• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

UK - Election 2015

LondonJohn said:
It's a pan-broadcaster exit poll (which is one of the reasons it's so powerful, since it's an amalgamation of resources).

It looks primarily at the top 100-120 marginal seats, but with lesser surveys conducted in the next 100 or so seats as well. It uses a combination of purely local (i.e. seat-by-seat) absolute analysis and trend analysis to predict the outcome.

The main (real) reason why one of the poll's custodians was unwilling to say where Ukip's two predicted seats would come from is that exit polls never call individual seats. But on top of that, it's likely that there are three seats in which the exit polling shows Ukip as having a decent chance of winning, and an algorithm and trend analysis applied to those raw numbers suggests that two out of the three (but not which two exactly) will be won by Ukip.

And in addition to that, the tiny number of seats being predicted for Ukip means that by definition, margins of error for the 2 Ukip seats could turn out to be disproportionately large (while the overall statistical significance of the poll remains very high).
Click to expand...
Seems the exit poll is on a constituency basis. thanks for playing :D.




and speaking of fails why do I recall them refusing to call the UKIP seats?
 
Last edited:
Giz said:
You mean the left-leaning press such as the BBC and Guardian who didn't want to accept what a liability Ed Miliband was, and therefore under-estimated the extent that voters would jump off the milibandwagon rather than see him as PM?
Click to expand...

Blaming a hostile media is a good way for a party to avoid admitting it simply fielded a bad slate of candidates.
 
Lothian said:
Seems the exit poll is on a constituency basis. thanks for playing :D.
Click to expand...


That's what I said:

It looks primarily at the top 100-120 marginal seats, but with lesser surveys conducted in the next 100 or so seats as well. It uses a combination of purely local (i.e. seat-by-seat) absolute analysis and trend analysis to predict the outcome.
 
LondonJohn said:
That's what I said:

It looks primarily at the top 100-120 marginal seats, but with lesser surveys conducted in the next 100 or so seats as well. It uses a combination of purely local (i.e. seat-by-seat) absolute analysis and trend analysis to predict the outcome.
Click to expand...
We can all play the Cherry picking game

LondonJohn said:
exit polls never call individual seats.
Click to expand...
 
dudalb said:
Blaming a hostile media is a good way for a party to avoid admitting it simply fielded a bad slate of candidates.
Click to expand...


I think in this instance the charge is that the media were not hostile enough* to Ed Miliband at a personal level, meaning that they (the media) tended to overstate Labour's chances.


* Excepting The Sun, which carried on a relentless anti-Ed-Miliband campaign throughout.............
 
Lothian said:
We can all play the Cherry picking game
Click to expand...


*sigh*

When I say "exit polls never call individual seats", I mean that exit polls do not say which party is predicted to win each particular seat.

The analysis that the BBC has just done (on a seat-by-seat basis) is an iteration of the exit poll, by applying the seat analysis indicated by the exit poll to the overall picture of key marginals.
 
angrysoba said:
The IRA? Heaven forfend. Freedom fighters. Loveable rogues, dear chap. Martin McGuinness just liked wearing a balaclava for fancy dress parties, etc...
Click to expand...
Ball masqué, it's called. :)

angrysoba said:
That'll be it. Sick of voting Lib Dem and ending up with a Tory agenda, voters decide to cut out the middleman.
Click to expand...
Yes, and the left-wing of LibDem defected to Labour. At least, that's what the constituency-detailed forecast of a few minutes ago showed.
 
LondonJohn said:
*sigh*

When I say "exit polls never call individual seats", I mean that exit polls do not say which party is predicted to win each particular seat.
Click to expand...
Call me old fashioned but things like 'Labour hold' and 'Conservative gain' with the constituency name underneath the "Exit Poll banner" seemed to say which party is predicted to win each particular seat.
 
Lothian said:
Call me old fashioned but things like 'Labour hold' and 'Conservative gain' with the constituency name underneath the "Exit Poll banner" seemed to say which party is predicted to win each particular seat.
Click to expand...


I'll repeat slowly:

The exit poll did not call the election on a seat-by-seat basis.

The exit poll called the election on an overall seat basis only (even though that overall seat calculation was based significantly upon individual marginal seat polling).

