Here in the UK we’re about to go to the polls to elect some sort of government in just a few weeks. Nate Silver’s FiveThirtyEight team are naturally on the case in providing their famously accurate election forecasts. They were kind enough to explain again the methodology being used in this blog post by Ben Lauderdale.
Go there and read it in full for the clear and interesting explanation, but in super-quick summary it is starting with their famed method of analysing poll results over time and adjusting for the historic bias each poll has shown vs reality, both in terms of source and time-left-before-election.
What the average poll says now is not the best guess of what will happen in the subsequent election…We can estimate how the relative weight to put on the polls changes as elections approach and use that in our forecast.
But it soon becomes more complex. In their view, due to the increasing influence on the results by parties that nationally have a low share of vote but with high regional variance, applying a uniform swing to the whole country based on national polls doesn’t work.
However, constituency-level polls are not frequent or numerous enough to include in the above. They did manage to get some, but, being relatively sparse, are developing a model around them.
We use a multilevel regression model to describe how vote intention at the constituency level depends on a variety of factors, including region, incumbency, constituency demographics and results in the last election. We then reconcile the constituency-level vote intentions we get from this data with the national-level forecast that we constructed using the national polls, by applying a swing model that we built from the historical record of constituency vote share swings from election to election.
I’m looking forward very much to seeing how it goes, even if I’m not greatly keen on the result they predict today! Follow their predictions here.
Their full description of their model includes a lesson on the importance of phrasing survey questions. Apparently people do not answer “If there was a general election tomorrow, which party would you vote for?” in the same way as “Thinking specifically about your own parliamentary constituency at the next general election and the candidates who are likely to stand for election to Westminster there, which party’s candidate do you think you will vote for in your own constituency”.
Dabbling with Data 