Urban75 Home About Offline BrixtonBuzz Contact

Jeremy Corbyn's time is up

The old methodology basically meant pollster's were always fighting yesterday's battles. They would apply adjustments to a new poll based on their estimates of why polls at the last election were not the same as the last election result. As such, if the reasons for the difference between the polling and a true result shifted, their adjustments would become inappropriate. Since those reasons are multifaceted, self-contradictory, chaotic and highly prone to drift, it was very hard to get the adjustments right.

YouGov's new methodology does not work in the same way. It is more akin to pricing motor insurance. They identify all the independent factors they can that help define voting intention (e.g. age, sex, income). They then use current polling to calculate the effect of each of the independent factors. Finally, they split the demographics in each area by the independent factors and apply their modelled factor effects to estimate the vote.

This doesn't rely on crude adjustments, because it is directly modelling the effects that cause the need to make such adjustments. It's much more reliable. The downside is that it needs a lot of data to make it work. Motor insurers, who use similar models, typically have millions of data points. However, YouGov seem to be able to get reliable voting predictions based on "just" 50,000 data points. But this is still a big increase from the old methods of asking more like 1000 people.

On the plus side, however, they can use every single person who will answer questions, because corrections for biases are built into the modelling methodology. So they don't need to throw away data, like current methods need to.
ah, now THAT is fascinating - I always knew there would - some day - be a real dividend from having a statistician in the house! :p
But do YouGov only conduct Internet polls, or do they do telephone and face to face?
 
I don't think that's true, there's loads of things sustaining them - one of which is clearly fear of a Corbyn government.
agreed. plus, the only thing that is stopping them from Dumping May is that she is being fingered to fully cop the blame when Brexit goes massively tits up.
And, if a mole of mine is on the button, the 'stop Boris' campaign isn't quite ready yet.
 
yes I thought so. won't that give their polls an anti - wrinkly bias?
If th bye got their new method right, it won’t matter, because data will only be used to parameterise the parts of the model it is relevant for, not as surrogate data for the parts it is not.
 
Have to say I was quite surprised at the extent of anti-Corbyn comments over on the 'Corbyn's anti-war position' thread.

If he's poorly thought of on these boards surely his time is up....
 
Have to say I was quite surprised at the extent of anti-Corbyn comments over on the 'Corbyn's anti-war position' thread.

If he's poorly thought of on these boards surely his time is up....
Oh, I don't think his time is up. It clearly isn't, whatever I think of his approach to Syria.
He's never been more secure. Most of the criticisms of his approach to foreign policy in that and this thread are coming from outside the party. People like me who have no say or effect on his position in the party.
 
But if the perception within the party is that he is out of step with the broader public opinion then they'd start stoking the fires of discontent again....wouldn't they.
 
But if the perception within the party is that he is out of step with the broader public opinion then they'd start stoking the fires of discontent again....wouldn't they.
they'll do that whatever because they don't like him, the red tories. there's a whiff of smoke even if his position's secure. because they're getting ready for the next time he seems vulnerable.
 
But if the perception within the party is that he is out of step with the broader public opinion then they'd start stoking the fires of discontent again....wouldn't they.
Is that the perception within the party? If so that perception is wrong, support for action against Syria varys with the the of action proposed and the wording of the question but there's a pretty solid block of the public that are opposed to action of any sort.
Looking at the situation overall, headline voting intention polls continue to show Labour and Conservative neck-and-neck on average. On Syria, differently worded questions produced results that vary from clear opposition to just slightly more opposition than support, but it’s clear the public did not whole-heartedly support military action in Syria.
While there's no breakdown of the vote by party support I'd be surprised it wasn't a very strong correlation between Lab voters and opposition to action.
 
Anti-Semitism debate in the Commons, with Corbyn coming under fire from all sides. Heavily noted by observers that he did not stay to hear it all, though to be fair it would be truly shocking if he hadn't heard it all by now.

The social media aspect is very frustrating. I have not seen any analysis of the anti-Semitic tweets and posts that indicates how many are directly attributable to individuals who can therefore be identified as members or not and dealt with. Equally much of it must be impossible to trace to individuals but nevertheless a research company could have something to say about its origins. I don't know how well this was covered by Shami in her report though.
 
Back
Top Bottom