How good is MRP at understanding what is going on in specific constituencies?
Ben Lauderdale is one of the smarts behind the YouGov MRP model this time and last time.
Here’s what he said on Twitter:
Me: My super-rough rule of thumb is to look at the average sample size in each seat (about 130 in this case); see how many multiples you need to get up to 1,000 (8 in this case); and then concluded that if a factor is present in that many seats then MRP should be able to cope.
Ben: That is generally the calculation I do as well!
The understanding which MRP-based data can give about individual constituencies can be supplemented, of course, by constituency polling.
So basically, Mark, they are working off sample sizes of 130 in each constituency. It would be extremely hard with that model, to ensure you had a reasoned representation by ward, gender, age, likelihood to vote etc. So these figures are VERY rough and not to be trusted, especially as we are 2 weeks out.
Toby – it sounds like you’ve misunderstood how MRP works: as long as the factors in a constituency are ones that apply in multiple seats (e.g. who holds it, are there lots of students, is it in Wales, etc.) then MRP can – if done right, obviously – model accurately the impact of those factors. All the factors you mention are, therefore, exactly the ones that MRP can be extremely good at catering for. There’s a good description of how it works and why that’s the case here: https://www.newscientist.com/article/2224783-what-is-mrp-and-can-it-predict-the-result-of-the-uk-general-election/
It always surprises me that so many people have little interest in politics even at election time. They see politicians on the news programmes or read the headlines in whatever newsrag they buy. the tory party always comes off best because they are well dressed, have learned to construct sentences that sound good, even when they contain no sense whatsoever. I think the main function of private education is to teach the well off children of the Establishment to speak properly so they can shine in debates whist talking rubbish.
MORAL Stop private education or it will continue for generation after generation!
Question. Bearing mind Windrush, May’s Hostile Environment, the chaotic roll out of Universal Credit ( To name but three), how can anyone who professes to uphold any sort of religious beliefs even consider voting Tory?
The other MRP predictions just released are from Best for Britain’s https://tacticalvote.getvoting.org/ site. They have changed their recommendations for tactical voting compared with the start of the campaign as a result and it’s not good news for us, at least in my region, North West England. One point on this one is that I think the sample sizes are smaller than YouGov’s.
In theory, constituency polls should act as tests of MRP, because they are independent samples. However as they are generally taken on different dates from the MRP (which itself is taken over a longer time period than conventional national polls), you have to correct for the movement in the polls over the time period.