Online, there are no city councils, brick-and-mortar libraries, or recreational sports leagues. Yet communities form all the same.
Saif Shahin, SOC professor and Internet Governance Lab faculty fellow, researches the links that bind and build these digital groups, including those founded on hateful beliefs. In a recent paper, “White Twitter: Tracing the Evolution of the Alt-Right in Retweets,” Shahin dug into nine years’ worth of retweets to understand the transformation of White Nationalism on social media ahead of the 2016 election.
Shahin sat down with American as the inaugural guest on the 30 Minutes On podcast to recap his research and expand on the past, present, and future of social media dynamics.
Listen to the podcast, or read the full transcript below:
Andrew Erickson: Hello and welcome to 30 Minutes On a podcast from American magazine. I’m your host, American magazine staff writer Andrew Erickson. We’re launching this podcast as another way to connect with our readers, tell stories, and much like in our 3 Minutes On… page in the magazine, dive deep into interesting subjects by speaking with experts who are members of the AU community.
Today, in our inaugural podcast, we’re spending 30 minutes on social media with School of Communication professor Saif Shahin. He’s a faculty fellow in SOC’s Internet Governance Lab, and an associate editor at the Journal of Information Technology and Politics. As a researcher, he examines the big questions behind big data, digging into issues intersecting digital media and culture, including media sociology, social computing, social justice, and global media and politics. Last year, Shahin and a colleague, Margaret Ng, a researcher at the University of Illinois at Urbana-Champaign, researched the evolution of the alt-right on Twitter. They compiled more than 300,000 retweets between 2009 and 2016 to learn more about the technological and social elements behind White Nationalism, and how the alt-right managed to carve out a more pronounced role in American politics ahead of the 2016 election. We talk about that research, the evolution of social media, and what online interaction might look like in the future. Here’s that interview. We hope you enjoy.
Professor Shahin, thank you for joining us and agreeing to be the inaugural guest on the 30 Minutes On… podcast, we’re so excited to have you.
Saif Shahin: Thank you for having me, Andrew. I’m delighted to be on the show.
AE: So, I just wanted to start off—last year you published an article and in January presented it with your colleague at the Hawaii International Conference on System Sciences, “Tracing the Evolution of the Alt-Right in Retweets.” Can you talk me through how you and your colleague decided to look into this, and this specific time period as it relates to Twitter and the alt-right?
SS: Yeah, so in this study we look at what is broadly called the alt-right movement, more specifically the White Nationalist movement between 2009 and 2016. The time period is pretty self-explanatory, really. The whole movement really started in reaction to Barack Obama’s election in 2008 as the first black president of the country, and there’s this conservative ideologue, if you will, called Paul Gottfried. He gave a speech about two to three weeks after Obama’s election in which he, for the first time, used the term the alt-right, or the alternative right, at least in the way we understand it right now. That speech pointed the failures of the quote-unquote neo-conservatives, who were in power under George W. Bush, failures in terms of leading America to a point where it elects a black president and, hence, a need for conservatives to return to their roots, specifically white, Christian, American roots. Some of them even, in fact, called themselves paleoconservatives, in contrasts to neo-conservatives. But, yeah, that’s really when the whiten nationalist movement, or the alt-right movement as we think of them right now, began. And, so, in our study, we look at this movement and its spread online, and on Twitter specifically, from the beginning of 2009 to the end of 2016. So the argument, really, is that this movement wasn’t anywhere, and it didn’t even exist before this period. It started off, like any movement, as a very fringe movement, and then within this eight-year period, it moves from the fringes to the center stage of American politics. And so what we do in this study is we take a year-on-year look at how this movement grew, what factors could potentially explain its growth. And so yeah, that’s what this study is about.
AE: And why Twitter specifically, and why retweets? What can retweets tell us that replies or favorites can’t tell us or that can’t be captured?
SS: Twitter has been a favorite for political conversations broadly since its outset, but also for conversations that lie at the intersection of race and politics. So, we all know about ‘Black Twitter’ as a phenomenon and people started talking about ‘Black Twitter’ as early as 2008, 2009, when a report from the Pew Research Center talked about ‘Black Twitter.’
AE: And that’s in the infancy of Twitter.
