Nicest Troll in Town

 

[Archival Audio Collage]

Chris Bail

I think we have kind of an outdated idea of what social media really does in the world. You know, we have this idea that we’re all individuals out searching for information.

Robert Pease (host)

That’s Dr. Chris Bail, Director of the Duke University Polarization Lab and author of the new book, Breaking the Social Media Prism: How to Make Our Platforms Less Polarizing. After some innovative research into what keeps us scrolling, he’s come to a new way of understanding online behavior. 

Chris Bail

We don’t think that social media is a competition of ideas. It’s a competition of identities.

Robert Pease (host)

What’s your online identity? Is it your better self, your alter ego, or an ongoing experiment to see what harvests some status and self-esteem? This is the Purple Principle, a podcast about the perils of polarization. I’m Robert Pease.

Emily Crocetti (host)

And I’m Emily Crocetti. Our guest today is going to challenge what you think you know about social media and its polarizing trends.  

Robert Pease (host)

Oh, If only the reds had a little blue in their feeds, and blues a little red. But it turns out, it’s not that simple.   

Emily Crocetti (host)

Dr. Chris Bail has been leading a team of researchers on social media for several years now. Let’s start off with a brief intro to that team and a very quick tour of Duke University’s Polarization Lab. 

Chris Bail

Sure, you’d see a pretty diverse group of people. So we are social scientists like me, sociologists, political scientists, but also data scientists, computer scientists, and statisticians. There’d be a lot of computers around. We would be talking an awful lot about social media and how it divides us and what we can do to reverse the course.

Robert Pease (host)

It’s interesting.You have such a wide variety of people and backgrounds. What are the challenges in working across so many disciplines ?

Chris Bail

There are challenges, but I think there’s many more advantages. So this question, political polarization on social media, is an inherently interdisciplinary problem. We can’t imagine approaching it without the theoretical perspective of say, social psychology, to understand things like the spread of misinformation or intergroup attitudes. But at the same time, we need to run experiments where we build – but more recently we even created our own social media platform to do research. And, you know, these are the types of things that are fundamentally questions about data science and computer science.

Robert Pease (host)

We do want to hear about the bots, and your prototypes and other tools. First though, you have a different view of what attracts or retains people on social media. Can you tell us about that?

Chris Bail

We have this idea that we are all individuals out searching for information, and we log on and we find some information we like, and we might share it with our friends, and the rest of the time we look at cute cat pictures or whatever. But really what I discovered in the course of writing this book is that in this era of increasing isolation – not only the pandemic, but even before the pandemic – social media is really shaping how we understand each other and ourselves. One of our most basic human instincts is to develop identities that make us feel good about ourselves. And that creates a different dynamic. We can experiment with more extreme versions of ourselves. We can selectively share parts of our lives, emphasize some parts, leave out others. But then social media also gives us powerful new tools for tracking what other people think of us. So we have now built into our platforms ways of monitoring likes, follows, engagements, and so on. And so what happens, as I’ve come to conclude after running many experiments and interviewing hundreds of people as they use social media, is that this social media prism, this thing that kind of distorts how we see each other and what we think of ourselves and each other, ultimately leaves us with a very biased view of what the other side thinks. If you’re a Democrat, what Republicans think, and if you’re a Republican, what Democrats think.

Robert Pease (host)

And that bias is so important and feeds into the title and main metaphor for your book, “breaking the social media prism.” Tell us about why you chose that metaphor and why it better describes our relationship to social media?

Chris Bail

So, I think the reigning metaphor right now is the echo chamber, or maybe the filter bubble. And so here, the idea is that the primary reason that social media is polarizing us is because it’s segregated us and trapped us inside these bubbles, or echo chambers where we’re insulated from people who don’t agree with us. And if our views go unchallenged, they get stronger and stronger. And actually, I also thought this was probably what was going on three or four years ago when we founded the Polarization Lab. We set out a pretty large study to try to take people outside of their echo chambers and see what happens. We actually recruited more than a thousand Republicans and Democrats who use Twitter, and we paid them to follow bots that we created in the lab that retweeted messages from people from the other side. And the idea was, of course, that if you take people out of their echo chamber, they should become more moderate. You should see their views getting challenged. They should become more self-reflective and they should kind of adapt their views. And unfortunately, what we found is more or less the opposite. So nobody became more moderate when they followed these bots for a month that exposed them to the other side. And some people, and particularly Republicans, became much more polarized. They kind of doubled down on their preexisting views.

[Archival Audio Collage]

Robert Pease (host)

Well, Emily, somehow that’s remarkable but not completely surprising at the same time. We learned from Dr. Abigail  Marsh of Georgetown back in Season One that we’re not so different from a herd of musk ox who cluster together when they perceive a threat.

