In some ways the polls are all over the shop, and yet in other ways they are beginning to coalesce. But first off, they are now moving at least, just not quite how everyone expected.
Moving polls
History shows us that usually the polls narrow during the actual campaign and typically towards the end of it. But if one takes a look at the Labour poll lead over the Tories, it’s been pretty much static. However, if we get under the bonnet of the numbers a little more, we can see the polls have moved around quite a bit. The leader, Labour, has drooped by around three points, which is fairly normal. But the Tories have slipped by four points. Normally the leader goes down and the underdog goes up, so that’s unusual but not perhaps unexpected in the situation the Tories find themselves. Who has benefitted? Principally Reform with the Lib Dems moving up a couple of points too but still below where they usually end up in General Elections. The net result? Labour’s poll lead of around 21 points hasn’t budged.
But the political parties have in fact moved around by seven points, which is significant. And with Labour needing to be 12.5 points ahead in the polls to achieve a one seat majority, in a normal election, this would have meant that potentially the polls by now would have moved uncomfortably closer to that number for Labour. The late rise of Reform is screwing with polling normality.
Erratic polls
But something else is going on too. The polls have been somewhat erratic. On the same day this past week two pollsters gave the Labour lead at 21 points whilst predicted it as only 11 points. On another day, two MRP polls, the alleged but controversial new gold standard, told us that Reform would win zero seats and also 30 seats. You read that right: 30 seats! In Clacton alone, they told us that Farage wouldn’t win and at the same time would not only win but achieve the largest ever swing in UK election history. That’s just weird and way outside any standard margin of error.
Herding
We are also perhaps seeing the signs of pollster ‘herding’, where the pollsters play with their models to end up at the overall average so they can’t be more wrong than their competitors on election day. The current poll of polls average is a Labour lead of 21 points. The group of pollsters that have been above the poll of polls average have moved from 25 at the start of the campaign to now 22. Those that were below the poll of polls average have moved from a 16 point Labour lead (and they even had some early polls at 12-14 points previously) up to an average of a 19 point lead. Hmmm…slightly suspicious. Watch this space. Of course, this could just be the unusually large percentage of ‘don’t knows’ finally reporting they’ve made their minds up. But anecdotal reports from all parties’ ground campaigns show that the percentage of ‘don’t knows’ is still very high, around 25% or more.
Difficult polling history
The 2015 election result was not what the pollsters had predicted; Cameron won a majority. It was a shock to the whole polling industry which reacted by engaging a team of outside experts to look at what went wrong. In the end, the pollsters made the changes recommended by the inquiry, plus others of their own covering sampling, questioning and weighting approaches. As it turned out, at the following 2017 General Election, had they not done this and instead carried on doing exactly as they had been in 2015, then the 2017 polls would have been more accurate! Go figure.
As we keep saying, polling is a black art not an exact science. Let’s see where we end up.
Comments