SS: Yes, absolutely. Twitter launches in 2006, and two to three years later we have this phenomenon, so there’s this very close connection between racial identification, if you will, and politics, which plays out on twitter and has since its outset. And anecdotally we also know that Trump uses Twitter all the time. Also, in 2016, I found a study which looks at the presence of extremist groups on Twitter. So this study, this was more of a white paper that came out of George Washington University’s project on extremism. Which, I think it was titled ISIS versus Nazis, something like that, so there has been interest in extremism online, but specifically Islamic extremism online, and on Twitter. How Islamic extremist terrorist groups use Twitter for various purposes, and so this study talks about how non-Islamist groups such as White Nationalists are beginning to do that. So I took a look at that study, which had identified a few of these groups, and what I found really interesting was that we tend to think of the alt-right movement as this one consolidated movement for many people, just lunatics, really. That was the perception for a lot of people. But what this study showed was that there were different kinds of alt-right groups. There was the Ku Klux Klan, there were Nazis, there were Neo-Confederates, and there were also supporters of these cult-like figures such as Doctor David Duke. And so this study basically identified Twitter accounts that were propagating the alt-right ideology in broad terms. And this happened to come out in 2016, and we all know what happened in 2016, and so it made me think about what happened in the years preceding 2016. So, how this movement had basically grown and what could be the potential reasons for its growth, as it happened online, that’s what led me to conduct this study.
Retweets are a good means of identifying the growth of a particular movement in this case, but it would be something else as well. People can interact with tweets in a variety of ways. They could like it, reply to it, things like that. Retweets specifically are typically understood to imply support. You basically retweet something because you want more people to read it. People who follow you, who know you, you want them to retweet as well. There’s a difference between just retweeting and retweeting with comments as well. If you disagree with something, you might still retweet it but retweet it with a comment and say what ever you want to say about that.
AE: I feel like we always see that statement in somebody’s Twitter bio of ‘retweets don’t equal endorsements,’ but they often do.
SS: The reason why people say that is because retweets typically do imply endorsement. Not always, of course, but in general, yes. And the other thing is that retweets are also a sort of identity signal, if you will. When you retweet, you often think about what people are going to think of you as a person. When you are posting online in general, that happens, but specifically with retweets as well. And so in research literature, it is known that retweets can lead to the formation of communities and groups. You see something, you want more people to see that, you retweet that. Communities don’t form around one single retweet, but if you retweet a particular idea expressed in similar tweets or if you retweet a particular organization or individual a lot, when a lot of people start doing that, then that’s how communities start forming on Twitter. And so, for the purposes of this study, that’s exactly what I was interested in, looking at how retweets posted by the accounts identified as belonging to different alt-right constituencies on Twitter were being retweeted and how these retweets grew over time between 2009 and 2016. So that’s the reason why I chose to look at retweets with my colleague.
AE: By 2016, looking at the charts you and your colleague compiled, we’re talking retweets in six figures and David Duke, kind of centralizing around him, we’re talking about 100,000-plus retweets. How do you and your colleague make sense of so many tweets and how do you begin to analyze something so complex and something in such large numbers?
SS: Yeah, I mean, data analytics has become very important in our lives as we have more and more data that we produce and have to live and work with. Data analytics as a field has grown significantly. I actually teach courses related to data analytics here at American University. In this paper specifically what we did was use a method called social network analysis, which identifies connections across Twitter users through retweets, in this case. So, one person retweets another person, and that’s a connection. A lot of people are retweeting the same twitter accounts, then they end up having these connections that form a community or a group. Social network analysis can then show you how these groups grow over time. Who are the people who are at the center of these groups? And when I say people, it could be individuals, it could be organizations as well. Anyone who has a Twitter account, basically. We use social network analysis to look at how these groups emerged and grew but also the cross-linkages across these groups. How much are people who are retweeting, say, Ku Klux Klan a lot, also retweeting Nazis or Neo-Confederates, and so forth? How do these trends change over time? Initially what we found is it was the Nazis and the Neo-Confederates who were kind of the key groups or clusters as they were called in network analysis, and then eventually it was Doctor David Duke who emerged as a significant third group and eventually took over the conversation. So by 2016, the whole network kind of coalesced around Doctor David Duke and that’s kind of when it became so big. So one of the key arguments that me and my colleague make in the paper—we argue that there is significant activity in 2015 as well, there’s growth in 2015, and then in 2016 it’s just basically going off the charts.
