This year’s general election result took almost everyone by surprise, including the pollsters, forecasters and other experts. On 3 June, Joe Twyman, Dr Ben Lauderdale, Dr Rosie Campbell, Professor Justin Fisher and Professor Matt Goodwin took part in a roundtable to discuss where the predictions went wrong and lessons for 2020. David Ireland offers an overview of the event.
The exit poll that came out at 10pm on 7 May took almost everyone by surprise. Over the course of Friday morning, the scale of the Conservative majority revealed itself, showing that even the exit poll had underestimated the Conservative support. What happened? How did the polls get it so wrong and what are the lessons for 2020? This blog highlights the key issues from a recent roundtable on GE2015 hosted by UCL’s Department of Political Science and the Constitution Unit and chaired by Dr Jennifer Hudson.
Joe Twyman, Head of Political and Social Research, YouGov
As one of many pollsters who had long predicted a hung parliament, Joe acknowledged YouGov didn’t get it right this time. He also, rather humorously, showed the range of Twitter abuse directed at him as a result.
Voting intention remained tightly balanced in the months leading up to the election, but YouGov’s polling revealed that the ‘fundamentals’ may not have been given enough weight in predicting vote share.Importantly, no party had ever come from behind on the economy and leadership to won an election before, and this election was not to be the first. The economy remained the single most important issue, and here, the Conservatives were significantly ahead. Similarly, Miliband never got close to Cameron on party leader ratings.
The failure of the polls was the failure to get the ‘horse-race’ between the two main parties right. But the collapse of the Liberal Democrats, the rise of the SNP in Scotland and the emergence of UKIP were all correctly predicted by YouGov. Joe also addressed the issues of private polls completed on behalf of the parties which claim to have predicted the correct result, but which cannot be verified as they were not published either at the time or after the event.
The process of understanding what went wrong has already begun. The British Polling Council (BPC) has launched an inquiry with the cooperation of its member organisations. Joe indicated it is likely to be a combination of different issues led to the inaccuracies across the industry.
Dr Ben Lauderdale, LSE
There were more polls in the year leading up to the 2015 general election than any other, including constituency-level polls funded by Conservative Peer Lord Ashcroft. This provided ample data for forecasters like Ben and colleagues at Election Forecast UK. Like the pollsters and other forecasters, their predictions were inaccurate, and they have begun the process of establishing why.
For Ben, the most frustrating thing about the results is that the eventual number of seats won by the Conservatives was outside of the margin of error calculated in his model. It is clear that the national polls were wrong, but election results indicate that constituency-level polling understated the Conservative vote share in Conservative marginal seats, whilst overestimating them in seats where the Tories were uncompetitive. Thus, forecasts underestimated Conservative support in seats where it mattered: in key marginals. Constituency polls were widely believed to help make forecasts more accurate, but their inclusion provided only very slight improvements on overall seat predictions; something Ben, as part of the BPC enquiry, is looking into.
Dr Rosie Campbell, Birkbeck, University of London
Outside of two largest parties, this election was an election of optimism. Yet only one of them was able to convert this optimism into an increased number of seats in parliament. Rosie explored the socio-economic differences between the Scottish Nationalist Party newcomers and the established Westminster parties to establish how different these supposed outsiders really are. In terms of occupational background, the SNP cohort are similar to Labour, with a similar proportion of their MPs coming from jobs in education, the NHS and manual work. One clear point of difference is that the SNP has the largest proportion of MPs whose previous job was a political researcher, indicative of their strength in Holyrood. This background contrasts with the Conservative party MP’s who are much more frequently from business and legal backgrounds. Their education is both similar, in that the vast majority hold at least an undergraduate degree, but it is also different, in that no SNP MP attended Oxbridge.
The SNP has also contributed to increasing the number of female MPs within the House of Commons, providing 20 female MPs – 36% of the total number of SNP MPs – out of the 191 female MPs now in Parliament. With 99 female MPs, the Labour Party is particularly responsible for driving the proportion of women in Parliament upwards. This represents 43% of all Labour MPs.
Professor Justin Fisher, Brunel University London
Equally important for understanding this election is party finance and electoral spending. Justin identified the ‘hyper-local’ nature of the campaign and the large financial advantage the Conservative Party enjoyed during the campaign, which allowed them to spend early and often. But contrary to the received wisdom, the vast majority of parties’ expenditure was not on the campaign itself, but on support activities (rent, salaries, etc.).
Labour’s late surge in donations is a direct consequence of the unions withholding money in the fallout from the row over the selection of the Labour Party candidate for Falkirk. Ultimately, however, this money came too late. Whereas the Conservatives were able to plan well ahead of the short campaign, with polling and telephone voter identification in target seats, Labour were unable to match their activity.
The Conservatives were the only party to utilise a national press campaign or use billboards. Direct mail was also used to target voters in marginal seats with tailored messages. Whilst social media played a greater role in this election, than in any previous one, it is still seen as less important than face-to-face contact from knocking on doors.
Professor Matt Goodwin, University of Kent
Does UKIP’s solitary MP after this general election signal a disastrous campaign and the end of the party as a force in British politics? Not according to Matt Goodwin, who argued that the 120 second place finishes in this election reveals a crucial insight into the changing nature of British politics, and the long-term prospects of Nigel Farage’s divisive party, than the number of seats won.
Crucial to his analysis is UKIP’s dominance of two key issue areas: immigration and Europe. The ‘ownership’ of the issue of immigration is particularly worrying for the Conservatives, for whom the issue has been a traditional strength. Since 2010, their dominance on the issue has been taken over by UKIP who are now seen by the electorate as being ‘best’ on the issue.
The rise of UKIP doesn’t affect just the Tories, as 45 of UKIP’s second place finishes came in Labour safe seats, including in the North of England. The Labour Party did not originally perceive UKIP as a threat, and paid much less scrutiny to the UKIP threat than did the Conservatives. UKIP is a party who can use their dominance on key issues to gain 10-15% percent of the vote, and it both parties will have to become accustomed to their presence and adapt.
The shock of 7 May has now been replaced with the task of explaining and understanding what happened. The panel highlighted the hurdles that Labour, in hindsight, had little chance of overcoming. Not only were they significantly poorer off than the Tories, they were perceived as less competent on the economy, and Miliband came a distant second to Cameron on leader ratings. These factors, combined with a seemingly better organised Conservative campaign, meant that the Conservatives were able to increase their vote share and seats and bring back single party, majority government to Britain when even the experts were least expecting it.
About the Author
David Ireland is a Research Volunteer for Parliamentary Candidates UK and is currently completing an MSc in Security Studies within UCL’s Political Science department. His work on the project has included compiling data on 2015 parliamentary candidates and measuring career trajectories within the House of Commons. Outside of PCUK, his research examines the impact of international integration on civil violence in the developing world.