The BBC has taken the exit poll's numbers, and applied them to the key marginals on a proportionate basis. In other words, they apply the net Tory gain implied by the exit poll onto the list of key marginals (taking into account the party from whom the Tories might be hoping to gain the seat), and likewise with Labour, the SNP, Ukip and so on. It's not hard to do that calculation overlay onto the overall seat projection given by the exit poll.

That's how the BBC has come up with its own prediction on how each individual marginal seat will show. It's not the raw output from the exit poll. It's an iteration of the exit poll (ETA: and in fact it's essentially an attempt to reverse-engineer the exit poll into a seat-by-seat output)
 
Last edited:
LondonJohn said:
I'll repeat slowly:

The exit poll did not call the election on a seat-by-seat basis.

The exit poll called the election on an overall seat basis only (even though that overall seat calculation was based significantly upon individual marginal seat polling).

The BBC has taken the exit poll's numbers, and applied them to the key marginals on a proportionate basis. In other words, they apply the net Tory gain implied by the exit poll onto the list of key marginals (taking into account the party from whom the Tories might be hoping to gain the seat), and likewise with Labour, the SNP, Ukip and so on. It's not hard to do that calculation overlay onto the overall seat projection given by the exit poll.

That's how the BBC has come up with its own prediction on how each individual marginal seat will show. It's not the raw output from the exit poll. It's an iteration of the exit poll.
Click to expand...
If that is correct then it would be the marginals with the lowest Labour majorities that that would go blue and the highest would remain red. No Tory seats would go red. That is not what was shown. The seats changing were dotted around.
 
Another 10 or so results, and I'll happily slide off to bed safe in the knowledge of the overall outcome.
 
Lothian said:
If that is correct then it would be the marginals with the lowest Labour majorities that that would go blue and the highest would remain red. No Tory seats would go red. That is not what was shown. The seats changing were dotted around.
Click to expand...


That's because it's not as simple as that. Firstly, net gains do not equal gross gains. They equal gross gains minus gross losses.

And secondly, the actual percentage spreads by party from 2010 make a significant difference. For example, imagine two constituencies, A and B. The 2010 results were as follows:

A:
Lab: 48%
Con: 47%
LibDem: 5%

B:
Lab: 39%
Con: 37%
LibDem: 24%

Even though A was a more marginal constituency in 2010, statistics (and common sense!) suggest that the Conservatives were more likely to win B than A in 2015 (by taking a proportion of the larger LibDem vote in that seat).

That's how this sort of thing works. I'll repeat one last time: the exit poll did not give output on a seat-by-seat basis. The BBC have worked that out for themselves in line with the exit poll predictions.
 
Putney's a Con hold with a 1.7% increase in the Tory vote.
 
LondonJohn said:
That's because it's not as simple as that. Firstly, net gains do not equal gross gains. They equal gross gains minus gross losses.

And secondly, the actual percentage spreads by party from 2010 make a significant difference. For example, imagine two constituencies, A and B. The 2010 results were as follows:

A:
Lab: 48%
Con: 47%
LibDem: 5%

B:
Lab: 39%
Con: 37%
LibDem: 24%

Even though A was a more marginal constituency in 2010, statistics (and common sense!) suggest that the Conservatives were more likely to win B than A in 2015 (by taking a proportion of the larger LibDem vote in that seat).

That's how this sort of thing works. I'll repeat one last time: the exit poll did not give output on a seat-by-seat basis. The BBC have worked that out for themselves in line with the exit poll predictions.
Click to expand...
Thanks for confirming, as I said, that the Exit poll 'predicted number of seats' is calculated on a seat by seat basis.
 
Look at the UKIP candidate in Tooting: De Skuba Skwirszynski. There's another filthy Polish immigrant trying to steal English jobs and live off the English taxpayer!
 
ddt said:
Look at the UKIP candidate in Tooting: De Skuba Skwirszynski. There's another filthy Polish immigrant trying to steal English jobs and live off the English taxpayer!
Click to expand...
There have been plenty of interviews that I've heard with British folk of Pakistani heritage (i.e. with Pakistani parents) who are voting UKIP, because the country is full and immigration has got to stop. I just don't know where to begin.
 
You must log in or register to reply here.

Back
Top Bottom