So, our argument is that Trump didn’t create White Nationalism as is sometimes assumed, but what Trump did was, when he emerged in 2015 as a potential candidate, he enabled different White Nationalist constituencies to come together and to rally around him. So he became a figurehead. They were able to shed their own differences and find this person around whom they could rally, and that is what made the movement so strong and powerful, at least online. It will be interesting to see how these trends happened offline, which is something that we do not cover, but our study, our online study, definitely indicates that this is a possibility offline as well, where different varieties of White Nationalists started to converge around Trump in 2015, 2016, and that is what made them as powerful as they became and enabled them to basically shape the election.
AE: And was there anything about how these groups connected or about the scale of it that surprised you and your colleague?
SS: Well, really, it was surprising for us to see how marked these differences were, at least initially. Because just like everybody else, we thought that these White Nationalists were one constituency at the far right of the political spectrum. So it was really revealing because in general terms we know that, we hear about echo chambers online, everybody knows and talks about them, but we think of echo chambers primarily as Democrat slash liberal versus Republican slash conservative. So, two large echo chambers is what we think about. But what we found was that not just the conservative echo chamber broadly but the very far right end of the conservative echo chamber is itself splintered into a lot of these factions to begin with at least. So echo chambers are much more deeply entrenched than we typically talk about. That was something that we found that was very interesting. Just the fact that the differences across Ku Klux Klan and Nazis and Neo-Confederates, we would think of them as belonging to one constituency, but they didn’t, and these differences were real. So, people who were retweeting Nazis a lot were not retweeting Neo-Confederates, and vice versa.
That was the first thing that was surprising for us, the entrenched nature of differences among White Nationalists. And then also the process through which they jelled eventually. They came together and that was less surprising, more revealing. We could very clearly see how Trump’s emergence changed the lay of the land, if you will and, like I said, enabled these splintered groups to come together. Also, the scale, eventually in 2016, like you said, there was significant growth in 2015, too, but then in 2016, it’s like several, several times larger.
AE: It just exploded.
SS: It just explodes in 2016.
AE: Obviously we have another election coming up in November. It feels like the leadup to it has been decades long. What can we expect, not just on the far right and on the fringes of far-right extremism, but what can we expect on the left, the right, and what are seeing as far as social media trends that maybe could shape this 2020 election?
SS: That’s very hard to say right now. I haven’t done a study so far on what’s going on, but through the bits and pieces of analysis that I have done mostly through working with my students in the classes that I teach on data analytics, I think one big factor, especially among Democrats, the Democratic primaries that are going on right now, I think one thing that has surprise me perhaps, although it shouldn’t really be a surprise, is that a lot of Republican support for Trump is not just pro-Trump, but it’s also anti-Democrat or anti-Democratic Party in nature. I see people with bios that very explicitly talk about them being a former Democrat or how much they hate the Democratic Party and things like that. So that is something that I don’t know how much of it is recognized by the Democratic Party as an institution and by Democratic Party supporters. And that is something that I feel can play a significant role in the general election. Recognizing that, if you think of it from a Democratic Party perspective, is something that should factor in the primaries as well. So my gut feeling is that a candidate who is least associated with the Democratic Party would have a chance of getting more of the swing votes in the general election. That is something that, at least as far as social media trends go, is something that has struck me as being quite significant.
AE: Switching gears a little bit, I know you started out as a journalist. You wrote about politics, defense, social issues, and you were the news editor at Mint, which is India’s second-largest business newspaper. Can you walk me through your journalism career, how you got your start, and how you made your way as a journalist?
SS: Yeah, I started writing for newspapers in India. I’m from India originally. I started writing for newspapers in India while I was in college. I was doing my undergraduate degree in journalism, did some work on television as well, and eventually moved to working for online news organizations. Moved around quite a bit, studied in England, did my master’s there, so used to write for BBC’s website and other local publications while I was there. I also worked in the Middle East for a few years in between, then went back to India. But that was around the time when I started thinking about moving away from journalism and building a career in academia. So, some of it had to do with the fact that as a journalist, you work on these daily stories.
AE: Headlines and deadlines.
SS: Headlines and deadlines, exactly. After a few years, it starts to be a bit too much. Intellectually, it was no longer satisfying for me that I would write an 800-word story—and I know you can still do longform journalism, but it’s rarer and not everyone gets to.
AE: With daily deadlines it can be hard, yeah.
SS: So that was one big reason. The other reason was my general interest in politics and international relations as well. I wanted to study these things more, you know, learn and understand them better intellectually. And so that’s what made me think about moving into academia. So, yeah, I moved to the US, joined the PhD program at the University of Texas at Austin, briefly worked at Bowling Green State University, and have been here at American University since the fall of 2018.
