Tactical voting could swing this election – but we need to get it right

Zoe Williams
Photograph: Alamy Stock Photo

It is rare, in a first-past-the-post electoral system, to live in a constituency where you can vote exactly as you choose and know you’ve made an impact. In a safe seat – which is most of them – you’re either piling up votes needlessly for your candidate or barely making a dent in the majority of your opponent. Tactical voting is thus nothing new, whether it is done directly by voting for a party you don’t support, while knowing your own preference hasn’t a hope, with a mind to stopping something worse, or indirectly, via the relatively recent phenomenon of sites such as Swap My Vote.

Related: Tactical voting website criticised for 'bogus' advice

Rarely has interest in tactical voting been as intense as it is for December’s election, which is partly attributable to polarisation. These Conservatives, to mangle the M&S advert, are not just any Conservatives: they are extraordinarily extreme, economically reckless, socially contemptuous (see Jacob Rees-Mogg and his views on the victims of Grenfell) and instinctively ethno-nationalist Conservatives. More people, across a greater range of political standpoints, will go to greater lengths to prevent a Boris Johnson government than one led by David Cameron. And on the other side of the coin – although it is facile to draw a direct equivalence of extremism – Jeremy Corbyn’s Labour is considered far more dangerous by the right than Ed Miliband’s ever was.

Naturally, it is also Brexit that makes this election more consequential – whatever the outcome, many of us – give or take half the country, ho hum – will be deeply unhappy with it. So the understood norms of tactical voting – that it will be limited to marginal constituencies, that only 5-10% of voters will engage in it – will doubtless be stretched. But psephologist Chris Hanretty warns against exaggeration: it’s not an issue for the Tories because there’s no challenger (the Brexit party is more of a bugbear; having no MPs, it presents no historical or quantifiable threat). So this is really only the preserve of Labour and Liberal Democrat supporters, with homeopathic quantities of Greens, which establishes the ceiling. Within that, there will be some who would never switch, and some who will cross the divide and go Labour to Tory (also bear in mind those who voted Lib Dem before but who are pro-Brexit). “Let’s say you can double it, there will still be no more than 15 to 20% of the electorate voting strategically,” Hanretty predicts.

Yet that’s enough to change the result. A study four years ago found that in 1997 tactical voting helped to elect nine Labour MPs and 21 Lib Dems (what’s surprising about that research is, given the Labour landslide, people always assume Blair did the best out of the deal). Yet that same paper had a sting in its tail, estimating that up to 30% of tactical votes went in the wrong direction, away from the party that stood a good chance of winning, towards one that didn’t.

There is a huge amount of animus and posturing flying about right now: Lib Dems castigating the Labour party for their refusal to commit to remain, Labour supporters vowing never to support any party that was once (pretty recently, granted) in coalition with the Conservatives. Most of that is noise rather than signal. For any voter opposed to austerity, mindful of the climate crisis and anti-Brexit, the overwhelming impetus will be to avoid a Conservative majority, and for that they need not soul-deep certainty about the characters of Jo Swinson and Corbyn, but reliable, uncontested information.

Typically, this starts with identifying marginals, but it is a little more complex than simply counting up majorities to find those under 5,000 (then there are ultra-marginals, with three-digit majorities). Context matters. “The SNP is defending a number of small leads,” Hanretty says, “but the votes are breaking SNP, so we should be removing them from our marginal calculations.” It is widely assumed that both the major parties are facing total wipeout in Scotland. Most tactical voting sites are working on the basis that there are 150 seats in which voting tactically could be decisive, and mapping from there, using MRP (multi-level regression and post-stratification) data.

This matters a lot when you’re trying to predict the vote in a given seat, which is the only way to make this decision. Large, national opinion data used to be a reasonable way to forecast seats, but not any longer. People still do it, because the alternative is so expensive, but it just doesn’t seem to tell us very much; the most accurate predictions, from the referendum to the Peterborough byelection, were all made using MRP.

This is where you take a number of data points – how someone voted in 2017, in 2015, in the referendum, their sex, their level of education – then you figure out the number of people in that bracket in each constituency (this is the post-stratification). The samples are typically vast – tens of thousands – which yields millions of data points. One pollster talked of having 2.5 million voter personas.

There is some discomfort around time lag when our politics is so volatile – you’re getting a picture of the nation six months ago (if the EU elections are your last solid data point), so it’s hard to keep up with campaign effects. Yet those in the business of making predictions broadly agree on the value of the research.

No consensus has so far emerged on how to interpret it, however. Best for Britain came out early, with suggested votes for remainers in each constituency; roughly twice as many Labour candidates as Lib Dems (350+ versus 170+), but still, some of the suggestions are wildly favourable to the latter. Cities of London and Westminster is a clear Tory-Labour marginal, and yet the advice was given to vote Lib Dem before the Labour candidate had even been chosen. In many ways, it would be as premature to get into mud-slinging as it was to produce the advice in the first place, since there is nothing to stop Best for Britain revising their advice all the way from now until 12 December. Tactical Vote 2019 – with an explicit “stop the Tories” rather than “stop Brexit” agenda – currently seems the most credible. Other dogs are yet to bark (People’s Vote, most obviously, which is mired in its own crisis).

Urgently needed is a site for the tactical climate vote – bespoke calculations not necessarily for Green voters but for anyone with the environment as their top priority. This will most likely push voters towards Labour wherever it’s realistic, since its climate policy has never been better. The overarching necessity, of course, is that these organisations find some way of dovetailing their suggestions; there would be nothing less tactical than everyone voting strategically in competing directions. And if the progressive parties can’t cooperate on that, God help them in their dream parliament.

• Zoe Williams is a Guardian columnist