BREXIT graphs then and now

Short post showing graphically what the national survey of 944 respondents (quota sampled) gives in terms of model predictions of the vote.


Red is the cumulative vote likelihood amongst the actual voters in last year’s EU referendum, blue is that for those who claimed they’d vote in a hypothetical April 2017 rerun – including up to 5 million additional voters (nationally) splitting 3.5 million REMAIN – with weightings to reflect likely turnout based on attitudes.


The effects are quite easy to see. Though the model slightly over-predicts the LEAVE vote when looking at the national distribution, the mean probability matches the observed split in last year’s vote. Thus the movement in the curves is of most interest – particularly:

  • Where it crosses the 50% chance of voting LEAVE line – above that a respondent is increasingly like to vote LEAVE
  • How the huge net extra REMAIN block heavily shifts it downwards for people up to around 60% – which led to my prediction of a REMAIN win of 51% nationally and in the East Midlands
  • How the East Midlands presents a real problem for Labour MPs there. The ‘top 45%’ of voters have LEAVE vote likelihoods greater than around 70% and in essence there are two curves here – for REMAINers (the bottom 0 to 45%) and HARD BREXITeers (the top 55% to 100%). There’s quite a steep gradient in the middle 10% reflecting sensitivity in their vote (partly driven by lower turnout, as you might expect, if you’re quite unsure of your how to weight your competing attitudes to form a single vote).


The ‘extra’ voters in the East Midlands causing the shift in the blue line actually came primarily from

  • Previous non voters in the 2015 general election
  • REMAIN Conservatives – in the East Midlands among those who voted Labour in 2015 there is a LEAVE majority, whilst among those who voted Conservative there is a REMAIN majority (though that reflects, in part, that a bunch of LEAVErs defected to UKIP in 2015, but they didn’t vote again in the 2017 GE and anecdotal evidence given to me suggests they voted LABOUR due to disgust at the Tory campaign and a deliberate downplaying of BREXIT by Labour candidates in the East Midlands. The Tory leaflets round here certainly were more nationalist than I’d seen in years. The Conservative vote in Gedling (and Broxtowe where Anna Soubry – REMAIN ‘character extraordinaire’ – clung on, just) is quite ‘Tory Wet’ in nature.

The East Midlands is highly polarised and Labour 2015 voters continue to lean heavily toward HARD BREXIT – see how closely the red and blue lines are at the top. Labour MPs are therefore understandably twitchy about Labour’s National ‘fudging’ of Europe (as the Guardian reported) – and which I noted, was necessary for Corbyn to get the young REMAIN people on board. But it’s a whole different kettle of fish round here, and my local Labour MP (who I spoke to yesterday) immediately recognised the ‘Mansfield effect’ in the top part of the graph – where a heavily BREXIT Labour ultra-safe seat went Conservative for the first time ever when the Conservative hard line in the general election campaign and ‘too REMAIN’ a local Labour campaign caused a direct swing from Labour to Conservative. Interesting times.