Twitter Outrage, Charted: The Partial Anatomy of the #FuckPhyllis Trend, or Why I Don’t Trust BuzzFeed
by michael_simeone
On Sunday, January 26, Chancellor Phyllis Wise of the University of Illinois at Urbana-Champaign sent out notice that despite below-freezing temperatures, classes at the University would resume according to schedule on Monday morning. The prospect of attending class in these conditions upset many students at Illinois, prompting an uproar on Twitter reported by BuzzFeed and the creation of the phony @ChanPhyllisWise twitter handle as well as the #FuckPhyllis hashtag. As evidenced by a dozen or so embedded tweets, the #FuckPhyllis trend contained despicably racist and sexist insults directed at the Chancellor. Buzzfeed also reported that students turned to change.com to produce a petition calling for classes to be cancelled, which garnered over 7,000 signatures over night. According to social media analytics service Topsy, the #FuckPhyllis hashtag only included some 2,000 tweets at the height of its activity. And while there is no disputing how reprehensible the conduct of these students is, we should look closer at this trend: who is tweeting, when, about what, and in agreement or to echo whom? BuzzFeed calls as much attention to the hate speech as possible (seemingly implicating every petition signer in the social media hate-fest), but to whom did this hashtag, now effectively dead, ultimately belong?
When considering the scale of Twitter, 2,000 tweets is a very small number of communications. But the reason #FuckPhyllis is so interesting (and likely the reason Buzzfeed paid attention) is how rapidly it trended:
Today the hashtag sees barely any activity at all. But in the lifespan of this particular social media spike there may be some interesting patterns surrounding the network topology of different communities of tweeters. It’s important to suspect that many of the offensive tweets have since been removed by their authors, but many do remain. Still, this is a hole in our data. Assuming many of the most vile things are now gone, all we can do is examine hostile vs. corrective communication networks rather than a deep look at racist tweets of unknown quantity and verbiage.
What follows is a quick look into the #FuckPhyllis hashtag, the twitter activity surrounding the topic on January 26, 2014, starting at around 10pm CST all the way up to 5pm CST the following day. I used NodeXL to pull tweets that used the #FuckPhyllis hashtag, and so the limitations of the Twitter API are present: not every tweet is included in this analysis. Also, embedded tweets appear to refer to UTC rather than CST (making tweets from the 26th appear to be dated on the 27th of January) Still, there are some general trends in the hashtag that are illuminating.
First, lets take a look at a graph of the tweets and retweets from the two hours of the life of the hashtag:
In this graph, lines or edges indicate a retweet, and loops indicate tweets that are likely responses to tweets by the same user. Thus, the appearance of many loops lumped together shows people holding conversations by replying to their own tweets. Each node is a tweet. Node colors are determined by a modularity algorithm (Girvan-Newman) that groups nodes by shared connections. What is shocking about this graph is how many tweets are actually reacting against the original negativity of the #FuckPhyllis trend. The node highlighted in red, with by far the most retweets, is @suey_park’s retweet about white privilege from U of I student Briana Walker (who is synthesizing comments made by Park earlier on):
what does #fuckphyllis have to do with the cheif? white supremacy’s blind quest to defend itself. (pro-@suey_park, anti-racist for life)
— Brianna Walker (@bwalkerLIS) January 27, 2014
In fact, nearly every tweet that has a truly graphical property–retweeted by overlapping communities of agents who in turn are retweeted–rather than a simple hierarchal layout (see the red subgroup in the lower left of the above graph) contains or retweets a message of anti-racism. This sentiment ranges from simple eye rolling to more sophisticated thoughts on white privilege at U of I. Within a few hours of the first hateful tweet, social media conversation was dominated (centered on) by reprimands and anit-racist commentary. The graph demonstrates that it is the anti #FuckPhyllis tweets that have a high Eigenvector centrality (used in network analysis as one way to study influence), or in other words a high number of retweets and mentions that also happen to be retweeted and mentioned. The influence of anti-racist tweets appears to outstrip that of anti-Wise tweets.
