Friday 27 March 2015

Latest UK-Elect General Election Forecast

UK-Elect General Election Forecast, March 25th 2015.
Hung Parliament - Labour Short By 46 + SNP (44) = ??
Labour is again forecast by UK-Elect to be the largest single party (although the margin is fairly small), and the Scottish National Party is again forecast to have a key role in deciding who will form the next UK government. Overall, the figures of 323 seats, the combined total for Labour + SNP MPs in this prediction, would be likely to prove just enough to form a government if the two parties can come to an agreement.
The reason for this change since the previous forecast is mostly due to a small increase in the predicted Labour vote relative to that of the Conservatives that may - or may not - turn out to be significant.
The margins, however, are extremely narrow, with a real chance of predictions taking too much account of transient movements in the opinion poll averages. A lot can still change before May 7th...
The UK-Elect "adjust target percentages when forecasting" option was set for this forecast, adjusting the percentages to represent what we currently expect to happen on May 7th, rather than just using the current poll percentages as a target. This option has the affect of adjusting the input opinion poll percentages to take account of what happened in many past elections as the date of the election approached - i.e. that the support levels for the parties returned part-way towards their previous totals.
Interestingly, this is also the first recent UK-Elect forecast to predict zero seats for UKIP, although it would have been possible to deliberately "special-case" this part of the prediction by increasing the incumbency support levels for UKIP because of the anticipated personal support for its incumbent MPs and its party leader (Nigel Farage). It is arguable whether or not this should be done, as most experts DO expect UKIP to win one or more seats due to this personal support, but the level is difficult to predict based on past elections. However, any further polls taken in constituencies such as Clacton and South Thanet will automatically feed into future UK-Elect predictions of the number of UKIP MPs, so it would be no surprise if this number were to increase again.
This forecast was created using updated UK-Elect v9.3 functionality. The method chosen this time was the UK-Elect v9.3 method, combining separate forecasts for Scotland, Wales, London and GB, and using the latest by-elections, constituency opinion polls and known candidate lists as part of the input. Incumbency support was disabled (although the v9.3 method does include some built-in incumbency support). Tactical voting was disabled as it is not clear to what extent it will occur in May. Multiple iterations were used to better achieve the correct target percentages.
The GB percentages input for this forecast were Con 33%, Lab 33%, UKIP 14%, Lib Dem 8%, Green 6%. For Scotland the percentages used were SNP 46%, Lab 27%, Con 15%, Lib Dem 4%, UKIP 4%, Green 4%, for Wales the percentages used were Lab 39% Con 25%, UKIP 14%, Plaid Cymru 10%, Green 6%, Lib Dem 5%, and for London the percentages used were Lab 43%, Con 33%, UKIP 10%, Lib Dem 7%, Green 6% Other parties votes were not specifically set. Note that the final forecast percentages differ from the input percentages due to the methodology used (including adjusting for the number of days until the election) - e.g. the final UK target percentages used were Con 33.44%, Lab 32.47%, UKIP 12.56%, LD 9.98%, Green 5.33%.


Party Seats Change
Labour 279 +22
Conservative 271 -32
SNP 44 +38
Liberal Democrat 33 -23
UKIP 0 -2
DUP 8 -
Sinn Fein 5 -
SDLP 3 -
Green 1 -
Plaid Cymru 3 -
Others 3 -3
Labour Short By 46 - Hung Parliament See UK-Elect Latest Forecast for the UK-Elect 'Latest Forecast' page.

UK-Elect Election Forecast Maps
Forecast for Eastern England Forecast for Scotland
Forecast for Wales Forecast for UK
Forecast for London Forecast Gains
Click on image to enlarge. See also more maps.
 
Additional UK-Elect generated maps and screenshots
Forecast Losses Forecast 2nd Place UK - Coloured by most significant 'Swing To' percentage
Screenshot - start of a guided forecast Screenshot - Scottish constituencies Screenshot - configuring gains
Hover cursor over map for more information, click on image to enlarge
Notes: The forecast base was the 2010 General Election, although gains are compared with the current situation (March 2015). The forecasting was made using the UK-Elect v9.3 method, on a separate regional basis for Scotland, Wales, London, and Great Britain, with many other factors taken into consideration, including the number of days until the election, by-election results, local constituency opinion polls, and the changes in the national poll percentages since the by-elections and constituency opinion polls. Additional enhanced support for incumbent parties was not specifically enabled for this forecast, as the UK-Elect v9.3 method includes some built-in support for it. If enabled, it would have had the effect of further raising the support levels for incumbent candidates, particularly for MPs belonging to the smaller parties (Lib Dems, UKIP etc.)
See UK Election Forecasting Theory, Techniques and Controversial Discussions and UK Election Forecasting - A detailed explanation of the techniques used by UK-Elect for more details of UK-Elect forecasting techniques, or if you have a Windows-based PC then try some forecasting yourself using the UK-Elect Trial Version
UK-Elect v9.3 users will be able to mostly reproduce the above forecast (if they are quick!) by doing a "Guided Forecast" and specifying the separate percentages for Scotland, Wales, London and the overall GB percentages. Note that the method used was the UK-Elect v9.3 method with the additional default UK-Elect settings for that method enabled. Incumbency and tactical voting support were disabled, but date-specific adjustments (adjusting the calculation target percentages depending on the number of days until the election) were used for this forecast.
Results from Northern Ireland are based on those of the last election and included for completeness only.Suggestions and Corrections: UK election forecasts are sometimes very controversial. To notify us of any suggested change to this one, or to let us know of any part of it that is just dead wrong, please email us on support@ukelect.co.uk.

No comments:

Post a Comment

Comment is open to all feel free to link to this blog.