Feedback
  • Douglas Alexander | When it comes to the referendum, let’s remember “in unity isstrength”

    • source icon
    • 11:36
  • Mark Wallace | Kelly and Baldry decide to step down – why are MPs really leavingthe Commons?

    • source icon
    • 11:36
  • Political Scrapbook | Bye bye! The (often non-existent) excuses given by quitting Tory MPs

    • source icon
    • 11:01
  • LabourList | Miliband: I want my Cabinet to have 50% women

    • source icon
    • 10:44
  • After careful thought over summer recess & discussions with my family, I have de...

    • source icon
    • 10:35
  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers

  • PoliticsHome | Only the latest five entries on the PhiWire are visible to non-subscribers



Top 5 Political Stories

  • 1
  • Two more Tories to stand down in 2015
  • 2
  • Cameron to outline terror strategy
  • 3
  • Tory MPs mull personal EU quit pledges
  • 4
  • Lib Dems to widen free school meals
  • 5
  • Bercow faces MPs over Clerk choice

Other Top Stories

  • Darling set for referendum police talks
  • Subscription Only
  • CBI backs 'single hub' aviation strategy
  • Subscription Only
  • UK to push Nato allies on defence spending
  • Subscription Only

New this morning

  • Umunna rues Brown effect on Labour
  • Subscription Only
  • US calls to arm Ukrainian troops
  • Subscription Only
  • Lib Dems to set out green pledges
  • Subscription Only
  • Manufacturing growth to slow
  • Subscription Only
  • Conservative website urges migration curbs
  • Subscription Only
  • SAS commander to advise PM
  • Subscription Only
 Politics Home News

Methodology

A new way to project the general election

Back to the poll centre

 

Methodology

 

Rob Ford, Hallsworth Research Fellow in the Institute for Social Change at the University of Manchester, explains the methodology of the Polling Centre projection below. 

At any given point, there are many polls in the public domain, which paint different, often contradictory pictures of public opinion. It is difficult and confusing to sort out the true pattern of public opinion and to separate real change in the views of voters from movements that are a consequence of random sampling error or the result of differences in the methods that different pollsters apply to estimate public opinion.

We aim to extract the signal from the noise by employing statistical techniques that can combine different sources of polling information while acknowledging and accounting for their differences.

1.  Electoral sentiment

The method we use involves two steps. The first step produces an estimate of current electoral sentiment by pooling all the currently available polling data, while taking into account the estimated biases of the individual pollsters (“house effects”), the effects of sample size on the likely accuracy of polls, and the effects of the sampling decisions pollsters make, which mean their samples are not truly random (“design effects”).

To estimate “house effects”, we make use of the 2005 election result—our model treats the 2005 result as a reference point for judging the accuracy of pollsters, and adjusts the poll figures to reflect the estimated biases in the pollsters figures based on this reference point.

The house effects are treated as fixed over the electoral cycle. We have examined whether there is any evidence of significant change in the pattern of house effects over the past five years and find little evidence that the relative pattern of results delivered by the different pollsters have changed over the past five years, with two important exceptions.

Populus and Ipsos-MORI both made major changes to their methodology since 2005: Populus introduced weighting by past vote in March 2007, while Ipsos-MORI switched to telephone polling in April 2008 and also introduced weighting by sector of employment (public sector or private sector).

We incorporate the effects of these changes by estimating two house effects for these polling houses, one before and one after the change. Pollsters such as Angus Reid and Opinium, which only commenced polling after 2005 are judged against the overall pattern of information provided by the longer running pollsters, and adjusted accordingly.

We do not believe these house effects are the result of polling houses adjusting their samples or results to fit their ideological preferences – as their reputations depend on accurate polling this is very unlikely. Rather they are the results of the various decisions polling houses make to attempt to gather a representative sample of voters in a short time frame at reasonable cost.

In general, we find that the “house effects” are relatively small and historical analysis we have conducted suggests that they have become smaller since 1997, with pollsters gradually improving the accuracy of their techniques.

