elections – The Policy and Internet Blog

Stormzy 1: The Sun 0 — Three Reasons Why #GE2017 Was the Real Social Media Election

After its initial appearance as a cynical but safe device by Teresa May to ratchet up the Conservative majority, the UK general election of 2017 turned out to be one of the most exciting and unexpected of all time. One of the many things for which it will be remembered is as the first election where it was the social media campaigns that really made the difference to the relative fortunes of the parties, rather than traditional media. And it could be the first election where the right wing tabloids finally ceded their influence to new media, their power over politics broken according to some.

Social media have been part of the UK electoral landscape for a while. In 2015, many of us attributed the Conservative success in part to their massive expenditure on targeted Facebook advertising, 10 times more than Labour, whose ‘bottom-up’ Twitter campaign seemed mainly to have preached to the converted. Social media advertising was used more successfully by Leave.EU than Remain in the referendum (although some of us cautioned against blaming social media for Brexit). But in both these campaigns, the relentless attack of the tabloid press was able to strike at the heart of the Labour and Remain campaigns and was widely credited for having influenced the result, as in so many elections from the 1930s onwards.

However, in 2017 Labour’s campaign was widely regarded as having made a huge positive difference to the party’s share of the vote – unexpectedly rising by 10 percentage points on 2015 – in the face of a typically sustained and viscious attack by the Daily Mail, the Sun and the Daily Express. Why? There are (at least) three reasons.

First, increased turnout of young people is widely regarded to have driven Labour’s improved share of the vote – and young people do not in general read newspapers not even online. Instead, they spend increasing proportions of their time on social media platforms on mobile phones, particularly Instagram (with 10 million UK users, mostly under 30) and Snapchat (used by half of 18-34 year olds), both mobile-first platforms. On these platforms, although they may see individual stories that are shared or appear on their phone’s news portal, they may not even see the front page headlines that used to make politicians shake.

Meanwhile, what people do pay attention to and share on these platforms are videos and music, so popular artists amass huge followings. Some of the most popular came out in favour of Labour under the umbrella hashtag #Grime4Corbyn, with artists like Stormzy, JME (whose Facebook interview with Corbyn was viewed 2.5 million times) and Skepta with over a million followers on Instagram alone.

A leaflet from Croydon pointing out that ‘Even your Dad has more Facebook friends’ than the 2015 vote difference between Conservative and Labour and showing Stormzy saying ‘Vote Labour!’ was shared millions of times. Obviously we don’t know how much difference these endorsements made – but by sharing videos and images, they certainly spread the idea of voting for Corbyn across huge social networks.

Second, Labour have overtaken the Tories in reaching out across social platforms used by young people with an incredibly efficient advertising strategy. There is no doubt that in 2017 the Conservatives ran a relentless campaign of anti-Corbyn attack ads on Facebook and Instagram. But for the Conservatives, social media are just for elections. Instead, Labour have been using these channels for two years now – Corbyn has been active on Snapchat since becoming Labour leader in 2015 (when some of us were surprised to hear our teenage offspring announcing brightly ‘I’m friends with Jeremy Corbyn on Snapchat’).

That means that by the time of the election Corbyn and various fiercely pro-Labour online-only news outlets like the Canary had acquired a huge following among this demographic, meaning not having to pay for ads. And if you have followers to spread your message, you can be very efficient with advertising spend. While the Conservatives spent more than £1m on direct advertising with Facebook etc., nearly 10 million people watched pro-Labour videos on Facebook that cost less than £2K to make. Furthermore, there is some evidence that the relentless negativity of the Conservative advertising campaign actually put young people off particularly. After all, the advertising guidelines for Instagram advise ‘Images should tell a story/be inspirational’!

On the day before the election, the Daily Mail ran a frontpage headline ‘Apologists for Terror’, with a photo of Diane Abbot along with Corbyn and John McDonnell. But that morning Labour announced that Abbot’s standing aside due to illness. The paper circulating around the networks and sitting on news-stands was already out of date. Digital natives are used to real-time information, they are never going to be swayed by something so clearly past its sell-by-date.

Likewise, the Sun’s election day image – a grotesque image of Jeremy “Corbinned” in a dustbin was photoshopped to replace Corbyn with an equally grotesque photograph of May taking his place in the dustbin, before the first editions landed. It won’t have reached the same audience, perhaps, but it will have reached a lot of people.

It will be a long time before we can really assess the influence of social media in the 2017 election, and some things we may never know. That is because all the data that would allow us to do so is held by the platforms themselves – Facebook, Instagram, Snapchat and so on. That is a crucial issue for the future of our democracy, already bringing calls for some transparency in political advertising both by social media platforms and the parties themselves. Under current conditions the Electoral Commission is incapable of regulating election advertising effectively, or judging (for example) how much national parties spend on targeted advertising locally. This is something that urgently needs addressing in the coming months, especially given Britain’s current penchant for elections.

