9 Concrete, Specific Things We Actually Know About How Social Media Shape Elections

What studies sanely say about democracy in the age of Twitter, Facebook and YouTube

President Obama speaks Monday at a campaign event in Columbus, Ohio. Social media allow candidates to talk up their platforms without journalism's "filter" (or a teleprompter's "filter," for that matter). (Reuters)

Nearly a quarter of American time spent on the Internet is spent on Facebook. 15 percent of online Americans tweet. Two-thirds of Americans use a social network.

And there's an election in November.

Opinions about the Internet's potential effects on politics range from the thoughtlessly utopian to the haughtily dour. We don't often get a reality check.


How do social media actually shape elections? What does research say about how social media have affected elections in the past? When you take public discourse and run it through the Twitter spaghetti maker, how does it look?

Some of these may seem obvious, but the goal is sanity over bloviation. Having looked at the literature, here are the nine things I think we can definitively say about how social media can shape elections, based on the way it already has done so, in elections around the world:

N.B. As always, let's just accept the limits of the phrase social media. I'm using it to mean the Web 2.0 Friends: Twitter, Facebook, YouTube, maybe a little Pinterest on the side. All media are social; all media are user-contributed (ditto all "knowledge") and "social media" works well enough to refer to both.

1. People and campaigns mostly use social media for dissemination, not dialogue.

"Neither campaign made much use of the social aspects of social media."

So says the Pew's study, released last week, on how the American Presidential campaigns are using the web in 2012. The abstract of that study describes social media "mainly as a way to push messages out" without having to pass those messages through a journalistic filter. It also noted that the only place where campaigns used non-professional, "citizen" content was in the Obama campaigns's news blog. Everywhere else, professionally-made content reigned. Twitter accounts even rarely or never retweeted ordinary voters:

[J]ust 16% of Obama's tweets over the two-week period studied were retweets. The Romney campaign had just one retweet during this period -- something from Romney's son Josh.

Pew just looked at candidates, though. So to understand how social media shapes an entire election's playing field, you have to look globally.

Many fewer Swedes use Twitter than Americans (someything like one to eight percent of Swedish internet users), but that may make studying the country's tweets easier. In the country's 2010 Twitter elections, Anderse Olof Larsson and Hallvard Moe studied all global tweets tagged with #election2010 (in Swedish, #val2010). They found that dissemination still trumped. A full 60 percent of the tweets were "singletons" -- tweets that weren't an @-reply or retweet. Only seven percent of tweets tagged #val2010 were @-replies. This could be a sign of simply what tweets get hashtagged, but I think it gestures, too, to the fact that most of what happens on Twitter is broadcast.

2. Campaign websites remain the hub of US Presidential campaigns.

This is a Pew finding, and it ties right into how campaigns use social media for broadcasting. Write the study's authors:

Even if someone starts on a campaign's social network page, they often end up back on the main website-to donate money, to join a community, to volunteer or to read anything of length.

They add that this has heightened over the course of the campaign: A July redesign of the Obama 2012 site made it so that rather "than sending users to the campaign's YouTube channel, the video link now embeds the campaign videos directly into the website, where the only videos are the ones Obama wants you to see."

3. But non-major parties tend to converse more on Twitter. Especially pirates.

In their study of the Sweden 2010 election, Larsson and Moe made networks of who was retweeting and replying to whom. While "elites" (meaning, here, the major political parties) dominated both networks by volume, the "non-elite" (minor) parties appear more active in both. Younger, more upstart parties -- like the Feminist Party or the Pirate Party -- used Twitter more to chat with or retweet ordinary voters. Minor parties don't address the huge social audiences faced by the larger parties, and they may just be more familiar with the street:

So smaller parties can harness the benefits of chatting and retweeting ordinary citizens.

A comparable situation in the US? Buddy Roemer. A former Democrat who ran in the Republican Presidential primaries, Roemer was a ubiquitous and prolific force on Twitter last fall, retweeting and linking and replying and conversing. He's now slightly less active, but the lesson remains: Non-major candidates can generate cachet by using Twitter like, well, Twitter.

4. Elite journalists converse, too -- with each other.

Among the top Twitter users in that Sweden 2010 election (which we'll keep talking about for the next few points), there were a couple of elite journalists -- think of them like @MillenniumMagazineBen.. They all used Twitter to converse, but almost all that talking happened within their closed circle of professional journalists. We call Twitter a tool for conversation, but it's as much a theater of dialogue.

5. But there's a potential for nonelite, anonymous users to succeed.

Three of the 10 most active Twitter users in Sweden's 2010 election were anonymous, writing satire or sardony. (@PourMePåsköl.) Anonymous users can rise to the top on Twitter, the study's authors highlight, which satisfies their interest in seeing diverse and nonelite voices find an audience. But, as the Twitter accounts are anonymous, it's unknown if they had a "headstart" by actually being elites in the first place.

Note that this has always happened: The Publius of the Federalist Papers was an anonymous user, producing user-generated content for national distribution networks. Publius could write anonymously because text makes anonymity really easy. Twitter's innovation, then, is that it allows anonymous writers to participate at the speed of discourse: to share a link before anyone else, to snark during Presidential debates, to fret while election results flow in.

