming rule rather than noise associated with ongoing conversations. The tweets were then analysed using text mining and social media monitoring software, SAS Text Miner V.12.135 and NUVI,36 and findings were interpreted. NUVI is a social listening tool that allows for monitoring of messages from Twitter using basic aggregation tools based mainly on frequency and reach of users, posts and keywords. With NUVI, the tweets were analysed for trending hashtags, top influencers (determined primarily by number of followers and frequency of mentions and retweets) and location of tweets by state and per capita (analysed with a combination of geotagging and specific locations associated with the Tweeter’s bio). The use of publicly available data in this study did not require approval from our institutional review board.

Text mining

The tweets’ textual content was analysed using SAS Text Miner V.12.1.35 SAS Text Miner is an algorithm-driven statistical software used to uncover and understand information. SAS Text Miner provides the ability to parse and extract information from text, filter and store the information, and assemble tweets into related topics for analysis to gain insights from the unstructured data.16

After the tweets were separated, the initial step was to extract, clean and create a dictionary of words using a natural language processor. This includes detecting sentences, identifying parts of speech and stemming words. Using a Text Parsing node, each message was divided into individual words. These words were listed in a frequency matrix and words that contributed little to the understanding of the topic, such as auxiliary verbs, conjunctions, determiners, interjections, participles, prepositions and pronouns, were excluded from the analysis. Following, a Text Filter node was used to exclude words that appeared in less than four messages, as a conservative measure to reduce noise. A single author with knowledge of the subject matter visually inspected and manually removed irrelevant terms. The words initially included (and excluded) in the analysis were visually inspected to ensure accuracy and identify unrecognisable symbols and letter groups for exclusion.

With the inclusion criteria set, a Text Topic node was used to combine terms into 10–15 topic groups. This clustering divided the document collection into mutually exclusive groups based on the presence of similar themes using expectation maximisation clustering. After visually examining each of the created topics, a nine-topic solution most clearly illustrated the main themes, produced the final mutually exclusive topic groups (eg, original tweets appeared in only one topic) and had a final group of tweets that represented over 1% of the captured conversation, to further reduce noise. Lastly, the researchers reviewed the individual tweets of the final topic groups to interpret the final themes. This was accomplished by individually reviewing the actual messages from each cluster or topic to arrive at the description that is now contained in the tables for tweets. Valence of each topic was also determined by the researchers rather than by software sentiment analysis due to the inability of the algorithms to deal with the complexity of informal and specialised language in this sample.37 38

Results

The nine topics generated reveal initial reactions to whether the FDA’s e-cigarette regulations will benefit or harm public health, how the regulations will impact the emerging e-cigarette market and efforts to share the news. Five of the 10 topics were almost exclusively links to articles from the press to recap newly announced regulations, citing both individuals from the industry and FDA officials, and a video produced by a vaping advocate proposing the government’s alternative ‘hidden’ agenda to harm public health and support Big Tobacco with the new regulations. Table 1 contains the topics, number of tweets included and a description of each topic’s theme.

Table 1

Electronic cigarettes tweets by topic

The other half of topics represented unique reactions to the FDA e-cigarette regulations. These topics were dominated by negative or mixed reactions, meaning the individual tweets within each topic were comments against regulation or a combination of supportive and unsupportive reactions. No topic contained tweets that were exclusively positive or in support of the regulation. Supportive tweets for the new policy included supporters’ opinions that the government’s efforts to keep minors safe are headed in the right direction. A smaller sample of supporters shared this sentiment but commented that the FDA’s decision did not go far enough—there is no regulation of flavours. The majority of individual tweets, however, were negative comments and links to news that challenged the FDA’s deeming of e-cigarettes as tobacco products, questioned what the new regulation entails and means for vaping consumers and the e-cigarette industry, such as concerns about availability and costs, framed the policy announcement as the onset of a regulation ‘war’ and requested to know more information.

Tweets also contained comments about initial reactions and lawsuits to challenge the regulation, and opinions about how the regulations will not be effective. There werecalls for politicians to act against the new regulations and suggestions that this policy represents the teaming of government and industry, such that the FDA deeming will work only to enhance the power of Big Tobacco by causing more people to smoke cigarettes, ensuring taxes are collected and shutting out small businesses. Broadly, the topics contained a negative review of the regulation, although there were some positive comments throughout. While many tweets framed the regulation as government over-regulation, whether this should be applauded or be detested varies among the posts.

Public reactions were also revealed in the hashtags used during this initial period, which focused on sharing news and general disapproval of the regulation. The top trending hashtags from this search included #vaping (appears 358 times), #ecigs (334), #vape (329), #ecigarette (119), #vta (64), #notblowingsmoke (62), #news (58), #vapelife (39) and #ecigssavelives (38), with many representing words and phrases commonly associated with vaping advocacy and support for use of e-cigarettes.