The secret and often dark world of personalized political advertising on social media, where strong undercurrents of support remain hidden to the outside world, is one reason why polls fail to predict election results until after the election has taken place. Having the data to understand the social media election would also explain some of the volatility in elections these days, as explored in our book Political Turbulence: How Social Media Shape Collective Action. By investigating large-scale data on political activity my co-authors and I showed that social media are injecting the same sort of instability into politics as they have into cultural markets, where most artists gain no traction at all but a (probably unpredictable) few become massively popular – the young singer Ed Sheeran’s ‘The Shape of You’ has been streamed one billion times on Spotify alone.

In 2017, Stormzy and co. provided a more direct link between political and music markets, and this kind of development will ensure that politics in the age of social media will remain turbulent and unpredictable. We can’t claim to have predicted Labour’s unexpected success in this election, but we can claim to have foreseen that it couldn’t be predicted.

Stormzy 1: The Sun 0 — Three Reasons Why #GE2017 Was the Real Social Media Election

Could Voting Advice Applications force politicians to keep their manifesto promises?

In many countries, Voting Advice Applications (VAAs) have become an almost indispensable part of the electoral process, playing an important role in the campaigning activities of parties and candidates, an essential element of media coverage of the elections, and being widely used by citizens. A number of studies have shown that VAA use has an impact on the cognitive behaviour of users, on their likelihood to participate in elections, and on the choice of the party they vote for.

These applications are based on the idea of issue and proximity voting — the parties and candidates recommended by VAAs are those with the highest number of matching positions on a number of political questions and issues. Many of these questions are much more specific and detailed than party programs and electoral platforms, and show the voters exactly what the party or candidates stand for and how they will vote in parliament once elected. In his Policy & Internet article “Do VAAs Encourage Issue Voting and Promissory Representation? Evidence From the Swiss Smartvote,” Andreas Ladner examines the extent to which VAAs alter the way voters perceive the meaning of elections, and encourage them to hold politicians to account for election promises.

His main hypothesis is that VAAs lead to “promissory representation” — where parties and candidates are elected for their promises and sanctioned by the electorate if they don’t keep them. He suggests that as these tools become more popular, the “delegate model” is likely to increase in popularity: i.e. one in which politicians are regarded as delegates voted into parliament to keep their promises, rather than being voted a free mandate to act how they see fit (the “trustee model”).

We caught up with Andreas to discuss his findings:

Ed.: You found that issue-voters were more likely (than other voters) to say they would sanction a politician who broke their election promises. But also that issue voters are less politically engaged. So is this maybe a bit moot: i.e. if the people most likely to force the “delegate model” system are the least likely to enforce it?

Andreas: It perhaps looks a bit moot in the first place, but what happens if the less engaged are given the possibility to sanction them more easily or by default. Sanctioning a politician who breaks an election promise is not per se a good thing, it depends on the reason why he or she broke it, on the situation, and on the promise. VAA can easily provide information to what extent candidates keep their promises — and then it gets very easy to sanction them simply for that without taking other arguments into consideration.

Ed.: Do voting advice applications work best in complex, multi-party political systems? (I’m not sure anyone would need one to distinguish between Trump / Clinton, for example?)

Andreas: Yes, I believe that in very complex systems – like for example in the Swiss case where voters not only vote for parties but also for up to 35 different candidates – VAAs are particularly useful since they help to process a huge amount of information. If the choice is only between two parties or two candidates which are completely different, than VAAs are less helpful.

Ed.: I guess the recent elections / referendum I am most familiar with (US, UK, France) have been particularly lurid and nasty: but I guess VAAs rely on a certain quiet rationality to work as intended? How do you see your Swiss results (and Swiss elections, generally) comparing with these examples? Do VAAs not just get lost in the noise?

Andreas: The idea of VAAs is to help voters to make better informed choices. This is, of course, opposed to decisions based on emotions. In Switzerland, elections are not of outmost importance, due to specific features of our political system such as direct democracy and power sharing, but voters seem to appreciate the information provided by smartvote. Almost 20% of the voter cast their vote after having consulted the website.

Ed.: Macron is a recent example of someone who clearly sought (and received) a general mandate, rather than presenting a detailed platform of promises. Is that unusual? He was criticised in his campaign for being “too vague,” but it clearly worked for him. What use are manifesto pledges in politics — as opposed to simply making clear to the electorate where you stand on the political spectrum?

Andreas: Good VAAs combine electoral promises on concrete issues as well as more general political positions. Voters can base their decisions on either of them, or on a combination of both of them. I am not arguing in favour of one or the other, but they clearly have different implications. The former is closer to the delegate model, the latter to the trustee model. I think good VAAs should make the differences clear and should even allow the voters to choose.

