Will Scotland leave the United Kingdom? Will the U.K. leave the European Union? And how could polls be so wrong? Those are among the questions being asked in the wake of a surprising May 7 U.K. national election. Brian Gaines is a political scientist at Illinois who has taught British politics and supervised a program that sent undergraduates to intern in the British Parliament. He also studies elections, public opinion and polling. He spoke with News Bureau social sciences editor Craig Chamberlain about what happened and why.
Perhaps the biggest surprise was that Scottish voters – who rejected independence from the U.K. last fall – gave 56 of their 59 parliament seats to the pro-independence Scottish National Party. How should we read this?
The SNP pivoted from defeat in September’s referendum very skillfully. Nicola Sturgeon, who took over the party immediately after that vote, deserves full credit for rallying activists and swelling membership figures. But it would be unwise to read the result – unprecedented though it was – as an overwhelming embrace of independence. The SNP’s impressive surge in votes swelled into 95 percent of the seats because of the first-past-the-post system, in which each seat is awarded based on a plurality, and not necessarily a majority, of the votes. In a multiparty system like that in the U.K., plurality elections can be hyper-responsive to voting shifts. A modest decline in support could easily cost the SNP most of those seats next time.
Consider, too, that the pro-union (anti-secession) vote was split many ways. Many of the new SNP members of Parliament did not win a majority of the vote in their constituencies. Moreover, it is easier to vote for a pro-secession MP than for actual secession. The separatist Bloc Québécois did so well in the 1993 election in Canada that it became the official opposition to the government. But Quebec is still part of Canada, and there are now only two BQ MPs.
The election gave Prime Minister David Cameron a second term and more seats for his Conservative Party. Yet it doesn’t sound like it will make his life that much easier. Why?
He will no longer depend on a coalition, as he has the last five years with the Liberal Democrats, but he enjoys only a slim majority. He doubtless knows how John Major’s 1992 triumph degenerated quickly, with a fractious Conservative Party eventually losing its majority. First and foremost, Cameron will have to manage his own MPs, many of whom still regard him with some suspicion.
What do the election results mean for the U.K. in the European Union?
For now, Cameron must time and then manage a promised referendum on EU membership. It is premature to assume that the public will vote to leave, and exactly what Cameron can demand, promise and expect in any renegotiation of terms with the EU is far from clear. The EU, meanwhile, remains preoccupied by the prospect of Greece having to quit the Euro, and maybe the EU, in the near future.
Both this election and the Scottish independence vote last fall were far from the close votes that polling suggested. Is this a problem of polling in the age of cell phones and the Internet? Or just a lot of people waiting until late to make up their minds?
In 1992, polls agreed that Labour would win, but the Conservatives greatly exceeded forecasts and formed the government. The polling industry was thrown into chaos, and, ultimately, British pollsters changed the way they drew samples. There might be similar lessons about logistics this time, but I suspect that bad forecasts in 2015 were less about who answered the questions and more about statistical models of how votes convert into seats.
Making that conversion is pretty easy in the U.S., with its two-party system. Republican gains are Democratic losses, and vice versa, with a slight wrinkle in that turnout also shifts. In multiparty systems, however, it is hard to figure out the breakdown of seats based on national polls when each seat is awarded by plurality or “first past the post.” Statistical models are hardest to work out when (a) the vote is highly fractionalized (there are many candidates competing for most seats); (b) fractionalization is changing; (c) “swings” in support vary by region; and, (d) amidst this chaos, voters become less attached to parties and more prone to late decisions. All of those conditions held in this election.