One of the most popular remaining anti-Wise tweets (although this contains zero racist overtones) is represented by the aforementioned red subgroup demonstrates several retweets of the following:
One for NIU, two for you ISU. You go ISU. ….And none for U of I #fuckphyllis
— Bridget Anselmo (@bridget_elyse02) January 27, 2014
Even though the above tweet was retweeted a total of 64 times in its total lifespan, the network reveals that it has little influence over the general social media conversation as it evolves. Now in the upper left of the next graph, this particular tweet remains marginal in the following 17 hours worth of tweets:
The dark blue component in the middle of this second graph continues to feature @suey_park (although this time for a different but similarly anti-racist tweet). @suey_park’s tweet again has the most mentions and retweets by those who are also retweeted and mentioned. As before, the graph is predominantly an anti-racist backlash, and the most retweeted of the anti-Wise messages is not overtly racist or sexist.
By comparison, here is a graph of tweets that include “ChanPhyllisWise”, the title of the now deleted bogus twitter account used to mock the Illinois Chancellor: Similarly, this graph has been clustered by modularity. The majority of tweets shown on this graph demonstrate the kind of vitriol reported by Buzzfeed, and here are the two central dark blue and light blue tweets that rest in a pair at the middle of the graph:
I don’t even go to U of I and I find @ChanPhyllisWise hilarious You guys have support here in Minnesota #fuckphyllis
— Adam Cox (@agregcox) January 27, 2014
Unhelpful College Chancellor @chanphylliswise pic.twitter.com/HAFdAPs4uw
— Clare Bonistalli (@Cbono93) January 27, 2014
Both of these tweets demonstrate a relatively high total degree centrality (retweeted more than peer tweets) but a comparatively low Eigenvector rating (retweeted, but not by those who were themselves retweeted). Furthermore, we can see the topology of this graph to be markedly different from the first two. Where as the first two showed resonance around a small set of tweets, the third shows significantly more isolated conversations or tweets that were not retweeted at all. The conversation that was significantly more sexist, racist, and hostile in tone is also one that features more fractured conversations, less information exchange, and lower connectivity among all nodes.
Conclusions
I’m tempted to hypothesize that conversations that feature this kind of hostility in social media have a performative quality to them, and users appear to want to one up one another. The network topology of the third graph, for what little information we have, suggests self interest and not very much consensus beyond using the same hashtag and adopting an insulting tone. The first two show preferential attachment to @suey_park and some consensus about who is “right.” It would be interesting to listen in on similar (and sadly inevitable) Twitter trends as they emerge and again compare the topology of hostile and corrective social media networks and see how they stack up. From their networks of communication, we can see a difference between principle and anger as motivating principles for social media use. As Christopher Simeone put it when presented with this case, “principled actors who see themselves as part of a bigger cause or purpose behave differently than those whose only uniting principle is rage and self interest.” Or perhaps we don’t have enough information to tell yet.
What’s interesting, however, is that the BuzzFeed article, while calling attention to the racist, sexist, and hateful things U of I students tweeted, greatly over-represents the salacious portions of this social media trend. We’ll never know exactly how many tweets were deleted out of shame, but analysis of this tend shows a swift and harmonic response that obliterated anti-Wise sentiment and replaced it with a new conversation about white privilege. Yes, U of I students filled out a petition that got 7,000 signatures, but that does not equal 7,000 racists. Clicking a bubble that tries to get one out of class is very different from taking to social media to spread hate. The real A missing piece of the story is the response and unity of response to #FuckPhyllis.
Source files from NodeXL:
fuckphyllis hashtagfuckphyllis ChanPhyllisWise
Update 1/28/2014 9:50 MST: I want to encourage everyone to see Kevin Hamilton’s comments below, as they raise some important concerns, and I feel the need to clarify. I do not believe that the University of Illinois is a place free of racism and white privilege or a place where anti-racism somehow excuses acts of racism. I do, however, see the data discussed above as showing a contrast in approaches to communication that may map on to angry vs. principled tweeting, and, crucially, how divided the University can be when it comes to issues of race. While there is persistent and unjust quarter given to racial intolerance on campus (highlighted below by Kevin’s comments), this should not obfuscate that there are principled actors at U of I, and that the story of the place is deep and storied division on racial issues, not thorough moral decay.
