Methodology
National polls do not convert easily on a one-to-one basis to seats in Parliament. For example, in 1997 the Labour Party won 63% of the seats on 43% of the vote. Our model takes the simple principle of “uniform national swing”—the idea that support for parties rises and falls across all constituencies in the nation by the same magnitude—and augments it with specific regional polling, where it is available, from Scotland, Wales, London and so on.
For example, the vote share of the Conservative Party in Macclesfield in any given general election—a seat that they have held since 1918—is modelled as a function of its performance in the prior general election, as well as how the Tories are performing in both nationwide polls and specific polls from the north of England.
To estimate seats in Parliament from polling, first we trained a model using 9,398 individual constituency-level election results between 1959 and 2019 along with polling data (polling trend lines for each election cycle were estimated using the
MGCV package in
R). We fit a multinomial logistic regression model using the
LASSO method, a statistical technique that eliminates or reduces the impact of certain variables in order to maximise accuracy on unseen data.
The point estimate errors from a leave-one-out cross-validated version of this model were fed into another model that estimated the covariance between every party and every individual constituency between 2010 and 2019. For example, if Labour gains one vote in Macclesfield, there is a high chance that they will also gain another vote in Hereford, but the party is less likely to see a concurrent gain in support in Richmond Park. After 10,000 simulations of these correlated vote shares, we estimated the 95% confidence intervals for every party's vote share in each of Great Britain's 632 constituencies (ie, excluding Northern Ireland). We have assumed that every major political party will be standing a candidate in each constituency.
Our model does not take into account the local effects of parties' campaigns in individual constituencies nor tactical voting. As the date of the next election is uncertain—it must occur before the end of January 2025—our prediction is a “nowcast” based on a scenario in which the election occurs tomorrow. Once the election is called we will update our model to incorporate uncertainty over the six-week election campaign. Until then we will frequently update our estimates to include the latest polling data.