Ed.: I guess Trump is a contrasting example of someone whose campaign was all about promises (while also seeking a clear mandate to “make America great again”), but who has lied, and broken these (impossible) promises seemingly faster than people can keep track of them. Do you think his supporters care, though?

Andreas: His promises were too far away from what he can possibly keep. Quite a few of his voters, I believe, do not want them to be fully realized but rather that the US move a bit more into this direction.

Ed.: I suppose another example of an extremely successful quasi-pledge was the Brexit campaign’s obviously meaningless — but hugely successful — “We send the EU £350 million a week; let’s fund our NHS instead.” Not to sound depressing, but do promises actually mean anything? Is it the candidate / issue that matters (and the media response to that), or the actual pledges?

Andreas: I agree that the media play an important role and not always into the direction they intend to do. I do not think that it is the £350 million a week which made the difference. It is much more a general discontent and a situation which was not sufficiently explained and legitimized which led to this unexpected decision. If you lose the support for your policy than it gets much easier for your opponents. It is difficult to imagine that you can get a majority built on nothing.

Ed.: I’ve read all the articles in the Policy & Internet special issue on VAAs: one thing that struck me is that there’s lots of incomplete data, e.g. no knowledge of how people actually voted in the end (or would vote in future). What are the strengths and weaknesses of VAAs as a data source for political research?

Andreas: The quality of the data varies between countries and voting systems. We have a self-selection bias in the use of VAAs and often also into the surveys conducted among the users. In general we don’t know how they voted, and we have to believe them what they tell us. In many respects the data does not differ that much from what we get from classic electoral studies, especially since they also encounter difficulties in addressing a representative sample. VAAs usually have much larger Ns on the side of the voters, generate more information about their political positions and preferences, and provide very interesting information about the candidates and parties.

Read the full article: Ladner, A. (2016) Do VAAs Encourage Issue Voting and Promissory Representation? Evidence From the Swiss Smartvote. Policy & Internet 8 (4). DOI: doi:10.1002/poi3.137.

Andreas Ladner was talking to blog editor David Sutcliffe.

Is Left-Right still meaningful in politics? Or are we all just winners or losers of globalisation now?

The Left–Right dimension — based on the traditional cleavage in society between capital and labor — is the most common way of conceptualizing ideological difference. But in an ever more globalized world, are the concepts of Left and Right still relevant? In recent years political scientists have increasingly come to talk of a two-dimensional politics in Europe, defined by an economic (Left–Right) dimension, and a cultural dimension that relates to voter and party positions on sociocultural issues.

In his Policy & Internet article “Cleavage Structures and Dimensions of Ideology in English Politics: Evidence From Voting Advice Application Data”, Jonathan Wheatley argues that the cleavage that exists in many European societies between “winners” and “losers” of globalization has engendered a new ideological dimension pitting “cosmopolitans” against “communitarians” and that draws on cultural issues relating to identity — rather than economic issues.

He identifies latent dimensions from opinion data generated by two Voting Advice Applications deployed in England in 2014 and 2015 — finding that the political space in England is defined by two main ideological dimensions: an economic Left–Right dimension and a cultural communitarian–cosmopolitan dimension. While they co-vary to a significant degree, with economic rightists tending to be more communitarian and economic leftists tending to be more cosmopolitan, these tendencies do not always hold and the two dimensions should be considered as separate.

The identification of the communitarian–cosmopolitan dimension lends weight to the hypothesis of Kriesi et al. (2006) that politics is increasingly defined by a cleavage between “winners” and “losers” of globalization, with “losers” tending to adopt a position of cultural demarcation and to perceive “outsiders” such as immigrants and the EU, as a threat. If an economic dimension pitting Left against Right (or labour against capital) defined the political arena in Europe in the twentieth century, maybe it’s a cultural cleavage that pits cosmopolitans against communitarians that defines politics in the twenty-first.

We caught up with Jonathan to discuss his findings:

Ed.: The big thing that happened since your article was published was Brexit — so I guess the “communitarian–cosmopolitan” dimension (Trump!) makes obvious intuitive sense as a political cleavage plane. Will you be comparing your GE2015 VAA data with GE2017 data? And what might you expect to see?

Jonathan: Absolutely! We will be launching the WhoGetsMyVoteUK Voting Advice Application next week. This VAA will be launched by three universities: Oxford Brookes University (where I am based), Queen Mary University London and the University of Bristol. This should provide extensive data that will allow us to make a longitudinal study: before and after Brexit.

Ed.: There was a lot of talk (for the first time) after Brexit of “the left behind” — I suppose partly corresponding to your “communitarians” — but that all seems to have died down. Of course they’re still there: is there any sense of how they will affect the upcoming election — particularly the “communitarian leftists”?