Oh, and 1/28/2014 11:20 MST: I’m a U of I alum (2011)
Update 1/29/2014 4:32 MST: The joint statement published today by U of I President and Board Chairman
Update 2/3/2014 11:12pm MST: Part 2, pertaining to social media stories and where they go wrong
suffen.us 
Weird question, but is there any chance of posting the NodeXL data?
They should be posted under “source” at the bottom. Thanks!
…well, that was a late-night reading comprehension fail. Thanks!
Michael, I’m having trouble understanding your conclusions and correlations here. Are you looking to demonstrate the relatively small number of tweets within this event that contained hateful language? You also seem to be looking to assert some kind of conclusions about the motivations of different speakers based on how their tweets were received. All in all, this seems like a version of the argument made by many (including Champaign’s Mayor), that there are more anti-racists than racists present in this social event. If so – and I may be wrong – I question the usefulness of such an approach. I don’t think many would question the “performative” nature of the racist tweets, any more than they would question the “performative” nature of racist-themed frat parties. I’m disturbed not by the number of racist Illinois students on Twitter, nor by what social media activity reveals about peoples hearts (very little), but by the fact that these sort of “performances” even *occur* to someone in our community as acceptable. A ground already existed for the racist tweets to fall on, a ground that has revealed itself multiple times in student use of social media on this campus (see the racist memes two years ago). To understand the problem in terms of speakers, and their quantity, and not in terms of listeners, and their readiness, is to miss the point. The Buzzfeed article, I would argue, from having watched it form, was not an attempt to capture “buzz” in a quantitative way. Rather, it was an instrumental attempt to use the “perception of buzzworthiness” associated with that site to make visible a practice that would otherwise have benefited from obscurity and ephemerality. I saw that article not as evidentiary argument, but rather a judgment in public. We deserve judgment here at Illinois, not for how many racists we have in our midst, but for how structural racism allows repression to exist *without* individuals having to experience the ugly business of explicit, personal convictions about the relative worth of persons.
These are great points, and I’m grateful to have them added to the conversation. I agree with you about Illinois deserving judgement, as well as, in context, how social media can only tell us so much. And I certainly am not trying to say that the reaction exonerates the community, especially given its past. However, the data from twitter suggests that Illinois is a place that is actively grappling with this problem, not a place so rotten that it is beyond redemption. I know you are not maintaining the latter, but I saw here that the story of Illinois seems more like a place that is deeply divided over race, not one where racism enjoys full immunity.
Actually, I thought one of the strengths of Michael’s analysis was, that it doesn’t just count speakers, but tells us something about the nature of the “listening” that was or wasn’t involved. If the racist discourse didn’t produce a lot of retweeting and community-formation, that seems to me a relevant fact. I’m not saying it proves racism is a big, little, or medium-sized problem — just that it’s worth knowing.
The article is one of the better quality examples of Buzzfeed journalism. That’s not saying much, but you still owe it a close reading. It doesn’t “greatly over-represent the salacious portions” or leave the impression that there were 7000 racists on the petition. It says that the trend started with people who were critical of the decision and it never generalizes from this criticism to racism/sexism as if the latter were just built in: “After that, for some, the attacks took a racist turn.” Also, it quotes the introductory paragraph of the petition, so the reader can see that it doesn’t contain racist or sexist terms. You don’t evidence your claim that they’re exaggerating in the first place, you announce that they’re exaggerating and then you “disprove” it. Since they don’t seem to be exaggerating this raises the question of why you want to do this. “The real story is the response and unity of response to #FuckPhyllis” – why? Suppose there are two stories, a trend and another trend, why is only one “real”? And why is it even a question whether speakers “are racists” or not and whether they feel a shame that is invisible and unknowable or not? There are two tendencies here, to count heads and to speculate on what people were feeling or who they are. Both tendencies are quite familiar and they contribute on the side of actively NOT helping people understand racism and sexism.