Emily Crocetti (host)

And we’ve seen that all-important identity dimension in our two recent episodes on cults and conspiracies. First from Rachel Bernstein, the host of Indoctrination, who emphasized social identity and group membership as a huge element of cult-like appeal on and offline.  

Rachel Bernstein

And you have instead people who say, I want to be a part of something, I need to connect with other people who feel the same way. I want to go to marches. I need to know that we’re all speaking the same language. Those are the people who join cults. Those are the people who need to be surrounded by it so it becomes their entire life and their whole social circle.

Robert Pease (host)

And that same dynamic was clearly at work in our episode on the documentary Behind The Curve, about the large and highly social flat earth community. Especially in the case of the film’s protagonist, Mark Sargent. 

Mark Sargent

This is a flat earth meetup. And they flew me down for this because I’m infamous. Because – no, because I’m real. That’s what everyone said. You know, they want to, they want to shake my hands and, and make sure that they, you know, touch me. That sort of thing. Like, “Oh yeah, he’s a real guy. He’s not just a voice on a video.”

Robert Pease (host)

Is that what we’re looking for online? Our 15 thousand or 15 million clicks of fame? In his book, Breaking the Social Media Prism, Chris Bail describes Ray, a mild mannered guy offline. But online he’s achieved a degree of fame, or at least notoriety, through an extreme alter ego. 

Chris Bail

So in the book I write about extremists. And, you know, extremists are difficult to study. They don’t like to respond to the surveys that social scientists like me might normally use. They’re often not very excited about talking to journalists. And if people really are so different online and off, as I’ve been arguing so far, then probably the most important parts of extremism are kind of hidden from our view. And what I thought we needed was a way to track people both online and off. And so one of the new things that I did in the research for this book is interview people before and after they follow our bots and try to track and see concretely; hear the stories of people as they change in order to solve this puzzle. And what I found is that with this extremism, people are getting something online that they can’t get offline, and that is social status. 

The story that sticks most vividly in my head is the story of a guy. I called Ray in the book. And, you know, Ray’s an interesting character. When we first interviewed him, he seemed like the nicest, kind of most civil guy. He went out of his way to kind of denounce racism. He kept emphasizing that  he liked to reach across the aisle. Even though he was a conservative, he had Democratic friends and so on and so forth. And then one of the unique things we can do from our research is actually look at what Ray looks like online. And Ray is completely different. In  fact, Ray was one of the most extreme social media users we looked at in the entire study. I mean, stuff so vile that I can’t even describe it. And so this kind of question arises, you know, why is this guy so different in his online life and his offline life? Why is he the type of guy who goes kind of out of his way to distance himself? In fact, at one point, he told us, so many of these people on social media are big losers. You know, they’re yelling at everybody, but really they just live at home with their mother. It turns out that Ray lives at home with his mother. And so how are these extremists hiding in plain sight? And why? We think that one of the answers is that social media is providing them this new kind of status. And it is cult-like, it is reinforced by other extremists. It’s not a meaningful kind of status. Most of the people who like Ray’s posts are other extremists. And I don’t think they would have him over for dinner or be friends with him offline. But Ray gets something out of it. He gets so much that he’s created this completely different life for himself online.

[Archival Audio Collage]

Emily Crocetti (host)

And those Rays or alter-Rays of the online world spend huge amounts of time and energy generating extreme content. 

Robert Pease (host)

One of Chris Bail’s favorite data points is that only 6% of Twitter users generate the majority of content on that platform. 

Emily Crocetti (host)

Which we might have guessed. But still kind of incredible. 

Robert Pease (host)

And that really crowds out moderate voices on Twitter, as on other platforms.  

Emily Crocetti (host) 

But moderates are threatened online in a darker way as well. Dr. Bail describes another study participant, Sarah, blasted both for expressing an opinion and for an expression of sympathy. 

Robert Pease (host)

Before we enter Sarah’s own social media prism, let’s get a better sense of how Chris Bail and the polarization team studied online reactions of a thousand participants to opposing political viewpoints and then held personal interviews with a select group.

Chris Bail

We’re at an interesting point in social science right now, where, as I mentioned a little bit earlier,  we need data science to get creative and study online spaces. And I had been studying social media for about five or six years. I did a lot of early work on Facebook and we could collect a large amount of data, which for social scientists was really exciting because normally we can only survey a thousand people, maybe 2,000 people, or interview a couple dozen people. And all of a sudden we can get a couple million tweets in a few minutes. So when I first set out to study echo chambers, I thought, okay, well this is going to be easy. We can just get all the data and we can look at what people think, and we can see who they’re exposing themselves to. And we can do that pretty easily with some social media data. And we could try to figure this thing out. But the problem is, there’s a chicken and egg problem, right? So if it’s the case that I  expose myself to people who are like-minded, then what I might be actually measuring when I go to do the analysis is not that my social media connections are changing my mind, but that I’ve arranged my social media network in such a way that it just reinforces my views in the first place. So in other words, what we really needed to do is conduct an experiment. And we wanted some people to follow these bots that we built that expose them to people from the opposing party. And then we wanted another group of people not to do that. And then we surveyed them about a lot of different political views before and after following the bots. And then we look at whether the people who are following the bots changed their views more than the people who didn’t follow the bots changed their views. And sure enough, this is how we found that  taking people outside their echo chamber actually seems to make them more polarized and not less.

