Here are some further observations about the British election, based on further analysis over the weekend.

3. How did the Conservatives win so many marginal seats (continued)?

In my earlier blogpost I wrote “a look at the election returns as they came in has convinced me (though I am open to persuasion based on a closer review of the numbers) that United Kingdom Independence Party numbers held up well in safe Labour and perhaps safe Conservative districts, but tended to go over to Conservatives in marginal seats.” I’ve now gone over the numbers and have become convinced I was wrong. My analysis is confined to England and Wales, since Northern Ireland has its own parties and the Scottish Nationalists won 56 of the 59 seats in Scotland, leaving one each for Labour, Conservatives and Liberal Democrats. What I found was that in seriously contested seats the Ukip percentage was, on average, almost exactly the same as in the United Kingdom as a whole, as was the Labour percentage, while the Conservative and Lib Dem averages were significantly higher than in the United Kingdom as a whole (though the Conservative percentage was about equal to their percentage in England and Wales). I class as seriously contested 123 districts identified by the UK Polling Report as electoral battlegrounds, plus six Labour-held seats not on their list which were won by Conservatives (Bolton West, Vale of Clwyd, Plymouth Moor View, Derby North, Telford and Morley & Outwood, which saw the biggest upset of the night, the defeat of shadow chancellor Ed Balls). UK Polling Report’s selection of electoral battlegrounds was based on the pre-election polling, which of course proved to be wrong; you can see the projections for each seat in the UK on the Electoral Forecast website, The projections were that of these 129 seats, Conservatives would win 48, Labour 53, Lib Dems 27 and Ukip 1 (the seat of the former Conservative MP Douglas Carswell, who resigned his seat after switching parties and won it in a by-election). The actual results were very different: Conservatives won 99 of these seats, Labour 22, Lib Dems 7 and Ukip the correctly predicted 1. If the projections based on pre-election polling were right, Conservatives would have won 280 seats, Labour 261, Lib Dems 27 and Ukip 1. The actual count was Conservatives an absolute majority of 331, Labour 232, Lib Dems 7 and Ukip 1. The results in these 129 seats taken together made the difference between a governing majority and a hung parliament in which neither of the major parties, even with its obvious coalition partner (Lib Dems for Conservatives, Scot Nats for Labour, though explicitly), would have had a majority of 323 (the speaker and the four Northern Irish Sinn Fein members don’t vote for parties). A big, big difference.