The calm after the storm has settled in London. The UK has had its third national election in as many years after Conservative Prime Minister Theresa May called a snap election to shore up her party’s position ahead of the upcoming Brexit negotiations. Following a series of polls showing May enjoying a 20-point lead over Jeremy Corbyn’s Labour party, May was convinced to call an election to improve her party’s position in the House of Commons so as to give the Prime Minister a strong hand when negotiating the exit deal between Britain and the EU. The gamble, it is clear, has failed spectacularly. The official BBC exit poll, based on the methodology of the University of Strathclyde’s most famous polling expert John Curtice, predicted that the Conservatives would lose their overall majority in the Commons and that Labour were set to gain seats. The predication, a hung parliament, proved accurate.

For the majority of viewers, journalists and politicians, the only response was bewilderment. It simply wasn’t possible that the Conservatives could have lost seats, all but one of the polling agencies were predicting a huge win for May. So how was it that YouGov, the one agency that predicted Labour gains and a hung parliament, got it right while the rest missed the mark? And how did social media play a role in this short campaign? This blog will argue that a combination of people unwilling to admit to their intention to vote for the Conservative party, online culture and the social desirability of left-wing views online are part of the reason that Labour’s campaign succeeded in clawing back the party’s 20-point deficit and securing their largest share of the popular vote since 2001.

The Pollsters

To get a flavour of the polling numbers around this election, let’s take the aggregate scores from when the election was called, midway through and the last polling data available before the election. The data collected below is from members of the UK Polling Council and as such does not include data from unaffiliated polling organisations and companies like SurveyMonkey and others. It presents polling results in chronological order for each part of the campaign.
Theresa May called a general election on the 18th April 2017, the aggregate polling data on that day points to an 18.75 point gap.

PollsterDateCON (%)LAB (%)LD (%)UKIP (%)Grn (%)Con Lead 
Opinium/Observer428454526119319
YouGov/Times428444824127224
ICM/Guardian428434625118421
ICM/Guardian4284244261211418
Opinium/Observer42838382971459
ComRes/Independent/S. Mirror428384625119421
YouGov/Times4283844231210421
YouGov/Times4283142251111317

Average Conservative Poll Lead: 18.75 points.

Table 1: Polling data from April 2017 (Source: UK Polling Report)

Now for the polling results from approximately midway through the election campaign.

PollsterDateCON (%)LAB (%)LD (%)UKIP (%)Grn (%)Con Lead 
Panelbase/42870473375314
Kantar/42870472986418
YouGov/Times428654630115216
YouGov/Sunday Times42867493193218
ICM/Guardian428694828106320
ORB/Sunday Telegraph42866463287214
Panelbase/42864483185217
ComRes/Ind on Sun/S. Mirror428674830105318

Average Conservative Poll Lead: 16.875 points.

Table 2: Polling data from May 2017 (Source: UK Polling Report)

Finally, the final eight polls held in the run up to election day (the final poll from Survation has been omitted here because it produced highly anomalous results: a gap of one point total, although this is interesting it gives a skewed picture of the main polling results and as such has been replaced by an Opinium poll from the previous week).

PollsterDateCON (%)LAB (%)LD (%)UKIP (%)Grn (%)Con Lead
Panelbase/4289344367528
Kantar/4289343387425
ICM/Guardian42893463475212
YouGov/Times42893423510527
ComRes/Independent42893443495210
BMG/Herald42893463385313
Ipsos-MORI/Evening Standard4289344367428
Opinium/4289243368527

Average Conservative poll lead: 8.75 points

Table 3: Polling data from June 2017 (Source: UK Polling Report)

What is clear from the data presented above is that the Conservative poll lead was narrowing considerably in the run up to the election. That said, the results still underestimated the Labour results and overestimated the Conservative performance. The final results gave the Conservatives 42.4% of the popular vote and Labour 40% of the popular vote.

 

Figure One: UK General Election Results 2017 – % Popular Vote (Source: BBC)

So what went wrong? And how did the exit poll get the figure right?

Methods, Methods

Following a difficult few years for polling companies, including missing the Brexit vote and Cameron’s overall majority in the 2015 general election, the polling agencies have revisited their methods. One of the key issues they have sought to address is the oversampling of graduates and young people; groups much more likely to vote Labour and against Brexit. The chief reason for this is that, especially young people, turn out to vote in lower numbers than older demographics. As such, the polls for this election have sought to address this. The notable exception to this is YouGov, who continued to include younger people who simply said that they intended to vote in their data, while other polling companies used turnout numbers by demographic to aggregate their data. YouGov, meanwhile, sought to find ways to poll those less interested in politics than their typical previous participants. After their most controversial poll which predicted a hung parliament with 310 seats for the Conservatives (actual number: 316) and 255 seats for the Labour Party (actual number 262), Anthony Wells, director of the social and political research team at YouGov wrote:

“After we had weighted our sample, taken account of how likely people say they are to vote, and weighted down the answers of those people who didn’t vote last time, we were left with a sample that implies turnout of 51% among people under 25 and 75% among people aged 65+; a turnout gap of 24 points between young and old.”

