As the attention of American political junkies turn to May 6 general election in the United Kingdom, we want to highlight the polling averages and projections of our colleagues at PoliticsHome. We have created a page that displays the PoliticsHome interactive chart of their polling average and seat projections over time and will, for the next 10 days, display their latest projections in the right column of our site. I have also reproduced both below.
Regular readers will recall that we partnered with PoliticsHome during the fall of 2008 and included a miniature version of their Campaign08 website on Pollster.com. Alas, PoliticsHome has put their US website into mothballs, but they certainly know the politics of the UK. Don't be fooled by the different look and feel of their chart. The underlying methods for poll aggregation and seat projection are quite sophisticated.
PoliticsHome offers much more detail -- and an impressive interactive constituency map -- on their "Poll Centre" page (note: a "constituency" is what we typically call a "district"). There you will find a detailed explanation of their methodology written by Rob Ford, a Hallsworth Research Fellow in the Institute for Social Change at the University of Manchester.
The gist is this: They first aggregate all of the UK national vote preference polls to create the "estimate of current electoral sentiment" plotted in the chart above. Their pooled estimate takes 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").
Next they create a projection of how the national vote preference will likely translate into seats in Parliament won by each party. Like the models of Simon Hix and Nick Vivyan (h/t MonkeyCage) and Nate Silver and his colleagues at FiveThirtyEight, they do not assume a "uniform shift." As Ford explains:
This simplified [uniform shift] 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 deal with this uncertainty problem, their model incorporates calculations of the probability of "that each party will win a given seat" based on historical election data from 2001 and 2005, a process that Ford explains in more detail.
They also deviate from "uniform swing" in two ways:
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.
From my vantage point, what differentiates this model is that its deviations from "uniform swing" are based on hard empirical evidence from regional polls rather than "educated guesses" about cross-party defections. Also, where the PoliticsHome model makes exceptions for individual seats ("to allow for factors that do not fall within the scope of the statistical analysis"), Ford provides clear documentation and explanation for each.
PoliticsHome updates its election projection on a weekly basis, so we have two more updates to go. They posted earlier today on the "state of play" before this afternoon's debate.
Update: While Rob Ford wrote the methodological brief posted on PoliticsHome, I am told that the PoliticsHome model is the work of a team of political scientists that also includes Will Jennings, Mark Pickup and Chris Wlezien.
Follow Mark Blumenthal on Twitter: www.twitter.com/MysteryPollster