Are too many public service workers included in samples?
The overwhelming characteristic of British political opinion polling for decades has been the constant over-statement of Labour position in relation to the other parties. This came to a head at the London Mayoral election where the three phone pollsters were producing very different figures from the online firm, YouGov.
To their great credit ICM and Ipsos-MORI did something about it. The former made some adjustments to its approach while the latter immediately announced a review of its methods.
We now have news of the first stage of the MORI changes which were included in the calculations for the national David Davis poll in the Independent on Tuesday and the UNISON poll covered yesterday. Julia Clark, the firm’s head of political polling, has emailed me to say: “Our review is ongoing, but one thing it has uncovered is that our samples have been tending to somewhat over-represent public sector workers. Therefore we will now include a weight for public sector workers moving forward…”
Although I have not got any specific data it’s a reasonable assumption that those who get paid out of the public purse might have a different political perspective than those who don’t. May be this is why polls, not just from MORI, have been pro-Labour.
The question this raises is quite why a disproportionate number of public sectors workers answer the phone and agree to take part in interviews when the pollster calls. Is there something in the nature of the people employed or their working patterns that causes this to happen?
Ipsos-MORI, of course, is the only one of the mainstream national pollsters which does not take steps to ensure a politically balanced sample. This is a controversial technique and involves asking how people voted last time and taking account of these responses when finalising the figures. It will be interesting to see how this affect future surveys from the firm and how the other firms will react.