Sunday, April 26, 2015

Campaign Week 4: No Consensus Among Pollsters

This election is going to be very difficult to predict.  We have three parties going from practical obscurity in 2010 to national prominence in this campaign, and it's very difficult to tell how spread out or concentrated their newfound support will be.

This alone makes coming up with a seat projection difficult enough.  What makes things even more challenging is the fact that polls are, quite simply, inconsistent with each other.


On aggregate, there's been little change since last week.  The seat ranges for the two main parties are almost fully overlapping each other, meaning it's impossible to say with any degree of certainty which party will finish ahead.
Average of opinion polls released up to Saturday evening
In percentage terms, the difference between the two main parties is about 0.3%.  However, you wouldn't know it if you were looking at the newspaper headlines.  Some polls over the past week have shown a four-point lead for the Conservatives, while others have shown a three-point lead for Labour.  It's hard to tell whether one is right, the other is right, or we should really be splitting the difference between them.

It's not just been the last week where it's been like this.  We have seen consistent house effects amongst the pollsters throughout the course of the campaign.

Poll Average since the prorogation of Parliament

The house effect for a pollster is how they differ from the consensus.  As you can see, there is not even a consensus amongst the pollsters.  Here is how each of them differs from the poll average:
Lord Ashcroft: 22 seats to CON
ComRes: 24 seats to CON
Populus: 24 seats to LAB
Survation: 18 seats to CON
TNS-BMRB: 20 seats to LAB
YouGov: 24 seats to LAB
ICM: 23 seats to CON




























I hear what you're asking: why not look at historical performance to determine which poll to look at?  Well, of the above pollsters, ICM and Populus were joint closest to the actual result in the 2010 election.  However, they too are on opposite ends of the house effect spectrum.

All in all, the only certainty coming in two weeks' time is a hung parliament and an unstable government.  Nobody can say with any certainty what will happen beyond that.

No comments:

Post a Comment