Jonathan: Well this is the very group that Theresa May’s Conservative Party seems to be targeting. We should note that May has attempted to appeal directly to this group by her claim that “if you believe you’re a citizen of the world, you’re a citizen of nowhere” made at the Tory Party Conference last autumn, and by her assertion that “Liberalism and globalisation have left people behind” made at the Lord Mayor’s banquet late last year. Her (at least superficially) economically leftist proposals during the election campaign to increase the living wage and statutory rights for family care and training, and to strengthen labour laws, together with her “hard Brexit stance” and confrontational rhetoric towards European leaders seems specifically designed to appeal to this group. Many of these “communitarian leftists” have previously been tempted by UKIP, but the Conservatives seem to be winning the battle for their votes at the moment.

Ed.: Does the UK’s first-past-the-post system (resulting in a non-proportionally representative set of MPs) just hide what is happening underneath, i.e. I’m guessing a fairly constant, unchanging spectrum of political leanings? Presumably UKIP’s rise didn’t signify a lurch to the right: it was just an efficient way of labelling (for a while) people who were already there?

Jonathan: To a certain extent, yes. Superficially the UK has very much been a case of “business as usual” in terms of its party system, notwithstanding the (perhaps brief) emergence of UKIP as a significant force in around 2012. This can be contrasted with Sweden, Finland and the Netherlands, where populist right parties obtained significant representation in parliament. And UKIP may prove to be a temporary phenomenon. The first-past-the-post system provides more incentives for parties to reposition themselves to reflect the new reality than it does for new parties to emerge. In fact it is this repositioning, from a economically right-wing, mildly cosmopolitan party to an (outwardly) economically centrist, communitarian party, that seems to characterise the Tories today.

Ed.: Everything seems to be in a tremendous mess (parties imploding, Brexit horror, blackbox campaigning, the alt-right, uncertainty over tactical voting, “election hacking”) and pretty volatile. But are these exciting times for political scientists? Or are things too messy and the data (for example, on voting intensions as well as outcomes) too inaccessible to distinguish any grand patterns?

Jonathan: Exciting from a political science point of view; alarming from the point of view of a member of society.

Ed.: But talking of “grand patterns”: do you have any intuition why “the C20 might be about capital vs labour; the C21 about local vs global”? Is it simply the next obvious reaction to ever-faster technological development and economic concentration bumping against societal inertia, or something more complex and unpredictable?

Jonathan: Over generations European societies gradually developed mechanisms of accountability to constrain their leaders and ensure they did not over-reach their powers. This is how democracy became consolidated. Hoverver, given that power is increasingly accruing to transnational and multinational corporations and networks that are beyond the reach of citizens operating in the national sphere, we must learn how to do this all over again on a global scale. Until we do so, globalisation will inevitably create “winners” and “losers” and will, I think, inevitably lead to more populism and upheaval.

Read the full article: Wheatley, J. (2016) Cleavage Structures and Dimensions of Ideology in English Politics: Evidence From Voting Advice Application Data. Policy & Internet 8 (4) doi:10.1002/poi3.129

Jonathan Wheatley was talking to blog editor David Sutcliffe.

Five Pieces You Should Probably Read On: The US Election

This is the first post in a series that will uncover great writing by faculty and students at the Oxford Internet Institute, things you should probably know, and things that deserve to be brought out for another viewing. This week: The US Election.

This was probably the nastiest Presidential election in recent memory: awash with Twitter bots and scandal, polarisation and filter bubbles, accusations of interference by Russia and the Director of the FBI, and another shock result. We have written about electoral prediction elsewhere: instead, here are five pieces that consider the interaction of social media and democracy — the problems, but also potential ways forward.

1. James Williams: The Clickbait Candidate

10 October 2016 / 2700 words / 13 minutes

“Trump is very straightforwardly an embodiment of the dynamics of clickbait: he is the logical product (though not endpoint) in the political domain of a media environment designed to invite, and indeed incentivize, relentless competition for our attention […] Like clickbait or outrage cascades, Donald Trump is merely the sort of informational packet our media environment is designed to select for.”

James Williams says that now is probably the time to have that societal conversation about the design ethics of the attention economy — because in our current media environment, attention trumps everything.

2. Sam Woolley, Philip Howard: Bots Unite to Automate the Presidential Election [Wired]

15 May 2016 / 850 words / 4 minutes

“Donald Trump understands minority communities. Just ask Pepe Luis Lopez, Francisco Palma, and Alberto Contreras […] each tweeted in support of Trump after his victory in the Nevada caucuses earlier this year. The problem is, Pepe, Francisco, and Alberto aren’t people. They’re bots.”

It’s no surprise that automated spam accounts (or bots) are creeping into election politics, say Sam Woolley and Philip Howard. Demanding bot transparency would at least help clean up social media — which, for better or worse, is increasingly where presidents get elected.

3. Phil Howard: Is Social Media Killing Democracy?

15 November 2016 / 1100 words / 5 minutes

“This is the big year for computational propaganda — using immense data sets to manipulate public opinion over social media. Both the Brexit referendum and US election have revealed the limits of modern democracy, and social media platforms are currently setting those limits […] these technologies permit too much fake news, encourage our herding instincts, and aren’t expected to provide public goods.”