Among Dora’s apt points here is the question for Michael about the description of negative reaction to the hashtag as the “real story.” That’s where this post does, I believe, get more prescriptive. And it’s a point I think Michael might helpfully expand to help situate his study.
So my prescriptive tones are rooted in my valuation of social media data as relational. Twitter trends aren’t just numbers of tweets or hashtag counts; they have a structure that is interesting and can be context-dependent. So I don’t see this as two trends, one after the other. I see this as two trends, interlocked, and played out across several networks of tweets.
If this were to develop into a full study, I agree that the connection between the two graphs (#fuckphyllis and ChanPhyllisWise) should be receive more attention.
But my general bone to pick with making news out of trending social media and then picking out a dozen tweets (1 percent of the total) is that it doesn’t show any relationships among those speaking, which is the whole reason social media even produces trends in the first place.
I can honestly say that as a student at the University of Illinois, no one was particularly outraged at having to go to school and no one was even trying to change the Chancellors mind via twitter. It wasn’t to get anywhere and it wasn’t out of white privilege. It started as a joke and on a small scale no one would have noticed. No one would have flinched at the hashtag and it mostly contained innocent jokes referencing Frozen and bad decisions. Unfortunately, there will always be people who cross the line, no matter where in the world you are. Anti-racists latched onto those that were overtly racist and sexist and used it to fuel their own vendetta against racism as a whole. The problem is, just like these students were hiding behind a keyboard, so were Suey Park and all the others. The hashtag at the university was perpetuated for maybe an hour and a half, and now it’s been going for an extra two days because of all the attention brought to it by anti-racists. No, racism is not okay and it should never be excused and if I were the chancellor, I would aim to give some form of punishment to those students who clearly went way too far. I’m assuming that’s what the charts were showing, that in relation to the actual tweets from students, the majority of tweets and the perpetuation of it all came from others. I was sitting in class the next day reading response tweets and was personally offended by all the people calling us lazy and by saying equally as vulgar things as the original offending students did. We are by no means lazy or incapable of hearing the word no. Quite the opposite actually. Most people I know really try to avoid missing class even when we feel sick or hungover. We are a well respected university with hardworking students. The hashtag (not the hashtag itself, but the tweets accompanying it) had cleverness and wit behind them and that would’ve occurred even if the chancellor was a white male (and to be honest most of us didn’t know Phyllis Wise was Asian). If anything I think the charts show that counter-hate was actually perpetuated much more and thought acceptable because our campus is primarily white, but in reality it counter acted hate with hate and contrary to popular belief, racism can work both ways. The spreading of Suey Parks white privilege theory has caused more of an uproar than necessary. Simply reporting the offenders to the University would have sufficed. I appreciate this article because it articulates what my friends and I have been discussing here about the situation the last few days.
“Quite the opposite actually. Most people I know really try to avoid missing class even when we feel sick or hungover”
“Hungover ” … yup, those are U of I undergrads, for sure.
and as for “Unfortunately, there will always be people who cross the line, no matter where in the world you are. ”
… way to shrug off the content of what they were saying. A few bad apples indeed. Very bad apples, and totally representative of a part of UI undergrad subculture.
After reading this post, I think I have a much better grasp of what actually happened than I did from reading the Buzzfeed article.
For me that’s the bottom line. I don’t think the post exculpates anyone, or tells us how we ought to react. People are still free to form a range of different judgments about the broader context of student life at Illinois. But they can now do so in an informed way — knowing something about the nature of the social networks involved, and not just the texts of a few tweets. I see that as a straightforwardly good thing.
If our question is about whether the Buzzfeed article is a good picture of events as they happened, then I agree with Ted – Michael’s work tells the story better. I can imagine how such representations might help inform the strategies of journalists or activists in how they intervene or report. Looking to the work of people like Gilad Lotan, for example, we might grow to expect certain sizes of spikes for a given audience’s collective attention, regardless of subject, and to know when a spike has reached a point of unusual focus, thus constituting an event “worth looking at” from the perspective of journalism’s traditional charge of determining newsworthiness. Work like Lotan’s, though early, begins to suggest that certain social groups will be predisposed to large swings in focus, regardless of the subject in view.
