I've interviewed one of the heroes of last year's US elections — forecasting expect Nate Silver — for the books blog, but I thought CoffeeHousers might be interested in what he had to say about UK elections. Silver's attempt to predict the 2010 election didn't fare so well (his model significantly underestimated Labour and overestimated the Lib Dems) and he explained the difficulties facing those who try in 2015: the lack of constituency polls, the multi-party system and tactical voting. Here's what he told me:
‘I think you'll have a big increase in the number of poll-driven forecasts here. We need more polling here. If you don't know what's happening in individual constituencies that's a bit tricky. Especially if you have votes drifting off to Ukip and these third and fourth parties, that makes it more complicated than just assuming a uniform swing, potentially. And generally the volatility is higher the more parties are involved in the campaign, because voters can say “I want to vote for Ukip but they're not going to win in my constituency so I'll have to vote for the Tories instead.”’
You can read the full interview here.