MRP poll: seductive for the Conservatives, grim for Labour

It is billed as the poll that came closest to calling the unexpected result of the 2017 general election, although some of the buildup is helped by YouGov’s substantial marketing budget, the envy of rival pollsters.

But the result of its MRP poll – a recently developed technique called multilevel regression and post-stratification – for the Times, with its constituency by constituency model, makes seductive reading for the Conservatives and grim reading for Labour, predicting a big win for Boris Johnson and 359 seats, or 42 gains, for his party – and a majority of 68.

Labour, meanwhile, falls back to 211 seats – a result that would be in line with the disaster of 1983 – on the back of the analysis, which unlike regular polls uses a massive sample of 100,319 people to try to model the result in every single constituency in Great Britain.

YouGov did score a success last time: nine days before the last election the firm predicted that Theresa May’s Conservatives would lose their majority with 310 seats. In fact the party actually got 317, but it came at a time when few other pollsters were predicting anything other than a May win.

However, for this analysis it has been polling people over the past week – when regular polling firms would aim for the last couple of days – and during an election campaign every day matters.

A week ago, neither party manifesto was out, the leaders’ debate had not happened – and the Labour antisemitism row had not surfaced. Also, more importantly while the Conservative squeeze of the Brexit party has come early in this campaign, the Labour squeeze of the Liberal Democrats has yet to complete.

There are some things in the forecast that look curious. Labour is not forecast to gain any seats at all, which would surprise party workers in some targets where it has dozens on the ground, such as Southampton Itchen (Conservative majority 31), Hastings and Rye (Conservative majority 346) and Chingford and Woodford Green (Con majority 2,438).

The polling also says Labour would suffer in the critical “red wall” band that runs very roughly from north Wales to the Humber, where party insiders are already deeply concerned, with some of its workers reporting a sharp fall in those willing to say they will back Labour.

The modelling suggests the Conservatives would gain 44 seats from Labour, many in Britain’s smaller cities and large towns: Wrexham (Labour majority 1,832), Derby North (Labour majority 2,015), Great Grimsby (Labour majority 2,565), despite the residual presence of the Brexit party.

Even West Bromwich East, previously held for Labour by Tom Watson on a majority of 7,713, is projected to change hands, and interestingly most of the seats projected to switch are straight Conservative versus Labour battles. In only a handful, such as Kensington in London, does a Lib Dem revival split the anti-Tory vote.

This poll does at least bring some clarity: the British public is being presented a scenario in which Boris Johnson has five years in Downing Street to complete Brexit and strike a trade deal with Donald Trump and more.

At the last two elections the electorate was presented with such prevailing scenarios at the last minute, and on this occasion there are 15 days to go. But if something is going to change, it would have to be dramatic, and in those seats in Wales, the Midlands and the north of England.


Leave a Reply

This website uses cookies. By continuing to use this site, you accept our use of cookies.  Learn more