In the same article, however, he also wrote:

“My own expectation has been that the Conservatives will probably get a majority of around 70.”

Although hindsight is indeed 20:20 and these predictions look silly in retrospect, there are good reasons that the pollsters managed to underestimate the young vote so dramatically. The first is historical precedent. As Edward Phelps noted in a 2006 paper, turnout among the young has been in decline since 1992 in the UK. However, there are signs that youth turnout, that is voters aged between 18 and 34 has been steadily climbing since 2001, albeit the case that youth turnout lags behind the over 65 vote by a huge margin. As Abhinay Muthoo notes in a 2015 paper:

“…over the thirteen-year period from 1992 till 2005, youth turnout at UK general elections sharply and steadily declined by twenty-eight percentage points, from 66% in 1992 (and before) reaching down to 38% in 2005. It increased back to a 49% rate in the 2010 election but this is still down by seventeen percentage points on turnout rates over the 1964-1992 period”

In the event, turnout among the young outstripped expectations, reaching 57%. That said, in the over 65 age group, turnout was also up and around 80%. As such, it would be wrong to argue that Labour’s unexpectedly strong election results can be pointed simply at young people turning out in much better numbers than expected, although that undoubtedly helped. The British electorate, in its first-past-the-post system generally vote for one of the two main parties, Labour or the Conservatives. Although in recent years, younger voters particularly have voted for smaller third parties including the Liberal Democrats, the Greens and the Scottish National Party, their total share of the vote is much below that of the big two parties. This splintering of the vote has come at great electoral cost to the Labour party which is the traditional home of young voters in Britain. Meanwhile, older Labour voters and the far-right of the Conservative party base have been tempted to vote for UKIP – the driving force behind the Brexit vote. However, post-referendum UKIP are increasingly seen as a spent electoral force and its voters appear to have returned to their traditional parties for the most part.

As such, motivating and turning out the youth vote has vital to Labour’s success in this election. It is of course difficult to pinpoint the reason that more young people turned out to vote, although there are a number of reasons that spring to mind: Brexit, tuition fees, discriminatory housing and welfare policies which target the young, among others. However, save Brexit, this had not previously been enough to motivate young people to vote in the previous general election, and as such can only be understood to complete part of the picture.

Figure Two: Turnout 18-34 year-olds and over 35 in Britain – 1964-2010 (Source: Warwick Policy Lab/Muthoo).

That said, there was clearly something that the pollsters missed, with an honourable mention of the imagination and frankly, the bravery of YouGov who almost got the numbers right. How did they miss this turnaround in the fortune of the Labour party? A part of the answer can be found in looking at online culture and the differing ways that the Conservatives and Labour made use of social media as a campaigning tool.

Shy Tories & Online Virtue Signalling

There is a long-established trend in British politics of voters under-reporting the likelihood that they will vote for the Conservative party. These so-called Shy Tories are one of the reasons that polls typically overestimate Labour’s share of the vote (until this election the Labour share of the vote had been overestimated in every election since 1992, including Tony Blair’s landslide in 1997).

The data below shows the polled predicted performance based on the final four polling surveys each general election between 1992 and 2015, with the actual results and the difference between the two main parties.

Figure Three: UK polling data and election results 1992 -2015 (Source: YouGov, Gallup, ICM, Survation & BBC)

 

Figure Four: Difference Between Polled & Actual Results UK General Elections 1992-2015 (Source: YouGov, Gallup, ICM, Survation & BBC)

The polls show that typically pollsters underestimate the number of voters that will go on to vote for the Conservatives and typically overestimate that number for Labour. A notable exception is in 2010 where the Labour number was slightly underestimated, although it could be the case that, following 13 years of Labour government, voters were similarly unwilling to admit to their intention to vote for the party. This was, after all the election of ‘I agree with Nick’ from the Liberal Democrats.

At this point, it is important for the writer to acknowledge that she is a fully-paid up member of the British Labour Party. Hence, for an explanation of this phenomenon, let’s find an independent third party’s thoughts. Analysing the failure of the polls to predict David Cameron’s majority in 2015, Professor Peter Kellner wrote about the Shy Tory factor:

“This year, as in 1992, the Tories have a weak image. They are widely thought to be out of touch and for the rich. But, at the margin, there may be some people who both have a poor view of the party but nevertheless think it will run the economy better than Labour. They are “shy Tories” not because they are unwilling to admit their choice of party to a stranger but because they really would like to support someone else but, faced with a ballot paper in the privacy of the polling booth, simply can’t.”

Hence, it can be concluded that there is an element of shame among some segments of the electorate about voting for the Conservative party and under-report their intention to vote for the Conservatives. This point is hardly new. However, the phenomenon of a Shy Tory voter needs to be re-examined through the lens of social media in order to understand part of what happened in 2017.

U OK, Hun?