The top influencers, or the user accounts with the highest number of followers, mentions, tweets and retweets, were also captured and are shown in table 2. Notably all eight top influencers identified were actively against the FDA deeming of e-cigarettes. The most popular positions included political beliefs about free markets and limited government involvement, personal interest/enjoyment of e-cigarettes, and belief and advocates that e-cigarettes are a healthy alternative to tobacco cigarettes. Top influencers included two individual consumers, five proponents and one public policy organisation. Proponents were defined as tweeters who represent e-cig sales or marketing agencies, individuals who advocate e-cigs or tweeters who specifically identify themselves as vapers in their profile bio.25

Table 2

Characteristics of influencers

The location of the tweets is shown in table 3, which presents the states and districts with the highest number of mentions and the current (state-level) regulations and restrictions for e-cigarettes. The major regulations that were included were classification of an electronic cigarette as a tobacco product, excise tax placed on electronic cigarettes, smoke-free restrictions and permit requirements for sale of e-cigarettes.8 Overall, the locations with the highest volume of tweets during this period tended to have the least amount of regulations for e-cigarettes. The location with the highest number of mentions (New York) has the lowest amount of regulations of the top eight locations.8 The only major regulation New York had in place was the restriction of sales to minors which is instated in all 50 states and the District of Columbia.8

Table 3

Electronic cigarettes regulations by US location

Discussion

As the use of e-cigarettes continues to rise, it is important to understand how the public perceives the impact of regulation for consumers, industry and the general public good. Twitter, in particular, is a social media outlet capable of providing numerous messages to its users within a short amount of time. Twitter allows users to send up-to-date information, which can be especially helpful for consumers and health experts interested in finding out more about trends in e-cigarette use, attitudes towards e-cigarettes and the perceptions of regulations affecting usage. Given Twitter’s affordances, this study analysed tweets following the FDA’s deeming announcement for insights into public opinion.

This analysis reveals that in the days following the FDA’s announcement of the new deeming regulations, public reaction on Twitter skewed negatively, which directly contrasts recent findings from surveys that indicate support for e-cigarette regulations among US adults.10–12 While it has been reported that upwards of three-quarters of the population highly support e-cigarette regulations,10 11 the public conversation on Twitter—revealed through tweet content against regulation and trending hashtags associated with vaping advocacy—during this period was generally one sided in the other direction. Tweets included inaccurate information—such as the regulation functions as a ban on e-cigarettes—and introduced ideas that were very likely not the intention of the federal government—that this regulation was really to serve the interests of Big Tobacco. Additionally, trending hashtags revealed keywords and phrases most commonly used by vaping advocates. The differences between survey findings that adults generally support regulation and the contrasting themes on Twitter may by due to the heavy influence of people and businesses who use Twitter for e-cigarette marketing,19 29 as well as the presence of younger individuals as Twitter contributors.39 Younger individuals are more likely to use e-cigarettes, and are also more likely to speak positively about e-cigarettes on Twitter.40 Notably, Twitter users also tend to have higher education, higher income and are more likely to be urban residents.41

Increased awareness and support is needed among the public. While the conversation is new and individuals are seeking answers on Twitter, and elsewhere, there is an opportunity for public health advocates to actively influence social media conversations and create a more balanced public discussion. Twitter also provides a valuable venue for organisations like the FDA to identify and address potential misunderstandings.13 16 For example, in our study we found that many did not understand that one intention of the regulation was to prevent use by minors, although this was communicated from FDA sources in some shared articles. Clear messages from the FDA via Twitter are needed to directly address such misunderstandings. Furthermore, in accordance with CDC guidelines, the FDA should spend more time monitoring trends and discussions on social media to better understand public knowledge levels and identify future misunderstandings or myths.42 Overall, with the majority of the content against the FDA deeming, this important snapshot of public opinion reveals reactions that are more against than supportive of the new regulation, rightly or wrongly.

This study is not without limitations. Although Twitter data were a logical sample for the study of initial reactions, these findings are not generalisable to the US population. Twitter users tend to be younger and members of minority groups.39 Additionally, some of the top influencers identified in this study are considered proponents, and are generally more inclined to support e-cigs regardless of their specific motivation. Moreover, number of followers, mentions and shares determined top influencers; different algorithms or qualifications (such as how many retweets each influencer achieved) would likely highlight different individuals, proponents or organisations as the most influential and should be explored in future studies. However, analysing these tweets provides insights for how social media savvy individuals, who are likely influential for public opinion, are reacting to and sharing information about the FDA’s regulations.

Conclusion

Twitter’s open platform and didactic nature provides opportunities for public health advocates and consumers to disseminate information. Following the FDA’s deeming regulations, Twitter users shared news and voiced their thoughts and concerns in a relatively open online forum, which can be influential for public opinion. Twitter users did this by sharing press articles and general negative reactions to the new policy. Public health advocates should consider using social media outlets, such as Twitter, to better communicate the policy’s intentions, reach and potential impact for public good to create a more balanced conversation online.

What this paper adds

  • Twitter affords the ability to track public perception and attitudes towards regulation of novel tobacco products, such as e-cigarettes.

  • We report emerging themes of public discussion in e-cigarette tweets, trending hashtags and top influencers immediately after the announcement of the Food and Drug Administration (FDA)’s authority to regulate e-cigarettes as tobacco products.

  • Reactions to FDA’s deeming rule were more against than in support of the regulation. Negative reactions included misinformation about the intentions and implications of the FDA’s regulation of e-cigarettes as a tobacco product.

References

View Abstract

Footnotes

  • Contributors AJL and GBW conceptualised and designed the

    study. GBW collected the data. AJL, GBW and HMT

    analysed the data. AJL, GBW, HMT, EMG and JP interpreted the data and wrote the initial draft manuscript. All authors contributed to the manuscript revisions.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.