6. Discussion tends to happen around news events that are already being covered.

This is another finding of the Sweden 2010 study. Check the graph below: it graphs the number of tweets against time. All those big spikes are tied to televised debates, well-covered rallies or the election itself.


Social media may democratize who gets to participate, but its aggregate decisions about what's news align pretty well with what the mainstream media think is news.

7. So what remains special about social media is that nonelite users control distribution.

Last month, a team of researchers led by New Mexico State's Ivan Dylko released their study of YouTube videos about the 2008 American Presidential election in the journal New Media & Society. Dylko's team looked at the most popular political videos every month, and those featured on the front News and Politics pages. Then they divided the figures depicted into "elite" and "nonelite" piles: the elite being major parties and media organizations; the nonelite being everybody else.

And what did they find? Check this chart, which looks at the percentage of video content variously featuring, made and posted to YouTube by elites and non-elites (as they defined those two groups):


Elites -- the candidates, but also the broadcast journalists -- dominated as the subject of the videos and as the makers of the videos. But "nonelite" people had posted almost all the videos.

Here's another graph that makes that more clear. This graphs the percentage of footage in each video created by pros:


Most of the videos were mostly made by the journalistic elite. But, in the words of the study, "over one-third of the most popular videos in the sample relied on no traditional media content at all."

The lesson here is nuanced: most political content in the 2008 election concerned candidates and was filmed by network camera crews. Media democratization of the sort predicted by the most radical technofuturists wasn't going on -- but ordinary people controlled what content made it to YouTube, and a non-trivial amount of content was being filmed and posted in its entirety by regular people.

8. Participating more in political discussion online doesn't necessarily increase political knowledge.

In a study published this year in Computers in Human Behavoir, Occidental Colleges's Meredith Conroy looked at how participation in political Facebook groups correlated to offline engagement. That part of her study, alas, only examined undergrads at a single college -- which makes its findings of rather limited use -- but it also contained a whole second part, which analyzed the content of a wider swath of political Facebook groups.

Conroy's team wrote:

Our content analysis indicates that political Facebook group users, in general, often do not share much new information and the information they do share tends to be somewhat inaccurate, incoherent, or not very well supported with evidence. As a forum for people to easily engage and share their opinions, online political groups are beneficial; however, as a forum to learn new political information online political groups are ineffective due in part to low quality wall discussion.

That's more than semi-obvious, right? (Cf. Reddit, as reported by Gawker's Adrian Chen.) But it builds toward a central point:

9. The huge effect social media have in elections, then, is that they allow nonelites to frame and distribute content made by elites. For better or for worse.

This shows up across the board, in study after study. In Dylko's YouTube study of 2008, power came from ordinary people posting and distributing professionally-made videos. The study of Sweden 2010 noted the same. And in this election, the Pew study found that candidate-issued tweets fell or rose based on whether people retweeted them.

The biggest change that can occur, then, is that framing by social media can shift how the professional media itself frames stories. Social media feeds the loop of news judgement. A study of a media scandal in India, written by Maryland professor Kalyani Chadha, found Internet distribution of certain primary sources helped make the corruption they documented into a story in the professional media. And Lei Guo's study of Chinese citizen media in the same journal distilled a number of elements. Citizen media could frame and affect state-controlled media if it created a "turning-point" that could no longer be ignored. Citizen media, in other words, framed the media's telling of a story, which refracted back onto how the media told the story. But in that refraction, elements got lost: a major theme in the Chinese citizen media was individual vs governmnet, and state media didn't convey this.

* * *

Reviewing these few points, I don't have huge answers, but I want to add a few closing thoughts.

The first: Yes, I'm assuming there are certain norms that stretch across the world, that Swedish Twitter use is usefully generalizable to American Twitter use. I do so tripping and bounding over philosophers who think about this stuff full-time, including those who worry that American or Western ethics about society and privacy are coming to dominate "home-grown" ethics in other societies. But I think, given the conclusions above, Sweden is useful enough.

The second is that I keep thinking back to this quote, from Conroy's inquiry into Facebook. It asserts a particularly demanding definition of the term healthy democracy:

The fundamentals of democracy assume a knowledgeable public, one that is capable of representing its own self-interest effectively. A healthy democracy, then, should see tandem movement between political knowledge and political participation. Here we find that while online political group membership is correlated with offline political participation, we do not see an equally significant correlation with levels of political knowledge.

I'm not sure any democracy in history has been that "healthy" or seen participation and knowledge bloom together. But I think that quote's worth meditating on.

Why? Because all these points drive to a question that, for me, seems to be at the heart of how we talk about new media -- or democracy, or politics, or even the Constitution. It's:

How does speech, in some form, affect what happens in the world?

For a media critic or new media tech-writer, this, I think, is the question. Every inquiry flows from this.

And perhaps (finally) here's an answer -- in the form, of course, of a specific instance. Reviewing all these points, I kept thinking back to the Occupy movement. It was held (sometimes too perfectly) as the pristine embodiment of what social media looked like in the offline world. My editor, Alexis Madrigal, even wrote an algorithmic version of it. The camps have since disbanded, but still, nearly a year after it began, Occupy's shift of how we frame economic issues has remained. Debates about the country's economy mention its inequality in addition to its debt.

Long after the media and the immediate movement have faded, the frame lives on.