2.  Seat projection

The second step is to project our estimate of electoral sentiment across individual constituencies, to figure out what the balance of power in the House of Commons would look like if the vote shares reflected our current figures. Several websites already estimate this by applying “uniform national swing”: estimating the change in parties’ vote shares implied by current polling, applying this change in every seat and then adding up the seats won by each party under the new shares.

This simplified method gives us a general idea about the state of play politically, but it can be misleading for two reasons. Firstly, it assumes that the “swing” will be the same everywhere, which may not be the case. Secondly, it makes no allowance for uncertainty. If the UNS calculation indicates that the Conservatives will win a seat by 1%, the seat is allocated to them with the same certainty as another seat where the expected margin is 15%.

To improve upon the UNS model we use a method developed by David Firth and John Curtice and used by the BBC in their exit poll based forecasts in 2005. Our method applies the estimated national changes in vote to seat shares, but allows for random variation in the changes that actually occur in individual seats, and systematic variation in the pattern of swing.

We have estimated the degree of random variation in swing using historical electoral data from the 2001 and 2005; using these parameters, our change model estimates the probability that each party will win a given seat.

Our projection of overall seat shares is then a sum of these probabilities. If the Conservatives are very narrowly ahead, they may be assigned a probability of 60% - out of every 10 seats with such a probability, our model expects them to win 6. Where the Conservatives have a very solid expected lead, the probability might be 90%. In this case, the model expects them to win 9 out of every 10 seats.

We incorporate two important deviations from uniform swing. Firstly, there is fairly consistent polling evidence that the political landscape looks different in Scotland to the rest of Britain, perhaps in part due to the change in government at Holyrood in 2007. To allow for this, we employ a separate estimate of Scottish opinion derived from the most recent Scotland specific polls available.

Secondly, a series of polls have indicated that the level of swing from Labour to the Conservatives will be higher in marginal constituencies, a pattern which has also been observed in past elections. To allow for stronger performance in the marginals, we expect an extra two points of swing in seats where Labour hold majorities of between 6% and 14% compared with other seats.

To ensure the total pattern of change sums to what we expect, we also recalibrate the expected swing in other seats to correct for the deviations in Scotland and the marginals. Strong Conservative performance in the marginals is offset by slightly weaker performance elsewhere, and conversely weak Conservative performance in Scotland is offset by slightly stronger performance elsewhere. 

Using this method, we also produce seat by seat estimates of the parties’ chances in various seats. “Safe” seats are those the model suggests the leading party is nearly certain to win or hold, “likely” seats are those the leading party has a better than 75% chance of winning while “lean” seats are those where the battle remains close but our model suggests one party is just ahead, with a 50-75% chance of winning.

3. Adjustments

The projection for a handful of seats is subjectively adjusted to allow for factors that do not fall within the scope of the statistical analysis.

George Galloway will not be standing in Bethnal Green and Bow so this seat is adjusted to be 'likely Labour'.  (Under the statistical method alone it is categorised as 'likely other'.)

Blaenau Gwent has been adjusted to 'likely other' (Under the statistical method alone, it is categorised as 'safe other').  Wyre Forest is projected as 'likely other', but, as with Blaenau Gwent, caution is advised with this categorisation due to the unusual circumstances of the seat. 

Speaker Bercow is projected to win his Buckingham seat, but this is classified as 'other' rather than Conservative, since by convention the Speaker is non partisan and does not vote. 

The projection covers Great Britain only.  The 18 Northern Irish constituencies are excluded from the map and the statistical analysis, since the polls used for the analysis do not cover Northern Ireland. 

Further technical details of the methods we employ are available in the following academic papers:

Pickup, M and Johnston, R (2008) Campaign Trial Heats As Election Forecasts: Measurement Error and Bias in 2004 Presidential Campaign Polls. International Journal of Forecasting 24, 2: 272-284.

Curtice, J and Firth, D (2008) “Exit Polling in a Cold Climate: the BBC-ITV experience in Britain in 2005, Journal of the Royal Statistical Society, Series A: 171(3): 509-539

Back to the poll centre