Why the Tories Won

antrobus

bigheadxx wrote: »

There is a big deal being made about the exit poll but in 2010 they did not release an exit poll when polls closes.

Err, yes they did. At exactly 10.00pm as well. Just as they did in 2005 as well. (And before that as it happens.)

This was the 2010 exit poll prediction;

Conservative 307
Labour 255
Liberal Democrat 59
Others 29

People thought that one was 'wrong' as well. Because it was showing a seat loss for the Liberal Democrats, when the polls had been showing a big increase in their vote.

http://www2.warwick.ac.uk/fac/sci/statistics/staff/academic-research/firth/exit-poll-explainer/

onlyroz wrote: »

But why isn't there a precedent for exit polls being so different to pre-election polls?

See above.

antrobus

gadgetmind wrote: »

Lots more people in work so you're having to extrapolate the views of a whole load of workers from the few you manage to contact?

Possibly?

Early on, there was a difference between telephone polling (showing Con leads) and online polling (showing Lab leads) but that difference seems to have largely gone by the time the final polls were issued.

Online polling is (of course) characterised by the fact that people sign up for the specific purpose of making a few quid. I know that the likes of YouGov are perfectly aware of the fact that this means that their actual user base is somewhat 'unrepresentative', and take account of this in terms of constructing a survey panel.

However it has occured to me that if (as is likely) an employed professional on £45k a year gets more survey invitations and so makes more money, than (say) an unemployed factory worker on benefits, then there is a financial incentive for the former to pretend that they are in fact the latter.

Since the internet is a compartively 'new thing', it might well take time for people to discover that this is the case, so you might start off in a situation where online polling appears to be accurate, but gradually bias seeps in over time, as your survey base corrupts itself.

But that can't be true, because the telephone polls were just as wrong. Unless those pollsters were panicked into conforming with the herd.

catshark88

You Gov is probably entirely populated by MSE users from the "make a bit extra to pay debt/ mortgage" boards!

PS This is a joke and I am in no way suggesting that anyone on this forum would pretend to be anything that they are not.

gadgetmind

antrobus wrote: »

an employed professional on £45k a year gets more survey invitations

Maybe, but in most technical fields, that's the salary for degree + 5 years, and for no other reason than supply and demand.

an unemployed factory worker on benefits

If they are unemployed, then they are not a factory worker, and they have flexibility regards future career. If they get some study in - using their copious free time - then they could join the cohort discussed above, but they probably won't bother.

zagfles

antrobus wrote: »

Are you sure about that?

The Con vote increased in;

Taunton Deane, Yeovil, Bath, Cheltenham, Chippenham, North Cornwall, and Wells

(I only checked those 7. Then I stopped. That must be a big enough sample.)

Oops looks like you're right - I did check a couple of Cornish seats where the Tory vote had either gone down or stayed about the same and they still won the seat off the LDs. But it seems in other the Tory vote increased substantially.

I can understand people switching from the LDs to the Tories to keep Labour out in 3-way marginals, but not in SW seats where Labour have no chance whatsoever.

bigheadxx

zagfles wrote: »

Oops looks like you're right - I did check a couple of Cornish seats where the Tory vote had either gone down or stayed about the same and they still won the seat off the LDs. But it seems in other the Tory vote increased substantially.

I can understand people switching from the LDs to the Tories to keep Labour out in 3-way marginals, but not in SW seats where Labour have no chance whatsoever.

St Ives was the last to declare and the Tory swing from Lib Dem was substantial.

zagfles

chewmylegoff wrote: »

It may sound ridiculous to you but if you look at the results in individual constituencies that seems to be what happened. There are also some examples of UKIP polling reasonably well in former LD seats like Berwick and Brecon. Tories gained significantly in those seats as well, presumably what really happened is LD lost votes to Tory and Tory to UKIP.

Lib Dems had a wide and varied support base and not all of them were "anti Tory" by any means. Tory leaning LD voters would have been able to see the writing on the wall for the LDs in this election and voted accordingly.

I know they aren't all anti-Tory, I was just assuming the votes they lost were mainly from those that are. Those who aren't ought to be pleased with the LDs I'd have thought.

I can understand the point about people switching to keep Labour out, and that would make sense in the sort of 3 way marginals we have round here in the NW, but it makes no sense in places like Yeovil where Labour have no chance whatsoever.

zagfles

bigheadxx wrote: »

St Ives was the last to declare and the Tory swing from Lib Dem was substantial.

Eh? The Tory vote went down in St Ives!

http://www.bbc.co.uk/news/politics/constituencies/E14000964

Generali

zagfles wrote: »

In a national election? Isn't that illegal?

I voted twice in a local election legally.

Labour reformed the Corporation of London and, as many wards didn't have any voters, introduced giving votes to the companies that were in the wards. I exercised one of those votes as well as voting in my local area.

Generali

stator wrote: »

You don't honestly believe that one anarchist's graffiti represents an entire political group.

Of course not. It's what we economists call an 'example'.