No Magic Number

With 17 weeks to go until the 2015 General Election, people are starting to ask what I think will happen. Let’s be clear, this is not going to be one of those situations where I tell a stranger on a train from Paris what the outcome will be – exactly – more than a month before Election Day. I’m not even sure that I’d be willing to try and call it the night before.

And here’s why: with 650 parliamentary constituencies in the UK, half a dozen parties and surprises lurking everywhere, there just isn't enough data for me on a seat by seat level to make a prediction. And that is all that matters. The US has Nate Silver and his fantastic modelling because he can access data on each level that he needs. But we just don't have that kind of data.

Why? Well, we have a few parliamentary constituencies with under 25,000 voters in them – and the largest only has around 110,000. In contrast, the smallest US House of Representatives district has over 500,000 people in it (note: yes, apples to oranges, but you get the point). Research in these tight quarters is both tricky and expensive. But perhaps most importantly, were we all to regularly survey the voters of Orkney (the second smallest constituency), we’d probably have to speak with each voter there a couple of times between now and election day. It just wouldn't work. The end result of this is that we just do not have the data that is necessary to benchmark, diagnosis and trend watch reliably... At least not the way we watched the US Presidential election unfold poll by poll in 2012.

As a result, we get stuck talking about UK level data - the popularity contest. That is, we are talking about where the parties are as if each voter’s vote had the same weight. They don’t. But we're missing other ways of really watching the numbers unfurl over the weeks to come.

The UK level data can be misleading. The best example right now is, of course, UKIP. UKIP might be climbing in the national wide popularity contest, but they are not making headway in enough distinct constituencies to translate that into a serious number of seats.

Fundamentally, no one needs to win to win. For example, Labour don’t need to win a majority of the popular vote. They don’t even need the largest share of the popular vote. For many predicting the election, they see a narrow Labour majority or a Labour lead coalition. Basically, Labour can keep to the status quo and still quite possibly come out the largest party.

But there’s a long time between now and the general election. Nothing is fixed in this one. Which is, honestly exciting.

[Originally posted on LinkedIn]

If I thought there was no magic number before...