So the pollsters’ snapshots have been snapped for the last time. Here are the final numbers from the 11 houses:
The most obvious thing is how close the race appears, and the second most obvious thing is how similar the toplines look. Ten of the 11 pollsters have the Conservative-Labour spread within a point of even, while the straight average across all of them has the Tories 0.1 points ahead (due to rounding, both parties appear on 33.5%). All are within a point of 9% for the Lib Dems and 5% for the Greens. Where they differ more is on UKIP (11-16%, average 12.8%) and correspondingly on the combined Labour-Conservative vote (64-71%, average 67%). So for all the talk of a breakdown of the two-party system, the polls say that the two largest parties will have about the same share of the mainland vote as they did in 2010.
In terms of the debate between phone and online polling, the debate could well rumble on until the next election, simply because their final results are so similar. The modes are within 0.2 points of each other for the lead, 0.1 of each other for the Lib Dems an 0.2 for the Greens. Even with the very large sample sizes in final call polls, the differences are well within the margins of error. On UKIP the phone polls have 11.3%, the internet polls 13.7%, but with considerable variation amongst the online surveys. Even then, the variation is smaller than it’s been for much of the last few years.
So with partial exception of UKIP, they’re either all right or all wrong.
In Scotland, we got some new Westminster polls – You can read John Curtice’s roundup here. His poll of polls shows:
My piece from yesterday about the risks to polling error does seem to have provoked a bit of debate. One paragraph that a few people may have closed over was the following (on the first page):
Furthermore, the so-called shy Tory factor is really a statistical pattern, for which shyness or dishonesty are merely possible explanations. As will be discussed, there is evidence for the various theories as to its causes, but we don’t know, and may not know, how big their impacts are (or were) individually. We just know that polls have shown evidence of bias. Thus the terms “shy Tory factor” and “shy Tory effect” are used in a far broader sense than their literal meanings.
In other words, in the context of the piece, I’m talking about risk (from the pollsters’ perspective) that the Conservatives are ahead, rather than level, but where the differences actually arise is another question. As I explain on page 5:
Political weighting in particular has generally worked well in a stable political system, by ensuring that a weighted sample contains – in addition to all the demographic proportions – representative numbers of those who identify with each party, or voted for it at the last election, typically with an adjustment for false recall (that is, people reporting their past vote incorrectly). But more complex patterns are problematic – for example, a pollster can ensure that a sample contains the correct number of 2010 Lib Dems. But what if (as is likely) 2010 Lib Dems defecting to one party are very different from those going to another, or to the “don’t know” column, and have differential response rates?
The same problem potentially arises in reverse with UKIP – the party has taken support from most other parties, but if each party’s defectors to UKIP behave differentially to its loyal 2010 voters, UKIP support may be over- or underestimated. Worse, as far as the big picture is concerned, if the defector versus loyalist behavioural differences are themselves different between parties (in other words, the composition of UKIP support is mismeasured), they directly impact on estimates of support for other parties.
Hopefully that clears things up.
And so, the time has come to make a prediction. As per my piece yesterday, I feel that the risks of polling error are strongly weighted towards a Conservative lead being underestimated. But for the reasons that I gave, I also have to build in a chance that the patterns have broken down and that the polls are right after all. Here are the numbers, using my differential swing model:
So the main difference from other forecasts is a clear lead for the Conservatives, for obvious reasons. Realistically, I’d hope to be within 20 seats on the two main parties and a handful of seats on the others. We’ll find out soon enough.
NCP would like to wish all followers who standing in or otherwise involved with tomorrow’s general and local elections the very best of luck. And the same to everyone else making snapshots and/or predictions!