Opinion Polls

New Topic
This topic has been archived, and won't accept reply postings.
 The New NickB 16 Sep 2019

I’m usually a big advocate for the polling companies, within the published degrees of accuracy, they rarely get it wrong. Obviously  when votes are very close and can go either way, they are going to struggle.

However, the voting intension polls seem to be all over the place at the moment, swinging 5 or 6 points between polls, which are more or less daily. Is our politics really this volatile or is there something not quite right with some of the methodology.

 stevieb 16 Sep 2019
In reply to The New NickB:

“Opinion polls are a device for influencing public opinion, not a device for measuring it. Crack that,and it all makes sense.“  - Hitchens, I think Peter for once rather than Christopher. 

Election polls are affected by which parties are prompted for. Some polls will not prompt for the greens, or other smaller parties. Also, the question and any previous questions will affect the answer. 

MarkJH 16 Sep 2019
In reply to The New NickB:

> However, the voting intension polls seem to be all over the place at the moment, swinging 5 or 6 points between polls, which are more or less daily. Is our politics really this volatile or is there something not quite right with some of the methodology.

If you look at particular pollsters, then the results look much more consistent.  All the pollsters will use different corrections to try and eliminate known biases in their samples.  It looks like these differences (rather than the poll results) are the source of the discrepancies.

Removed User 16 Sep 2019
In reply to The New NickB:

I think the polls have been all over the place for months due to the ever changing Brexit situation. Do you visit Britainelects? They summarise all results from the reputable polling organisations.

 balmybaldwin 16 Sep 2019
In reply to The New NickB:

If you are referring to the recent ComRes and Opinionum (sp?) surveys, it appears to be due to the way they account for turnout.  ComRes are supposedly accounting for a 25% turnout among 18-35 yr olds, where as Opinionum are looking at 15%

Post edited at 12:06
OP The New NickB 16 Sep 2019
In reply to Removed User:

I’ve been using BritainElects for the summary, but also to link to the individual polls.

OP The New NickB 16 Sep 2019
In reply to balmybaldwin:

> If you are referring to the recent ComRes and Opinionum (sp?) surveys, it appears to be due to the way they account for turnout.  ComRes are supposedly accounting for a 25% turnout among 18-35 yr olds, where as Opinionum are looking at 15%

Turnout for 18-35 year olds in consistently about 50%, it was above 60% in 2017.

 balmybaldwin 16 Sep 2019
In reply to The New NickB:

Yes I mis read it's the weighting they apply... ie. 25% /15% of total turnout

OP The New NickB 16 Sep 2019
In reply to balmybaldwin:

> Yes I mis read it's the weighting they apply... ie. 25% /15% of total turnout

That makes more sense, although I’m interested in how they came up with those figures. The actual for the last three elections was 9% (2017), 19% (2015) and 14% (2010).

 krikoman 16 Sep 2019
In reply to The New NickB:

I gave up with opinion polls, after the two Corbyn leadership votes, and the last GE in 2017.

They were miles out on all of those, and the Brexit referendum wasn't very close either.

Call me jaded but I've come to see opinion polls as a direction some people would like us to vote in.

3
OP The New NickB 16 Sep 2019
In reply to krikoman:

I'm not sure about the Corbyn leadership polls, not something I paid much attention to, but you are wrong about the 2017 general election and Brexit, they were correct within the stated uncertainties. You would have more of a point if you talked about the 2015 general election.

 krikoman 17 Sep 2019
In reply to The New NickB:

> I'm not sure about the Corbyn leadership polls, not something I paid much attention to, but you are wrong about the 2017 general election and Brexit, they were correct within the stated uncertainties. You would have more of a point if you talked about the 2015 general election.


Weren't there polls showing the massive defeat of Labour, and the demise of the party altogether?

that may have simply been the press, mind, two years is a long time in Krikoman World.

 summo 17 Sep 2019
In reply to krikoman:

> Weren't there polls showing the massive defeat of Labour, and the demise of the party altogether?

I thought they had died, hence the lack of opposition in the commons for the past two years. The Tory remainers (or folk like chuka) who switched allegiance have arguably done more to oppose their own government, than the hibernating Labour party. 

5
 wbo2 17 Sep 2019
In reply to The New NickB:they'rve suffered a bit recently  as their data collection methods I.e. telephone polls needed s lot of reworking, and how they account for bias.  That's very true for party politics  - Brexit was well predicted,  and Trumps election was well inside the error bar.

 Not UK related but Nate Silvers 538 page is good to look at for information on how they 'poll the polls'

OP The New NickB 17 Sep 2019
In reply to krikoman:

> Weren't there polls showing the massive defeat of Labour, and the demise of the party altogether?

The story of the 2017 election was the rise of Labour in the polls, from a 20 point deficit to within a point or two by the week of the election.

 spenser 17 Sep 2019
In reply to The New NickB:

It also depends on which polling model is being used, Survation or YouGov (I can't remember which) use a constituency by constituency model sometimes which is much better for predicting the make up of the house of commons but is far more expensive than doing it on a region or nation wide basis which can't account for votes being stacked in safe seats or predict how swing seats will fall.


New Topic
This topic has been archived, and won't accept reply postings.
Loading Notifications...