AE: Is there any aspect of social media that caught us by surprise? It seems like the internet was always thought of as this way to bring us together, to bring us closer and build these connections, but at the same time we talk about bubbles and things that drive wedges between us. Do you think people, not just academics, but regular people, got it wrong and predicted this wrong about the impact of not only the internet becoming more mainstream, but social media platforms as well?
SS: So, this certainly happened, and it’s not the first time that it happened either. Even with television, some people saw television as leading to what one theorist called a global village. So, people across the globe, across nations becoming interconnected, and living as though they were living in a village where everybody knows everybody and things like that. That was the expectation from television at the outset and then eventually that’s not what happened. Toward the late 1990s, there were books talking about how television was leading to the breakdown of even local communities. People were just staying at home and watching TV and they were interacting less and less with each other in terms of human, face-to-face interactions and the social and political consequences of that phenomenon.
Similarly with the internet, initially, it was expected that the internet would break down all kinds of boundaries—physical boundaries, some people talked about the post-national turn where nations would no longer remain relevant in a world connected by the internet. Others talked about how internet allowed us to leave behind our physical limitations, bodily limitations, if you will, so basically the expectation that race would stop being relevant because online, people could be anything. So those were the expectations, and red flags were being raised right from the outset, but people who suggested otherwise were in the minority up until, I guess, 2016, when we realized that’s not exactly what was happening. In general, that was not the direction in which the world was moving and then, specifically, the role that internet and social media were playing in leading to a more fragmented, splintered society. I think it was quite a surprise to a lot of people. There are many people who have written books post-2016 who have acknowledged that they had very different opinions initially and they have very different opinions now.
AE: I see you use the word identity, or technosocial, a lot in your writing. How do technology and society come into play in your research? How have you come to define that word ‘technosocial’? And what does the future of studying something like that look like?
SS: So technosocial really relates to the idea that our social experiences are significantly shaped by technology, even as technology itself is shaped by the kind of society that we live in. It’s a mutually constitutive process, if you will. For instance, in our experiences on social media, as I was saying, we have not been able to leave behind our physical realities, be it race or religion or whatever, when we go online. Our online activities and actions are significantly shaped by who we are offline and, again, the paper that we were talking about, I called it ‘White Twitter’ because it’s focused on White Nationalism. We also touched briefly on the phenomenon of ‘Black Twitter,’ so my paper is a take on the phenomenon of ‘Black Twitter.’ So, like I said, it became apparent in the very, very early days of Twitter.
So these social activities, if you will, shape our online activities, and then what we do online, or in digital spaces, broadly speaking, technologically mediated digital spaces, they shape our lives offline as well in terms of the fact that social media as we now realize, played a significant role in the outcome of the 2016 election. Not just in terms of the growth of White Nationalism, which is what my research focused on, but in so many other ways that scholars have written and talked about and other commentators keep talking about in newsprint as well. Even before that, if we talk about the Arab uprisings in 2010, 2011, onward, they were significantly influenced by the fact that people certainly had social media. So even as our physical social realities shape our activities online, our online activities end up shaping our politics and our society, so that’s what technosocial means, this close interrelationship between technology and society.
AE: I realize it’s literally impossible to predict, but what do the next 15, 20 years of social media look like? What does the next million, billion-person platform look like?
SS: I wouldn’t only think in terms of how social media might change in the next few years, but how some of the underlying technology that enables social media is changing and what kinds of larger changes that it could bring. So I think one big change that we are going to see very soon is the fifth generation of mobile telephoning. Social media became as big as they did because of the emergence of the third and then the fourth generation in the aughts and then the 10s.
AE: Once people moved on from the Motorola Razr to the first few generations of the iPhone.
SS: Yeah, and with greater ability to transfer data and at much faster rates with the third generation and the fourth generation people started moving away from—earlier, the social aspect of the internet used to be places like chatrooms and bulletin boards and things like that, which were very sort of text based, organized around topics and themes and so forth. Not that social media didn’t exist previously—Myspace and all of that—but with Facebook and Twitter, that changed significantly, and the social aspect of online started becoming a lot more organized around our profiles on these platforms. So we moved on from these topic -and-theme-based platforms to more profile-oriented platforms such as Twitter. The other change was the growth of multimedia. Because of our ability to share data at much faster speeds, we moved on from just posting text to photos and eventually video. And then in the last two years, I would say, video-oriented platforms such as Tik Tok have exploded. That’s a result of, again, increasing internet speed.
