How researchers analyzed 15,927 tweets about e-liquids to understand public health conversations on social media
Imagine if you could listen in on thousands of daily conversations about controversial products, discovering not just what people are buying, but what they truly fear, desire, and believe. This isn't a futuristic fantasy—it's exactly what happened in 2018 when scientists turned Twitter into a digital laboratory to understand the world of e-liquids 1 .
In a groundbreaking study, researchers analyzed 15,927 tweets containing terms like "e-liquid" and "e-juice" to uncover the real public conversation happening about these products. What they discovered reveals as much about human psychology as it does about smoking habits, offering crucial insights for public health officials, regulators, and anyone interested in how social media shapes our choices 1 .
This research represents a new frontier in science—one where our digital footprints become valuable data for understanding public health trends.
The findings are particularly relevant today, as we continue to grapple with how social media influences everything from consumer behavior to political opinions.
When researchers sifted through the massive dataset of 2018 Twitter conversations, they discovered clear patterns in what people were discussing. The results revealed a complex landscape where marketing, pleasure, and health concerns collide 1 .
| Theme | Prevalence | Description |
|---|---|---|
| Promotional Content | 29.35% | Tweets focused on marketing, sales, and product promotions |
| Flavors | 24.22% | Discussions centered on flavor varieties and experiences |
| Person Tagging | 21.47% | Mentions of other users in conversations |
| Juice Composition | 17.61% | Talks about ingredients and what e-liquids contain |
| Cannabis | 16.83% | Discussions linking e-liquids to cannabis products |
| Nicotine Health Risks | 6.39% | Concerns about potential dangers of nicotine |
| Quit Smoking | 0.57% | Rare mentions of using e-liquids to stop traditional smoking |
Source: Analysis of 15,927 tweets about e-liquids from 2018 1
Perhaps most surprisingly, conversations about using e-liquids to quit traditional smoking were remarkably scarce—appearing in less than 1% of all tweets 1 . This suggests that on Twitter, at least, the harm reduction potential of e-cigarettes was not a dominant narrative.
One of the most significant findings was how frequently e-liquids were discussed in relation to cannabis, with nearly 17% of tweets making this connection 1 . This intersection between tobacco and cannabis products presents unique challenges for regulators and public health experts, suggesting that users may not clearly distinguish between these different substance categories.
As the study noted, this finding positions e-liquids "at the intersection of tobacco and cannabis use," warranting further investigation into how this perception might influence usage patterns, particularly among young people 1 .
To understand how scientists extracted meaningful patterns from thousands of tweets, we need to look at their methodological toolkit. This approach represents a growing trend in research called "digital ethnography"—studying cultural patterns through digital communication.
Researchers gathered all public tweets containing specific e-liquid-related terms ("e-liquid(s)," "e-juice(s)") posted between January 1 and December 31, 2018 1 .
Using specialized software and analytical techniques, they developed text classifiers to categorize tweets into different topics based on their content 1 .
The team employed pattern-matching methods to identify consistent themes across the dataset, similar to approaches used in other social science research 5 .
Finally, researchers interpreted the prevalence of different themes to draw conclusions about public discourse and attitudes 1 .
Data Collection
Classification
Analysis
This methodology demonstrates how traditional research techniques can be adapted to analyze digital content, creating new opportunities for understanding public opinion at scale.
| Research Tool | Function | Application in the E-Liquid Study |
|---|---|---|
| Text Classifiers | Automatically categorize text into predefined topics | Sorting tweets into themes like "Promotional" or "Flavors" |
| Pattern-Matching Software | Identify recurring patterns in large datasets | Finding consistent themes across thousands of tweets |
| Data Visualization Tools | Create visual representations of complex data | Illustrating prevalence of different conversation types |
| Statistical Analysis Programs | Calculate prevalence rates and significance | Determining what percentage of tweets fell into each category |
The implications of this research extend far beyond academic curiosity. Understanding these digital conversations is crucial for crafting effective public health strategies.
The study revealed that promotional content was the single most prevalent category of e-liquid tweets 1 . This finding raises important questions about how social media might be normalizing and encouraging uptake of e-cigarette products, particularly among non-smokers and youth.
"Twitter provides ample opportunity to influence the normalization, and uptake, of e-cigarette-related products among non-smokers and youth, unless regulatory restrictions, and counter messaging campaigns are developed to reduce this risk" 1 .
This research demonstrates the power of social media analysis as a real-time surveillance tool for public health attitudes and behaviors. Unlike traditional surveys that can take months to design and administer, social media analysis can provide immediate insights into emerging trends and concerns.
The relatively high percentage of tweets discussing flavors (24.22%) provides valuable information for regulators considering flavor restrictions 1 . Similarly, the modest but significant attention to nicotine health risks (6.39%) suggests both concerns and potential knowledge gaps among users 1 .
The 2018 e-liquid study pioneered an approach that has since evolved significantly. Recent advances in artificial intelligence have dramatically improved researchers' ability to analyze large datasets of public conversations 9 .
Where researchers in 2018 relied on traditional text classifiers, modern studies can leverage sophisticated tools like OpenAI's Generative Pre-Trained Transformer (GPT) to identify themes in public discourse 9 . Studies have demonstrated that these AI tools "performed at near-human levels, proving to be a highly useful tool for thematic analysis" 9 .
This doesn't eliminate the need for human researchers but rather augments their capabilities, allowing them to process much larger datasets while focusing on interpretation and application of findings.
As these datasets grow increasingly large, effective data visualization becomes crucial for identifying and communicating patterns 3 . Modern data visualization serves not just to present final results but to help researchers "gain insight and knowledge from data and its inherent patterns and relationships" during the analysis process 7 .
| Visualization Type | Purpose | Benefit |
|---|---|---|
| Thematic Maps | Show relationships between different themes | Reveals how topics connect in public discourse |
| Prevalence Charts | Display frequency of different themes | Allows quick comparison of conversation topics |
| Trend Lines | Track changes in topics over time | Shows how conversations evolve |
| Demographic Overlays | Connect themes to user demographics | Reveals which groups discuss which topics |
The 2018 e-liquid Twitter study represents more than just a snapshot of vaping conversations—it demonstrates a powerful new approach to understanding public health issues. By listening to unfiltered public conversations on social media, researchers can identify genuine concerns, misconceptions, and marketing influences that shape health behaviors.
As the study concluded, themes like "flavors, cannabis, health risks of nicotine, and composition warrant consideration as targets in future surveillance, public policy, and interventions" addressing e-liquid use 1 . This approach has since been applied to numerous other public health challenges, from climate change beliefs to vaccination attitudes 9 .
The digital conversation continues—and now, thanks to innovative research methods, we have the tools to understand what it's telling us about our health, our habits, and our society.
The challenge remains for public health officials, regulators, and educators to use these insights to craft more effective, targeted interventions that address the real concerns and motivations expressed in these digital spaces.