Populus Online

[Jack]

[manchesterman]

Contrast with YouGov poll showing Lab 4 pts behind. To me it all points to the fact that, whatever happens in politics nowadays, nothing is shifting the polls and the 2 main parties are effectively neck and neck!

[Deleted]

Dec 17, 2018 8:29:36 GMT @barnabymarder said:

YouGov has tended to be a bit out of line with the others for quite some time now but if anything the gap has grown. Other polls tend to show a Labour lead of between 1 & 4 per cent.

to the point they are the only pollster with a Tory lead since October

[John Chanin]

Dec 17, 2018 9:35:42 GMT @priceofdawn said:

Dec 17, 2018 8:29:36 GMT @barnabymarder said:

YouGov has tended to be a bit out of line with the others for quite some time now but if anything the gap has grown. Other polls tend to show a Labour lead of between 1 & 4 per cent.

to the point they are the only pollster with a Tory lead since October

I'm surprised we haven't had an article from Anthony Wells on UK Polling Report addressing this inconsistency, and the explanations for it.  Clearly YouGov have confidence in their method.  I suspect it's about turnout modelling.  Political opinion polling has got itself into a tangle by trying to predict election results rather than testing public opinion.  It does this by disregarding many opinions on the grounds that they are expressed by people who won't vote.  This seems to me to betray the purpose of opinion polling (the clue is in the name), and persistent claims from polling companies that they are not in the business of prediction, but only taking a snapshot of opinion.

There is a middle way, which is sometimes used, of asking people how strongly they hold their opinions, which YouGov often does with issue related questions.  But this promotes understanding rather than headlines, and that will never do.

[Deleted]

Dec 17, 2018 9:47:34 GMT John Chanin said:

Dec 17, 2018 9:35:42 GMT @priceofdawn said:

to the point they are the only pollster with a Tory lead since October

I'm surprised we haven't had an article from Anthony Wells on UK Polling Report addressing this inconsistency, and the explanations for it.  Clearly YouGov have confidence in their method.  I suspect it's about turnout modelling.  Political opinion polling has got itself into a tangle by trying to predict election results rather than testing public opinion.  It does this by disregarding many opinions on the grounds that they are expressed by people who won't vote.  This seems to me to betray the purpose of opinion polling (the clue is in the name), and persistent claims from polling companies that they are not in the business of prediction, but only taking a snapshot of opinion.

There is a middle way, which is sometimes used, of asking people how strongly they hold their opinions, which YouGov often does with issue related questions.  But this promotes understanding rather than headlines, and that will never do.

generally speaking YouGovs methods arent hugely different from anyone else. My guess would be sampling

[The Bishop]

As far as *this* poll is concerned, isn't this the first VI survey from Populus since the 2015 GE?

[thetop]

Dec 17, 2018 10:02:08 GMT @barnabymarder said:

I think they still tend to assume that Tory voters are more likely to vote by a significant margin, although that seems not to have been the case last year at all. Please tell me if this assumption is wrong.

Of course just because YouGov are different doesn't mean that they're wrong necessarily, though some will have their suspicions.

BES say so: www.bbc.co.uk/news/uk-politics-42747342 (using Tory voters and older people as interchangeable, because they increasingly are).

[The Bishop]

Though more recently, we have had new evidence suggesting there might have been a bit of a "youthquake" after all.

[thetop]

Dec 17, 2018 10:12:33 GMT The Bishop said:

Though more recently, we have had new evidence suggesting there might have been a bit of a "youthquake" after all.

Link?

[Deleted]

Dec 17, 2018 10:09:30 GMT thetop said:

Dec 17, 2018 10:02:08 GMT @barnabymarder said:

I think they still tend to assume that Tory voters are more likely to vote by a significant margin, although that seems not to have been the case last year at all. Please tell me if this assumption is wrong.

Of course just because YouGov are different doesn't mean that they're wrong necessarily, though some will have their suspicions.

BES say so: www.bbc.co.uk/news/uk-politics-42747342 (using Tory voters and older people as interchangeable, because they increasingly are).

last year the number of over 65s voting dropped marginally. As most 65+ vote Tory this hurt their voter turnout most.

Generally speaking 80% of leave voters which tend to be 65+ and Tory are committed to voting in the next election which is higher than the average.

One thing I've noticed is YouGov redistribute a larger number of don't knows.

[John Chanin]

Dec 17, 2018 10:27:27 GMT @priceofdawn said:

Dec 17, 2018 10:09:30 GMT thetop said:

BES say so: www.bbc.co.uk/news/uk-politics-42747342 (using Tory voters and older people as interchangeable, because they increasingly are).

last year the number of over 65s voting dropped marginally. As most 65+ vote Tory this hurt their voter turnout most.

Generally speaking 80% of leave voters which tend to be 65+ and Tory are committed to voting in the next election which is higher than the average.

One thing I've noticed is YouGov redistribute a larger number of don't knows.

Ah yes, this has been addressed.  YouGov made a big push to get people onto their panel who don't normally vote or respond to surveys.  This was in response to a previous BES finding.  Presumably they would therefore argue they are picking up a more representative sample of young people who don't know and don't care, whereas other pollsters get young people who are interested.  This is reflected in a higher percentage of don't knows. 

[The Bishop]

Dec 17, 2018 10:12:52 GMT thetop said:

Dec 17, 2018 10:12:33 GMT The Bishop said:

Though more recently, we have had new evidence suggesting there might have been a bit of a "youthquake" after all.

Link?

Sorry not sure, just something I saw referenced on Twitter.

I do recall thinking at the time that it was rather a small sample (of younger voters) for the BES to make such definitive conclusions about.

[Deleted]

Dec 17, 2018 11:41:43 GMT John Chanin said:

Dec 17, 2018 10:27:27 GMT @priceofdawn said:

last year the number of over 65s voting dropped marginally. As most 65+ vote Tory this hurt their voter turnout most.

Generally speaking 80% of leave voters which tend to be 65+ and Tory are committed to voting in the next election which is higher than the average.

One thing I've noticed is YouGov redistribute a larger number of don't knows.

Ah yes, this has been addressed.  YouGov made a big push to get people onto their panel who don't normally vote or respond to surveys.  This was in response to a previous BES finding.  Presumably they would therefore argue they are picking up a more representative sample of young people who don't know and don't care, whereas other pollsters get young people who are interested.  This is reflected in a higher percentage of don't knows. 

this is true and there last minute change in redistributing dont knows is why they got the GE wrong had they stuck with their methodology theyd have come closer

[seanf]

It turns out that Populus did not poll voting intention. These were unweighted cross breaks on other polling questions.

WRT polls generally, there's not much difference between them. Con and Lab are 38% each, or thereabouts.

[Deleted]

If anyone is intetested id recommend 326 politics on twitter. They do their own weighting so they take the raw data and map it themselves

[Deleted]

[The Bishop]

I think its in the wrong thread as well.....

[Tony Otim]

On UNS it would have the Lib Dems gaining Fife NE and holding their other 4, Labour holding Edinburgh South, the Tories holding Aberdeenshire West, BRS and DCT and the SNP taking the other 50 seats. (Boris wouldn't make any difference to seats)...