Fake News about our Fake News Study Spread Faster than its Truth… Just as We Predicted.
Sinan Aral is the Director of the MIT Initiative on the Digital Economy and author of The Hype Machine about how social media disrupts our world.
Several prominent journalists recently spread misinformation about the science of misinformation…
They claimed that a study my colleagues Deb Roy, Soroush Vosoughi and I published on the cover of Science magazine in 2018 had been debunked. At the time our paper was published, it was the largest longitudinal study of the spread of fake news online. Having analyzed all of the verified true and false news that ever spread on Twitter from 2006 to 2017, we found that false news spread farther, faster, deeper and more broadly than the truth in every category of information. The results were dramatic and widely discussed. The scientific impact assessment firm Altmetrics measured it to be the second most influential scientific publication of 2018 in any discipline and the work has motivated thousands of new studies on misinformation — and how to stop it — over the last several years.
But recently, Kai Kuperschmidt, a contributing correspondent for Science magazine, and Daniel Engber, a senior editor at The Atlantic, claimed that this study had been debunked and overturned in dramatic fashion by a newer study, published in 2021 by Johan Ugander and Jonas Juul, analyzing the same data. Kuperschmidt wrote, in an article for Science magazine, that our paper “used data on misinformation that had been fact-checked by independent organizations…” and that when Ugander and Juul “factored in this bias, the difference between the speed and reach of false news and true news disappeared.” Engber picked up on this thread, linked to Kuperschmidt’s article, and tweeted “I love this so much: Remember the Science paper showing that misinformation travels farther and faster on social media than the truth? It was wrong!”
News of the prominent debunking spread like wildfire. Engber’s tweet was retweeted 390 times and liked over 1200 times within a few days. The quote tweets cheerfully glorified the debunking. Dr Rohin Francis, Medlife Crisis on Twitter, tweeted “Absolute classic. That study everyone cited with righteous glee, that misinformation spreads faster than true information, was in fact misinformation.” His quote tweet was retweeted 68 times with over 250 likes.
Unfortunately, for us and for misinformation science, they were all wrong. After fact checking their claims, the journalists discovered that they had been the ones spreading misinformation.
When they talked to Ugander and Juul, they learned that the new study actually confirmed our work and replicated our findings: Fake news did reach more people than the truth, on average, and it did so while spreading deeper, faster, and more broadly through layers of connections. They also discovered that we had ourselves had double checked the generalizability of our results in a separate robustness data set of articles that had never been fact checked, which also confirmed what we had found.
Three separate replications had confirmed our results and, in fact, since we published our paper, many more studies have replicated our findings in a variety of data sets and contexts.
To their credit, both Kuperschmidt and Engber wrote detailed corrections explaining their mistakes and publicly apologized to us, and to Ugander and Juul, for so badly mischaracterizing both of our studies. Science published an official correction the next day. Engber wrote a detailed article about the mistake in The Atlantic titled “Sorry, I Lied About Fake News.”
Unfortunately, just as our paper predicted, the debunking of their false debunking got nowhere near the attention as their original false claims.
Engber’s correction only got 35 retweets and 48 likes. Dr Ronin Francis linked to the correction and tweeted “Huh. In the interests of trying to get to the truth, the plot thickens…are all of us who are writing or tweeting about this story…actually proving it right? Seems like it! Thoughtful attempt at correction.” But his correction only received 2 retweets. Kuperschmidt’s well worded correction faired better with 120 retweets. But that too was no match for his original misinformation, which was retweeted 332 times.
Ironically, since falsity spreads faster than the truth, many will now go on believing that our prominent paper has been debunked. This will likely affect the trajectory of the science of misinformation. That’s significant, because our study motivated much of the work dedicated to curbing misinformation in the last five years. Call it collateral damage in the war on misinformation.
But there’s actually more to the story…
Our study not only predicted that this would happen, but also gave an explanation for why it happens. Our paper showed that the speed of false news is correlated with the fact that it is more novel, surprising, shocking and anger inducing than the truth. Engber explained that this was, in part, why he was so drawn to spreading the false debunking. In his correction, he wrote “This was just the sort of thing I love: The science of misinformation is rife with mind-bending anecdotes in which a major theory of “post-truth” gets struck down by better data, then draws a last, ironic breath.” He was shocked, maybe even a little angry that our original paper had gotten the attention it did and he was happy to share this surprising new information with his followers. His followers expressed similar reactions: shock, awe and anger.
In fact, these are all anecdotal examples of what we called The Novelty Hypothesis. Human attention is drawn to novelty, we gain in status by spreading novel information because we are seen as being “in the know” or having access the “inside information” that may surprise our friends, family and social media followers. We are quick to react to information that is novel, surprising, shocking and anger inducing. That is, in part, why we are 70% more likely to retweet false news than the truth.
Engber and Kuperschmidt’s misinformation ended up proving our theory — and not just the headline result, but the mechanisms for why fake news spreads as well.
There’s another reason this type of falsity goes viral. The business models and algorithms of the social media platforms are designed to promote and profit from them. As I wrote in my book The Hype Machine Machine, the attention economy runs on engagement and social media platforms encourage engagement to maximize advertising profits. That’s a simplification and I go through all the empirical evidence for and against this claim in my book. But the basic contours of the argument are by now well accepted. Feed algorithms and friend suggestion algorithms combine to prioritize content that gets more (and thus encourages more) engagement. So it may be that the widely shared false debunking of our study was being promoted by the algorithms too.
Kuperschmidt and Engber’s corrections and apologies are commendable and encouraging displays of professionalism. But it turns out the whole ordeal is a masterclass in misinformation itself. Neither of them intended to spread misinformation! On the contrary — they both desperately wanted to correct their mistakes and get the truth out! And they tried.
But in social media, that’s just not how misinformation works. Falsity spreads faster than the truth and debunking rarely catches up to falsity… Well, in this case, the debunking of the debunking.
Maybe this truth, as I have laid it out here, will eventually spread farther than the falsity… I still have hope. As we all must when we pen our truths.