Journalism doesn’t always cover that which constitutes an aberration in a community’s rhythms, however. I would argue that the role of the Buzzfeed article was less reportage and more editorial, or “prophetic” (to use some old literary/anthropological language). Given that the “article” was in fact a serial, chronological presentation of screengrabs framed as events, my point might be a hard sell. But the piece ends with commentary quoted from emails from two individuals who were both part of the trend (and most likely, I believe, the people who alerted the Buzzfeed author to the events in the first place). Their commentary helps situate the piece more firmly in the genre of judgmental pronouncement.
So people are right to react to the article as a representation of a campus, and not just a report on certain individuals’ bad behavior. But this representation isn’t coming from the perspective of identifying spikes – the article wasn’t algorithmically generated based on activity, nor a report of a quantitative spike. The piece was a provocation based on a perspective of longer views of the subject.
Most of the discussion about this week’s events haven’t focussed as much on the accuracy of the Buzzfeed piece as reportage, and more on whether the article was right to see the hateful tweets as representative of the campus. This is the debate Michael steps into through his conclusions about the “real story” in the twitter stream. Was the Buzzfeed piece right to see in the hate-tweets a manifestation of much wider racism? Michael’s study seems to answer “no” – but based on analysis of this event, and not of the wider space in which it happened.
If we’re asking the question “What motivated the spike in attention around this hashtag” then we’re asking a different question, one analyses like this one might be very helpful in uncovering. But then we’ll still be left with the broader and much older questions of how media coverage of events does or doesn’t help us understand the underlying principles of a community or culture. And in those questions it becomes even more important to acknowledge how a community’s very organizing principles – good or bad – affect which stories get attention and achieve trustworthiness in the first place.
I would just add that my answer isn’t quite “no” to the question of the representative power of this social media trend. More like, “this social media trend could be indicative of widespread racism and entitlement among students (which could indeed be corroborated by past excellent and non-quantitative studies), but it does certainly demonstrate a committed group of people at the University committed to resisting it”
For sure – and, I would note that the article pointed to the same, quoting pretty much the same number of anti-racist tweets as racist tweets . It would be very interesting to do some similar studies that examine and compare the patterns of racist reactions to anti-racist trends on twitter to those of anti-racist reactions to racist trends. Maybe based on your tool-sharing I’ll give it a try…
I’d happily work with you on this
I would also like to see the number of retracted tweets (if there is a complete dataset somewhere) as a signal of personal delayed reaction. That is a powerful signal.
Agree very much. Twitter won’t release them, and I haven’t yet found anyone quick on the draw who collected them earlier on. Still holding out hope, however, and the relational tendencies among racist/anti-racist tweeting is something I want to keep looking at.
Again there is, not just now but always, a question about *why we’re having this conversation right now* and not some other one. That’s my question. Any blog post, suggests what is worth talking about. That’s why you write it. So I agree with Kevin, “If we’re asking the question ‘What motivated the spike in attention around this hashtag’ then we’re asking a different question” than “how can we understand what happened in a way that helps us with a critical analysis of racism and vice versa”? My question is why, when something racist has happened, you want to ask #1 when it seems obvious that #2 is more pressing. It is true that you certainly do not tell people to stop asking #2. But neither did you incorporate it. You yourself did not feel compelled to ask it as part of what you were doing. And I would say that if you did incorporate that question you would’ve had to write something quite different. To put this another way you make no reference to scholarship on racism, to the decades of deep work from ethnic studies and critical race theory. You assume that digital media analysis can go it alone and add something trenchant. The result suggests that it cannot.
Katie’s response above shows the *overlap* in thinking, among students who created overt racist tweets and those that were disgusted by them. This no digital media analysis will ever be able to approach, relational or not, without actively engaging the archive of critical race studies and that would change its assumptions to the very core.