Phil Howard discusses ways to address fake news, audit social algorithms, and deal with social media’s “moral pass” — social media is damaging democracy, he says, but can also be used to save it.

4. Helen Margetts: Don’t Shoot the Messenger! What part did social media play in 2016 US election?

15 November 2016 / 600 words / 3 minutes

“Rather than seeing social media solely as the means by which Trump ensnared his presidential goal, we should appreciate how they can provide a wealth of valuable data to understand the anger and despair that the polls missed, and to analyse political behaviour and opinion in the times ahead.”

New social information and visibility brings change to social behaviour, says Helen Margetts — ushering in political turbulence and unpredictability. Social media made visible what could have remain a country’s dark secret (hatred of women, rampant racism, etc.), but it will also underpin any radical counter-movement that emerges in the future.

5. Helen Margetts: Of course social media is transforming politics. But it’s not to blame for Brexit and Trump

9 January 2017 / 1700 words / 8 minutes

“Even if political echo chambers were as efficient as some seem to think, there is little evidence that this is what actually shapes election results. After all, by definition echo chambers preach to the converted. It is the undecided people who (for example) the Leave and Trump campaigns needed to reach. And from the research, it looks like they managed to do just that.”

Politics is a lot messier in the social media era than it used to be, says Helen Margetts, but rather than blaming social media for undermining democracy, we should be thinking about how we can improve the (inevitably major) part that it plays.

The Authors

James Williams is an OII doctoral candidate, studying the ethics of attention and persuasion in technology design.

Sam Woolley is a Research Assistant on the OII’s Computational Propaganda project; he is interested in political bots, and the intersection of political communication and automation.

Philip Howard is the OII’s Professor of Internet Studies and PI of the Computational Propaganda project. He investigates the impact of digital media on political life around the world.

Helen Margetts is the OII’s Director, and Professor of Society and the Internet. She specialises in digital era government, politics and public policy, and data science and experimental methods. Her most recent book is Political Turbulence (Princeton).

Coming up .. Fake news and filter bubbles / It’s the economy, stupid / Augmented reality and ambient fun / The platform economy / Power and development / Internet past and future / Government / Labour rights / The disconnected / Ethics / Staying critical

#5OIIPieces

Five Pieces You Should Probably Read On: The US Election

1. James Williams: The Clickbait Candidate

10 October 2016 / 2700 words / 13 minutes

2. Sam Woolley, Philip Howard: Bots Unite to Automate the Presidential Election [Wired]

15 May 2016 / 850 words / 4 minutes

3. Phil Howard: Is Social Media Killing Democracy?

15 November 2016 / 1100 words / 5 minutes

4. Helen Margetts: Don’t Shoot the Messenger! What part did social media play in 2016 US election?

15 November 2016 / 600 words / 3 minutes

5. Helen Margetts: Of course social media is transforming politics. But it’s not to blame for Brexit and Trump

9 January 2017 / 1700 words / 8 minutes

The Authors

James Williams is an OII doctoral candidate, studying the ethics of attention and persuasion in technology design.

Sam Woolley is a Research Assistant on the OII’s Computational Propaganda project; he is interested in political bots, and the intersection of political communication and automation.

Philip Howard is the OII’s Professor of Internet Studies and PI of the Computational Propaganda project. He investigates the impact of digital media on political life around the world.

#5OIIPieces

Can we predict electoral outcomes from Wikipedia traffic?

As digital technologies become increasingly integrated into the fabric of social life their ability to generate large amounts of information about the opinions and activities of the population increases. The opportunities in this area are enormous: predictions based on socially generated data are much cheaper than conventional opinion polling, offer the potential to avoid classic biases inherent in asking people to report their opinions and behaviour, and can deliver results much quicker and be updated more rapidly.

In their article published in EPJ Data Science, Taha Yasseri and Jonathan Bright develop a theoretically informed prediction of election results from socially generated data combined with an understanding of the social processes through which the data are generated. They can thereby explore the predictive power of socially generated data while enhancing theory about the relationship between socially generated data and real world outcomes. Their particular focus is on the readership statistics of politically relevant Wikipedia articles (such as those of individual political parties) in the time period just before an election.

By applying these methods to a variety of different European countries in the context of the 2009 and 2014 European Parliament elections they firstly show that the relative change in number of page views to the general Wikipedia page on the election can offer a reasonable estimate of the relative change in election turnout at the country level. This supports the idea that increases in online information seeking at election time are driven by voters who are considering voting.

Second, they show that a theoretically informed model based on previous national results, Wikipedia page views, news media mentions, and basic information about the political party in question can offer a good prediction of the overall vote share of the party in question. Third, they present a model for predicting change in vote share (i.e., voters swinging towards and away from a party), showing that Wikipedia page-view data provide an important increase in predictive power in this context.

