Three UK Election Projections, Raw Numbers and Seat by Seat

-edited

YouGov projects a Conservative majority of 68, Electoral Calculus 34, and a model of ComRes data yields 48.

Electoral Calculus has a nice User-Defined Election Projection Tool. For the above chart, I used the Electoral Calculus current model.

Electoral Calculus Seat Change Projections

Labour does not pick up any seats in the Electoral Calculus model. It projects a 34 seat Tory majority.

Note that Electoral Calculus projects Dominic Raab will lose his seat to the Liberal Democrats.

ComRes Projection

ComRes does not offer any projections but it does have region-by-region totals that I used to feed Electoral Calculus.

Electoral Calculus Projection From ComRes Data

Electoral Calculus Seat Change Projection From ComRes Data

This model projects Labour will pick up one seat from Plaid Cymru while losing 40 seats elsewhere.

This model also shows Tories doing better in Scotland than most believe. I suspect that will be the actual case based on falling poll numbers of SNP.

YouGov Analysis

The YouGov Election Centre updates its model periodically. It provides a seat-by-set CSV-file download which I took then added some cross-check calculations.

It is not easy to see on the raw YouGov CSV file which party won. I added a set of columns to show which party won each seat.

You can download my summation of their data from my Google Drive UK Seat by Seat Projection File.

There were 8 ties. I cannot tell from the CSV file which party won the ties other than Tories 2, Labour 4, and Liberal Democrats 2.

Results

  • YouGov Nov 19-26: Tory 68 Seat Majority
  • Electoral Calculus Nov 26-30: Tory 32 Seat Majority
  • ComRes Data Modeled by Electoral Calculus Nov 27-28: Tory 48 Seat Majority
  • Average: 49

I believe the average is about right.

Mike "Mish" Shedlock

Comments (16)
No. 1-10
Scooot
Scooot

Thanks for this. Hope it’s right.

Mish
Mish

Editor

YouGov was the only pollster that called for a hung parliament in 2017. They were universally laughed at for their call.

If you give a damn about who founded it, you are the fool.

The polling firms all want to be right. It's how they get business. Sheeesh.

Mish
Mish

Editor

Matt Singh: "How so these consistent differences arise? Contrary to the widely-peddled myth, it has nothing to do with the editorial leanings of the newspaper publishing it."

To that I will add ... Nothing to do with a polling firm paid to be right.

Mish
Mish

Editor

Ad Hominem attacks on pollsters who want to be right is damn foolish.

This does not make YouGov right, but it does make the attack on YouGov for political bias look more than foolish.

Mish
Mish

Editor

Germ did you delete your comments?

Apparently so.

Germ pointed to analysis condemning YouGov because they were founded by conservatives.

Curiously, this conservative-biased organization stood out against the crowd and predicted a hung parliament when everyone else predicted a Tory victory.

Germ is a coward who then deleted his comment and now lectures me.

Get real.

Mish
Mish

Editor

If I can recover these in entirety - I will

Mish
Mish

Editor

As amazing as it might seem folks, Germ is on to something assuming you believe all of the foillowing:

YouGov can hit the middle of a range. Which in turn implies YouGov knows what the other pollsters will do. Of course this implies that all the pollsters are in on this together for some nefarious reason. It's all rigged.

If you don't believe that it's naive. The proof speaks for itself. And of course it is important to think in 4-D or 8-D terms. Thus, YouGov in 2017 purposely predicted a hung parliament hoping to get lucky so that it could for the next 8 years pedal Torry-Biased garbage.

The proof, of course is visible in the supplied chart.

Carl_R
Carl_R

Obviously there is nothing to be gained by releasing polls that you know to be wrong. That said, there are pollsters who try to alter the outcome, but they do it via "push polls", not erroneous ones. Push polls have leading questions, designed to affect the opinion of the person being polled, rather than to determine his opinion. Note: I'm not saying YouGov is doing push polls. I'm agreeing with Mish that it would be absurd to think that any pollster would deliberately damage their own reputation by releasing incorrect numbers.

Scooot
Scooot

These polling companies don’t just poll elections, they do many other surveys too. Their accuracy is what drives their business. Election polling is a huge marketing exercise, getting it right is very important for their business.

lamlawindy
lamlawindy

Mish, I just checked out your spreadsheet with the individual UK constituencies prediction model. Very interesting. Keep up the good work!