This post was not meant to be an authoritative meditation on racism at the U of I. It is meant to fill in some holes left open by garden variety social media journalism.
I don’t think I ever wrote to answer the first question you outline. My interest is in the relationships among tweets, (not just that they simply exist or existed) which I believe is crucial to answer question 2. Yes, there is great scholarship and journalism out there that is pertinent to this situation. I hope to add to the conversation with a look at not just what was said, but a guess at who was listening and repeating what was said. These kinds of things, I hope, give us a better idea of what exactly we’re looking at when someone uses social media as an instrument for performing privilege or enforcing injustice.
I understand the distinction you’re making between existence of tweets and relationship between tweets. I was trying to say that those points about relationship assume that you know which are the racist tweets (the ones that contain slurs) and which are the non-racist tweets (the ones that object to the ones that contain slurs), so that there can be a relationship between two different trends. But people who object to overt racist slurs may not have any understanding of what constitutes racism or how it works; historically, they’re often invested in pushing a narrow idea of racism. To put it strongly the idea of a relationship between trends assumes that there is a non-racist position possible in the liberal institutional situation. Many scholars would say that there is not. I was not asking for an “authoritative” meditation on this kind of question, only for any incorporation of such issues at all.
I understand what you’re saying. I think to clarify: I don’t believe there’s a non-racist position, rather, an angry and aggressively sexist and racist collection of tweets (and indeed some that don’t seem racist and sexist but disclose a motivating contempt) and a corrective backlash.
And I really appreciate your point about the narrow definition of racism. A number of people retweeted this post to say “see, those people are in the vocal minority. U of I students in general do not think this.” This really bothered me and I wish I had been clearer. To me the social media shows a conversation starkly divided by racial controversy, but I agree, just because there is anti-racism does not mean that somehow the problem of white privilege has been solved on one side of the controversy line and not another.
I also saw some activity on Reddit/r/uiuc that basically said “we knew this would happen, so the chancellor could have prevented it by letting people know about classes earlier.” This corroborates the point you’ve been making about attempts to make racism as invisible as possible, or something that just is and is the responsibility of others to anticipate and manage rather than a systemic wrong.
[…] So, I believe that everyone must not rush into judgement when something starts trending on social media. This study by an Illini alum reveals that there were 2,000 tweets at the height of the #fuckphyllis … […]
[…] When I wrote about the racist, sexist tweeting reported by BuzzFeed, I wanted to call attention to the ways that copy/pasting or embedding tweets as a way of reporting or calling attention to events is an unhelpfully simplistic way of starting a public discussion about a given social media trend. We know that there is limited data within social media alone, but by reporting Tweets by merely listing half a dozen tweets implies a representative quality to a tweet that may not exist. For instance, the article in question headlines the Twitter activity at U of I with “After Being Denied A Snow Day, University Of Illinois Students Respond With Racism And Sexism.” The piece goes on to list 11 total tweets that ranged from cruelty to hate speech, followed by 9 tweets and a few comments by U of I alums that condemn the activity of the anti-Chancellor tweeters. 7 of the 11 spotlighted hate-tweeters have since deleted their accounts. But this approach of “balancing” an approximately equal number of pro and anti tweets, with each group ranging in intensity, is very limited. First, let’s appreciate the good things this story did: […]
[…] I am no stranger to Ms. Park. Earlier in the year she derided her alma mater’s entire student population for being racist because an obscene hashtag directed at our chancellor trended on Twitter. Later, it was made pretty clear that the number of tweets with the hashtag was not particularly high, and the number of racist and sexist tweets was even smaller. […]
[…] the negative reputation being built on their beloved university Twitter users started to retaliate, a lot. Posts Like “I’m sure all of you will look back at this in 10 years and be ashamed of […]
Thank you for any other excellent post. The place else
may anybody get that type of information in such a perfect way of writing?
I’ve a presentation subsequent week, and I am on the look for such info.
[…] a more in-depth approach to the Twitter discussion using social network analysis, take a look at Michael Simeone’s work on his blog. Simeone is a former University of Illinois researcher with the National Center for Supercomputing […]