This relationship is exaggerated in the case of newer parties — consistent with the idea that voters don’t seek information uniformly about all parties at election time. Rather, they behave like ‘cognitive misers’, being more likely to seek information on new political parties with which they do not have previous experience and being more likely to seek information only when they are actually changing the way they vote.

In contrast, there was no evidence of a ‘media effect’: there was little correlation between news media mentions and overall Wikipedia traffic patterns. Indeed, the news media and Wikipedia appeared to be biased towards different things: with the news favouring incumbent parties, and Wikipedia favouring new ones.

Read the full article: Yasseri, T. and Bright, J. (2016) Wikipedia traffic data and electoral prediction: towards theoretically informed models. EPJ Data Science. 5 (1).

We caught up with the authors to explore the implications of the work.

Ed: Wikipedia represents a vast amount of not just content, but also user behaviour data. How did you access the page view stats — but also: is anyone building dynamic visualisations of Wikipedia data in real time?

Taha and Jonathan: Wikipedia makes its page view data available for free (in the same way as it makes all of its information available!). You can find the data here, along with some visualisations

Ed: Why did you use Wikipedia data to examine election prediction rather than (the I suppose the more fashionable) Twitter? How do they compare as data sources?

Taha and Jonathan: One of the big problems with using Twitter to predict things like elections is that contributing on social media is a very public thing and people are quite conscious of this. For example, some parties are seen as unfashionable so people might not make their voting choice explicit. Hence overall social media might seem to be saying one thing whereas actually people are thinking another.

By contrast, looking for information online on a website like Wikipedia is an essentially private activity so there aren’t these social biases. In other words, on Wikipedia we can directly have access to transactional data on what people do, rather than what they say or prefer to say.

Ed: How did these results and findings compare with the social media analysis done as part of our UK General Election 2015 Election Night Data Hack? (long title..)

Taha and Jonathan: The GE2015 data hack looked at individual politicians. We found that having a Wikipedia page is becoming increasingly important — over 40% of Labour and Conservative Party candidates had an individual Wikipedia page. We also found that this was highly correlated with Twitter presence — being more active on one network also made you more likely to be active on the other one. And we found some initial evidence that social media reaction was correlated with votes, though there is a lot more work to do here!

Ed: Can you see digital social data analysis replacing (or maybe just complementing) opinion polling in any meaningful way? And what problems would need to be addressed before that happened: e.g. around representative sampling, data cleaning, and weeding out bots?

Taha and Jonathan: Most political pundits are starting to look at a range of indicators of popularity — for example, not just voting intention, but also ratings of leadership competence, economic performance, etc. We can see good potential for social data to become part of this range of popularity indicator. However we don’t think it will replace polling just yet; the use of social media is limited to certain demographics. Also, the data collected from social media are often very shallow, not allowing for validation. In the case of Wikipedia, for example, we only know how many times each page is viewed, but we don’t know by how many people and from where.

Ed: You do a lot of research with Wikipedia data — has that made you reflect on your own use of Wikipedia?

Taha and Jonathan: It’s interesting to think about this activity of getting direct information about politicians — it’s essentially a new activity, something you couldn’t do in the pre-digital age. I know that I personally [Jonathan] use it to find out things about politicians and political parties — it would be interesting to know more about why other people are using it as well. This could have a lot of impacts. One thing Wikipedia has is a really long memory, in a way that other means of getting information on politicians (such as newspapers) perhaps don’t. We could start to see this type of thing becoming more important in electoral politics.

[Taha] .. since my research has been mostly focused on Wikipedia edit wars between human and bot editors, I have naturally become more cautious about the information I find on Wikipedia. When it comes to sensitive topics, sach as politics, Wikipedia is a good point to start, but not a great point to end the search!

Taha Yasseri and Jonathan Bright were talking to blog editor David Sutcliffe.

Rethinking Digital Media and Political Change

Did Twitter lead to Donald Trump’s rise and success to date in the American campaign for the presidency? Image: Gage Skidmore (Flickr)

What are the dangers or new opportunities of digital media? One of the major debates in relation to digital media in the United States has been whether they contribute to political polarization. I argue in a new paper (Rethinking Digital Media and Political Change) that Twitter led to Donald Trump’s rise and success to date in the American campaign for the presidency. There is plenty of evidence to show that Trump received a disproportionate amount of attention on Twitter, which in turn generated a disproportionate amount of attention in the mainstream media. The strong correlation between the two suggests that Trump was able to bypass the gatekeepers of the traditional media.

A second ingredient in his success has been populism, which rails against dominant political elites (including the Republican party) and the ‘biased’ media. Populism also rests on the notion of an ‘authentic’ people — by implication excluding ‘others’ such as immigrants and foreign powers like the Chinese — to whom the leader appeals directly. The paper makes parallels with the strength of the Sweden Democrats, an anti-immigrant party which, in a similar way, has been able to appeal to its following via social media and online newspapers, again bypassing mainstream media with its populist message.

