In retrospect the American election seems rather an anti-climax. I hadn’t followed it at all closely – without the special interest of the tussle for the nomination between Obama &Hilary Clinton, followed by the prospect of a first black president, I was less interested in the details of the campaigning, which I lack the background knowledge of American politics to appreciate. With no ability to affect the result through a vote, I was content just to wait & see.
So I wasn’t following American blogs such as FiveThirtyEight this time round, & heard of Nate Silver’s triumph of prediction only after the event. Though this demonstration of the power of the statistical theory of the distribution of sampling proportions & their errors (what journalists have learned to call the margin of error on each individual poll, which, examined one by one, led to the belief that the result was too close to call) I am a little nervous about the idea that this constitutes a ‘magic formula’. Fisher’s dictum about ‘too good to be true’ floated in to my mind, & I wonder what may happen when the ‘only 10% likely’ result turns up in some future election.
On the positive side I hope that, as these ideas are absorbed more widely by election strategists there may be less emphasis on the idea that the modern election is decided by a relatively small number of swing voters in marginal constituencies. That is true only IF the larger, more statistically predictable, members of the electorate play their part & actually turn out to vote, which means they need to be courted too. (Has anyone ever calculated what would be the outcome of a UK election if only the 'targeted voters' bothered to take part?)
Related post