Emily Crocetti (host)

Could you tell us an anecdotal type of situation about one or two of your participants in the study that you came to know fairly well and how the feed of opposing viewpoints affected them?

Chris Bail

Let me give you another great anecdote. One of the most interesting people that we met in the course of our research was a young, moderate conservative woman who I’ve called Sarah. And Sarah was really genuinely a moderate. She went to an Ivy league school. Her father was a cop. She grew up in New York. But she has generally centrist views, maybe just right of center. She still reads the New York Times. She even reads the New Yorker sometimes, even though she definitely identifies more as a Republican than a liberal. And when we met her and we started asking her as we did with all the people we spoke with, “so tell us about your time on social media, what’s going on, who you have met, what’s been good, what’s been bad.” And she immediately went to two stories. The first story was about the time that she responded to a tweet from the NRA, kind of mildly supporting people’s right to own a handgun, saying that her husband had owned a gun and occasionally recreationally went to a shooting range. And then she told us that within minutes, her feed is just lit up with people telling her that they hope that their children find the gun and shoot them. I mean, just reprehensible stuff. Another story she told us was about the time that she was engaged in a bit of a disagreement with a liberal woman and in an effort to kind of de-escalate the tension in their conversation, she went to this user’s profile and she saw that this woman was a breast cancer survivor. And Sarah too was a breast cancer survivor. And so she thought, “well, I’m going to try to connect with this woman around this issue.” And she sent her a message saying,” I saw you had a double mastectomy, so did I.  And I just wanted you to know I’m glad you’re okay.” And then she says, this woman writes back, “I hope you die.” And then she tells this story of how this led her away from social media. This is what happens when you talk about politics on social media, I think were her exact words. And you know, this wasn’t just an upsetting night. This set Sarah on a course to really recalibrate a lot of her relationships. Because she comes from New York, a lot of her relatives are more liberal. But then she started unfollowing all her friends from college as well, all the liberal friends she met when she went to this Ivy League school.

And the result is that the very people whose opinions I think are most important for us to hear, the moderates, many of whom I think are your listeners, are invisible. You know, the data point I love to share is that about 6% of people on Twitter are responsible for the majority of the content about politics on Twitter. And that 6% of people have really extreme views. So the social media prism, it’s amplifying extremists, it’s fueling extremism. But it’s also muting moderates who really have no interest in status seeking on social media, right? And so the result of these two processes, emboldening extremists and muting moderates, again, I think leaves us all feeling much more polarized than we really are.

Emily Crocetti (host)

Sad but true. And Sarah is hardly alone there. Moderates are something like an endangered species online. Even their opportunities for procreation are limited. I’m thinking of that one  famous blog title, “Always Swipe Left on Moderates.”

Robert Pease (host)

Of course, why would anyone want to have moderate, rational discussion with their life partner on things like child education and medical coverage?  

Emily Crocetti (host) 

Exactly. So we also need some advice on our own online behavior regarding both identity and just plain sanity.   

Robert Pease (host)

Let’s hear more from Dr. Bail on his three important steps to neutralize the effects of the social media prism. 

Emily Crocetti (host)

And his hopes for more honest and civil online discourse promoted by an alternative social media platform developed at the Duke Lab, which they’ve named DiscussIt. 

[Archival Audio Collage]

Chris Bail

So in my book, Breaking the Social Media Prism, I have three ideas about how we can create this kind of bottom up movement to reverse the course. The first is that we need to learn to see the social media prism. So we need to learn to understand that when we say someone, when we see someone saying something extreme on social media, that this person is probably not a moderate member of the other party. They’re probably part of that 6% of all Twitter users who account for 73% of all tweets about politics and those people we know have disproportionately extreme views. So on polarizationlab.com, you can access tools that actually help you identify trolls and learn the language of trolls. We like to say, learn to speak troll. 

The second thing we can do is understand how the social media prism reflects ourselves. So is the person that we are projecting on social media the person that we want other people to see? And so learning to locate ourselves on this continuum and ask this question of, what is our civic responsibility, what do we want to be as a people? And how is social media going to get us there? And so we also have tools on polarizationlab.com that are described in my book that  actually will read your tweets and position you on a continuum that ranges from extremely liberal to extremely conservative.

