From the time of their creation, cryptocurrencies have existed in uneasy symbiosis with the media outlets that cover them. Despite the opportunities for mutual benefit, the crypto community and major news outlets have frequently found themselves at odds. After all, reports first fueled bitcoin’s rise from an idealistic experiment to blockchain behemoth. Today, as new cryptocurrencies compete for investors’ attention daily, coverage can make the difference between ascendance and irrelevance. As for reporters, crypto is subject matter worth salivating over. Which other beat could offer so many compelling characters and rags-to-riches narratives?
In this project, we set out to study the complicated and conflictual relationship between cryptocurrencies and publications that have covered them so intensely in recent years. Our analysis tracks the correlation between coverage on currency values over time and the sentiments different outlets display in their crypto reporting. To learn how the press and cryptocurrencies have profoundly influenced one another, keep reading.
Valuation vs. Coverage Volume
As we consider the historical trajectories of bitcoin’s market cap and coverage of cryptocurrency as a whole, the minimal degree of coverage prior to 2017 is quite striking. Aside from occasional flurries of press attention, such as when Australian engineer Craig Wright claimed to have created bitcoin in May 2016, the cryptocurrency attracted relatively little notice before its meteoric rise. Even as its historic run began to gain momentum, coverage spikes sometimes coincided with sudden downturns in value, such as when Chinese regulators moved to ban exchanges in September 2017. By December 17, 2017, however, a flurry of coverage reflected (and possibly propelled) bitcoin’s all-time high price of $19,783.21 per coin.
Of course, this peak was not to be sustained: Values dropped precipitously in the final weeks of 2017. Experts have attributed this sudden decline to a convoluted mix of factors, from the strong performance of alternative cryptocurrencies to people selling off bitcoin to pay for holiday presents. Whatever the cause, bitcoin’s downturn created a press frenzy: Articles peaked as the market cap plunged. The appeal of reporting disaster might explain more recent trends in coverage as well; in May and June 2018, for example, a slide in value correlated with a temporary increase in articles.
A Timeline of Tone
If time has intensified the coverage of cryptocurrency, the added attention has not been purely positive. In fact, our data suggest coverage of bitcoin’s infancy was mostly glowing and that subsequent articles have shifted toward neutrality overall. As recently as 2016, positive articles far exceeded negative ones, both in terms of volume and intensity. As coverage surged in mid-2017, however, articles expressing negative sentiment grew more common. This trend was fueled in part by grim prognostications from the likes of Warren Buffett and Mark Cuban, who guessed that “a bubble” was underway.
In the wake of bitcoin’s market cap plunge in the final days of 2017, negative articles multiplied – with cryptocurrencies falling 34 percent in the first month of 2018, cautionary tales of vanished wealth were common. As the number of articles picked up overall, average positive and negative sentiment values were also subject to far less fluctuation on a weekly basis. This finding could simply reflect an expanded sample size – with more articles written, outliers affect the average to a lesser extent. But it might also suggest the normalization of cryptocurrency: As the technology becomes a known entity, swings of praise and criticism may become less common.
Emphasis by Outlet
On the major news sites that covered cryptocurrency most often, coverage was highly variable. Business and finance outlets that skew toward a younger audience, such as Forbes and Business Insider, had coverage that was consistently above the overall median for positive sentiment: a combined 1,335 articles from these two outlets alone were more positive than the overall median sentiment of the sampled articles, while only 413 of their articles fell on the negative side. Some outlets that covered bitcoin often displayed a slightly more even mix of sentiments: CNBC published nearly 1,000 articles on cryptocurrency during the time period surveyed, with 52.9% being positive and 47.% being negative.
Still other outlets tended to report from a place of greater skepticism (The Wall Street Journal, Gizmodo) or outright criticism: Breitbart News and Raw Story, outlets on polar opposites of the American political spectrum, combined for 91 negative articles and only a single positive piece. That finding is somewhat surprising, as the site’s former executive chairman, Steve Bannon, has spoken in glowing terms of cryptocurrency’s populist potential. But both Breitbart and Raw Story’s critical tone is part of a larger pattern we observed among the most politically partisan outlets we surveyed.