It’s really important to look at the underlying technology. It’s the changes in those technologies that led to the emergence of social media to begin with, and the shift from ‘Web 1.0’ to ‘Web 2.0’ as we talk about in terms of the growth of user-generated content online. I think with 5G, that is going to change a lot. We will see a lot more VR, AR-based applications becoming a lot more popular. The internet of things will become much bigger than it already is. Again, 5G will not introduce these changes, it will just make them operate on a very different scale. And that, I think, will lead to wider changes in the kinds of applications that we use and the ways in which we use technology. So, yeah, I think that social media will continue to change and evolve. Something we are already seeing, like I mentioned, is the increased use of videos, and so with 5G, we might see more virtual reality-oriented social media applications. And then the other big thing I see happening is with artificial intelligence. So, again, that’s not something new. It has been around for a lot longer, but the use of artificial intelligence has been growing continuously and I think with 5G, we will see the ability to control more and more things more broadly. So that will make artificial intelligence-based applications more common. So those are some of the major changes that I see happening in the technology space broadly. And then, of course, they will have larger social and political consequences.
AE: Quickly, before we get out of here, what are you teaching? What are you working on now?
SS: I’m teaching a couple of courses this semester. One is called ‘Digital Media and Culture,’ in which we talk about the same things we were just talking about now: social media, society, its implications for culture, politics, how the two feed off each other, issues of identity and community. It’s amazing how much I learn from my students because they really are living these technologies. A lot of what we think of as digital media culture is the lived experiences of our undergraduate students.
I learn a lot from my students in those classes, and the other course I am teaching is called ‘Taming Big Data,’ and that is primarily about data analytics. So, I teach students how to mine data from social media platforms, from websites, but also from news archives and from other sources of large quantities of data, and then how to analyze them. I teach them a little bit of coding in that class. I teach them things like machine learning and sentiment analysis. Most students in that class are undergraduates, but I do have a couple doctoral students as well. The ability to work with and analyze large quantities of data, just becoming so central to pretty much everything that we do right now. It also helps them understand some of the larger social issues that pertain to what we call the datafication of our society, so the implications of all the data that we are generating and collecting and sharing for privacy for instance. Exactly how companies and governments are able to work with that data to learn as much as they are able to learn about us and make very precise profile about us. Being able to work with data allows them to not just know that intellectually but actually be able to see how that happens, and so that’s one of the cool aspects of that class.
I’m working on a couple of interesting projects. One is kind of along the lines of what we were speaking about earlier, which is, in fact, how nationalism hasn’t gone away but has, in fact, evidently become mor entrenched in 2020. Twenty years ago, people were talking about how internet was going to create a postnational society. Today, we have populist leaders around the world who are contesting on a platform that is built on nationalism. So even their appeal to racial or religious identities is closely tied to national identity. Trump’s ‘Make America Great Again’ slogan is one example of that. So, that paper explores how nationalism is expressed online and how deeply entrenched that is and what are the different ways in which that happens, how it relates to our other identities like race and religion and even partisan identities. So that’s one project I’m really excited about.
The other project is looking at online conversations about Europe’s General Data Protection Regulation, which was implemented in the middle of 2018, and purportedly that regulation was meant to help common users of the internet have more control over their personal data and put constraints on the ability of big tech to misuse users’ personal data, but what I found is that online conversations about these regulations have been taken over by technology companies, which is really important. Because instead of privacy rights activists or even lawmakers who worked to create their regulation, it is companies like Facebook and Google and a bunch of smaller technology companies who are driving the conversation about what this regulation is, its implications, its effects, and things like that. The conversation shifted from this regulation being about privacy rights to this regulation being bad for business. So that shift and how it happened online, that was interesting to see. I hope to publish that paper sometime soon as well.
AE: Great, well, I look forward to reading it.
SS: Thank you very much.
AE: Thank you so much for joining us and for agreeing to be our first guest. I appreciate it.
SS: Once again, thank you for having me. This was delightful.
AE: That will do it for the first episode of the 30 Minutes On podcast from American magazine. Thank you so much to our inaugural guest, Saif Shahin, for joining us and for chatting about his research in social media. Be on the lookout for our March 2020 magazine, which hits mailboxes soon. And we welcome your feedback. Please let us know what you think—about the magazine and the podcast—by emailing firstname.lastname@example.org. And be sure to follow us on Twitter and Instagram at @AU_Americanmag. Thank you for listening. We’ll see you next time.