The final thing we can do is to learn to talk with the other side more effectively, to engage in more productive conversations. So one of the first things we’ve done with our tools is create something we call “the Bipartisanship Leaderboard.” So with our large sample of Twitter users we’re able to identify prominent Republicans who often receive lots of likes from Democrats and prominent Democrats who often receive a lot of likes from Republicans.

And finally, we have tools that track hashtags that both Republicans and Democrats are talking about that we hope will allow people to find the issues where there seems to be room to connect. And so we hope that this tool will help people follow the conversation in the middle, the conversation that involves Republicans and Democrats, finding some common ground as difficult as it is to imagine that at this highly polarized moment in our country, there’s actually lots of examples.

Robert Pease (host)

Well, that’s good to know, but sometimes a little hard to get at. And we’d like to hear a little bit about Discuss It, which is your prototype or your lab’s alternative to Facebook, and how current social media platforms operate?

Chris Bail

So we need bottom up solutions. We need social media users to change their behavior. But in the long term, and if we want these changes to be sustainable, we also need top down reforms. We need to ask the following question: If we could redesign our social media platforms from scratch, how could we optimize for our better behavior and how could we incentivize better behavior? You know, why should we expect a platform that was created to allow Harvard undergraduates to rate each other’s physical attractiveness to serve as the public square for the 21st century? 

And the answer is that I think a lot of these platforms emerged to serve the idiosyncratic needs of small circles on the internet and really kind of cobbled together strategies to try to make them the forum for democracy that they’ve become. But nobody has ever done the research to say, well, what actually works and what doesn’t? So we decided to instead create our own social media platform. And so we kind of came up with this lame generic name, DiscussIt. We reached out to people and told them, could we pay you to help us test this new social media platform that we created called DiscussIt, where you’re going to engage in conversations with people about a variety of issues? Unbeknownst to the people at the time, we gave them an invite code to the platform that paired them with a member of the other party to have a conversation about one of two divisive issues, either immigration or gun control, that unfolded over about a week.

And what we discovered was really exciting. We discovered that people who engaged in anonymous conversation on this platform, structured conversation with a member of the other party, exhibited substantial decreases in polarized attitudes. And perhaps even more excitingly, we saw that this effect was even stronger for Republicans than Democrats. In our earlier study we had shown that taking Republicans outside their echo chamber on Twitter had made them much more conservative and it made Democrats a little bit more democratic. But in this new study, we were seeing the opposite thing. We were seeing that anonymous conversations seem to be even more important for Republicans.

And so we hope this is a first step towards identifying the design principles from scratch, identifying the design principles which will promote social cohesion instead of incivility.

Emily Crocetti (host)

That was our featured guest today, Dr. Chris Bail, Director of the Duke University Polarization Lab and author of the recent book, Breaking The Social Media Prism: How to Make Our Platforms Less Polarizing, published by Princeton University Press.

Robert Pease (host)

This book will challenge many of your beliefs about the online world, including that the solution is to completely disengage.

Emily Crocetti (host)

Dr. Bail says that’s a huge step backward, not a step forward.

Robert Pease (host)

Or that government regulation will have the intended effects of moderating polarization online.

Emily Crocetti (host)

Dr. Bail says we have a supply side problem. And the supply of hyperpartisan content will not be easily regulated or dissipated.  

Robert Pease (host)

But alternate platforms designed from the ground up for more civil discussion may have real potential at moderating, and mediating, our polar divide. Meantime, we suggest you read Breaking The Social Media Prism and evaluate your own online behavior, as well as those you bump into.

Emily Crocetti (host)

Are you spending your social media time simply looking for information? Or is there a whole lot more at work? 

Robert Pease (host)

Next time on the Purple Principle, we’ll log off social media and switch channels for the first episode of a two-part series on how Hollywood chooses its Presidents, particularly its  independent Presidents. 

Emily Crocetti (host)

We’ll meet with Rod Lurie, very likely the first TV creator to cast an independent as POTUS. That was Geena Davis as President Mackenzie Allen in the 2005 series, “Commander in Chief.”

Rod Lurie

And so what I attempted to achieve was actually achieved for a little while. I definitely made her an independent. And I tried to make the topics we were dealing with something that both sides could relate to. 

Robert Pease (host)

We hope you’ll join us for that episode, and if you enjoyed the show, please share our episodes with a friend. You can also subscribe to our newsletter, like us on Facebook, follow us on Twitter and Instagram, and leave us a review on Apple Podcasts. You’ll find all of these links in our show notes and on our website, purpleprinciple.com. This has been Robert Pease and Emily Crocetti for the Purple Principle Team: Alison Byrne, Producer; Kevin A. Kline, senior audio engineer; Emily Holloway, digital operations and outreach; Dom Scarlett, research associate. Our resident composer is Ryan Adair Rooney. The Purple Principle is a Fluent Knowledge production.