Will YouGov’s election model be vindicated when the results are in?

Opinion polling firm YouGov have raised eyebrows in the run up to Thursday’s poll with the results of their forecast model, which at the time of writing is forecasting that the Conservatives will only win 307 seats. This would deprive them of their parliamentary majority, and leaves open the possibility that Jeremy Corbyn could become Prime Minister if Labour could secure support from the smaller parties in the Commons.

The YouGov model is out of line with other election forecasts, which all predict that the Tories will have a majority in the next Parliament, albeit not the landslide majority which appeared likely a few weeks ago.

We will soon find out one way or another, of course, but the early results on Thursday night/Friday morning could provide a clue as to how the model is performing overall. The following tables have ranked the seats from 1 (the easiest) to 583 (the most difficult) where the Conservatives either came first or second at the 2015 election. The figures in brackets show what the YouGov model say will be the size of the Conservative majority, with a minus figure showing the majority of the winning party in seats where the model predicts that the Conservatives will lose. The colour of each cell shows the incumbent party (Conservatives blue, Labour red, the Lib Dems orange, UKIP purple, the SNP yellow and Plaid Cyrmu in green). The data was taken from the model released at the weekend.

The data used can be viewed as a Google Docs spreadsheet here.

The YouGov model predicts that the Conservatives will win 307 seats, with the smallest majority in High Peak, which is predicted to be a statistical dead heat with Labour. Of these 307 seats, nine are seats in which they are not the incumbent party (five from the SNP, three from the Liberal Democrats, and the former Ukip seat in Clacton).

If the model is correct, then the Home Secretary Amber Rudd would lose her Hastings and Rye seat to Labour. The 325th seat for the Conservatives would be the seat of the Liberal Democrat leader Tim Farron in Westmorland and Lonsdale. The 326th seat that would give the Tories a majority of one is the Gloucester constituency of Stroud, where the Conservative incumbent Neil Carmichael is defending a majority of 4,866, and bookmakers are offering odds of 1/25 that the Conservatives will win.

There are a number of constituencies where common sense would suggest that the model is lowballing the Conservative vote share. For example, the model suggests that the Conservative Hugo Swire will lose by nearly 11 points in East Devon to an independent candidate, despite having held the seat since 1997 and defending a majority of 12,261 (22.4%).

For those sitting up to watch the results into the early hours of Friday morning, the first seats to declare will provide an early clue as to how the YouGov model has performed. The following seats are the ones that the Press Association says are likely to be the first to declare in Great Britain. Comparing the results with the forecast model should provide an early read on how the rest of the night is likely to go.

Houghton & Sunderland South (expected 23:00 – forecast Labour hold, majority of 29.4%)
Con 28.4%           Lab 57.8%            LD 2.9%                 Ukip 8%                Green 1.6%

Sunderland Central (expected 23:30 – forecast Labour hold, majority of 24%)
Con 31.5%           Lab 55.5%            LD 4.7%                 Ukip 5.5%            Green 1.8%

Washington & Sunderland West (expected midnight – forecast Labour hold, majority of 27.7%)
Con 29.9%           Lab 57.6%            LD 2.9%                Ukip 7.6%            Green 2%

Swindon North (expected 00:30 – forecast Conservative hold, majority of 13.1%)
Con 50%               Lab 36.9%            LD 4.6%                Ukip 6.3%            Green 2.2%

Battersea (expected 01:00 – forecast Labour gain from Con, majority of 3.3%)
Con 40.5%           Lab 43.8%            LD 8.9%                Ukip 1.6%            Green 2.2%

Newcastle upon Tyne Central (expected 01:00 – forecast Labour hold, majority of 39%)
Con 23.8%           Lab 62.8%            LD 6.6%                Ukip 4.5%            Green 2.2%

Newcastle upon Tyne East (expected 01:00 – forecast Labour hold, majority of 38.5%)
Con 24.6%           Lab 63.1%            LD 6.1%                Ukip 4%                Green 2.2%

Nuneaton (expected 01:00 – forecast Conservative hold, majority of 8.6%)
Con 47.2%           Lab 38.6%            LD 4.7%                Ukip 7.3%            Green 2.2%

Putney (expected 01:00 – forecast Conservative hold, majority of 8.3%)
Con 43.8%           Lab 35.5%            LD 12.8%              Ukip 2.1%            Green 4.2%

After 1am, the results should start to come in thick and fast, but key races to watch which should declare before 2am include Bury North, which the YouGov model suggests could be a Labour gain from the Conservatives and would be Tory seat number 340, and Thurrock, which was tight Lab/Con/Ukip 3-way marginal in 2015 and where the Conservatives are defending a majority of 536.

If, four hours in to the count in the early hours of Friday morning, Labour have swept Battersea, Bury North and Thurrock, then we will be in territory where the Conservatives could be deprived of a majority in the House of Commons.

On the other hand, if the Conservatives have won all three of these, and are exceeding model expectations elsewhere, then they are virtually guaranteed a majority. The question will then be if the Tories are capable of winning a landslide by taking Labour seats in areas such as Don Valley, Hartlepool and Tony Blair’s former seat of Sedgefield.

Whilst there is much uncertainty regarding this peculiar general election, if Labour manage to pull off what would be one of the greatest shocks in British election history, the picture should start to emerge at an early stage in the count.

A qualified accountant and data analyst, interested in politics, economics and data. Twitter: @peterdonaghy