We had a choice between cooperation or competition and you won't believe what happened next.
We had a choice between cooperation or competition and you won't believe what happened next.
Issue 69: The story behind a recent paper--how we were forced to choose between competing with scientists doing similar research or cooperating with them
We had a choice between cooperation or competition and you won't believe what happened next.
Being scooped to a discovery is a terrifying prospect for many scientists. This occurs when two research teams are racing to a discovery only to have one team publish their findings faster.
By one analysis, scooped papers receive about one-quarter fewer citations than papers that are the first to report the same discovery, and it can be very difficult to publish a scooped paper in a top academic journal. This incentive structure can make science feel like a dog-eat-dog occupation. The mere thought of getting scooped is enough to ruin the day for many scientists (as you can see in the comic below).
This is why it was a shock when Jay and his PhD student Claire Robertson received an email from the editors of Nature Human Behavior telling them that another research group had submitted an extremely similar paper. At the same time. Analyzing the same dataset. With nearly identical results.
This was a shocking coincidence and something we had never experience—or even heard of before.
To make matters worse, both teams of scientists had submitted their manuscripts within three days of each other. We were then faced with a major moral and professional dilemma: We could race to publish our findings at another journal to avoid getting scooped (and give up our hopes of publishing in this prestigious journal) or cooperate with people we didn’t know to secure an unknown outcome while sharing the credit. Which would you choose?
As Claire put it, in a recent article:
I don’t have to imagine this, because this is exactly what happened during the publication process of our new paper, “Negativity Increases Online News Consumption.” My team (made up of myself, Dr. Philip Pärnamets, and Dr. Jay Van Bavel) and another team (Dr. Stefan Feuerriegel, Kaoru Schwarzenegger and Dr. Nicolas Pröllochs) had unknowingly spent months working on similar projects, both using the same incredible dataset from the Upworthy Research Archive (Matias et al., 2021)
The Upworthy Research Archive is a dataset containing years of real-life news consumption data from Upworthy.com. Upworthy.com used A/B testing from 2012-2015 to figure out what features of headlines would lead the most people to click on an article. For each article they published, they tested up to nine headlines via random assignment to Upworthy readers. So, for a story about the Supreme Court striking down Prop 8, which prohibited gay marriage in California, two possible headlines were “Wow, Supreme Court Have Made Millions Of Us Very Very Happy,” and “We’ll Look Back at This In 10 Years And Be Embarrassed As Hell it Even Existed.”
We decided to take the pathway to cooperation and combine our papers into one unified paper. With that, we were introduced to our new collaborators. Most of the credit for scientific papers goes to the first authors (who are usually students or postdocs eager to establish their independence) and last authors (who are often seasoned professors, overseeing the project). We quicly decided to distribute credit by having the two lead authors share the position (Claire and Nicolas) and the two last authors share that position as well (Jay and Stefan).
Our teams approached the theoretical questions from very different backgrounds, with our team coming from a background in social & cognitive psychology, the other team from the field of computer science. This created a lot of initial friction because our approaches to analyzing the same dataset were different—we had divergent theoretical and methodological approaches. Distilling two papers down to one unified narrative was frustrating for everyone involved.
But forcing everyone to get on the same page turned out to be highly fruitful in the long run. Due to the diversity of perspectives on our new team, our paper grew stronger theoretically and technically by bringing together the best of both groups. And the final results were extremely robust—we came to the same main results no matter how we decided to compute the data.
We analyzed news stories from Upworthy.com—a site known for viral stories and cringy clickbait headlines—to see what drove engagement with news stories. As one NPR article put it:
At its peak, the site, which is founded on a mission of promoting viral and uplifting content, was reaching close to 90 million people a month. Seemingly overnight, the newish company had changed the way the world wrote headlines, what it meant to create viral content and what kinds of expectations people bring to what they're clicking.
Since Upworthy was famous for A/B testing headlines for their news stories, we were able to analyze 22,743 randomized experiments they conducted on readers. These stories generated roughly 5.7 million clicks across more than 370 million impressions! This was the biggest and most impressive study our lab had ever conducted—by a long shot.
We found that negative language in news headlines increased the likelihood that a headline would be clicked on. For a headline of average length, each additional negative word increased the click-through rate by 2.3%. In other words, for any given news story, people are more likely to read it if you give it a negative headline.
We also found that positive language decreased the likelihood of a headline being clicked on. This is a pretty sad outcome for a website that promised to deliver uplifting, positive stories—it fell victim to the old psychological adage that bad is stronger than good, otherwise known as negativity bias. Or as they put it in the news business: “If it bleeds, it leads”.
Thankfully this saga about negativity also provided a more positive lesson: cooperation beats competition. In our case, the benefits of working as a team made the project vastly stronger and provided a paper that we hope will stand the test of time. We were also able to publish the paper in Nature Human Behaviour, where it came out this week to a very positive reception. Within a few days of publication, the paper has already been downloaded more than 26,000 times and is one of the most widely discussed papers in the history of the journal.
This underscores the value of cooperation and working in diverse teams. As we noted in “The Power of Us”, diverse teams are often more effective as long as they can operate under a common identity or work towards a superordinate goal. It is unfortunate that science does not reward cooperation more frequently, incentivizing teams to come together to analyze a rich dataset, solve a perplexing puzzle, or answer a big question. Instead, we are typically pitted against one another in a race to get research out quickly.
But doing important, breakthrough work increasingly requires teamwork—especially in science and other technical fields. And getting it right should be more important than getting it done first. Understanding and leveraging the science of identity and cooperation offers a powerful pathway to improve research and provide better answers to the hardest questions. Hopefully more people will make the same choice we did and opt for cooperation over competition in their own work.
News and Updates
Last week, Jay spoke on the Passion Stuck podcast with John R. Miles! They discussed:
Resolving political deadlock
Encouraging diversity of opinions and rallying for transformation
How to exhibit proficient leadership
Stimulating initiatives to tackle enduring global issues like climate change
Learn more about THE POWER OF US
If you like our newsletter, we encourage you to check out our award-winning book “The Power of Us: Harnessing Our Shared Identities to Improve Performance, Increase Cooperation, and Promote Social Harmony”. You can learn more about the book or order it from the links on our website (here). We keep the newsletter free, but are extremely grateful if you have a chance to purchase the book or buy it for a friend who wants to learn more about group psychology.
Catch up on the last one..
Last week, we released the second episode of our podcast on the solutions of polarization. Our special guests include pastor and political activist Evan Mawarire, who argues that “the loss of democracy is not an event, it’s a process”. He says that even small cracks in democracy must not go unchallenged — and draws on his personal leadership and community building as effective ways to protect democracy.