There is a difference, however: in the US, commercial media compete for audience share, so Trump’s controversial tweets have been eagerly embraced by journalists seeking high viewership and readership ratings. In Sweden, where public media dominate and there is far less of the ‘horserace’ politics of American politics, the Sweden Democrats have been more locked out of the mainstream media and of politics. In short, Twitter plus populism has led to Trump. I argue that dominating the mediated attention space is crucial. One outcome of how this story ends will be known in November. But whatever the outcome, it is already clear that the role of the media in politics, and how they can be circumvented by new media, requires fundamental rethinking.

Ralph Schroeder is Professor and director of the Master’s degree in Social Science of the Internet at the Oxford Internet Institute. Before coming to Oxford University, he was Professor in the School of Technology Management and Economics at Chalmers University in Gothenburg (Sweden). Recent books include Rethinking Science, Technology and Social Change (Stanford University Press, 2007) and, co-authored with Eric T. Meyer, Knowledge Machines: Digital Transformations of the Sciences and Humanities (MIT Press 2015).

Political polarization on social media: do birds of a feather flock together on Twitter?

Twitter has exploded in recent years, now boasting half a billion registered users. Like blogs and the world’s largest social networking platform, Facebook, Twitter has actively been used for political discourse during the past few elections in the US, Canada, and elsewhere but it differs from them in a number of significant ways. Twitter’s connections tend to be less about strong social relationships (such as those between close friends or family members), and more about connecting with people for the purposes of commenting and information sharing. Twitter also provides a steady torrent of updates and resources from individuals, celebrities, media outlets, and any other organization seeking to inform the world as to its views and actions.

This may well make Twitter particularly well suited to political debate and activity. Yet important questions emerge in terms of the patterns of conduct and engagement. Chief among them: are users mainly seeking to reinforce their own viewpoints and link with likeminded persons, or is there a basis for widening and thoughtful exposure to a variety of perspectives that may improve the collective intelligence of the citizenry as a result?

Conflict and Polarization

Political polarization often occurs in a so-called ‘echo chamber’ environment, in which individuals are exposed to only information and communities that support their own viewpoints, while ignoring opposing perspectives and insights. In such isolating and self-reinforcing conditions, ideas can become more engrained and extreme due to lack of contact with contradictory views and the exchanges that could ensue as a result.

On the web, political polarization has been found among political blogs, for instance. American researchers have found that liberal and conservative bloggers in the US tend to link to other bloggers who share their political ideology. For Kingwell, a prominent Canadian philosopher, the resulting dynamic is one that can be characterized by a decline in civility and a lessening ability for political compromise to take hold. He laments the emergence of a ‘shout doctrine’ that corrodes the civic and political culture, in the sense that divisions are accentuated and compromise becomes more elusive.

Such a dynamic is not the result of social media alone – but rather it reflects for some the impacts of the Internet generally and the specific manner by which social media can lend itself to broadcasting and sensationalism, rather than reasoned debate and exchange. Traditional media and journalistic organizations have thus become further pressured to act in kind, driven less by a patient and persistent presentation of all sides of an issue and more by near-instantaneous reporting online. In a manner akin to Kingwell’s view, one prominent television news journalist in the US, Ted Koppel, has lamented this new media environment as a danger to the republic.

Nonetheless, the research is far from conclusive as to whether the Internet increases political polarization. Some studies have found that among casual acquaintances (such as those that can typically be observed on Twitter), it is common to observe connections across ideological boundaries. In one such funded by the Pew Internet and American Life Project and the National Science Foundation, findings suggest that people who often visit websites that support their ideological orientation also visit web sites that support divergent political views. As a result, greater sensitivity and empathy for alternative viewpoints could potentially ensue, improving the likelihood for political compromise – even on a modest scale that would otherwise not have been achievable without this heightened awareness and debate.

Early Evidence from Canada

The 2011 federal election in Canada was dubbed by some observers in the media as the country’s first ‘social media election’ – as platforms such as Facebook and Twitter became prominent sources of information for growing segments of the citizenry, and evermore strategic tools for political parties in terms of fundraising, messaging, and mobilizing voters. In examining Twitter traffic, our own intention was to ascertain the extent to which polarization or cross-pollinization was occurring across the portion of the electorate making use of this micro-blogging platform.

We gathered nearly 6000 tweets pertaining to the federal election made by just under 1500 people during a three-day period in the week preceding election day (this time period was chosen because it was late enough in the campaign for people to have an informed opinion, but still early enough for them to be persuaded as to how they should vote). Once the tweets were retrieved, we used social network analysis and content analysis to analyze patterns of exchange and messaging content in depth.

