Political polling

[bi0boy]

The mean scores from 11 final polls from all the pollsters are:

Con: 33.727
Lab 33.727





_________Con Lab
Ipsos phone 36 35
Opinium online 35 34
COmRes online 35 34
TNS online 33 32
BMG online 34 34
YouGov online 34 34
Survation online 33 33
Ashcroft phone 33 33
ICM phone 34 35
Populus online 33 34
Panelbase online 31 33
AVERAGE 33.72727 33.72727

 

[Dogsauce]

As long as I've known Labour have done worse than expected leading to a lot of disappointment in my younger more tribal self. Do the polling numbers from previous elections back this up? I think they did better last time but it was still a defeat.

 

[belboid]

I'm tempted to bet on UKIP getting over 150 second places. Is it worth it?

hmm, they wont get any in scotland, only one or two tops in wales, which leaves them needing them all from Englands 550. Over 150 of those are marginals where they wont come close, so they'd be needing to get second in almost half the remaining seats. I don't see it, myself. 75-90 would be my guess.

 

[Brixton Hatter]

oh god, exit poll on C4 predicting:

Con 316
Lab 239
SNP 58
LD 10
UKIP 2
Other 25

 

[goldenecitrone]

A quick summary of the thread. From the sublime

Today's YouGov:-

CON 29% (31), LAB 42% (40), LD 11% (12), UKIP 12% (12); Govt app -40

Fieldwork 4th & 5th March.

Mods, might it be a good idea to sticky a poll thread for the run in to 2015? Just a thought.

to the ridiculous.

oh god, exit poll on C4 predicting:

Con 316
Lab 239
SNP 58
LD 10
UKIP 2
Other 25

 

[Brixton Hatter]

A quick summary of the thread. From the sublime



to the ridiculous.

Yeah WTF happened?

The pollsters will probably have some questions to answer...

 

[brogdale]

Yeah WTF happened?

The pollsters will probably have some questions to answer...

They certainly will, but those first numbers were from March 2013 tbf

 

[Brixton Hatter]

good point, well made

 

[Maurice Picarda]

If we can't trust the pollsters on General Election voting intentions, how can we trust them on Europe? Currently there's supposed to be a reasonable margin protecting us from a Brexit in 2017. But the scale of this polling disaster - everyone, from online panels to phone interviewers, presumably calibrating against each other - suggests that all bets on Europe are off.

 

[brogdale]

If we can't trust the pollsters on General Election voting intentions, how can we trust them on Europe? Currently there's supposed to be a reasonable margin protecting us from a Brexit in 2017. But the scale of this polling disaster - everyone, from online panels to phone interviewers, presumably calibrating against each other - suggests that all bets on Europe are off.

A Y/N question is considerably easier to poll with more accuracy tbh

 

[Maurice Picarda]

A Y/N question is considerably easier to poll with more accuracy tbh

Maybe. One takeaway from this evening is that the impact of newspapers may have been underestimated. Miliband tried to be the first party leader to cock a snook at Murdoch and still win. The viciousness of the press against Labour in the last couple of days was extraordinary, but we all hoped that their influence had waned with readership. Now, we have to double check what UK media owners have to gain or lose from Brexit.

 

[belboid]

A Y/N question is considerably easier to poll with more accuracy tbh

like in scotland? out by a similar margin, which can produce big shocks

 

[kabbes]

Back to the drawing board for the polling companies. An utter disaster for them.

 

[Supine]

I think there will be some unemployed ex pollsters kicking around soon.

No doubt they will get jobs in banks

 

[Santino]

Back to the drawing board for the polling companies. An utter disaster for them.

They must have each been using all the other polling companies to confirm the accuracy of their models.

 

[killer b]

someone mentioned 'clumping' somewhere upthread. looks like they were right.

 

[butchersapron]

Survavtion bottled out of publishing last minute poll that found:

CON 37%
LAB 31%
LD 10
UKIP 11
GRE 5

 

[brogdale]

Survavtion bottled out of publishing last minute poll that found:

CON 37%
LAB 31%
LD 10
UKIP 11
GRE 5

Bottling doesn't do that justice.

 

[bi0boy]

Survavtion bottled out of publishing last minute poll that found:

CON 37%
LAB 31%
LD 10
UKIP 11
GRE 5

It was certainly a sign of something wrong when the polls started converging and there was a suspicious lack of outliers.

 

[brogdale]

It was certainly a sign of something wrong when the polls started converging and there was a suspicious lack of outliers.

What was wrong was to fail to spot the cross-over point and subsequent overtake.

 

[Fez909]

Even if it was an outlier, why didn't they publish? You don't poll to see the results you want. You poll to find out what's going on.

It's for the analysts to decide whether to ignore outliers or whatever, not pollsters.

How many other "outliers" were not published?

 

[bi0boy]

Even if it was an outlier, why didn't they publish? You don't poll to see the results you want. You poll to find out what's going on.

It's for the analysts to decide whether to ignore outliers or whatever, not pollsters.

How many other "outliers" were not published?

It's a well known phenomenon, the omission of "obviously wrong" outliers. The expected number of outliers can be calculated, so it's usually easy to spot.

 

[butchersapron]

British polling council statement.

The final opinion polls before the election were clearly not as accurate as we would like, and the fact that all the pollsters underestimated the Conservative lead over Labour suggests that the methods that were used should be subject to careful, independent investigation.

The British Polling Council, supported by the Market Research Society, is therefore setting up an independent enquiry to look into the possible causes of this apparent bias, and to make recommendations for future polling.

 

[Dogsauce]

I wonder how much it influenced people's voting behaviour? How many thought Labour wouldn't lose a seat they gained last time and voted green?

 

[J Ed]

I wonder how much it influenced people's voting behaviour? How many thought Labour wouldn't lose a seat they gained last time and voted green?

Not enough to matter

 

[butchersapron]

I wonder how much it influenced people's voting behaviour? How many thought Labour wouldn't lose a seat they gained last time and voted green?

I would guess at fuck all.

 

[Roadkill]

The lib dems will spin it as a win. "They said we'd only get one, but we took Shetland AND Orkney!"

Just remembered this post. Prescient, perhaps...

 

[Sasaferrato]

Just remembered this post. Prescient, perhaps...

What a strange election. the pollsters got it incredibly wrong. The only one that was on the nose was the BBC exit poll, unbelievable as it was at that point.

 

[belboid]

What a strange election. the pollsters got it incredibly wrong. The only one that was on the nose was the BBC exit poll, unbelievable as it was at that point.

Except that was wrong too. Not generous enough to the tories

 

[butchersapron]

Except that was wrong too. Not generous enough to the tories

Indeed, was only right in terms of direction of march.