Banner

UK Election Forecasting Theory, Techniques and Controversial Discussions

This page will provide detailed information about the techniques and science used when forecasting UK-Elections, and will also highlight some of the past and continuing discussions on this sometimes controversial subject.

As a start, here are some links to some election-related commercial, academic, and other sources of information and debate.

Academic and Theoretical Sources

  • YouGov Election 2024 MRP Prediction Model Details MRP (Multi-level Regression and Post-Stratification) is a technique that assumes similar voters in different constituencies vote in a similar way. (It may be worth noting, as a slight oversimplification, that this is not actually always the case unless they are effectively subdivided by category again and again, until ending up with, for example, "formerly Conservative voting, working class, white, elderly, voters living in an ex council house in a North-East England constituency where that constituency is not a Conservative marginal." Of course, a standard opinion poll sample of about 1000 voters may contain very few such people, so significant changes of vote may not be easy to detect.) MRP models are usually fine-tuned to very accurately predict past elections, but are highly reliant on good data sources on a constituency-by-constituency basis, and changes to boundaries can be as much of a challenge to MRP methods as to other methods. The record of MRP models in forecasting elections ahead of time is very mixed, and different MRP models produce forecasts as widely varied as other forecasting methods, and vary in accuracy between elections (e.g. the YouGov MRP model forecast was one of the most accurate forecasts in 2017 and one of the less accurate in 2019). This link provides details about YouGov's current MRP model that uses polling data to identify common characteristics of voters in different constituencies.
  • MRP: what it is and why it may, or may not, be right Mark Pack discusses MRP in detali: what it is and why it may, or may not, be right at the next general election.
  • Breaking the Swingometer: Historical Precedents for Proportional Change Stephen Fisher of Elections Etc. considers whether traditional Uniform Swing based estimates are appropriate for the UK General Election of 2024
  • Why Does Everyone Disagree on MRP? James Kanagasooriam of UK Polling Report asks "Why Does Everyone Disagree on MRP?"
  • How Grumblers And Defectors Have Laid A Trap For Our Pollsters Peter Kellner explains his major doubts about the use of MRP by pollsters.
  • Election Forecasting in the UK Academic paper by Clive Payne from 2001, discusses UK Election forecasting generally, and reviews the various statistical methods used by the BBC to forecast different types of election in the UK in the previous thirty years.
  • Exit polling in a cold climate: the BBC-ITV experience in Britain in 2005 Academic paper by John Curtice and David Firth, provides an excellent discussion about election-poll related matters, drawing on the experiences of the BBC/ITN exit pollsters in 2005. This paper includes a mention of the Curtice/Firth method for calculating seat win probabilities.
  • Polls and the Vote in Britain Academic paper by Christopher Wlezien, Will Jennings, Stephen Fisher, Robert Ford and Mark Pickup discussing the evolution of electoral sentiment over British electoral cycles. Contends that polls from early in the electoral cycle contain much information about the final outcome.
  • Predictable and unpredictable changes in party support: A method for long-range daily election forecasting from opinion polls Academic paper by Stephen Fisher, outlines several methods for forecasting the next British general election on a daily basis from twenty months prior using opinion polls. Stephen's website is elections etc.
  • Back to Swing: Explaining uniformity and non-uniformity Article by David Voas in which he examines the evidence in support of Uniform National Swing, and finds it to be still quite strong.

Discussions and Debates on UK Election Forecasting (warning: can get heated!)

  • The 2010: UNS, Proportional Swing and all that... Useful comparison on the PoliticalBetting.com site: Dr Rob Ford, Dr Will Jennings, Dr Mark Pickup and Professor Christopher Wiezen reflect upon models and methods for projecting the vote at the 2010 British Election. Includes a summary showing how accurate the various methods (and several websites) were at predicting the 2010 election.
  • Ford: Our Model vs. 538 Political Scientist Robert Ford's explanation (on the www.Pollster.com site) of how the PoliticsHome UK poll tracking and seat projection model used in 2010 differs from the model developed by Nate Silver and his colleagues at FiveThirtyEight.com
  • Nerdfight - UK Election Methodology Defence by Nate Silver of the techniques he was using to predict the 2010 UK General Election
  • Wonk Wars: how does your model swing? Useful summary by Alex Barker on the FT Westminster Blog of the Nate Silver / Robert Ford debate (with further links, requires FT access)
  • UNS discussion on PolicalBetting Article by Andy Cooke on the Political Betting website about Uniform National Swing (Part 1)
  • UNS discussion on PolicalBetting Article by Andy Cooke on the Political Betting website about Uniform National Swing (Part 2)
  • UNS discussion on PolicalBetting Article by Andy Cooke on the Political Betting website about Uniform National Swing (Part 3)
  • Uniform Vs Proportional Swing - Which is Best? Comparison on the Liberal Democrat Voice site by Mark Pack, "Presenting a new analysis of the merits of the two main ways of converting party vote shares into seat number projections"

See Also:

Subscribe to our forecast email updates here

 
©Copyright 2026 UK-Elect