Conservative outlets’ disdain for crypto may seem odd given the Republican party’s opposition to regulation. But many conservative lawmakers have actually called for tighter controls on digital currencies, fearing they could lead to financial instability on a large scale. Interestingly, financial and economic publishers seemed to be more generous to bitcoin than outlets devoted to tech. In the past,crypto enthusiasts have accused financial experts of lacking the understanding to cover the technology accurately.
Our data indicate that some publishers have warmed to crypto’s prospects over time, gravitating toward positive coverage. Forbes and CNET were primary examples of this pattern, with a large number of predominantly positive articles appearing in the last year of the time period we studied. Consistent with our earlier results, however, the more common trajectory was for coverage to become increasingly negative. The average sentiment of Reuters, USA Today, and Gizmodo articles dropped substantially over time, for example – though that may be because their initial coverage of cryptocurrency was overwhelmingly positive.
When we consider the content of positive articles more specifically, we find the understandable celebration of digital currencies’ surging values: Noted bitcoin benchmarks included $9,000 and $12,000 per coin. But the charitable potential of blockchain enrichment was also visible, including some heartwarming stories about donations made possible by bitcoin’s rise. Conversely, negative articles emphasized security challenges and scams, consistent criticisms among those who disparage cryptocurrencies.
Hype or Huge Potential?
Our findings suggest that coverage of bitcoin has been anything but static: The tone and volume of reports have differed substantially, both over time and among various outlets. Whereas many publications have produced increasingly skeptical articles in recent months, others have actually become more bullish as the years go by. Whatever your own biases may be, the influence of press reports on cryptocurrencies is difficult to ignore. In traditional domains of finance, it’s long been understood that media can move the market. But in an industry in its infancy, the press may be particularly powerful.
Before you make investments based on the latest hot take you’ve read online, consider the ideological disposition of the article’s author, as well as of the publication itself. No one possesses perfect insight into bitcoin’s future, let alone the fate of cryptocurrencies more generally. As you form your own convictions, absorb as many authoritative perspectives as possible and be wary of extremes. Between hodl-ing forever and avoiding cryptocurrencies entirely, there are plenty of gray areas worth exploring.
This project made use of article link databases provided by mediacloud.org, an open source platform for studying media ecosystems. We surveyed 48 national and international media outlets for articles covering cryptocurrency from January 1, 2013, to July 31, 2018. Due to a low article volume, some sources were excluded from the analyses, and we only reported on articles published after May 1, 2013. All articles were downloaded and analyzed using the VADER (Valence Aware Dictionary and sEntiment Reasoner) sentiment analysis in conjunction with the Natural Language Toolkit (NLTK) library in Python, which provides a sentence-level analysis of long-form text and computes aggregate scores per article for positive, negative, and neutral sentiment. Our reported analyses include the full text of 7,527 online news articles.
For these analyses, we only examined positive and negative sentiments per article.
Each article received a score for both positive and negative sentiment. Across the dataset, articles rated by the VADER algorithm had a mean positive polarity of 0.1046 (s.d. = 0.0358) and a mean negative polarity of 0.0471 (s.d. = 0.02673). In order to make a meaningful comparison between these two scores for each article, positive and negative polarity scores were independently normalized according to a normal distribution. In order to determine whether an article was ‘Positive’ or ‘Negative,’ we compared the two z-scores and chose the higher one. For example, an article with a positive polarity score of 0.138 (z = 1.3538, 82nd percentile for positive polarity) and a negative polarity score of 0.101 (Z = 2.292, 98th percentile for negative polarity) would be rated as an overall negative article since it was had a higher negative z score than positive z score (and was in the 98th percentile of negative polarity scores overall).
Our initial article search was limited to mainstream media outlets and what was indexed by Media Cloud, meaning that while our database of articles was extensive, it is by no means comprehensive. Additionally, sentiment analysis of positive, negative, and neutral text is not necessarily a proxy for whether or not a publication supported cryptocurrency but rather reflects the type of language used in articles that discussed cryptocurrency, and our net scores (which were used to rate each article ‘Positive’ or ‘Negative’) were computed based on the mean, median, and standard deviations of only the articles surveyed in the study.
- Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014
Fair Use Statement
If you’d like to share our project findings and images with your own audience, you’re welcome to do so for non commercial purposes. Simply link back to this page to attribute our team appropriately. Unlike Satoshi Nakamoto, we appreciate a little credit for our work.