We found that overall people do tend to cluster around shared political views on Twitter. Supporters of each of the four major political parties identified in the study were more likely to tweet to other supporters of the same affiliation (this was particularly true of the ruling Conservatives, the most inwardly networked of the four major politically parties). Nevertheless, in a significant number of cases (36% of all interactions) we also observed a cross-ideological discourse, especially among supporters of the two most prominent left-of-centre parties, the New Democratic Party (NDP) and the Liberal Party of Canada (LPC). The cross-ideological interactions among supporters of left-leaning parties tended to be agreeable in nature, but often at the expense of the party in power, the Conservative Party of Canada (CPC). Members from the NDP and Liberal formations were also more likely to share general information and updates about the election as well as debate various issues around their party platforms with each other.

By contrast, interactions between parties that are ideologically distant seemed to denote a tone of conflict: nearly 40% of tweets between left-leaning parties and the Conservatives tended to be hostile. Such negative interactions between supporters of different parties have shown to reduce enthusiasm about political campaigns in general, potentially widening the cleavage between highly engaged partisans and less affiliated citizens who may view such forms of aggressive and divisive politics as distasteful.

For Twitter sceptics, one concern is that the short length of Twitter messages does not allow for meaningful and in-depth discussions around complex political issues. While it is certainly true that expression within 140 characters is limited, one third of tweets between supporters of different parties included links to external sources such as news stories, blog posts, or YouTube videos. Such indirect sourcing can thereby constitute a means of expanding dialogue and debate.

Accordingly, although it is common to view Twitter as largely a platform for self-expression via short tweets, there may be a wider collective dimension to both users and the population at large as a steady stream of both individual viewpoints and referenced sources drive learning and additional exchange. If these exchanges happen across partisan boundaries, they can contribute to greater collective awareness and learning for the citizenry at large.

As the next federal election approaches in 2015, with younger voters gravitating online – especially via mobile devices, and with traditional polling increasingly under siege as less reliable than in the past, all major parties will undoubtedly devote more energy and resources to social media strategies including, perhaps most prominently, an effective usage of Twitter.

Partisan Politics versus Politics 2.0

In a still-nascent era likely to be shaped by the rise of social media and a more participative Internet on the one hand, and the explosion of ‘big data’ on the other hand, the prominence of Twitter in shaping political discourse seems destined to heighten. Our preliminary analysis suggests an important cleavage between traditional political processes and parties – and wider dynamics of political learning and exchange across a changing society that is more fluid in its political values and affiliations.

Within existing democratic structures, Twitter is viewed by political parties as primarily a platform for messaging and branding, thereby mobilizing members with shared viewpoints and attacking opposing interests. Our own analysis of Canadian electoral tweets both amongst partisans and across party lines underscores this point. The nexus between partisan operatives and new media formations will prove to be an increasingly strategic dimension to campaigning going forward.

More broadly, however, Twitter is a source of information, expression, and mobilization across a myriad of actors and formations that may not align well with traditional partisan organizations and identities. Social movements arising during the Arab Spring, amongst Mexican youth during that country’s most recent federal elections and most recently in Ukraine are cases in point. Across these wider societal dimensions – especially consequential in newly emerging democracies, the tremendous potential of platforms such as Twitter may well lie in facilitating new and much more open forms of democratic engagement that challenge our traditional constructs.

In sum, we are witnessing the inception of new forms of what can be dubbed ‘Politics 2.0’ that denotes a movement of both opportunities and challenges likely to play out differently across democracies at various stages of socio-economic, political, and digital development. Whether Twitter and other likeminded social media platforms enable inclusive and expansionary learning, or instead engrain divisive polarized exchange, has yet to be determined. What is clear however is that on Twitter, in some instances, birds of a feather do flock together as they do on political blogs. But in other instances, Twitter can play an important role to foster cross parties communication in the online political arenas.

Read the full article: Gruzd, A., and Roy, J. (2014) Investigating Political Polarization on Twitter: A Canadian Perspective. Policy and Internet 6 (1) 28-48.

Also read: Gruzd, A. and Tsyganova, K. Information wars and online activism during the 2013/2014 crisis in Ukraine: Examining the social structures of Pro- and Anti-Maidan groups. Policy and Internet. Early View April 2015: DOI: 10.1002/poi3.91

Anatoliy Gruzd is Associate Professor in the Ted Rogers School of Management and Director of the Social Media Lab at Ryerson University, Canada. Jeffrey Roy is Professor in the School of Public Administration at Dalhousie University’s Faculty of Management. His most recent book was published in 2013 by Springer: From Machinery to Mobility: Government and Democracy in a Participative Age.

Political polarization on social media: do birds of a feather flock together on Twitter?

Conflict and Polarization

Early Evidence from Canada

Partisan Politics versus Politics 2.0

Read the full article: Gruzd, A., and Roy, J. (2014) Investigating Political Polarization on Twitter: A Canadian Perspective. Policy and Internet 6 (1) 28-48.