• Development Intelligence

#GE2019 – How did the pollsters do?

So at 2200 hrs last night the exit poll came in. Gathered from the responses of 20,000 actual voters outside 140 polling stations across the UK, as ever is has proven to be pretty accurate.

But after the last few election cycles when the pollsters haven’t always covered themselves in glory, now is a good moment to see how they fared this time. It should be remembered that not one pollster forecast anything less than a 5% Tory lead at any time during the election campaign and only one, Savanta ComRes, produced a poll below a 6% Tory lead and that was in their final eve of poll. In simple terms the pollsters broke in three directions:

Hung Parliament

  • Savanta ComRes – between 5-10% Tory lead, final poll of 5%

  • ICM – between 6-11% Tory lead, final poll of 6%

Reasonable Tory majorities

  • BMG – between 6-9% Tory lead, final poll of 9%

  • YouGov – between 9-17% Tory lead, final poll of 9%

Larger Tory majorities

  • NPC – final poll of 10%

  • Ipsos Mori – final poll of 11%

  • Panelbase – between 6-13% Tory lead, final poll of 11%

  • Survation – between 6-14% Tory lead, final poll of 11%

  • Deltapoll – between 11-15% Tory lead, final poll of 11%

  • Kantar – between 10-18% Tory lead, final poll of 12%

  • Opinium – between 12-16% Tory lead, final poll of 12%

  • Qrisously – final poll of 13%

With the final Tory poll lead actually being 12%, Opinium and Kantar were the winning pollsters with Savanta ComRes and ICM the losers.

Pollsters vary mainly because of their methodology; there’s as much art as science involved truthfully. And it should be remembered that all polls are a ‘nowcast’ not a ‘forecast’. But that is always such a good excuse the day after polling day!

We have been asked by several clients to give presentations in the New Year on the ‘post General Election political landscape’. If you feel this could be useful for your team, please get in touch.