As the 2020 Presidential election cycle accelerates, “universal healthcare” continues to be a polarizing topic––one that has garnered a great deal of attention and discussion within the Democratic presidential debates. The last debate, which took place January 14, 2020 at Drake University in Des Moines, Iowa, was no different. While the primary focus of moderators’ questions was on foreign policy and climate change, healthcare remained a significant component of the discussion.
In the hours leading up to, during, and after the debate, thousands of news articles and blog posts were published, along with hundreds of thousands of social media posts and other digital content on this topic. Within the context of the debate, the topic of healthcare generated 2,650 media articles and 85,750 social media posts were published between January 14 and 15, 2020.
The nearly 19 million words of healthcare-related articles and social media posts generated during these two days would take a team of ten people, reading 12 hours per day, more than ten days to consume. This is obviously untenable for campaigns, policymakers, and lobbyists when making key strategic and messaging decisions.
For more than a decade, technology has offered the ability to count, classify, and infer basic demographic attributes within this content. Where technology has fallen short is in its ability to provide unbiased context and convey impact on multiple sides of single issue, doing so within a timeframe that gives message makers and shapers the ability to react and plan in the critical hours during and after key news-making events.
In this analysis, we wanted to understand specifically how the issue of healthcare was framed and amplified during and after the debate, from both sides of the political spectrum. Our dataset was made up of the full text of 2,650 media articles and approximately 85,750 social media and blog posts between January 14-15, 2020. This corpus comprised the publicly available, verifiable news and other digital content around this topic.
Throughout the debates, the Democratic presidential field has advocated for an expansion of universal healthcare, ranging from expanding the Affordable Care Act, a public option, and “Medicare for All.” While the President and Republicans in both the House and Senate have criticized the Democratic proposals as bloated, costly, and unrealistic, Democrats have positioned their healthcare proposals as cost-saving, productivity-amplifying, and ultimately life-saving.
In real world reading, the impact an article has on a reader is highly dependent on their own point of view. If you’re a Republican and against universal health care, you would likely have a negative view of articles that frame universal healthcare in a positive light, and vice versa.
In order to address the challenge of analyzing reader perception across millions of words of text, Turbine Labs uses a form of Artificial Intelligence (AI) called Natural Language Processing (NLP), as well as advanced machine learning, to “read” each article from the perspective of a human reader—only 54,000 times faster. Each article is then assigned a score from zero to one, with the score of “1” indicating the article is unmistakably about the topic. This determines the Relevance of the article to the topic, which in this case, is healthcare policy among Democratic presidential candidates.
Once the Relevance of an article is determined, Turbine Labs uses a patent-pending process to analyze and score the emotion (often referred to as “sentiment”) of the text from predetermined points of view. For this content, we wanted to understand how Republicans and Democrats would generally perceive Democrats healthcare proposals. Each article is then assigned a “sentiment” score from that point of view, ranging from negative one (-1.000) to positive one (+1.000), down to the one-thousandth decimal. Each article can therefore have many sentiment scores, depending on the topic and the point of view.
Finally, Turbine Labs analyzes the likelihood the content was exposed to actual readers based on the inbound traffic to the article’s website, how often the article was shared on social networks, how often the article was referenced in other publications, and a number of other factors.
The results of this analysis, which take less than three hours to compile regardless of the volume of content, are plotted on a graph where each dot represents an article.
In the context of the Democratic debate, only 3% (77) of the 2,650 articles were scored as impactful––regardless of which side of the issue the reader was on. Moreover, just ten (<1%) of the articles were classified as definitively about the topic of healthcare policy.
For the remaining 2,500 articles, healthcare policy may have earned passing mentions or was simply used as a reference among other topics covered in the debate. But they were not the primary focal point for the journalists and bloggers who covered the event.
Making Decisions Based on Impact and Point of View, Not Noise and Inference
Actionable Earned Media Impact
Suppose you’re a lobbyist or policymaker opposed to the various Democratic healthcare proposals, such as Medicare for All, single-payer, or the Affordable Care Act (ACA). Would you have time to read all 2,600 articles to be adequately informed on the topic? Or even the 77 most impactful? Absolutely not.
Using Turbine Labs, in under 3 hours you would be able to determine that 31% of the coverage was positive to your position, while only 15% of the coverage was positive to the universal healthcare policies as seen on the table below.
In the same 3 hour timeframe, you would discover the article of the highest impact. In this case, that would be an article from Reason titled “Sorry, Bernie Sanders, Taiwan’s Single-Payer System Isn’t an Argument for Medicare for All.” You would also find that this article contained the most negative sentiment towards Democratic healthcare proposals. (Shown in the green dot on the right of the graph).
Now, suppose you’re on the opposite side of the issue and wanted to quickly identify highly impactful coverage that supported the Democrats healthcare proposals. While the New York Times authored the most relevant, positive sentiment article, it was an article published on Jacobin titled “What a Bernie Sanders Presidency Would Look Like” that was the most impactful.
Actionable Social Media Impact
On social media channels the conversation was more diametrically opposed. For example, those in favor of universal healthcare questioned why price was a concern with healthcare but not with the military, and praised the single-payer proposals of Sanders and Warren, while expressing disappointment with other candidates who had not made single-payer options a more prominent part of their campaigns. For example, Ana Kasparian, host and producer for The Young Turks, authored numerous posts in which she listed the numerous benefits offered by Medicare for All, such as “NO co-pays. NO deductibles. NO premiums. FULL coverage of everything including optical and dental.”
Those opposed included Liz Wheeler, host of the Tipping Point on the OAN Network, who authored three posts during the debate in which she alleged that Sanders wanted to “give illegal aliens free healthcare” and “abolish private insurance,” and that his proposal was “so expensive the government can’t afford it.” That content was the most highly viral, generating a combined 16.2K retweets and 41.2K likes.
If you’re a lobbyist or policy maker who supports the various Democratic healthcare proposals, consuming nearly 85,000 social media posts to understand context and point of view would be nearly impossible. However, using Turbine Labs, you would be able
Gaining an Edge
For political campaigns, policy makers, and lobbyists, it’s no longer enough to rely on feedback loops, filter bubbles and echo chambers of information to make well-informed decisions. In the 2020 election cycle and beyond, the ability to see both sides of an issue, in near real-time, will be critical to winning.
Separately, the sheer volume of data, channels and sources that have emerged over the last decade have eroded the efficiency of campaign and policy operations. Ironically, efficiency and effectiveness are critical to crafting and amplifying messages that move voters to the polling places and policies across the legislative floor.
With the ability to ingest, process, and synthesize hundreds of thousands of articles and social posts in a fraction of the time offered by traditional software or processes, Turbine Labs delivers actionable intelligence based on what has happened and what is happening next.
See the Full Picture – From Both Sides of the Debate
To complete this analysis, the Turbine Labs platform produced two Segment briefings on the topic of healthcare policy in the context of the Democratic Presidential Debate, which took place on January 15, 2020.
The first Segment was framed from the perspective of those who support universal healthcare policies. The second Segment was framed from the perspective of those who are opposed to universal healthcare policies.
To request copies of both studies, please complete the form below.