The first thing to make clear is that the Labour party and the Conservatives used social media differently. The Conservatives’ leader Theresa May, with her much older and less digitally engaged voter base, posted only 159 times on Facebook and Twitter during the general election campaign. Jeremy Corbyn, meanwhile, in his attempt to motivate younger voters, posted 925 times in the same time period. While this is enough to point to a very different approach to campaigning online, perhaps more interesting is the way the content posted by these accounts was received. May’s posts were shared 130,000 times over the course of the campaign, an average of nearly 818 times per post. Corbyn’s posts, by contrast, were shared 2.8 million times, or an average of 3,027 times per post. The same trend holds when comparing the Labour and Conservative official Facebook and Twitter accounts too. Labour’s posts received 1.7 million positive reactions, while the Conservatives’ posts received just short of 700,000 positive reactions. Taking as read, for the moment, that younger voters are simply more likely to support Labour and simultaneously be active users of social media there is something to say about the way users respond to content from the parties differently.

Taking the idea of Shy Tories into the digital realm means that even those social media users that intend to vote for the Conservative party are less likely than their Labour peers to interact with political posts that are explicitly pro-Conservative. Likewise, those same Shy Tory voters, who tell pollsters that they will vote for the Labour party, may also be tempted to interact with pro-Labour content online. In doing so, these voters amplify the pro-Labour message online.

The reason that Conservative voters, and particularly younger Conservative voters, may be tempted to interact with pro-Labour content online is to do with a social phenomenon called ‘virtue signalling’. Virtue signalling is a term coined by James Bartholomew, a British journalist. It refers to the act of performing one’s socially-acceptable values, even if one does not in fact hold those views. In 2016, Bartholomew wrote:

“By saying that they hate the Daily Mail or Ukip, they are really telling you that they are admirably non-racist, left-wing or open-minded. One of the crucial aspects of virtue signalling is that it does not require actually doing anything virtuous.”

Now, if any part of modern life is set up for virtue signalling it is social media. Firstly because of how straight-forwardly easy it is for users to demonstrate that they are ‘admirably non-racist’ with a few clicks. Secondly, the social pressure brought on by thought-silos of like-minded connections (echo chambers) being alienating to those privately considering voting for the Conservatives mean that the temptation is there for Shy Tories to engage with content from pro-Labour channels. That said, it is difficult to unpick the extent to which social media is able to convince undecided voters to turn out for one side or the other. It is easier however, to track how effective social media is in convincing your own core base to turn out and vote – a key issue for Labour’s younger voters. Although the data is currently missing on this (in part held back by Facebook’s black-box approach to advertising online) there is a strong case to be made by coupling strong social media performance and turnout. For example, previously Conservative-held Canterbury (Conservative majority 9,789) is home to a large student population. This constituency went Labour in 2017 (Labour majority 187) on the back of the student turnout. As such, there is an argument to be made that social media can effectively motivate your core vote to turnout if used appropriately.

On a slightly different note, the various social media channels have distinct tones and cultures. Twitter’s post limitations demand that users are pithy and as a result it is home to quick off-the-cuff remarks and ironic wit is considered the appropriate tone, for example. There are no social channels, however, where the tone is primarily about sober, considered messages or negativity. As such, the Conservative messaging ‘coalition of chaos’, ‘no deal is better than a bad deal’ and, of course ‘strong and stable leadership’, fall flat online. Meanwhile, Labour’s messaging more closely aligned with the tone of social media, including sharable videos with inspirational messages, popular hashtags like ‘#catsforcorbyn’ and the campaign slogan ‘for the many, not the few’, which has the double effect of enabling virtue signalling and hitting the correct tone for Facebook.

Time for a Rethink

So how did the pollsters miss Labour’s surge? At least part of the answer seems to be a failure to understand the role social media has played in changing the campaign landscape. With more young people preferring their social feeds to traditional media for their news, the role that social media can play in shaping election outcomes looks set to continue. YouGov, to their credit were able to tap into some of the things that led to the election result – most notably the increased number of young people coming out to vote. However, the reasons for young people turning out to vote are trickier to unpick. The shock Brexit result probably mattered, the tripling of tuition fees for university students under the Conservatives probably mattered too. The slow-motion car crash of the American presidency and May’s overly-friendly relationship with Trump might have mattered. However, this alone was not enough at previous elections to motivate young voters. Hence, the quality of the Labour party’s campaign online ought to be seen as one of the reasons they were able to outperform expectations. Other observers have written about the success of Labour’s ground-campaign – that is activists knocking on doors and talking to voters and helping to turn out the vote. This was absolutely essential to Labour’s relative success too. Likewise, the Conservatives’ inarguably terrible campaign performance did Labour more than a couple of favours in convincing Corbyn’s core vote to turn away from May’s party.

However, the role of social media should stand out as one of Labour’s big successes in this election campaign and a point of interest to social scientists looking to examine the 2017 UK election from all angles. As digital social communities evolve, social practices and norms evolve around them too. As such, studying social spaces online and especially the thought-silos of online communities is going to become more important to social scientists hoping to understand the ideas, values and communication among young people.