General Election Constituency Polls

[Davıd Boothroyd]

Jun 25, 2024 19:00:31 GMT graham said:

I am a little cautious re- this poll on the basis that many who declare an intention to 'Vote Labour' still see Corbyn as the Labour candidate.

It was commissioned by Corbyn supporters, so they made sure this was not possible. The question was "Now thinking about your constituency and the candidates standing, if the General Election were tomorrow, which party would you vote for? [IF already voted by post] Which candidate did you vote for?" and poll respondents were then given the candidate names followed by their party affiliations.

See: cdn.survation.com/wp-content/uploads/2024/06/25152633/Islington_North_2024-06-25_Tables.xlsx

[graham]

Jun 25, 2024 19:06:35 GMT Davıd Boothroyd said:

Jun 25, 2024 19:00:31 GMT graham said:

I am a little cautious re- this poll on the basis that many who declare an intention to 'Vote Labour' still see Corbyn as the Labour candidate.

It was commissioned by Corbyn supporters, so they made sure this was not possible. The question was "Now thinking about your constituency and the candidates standing, if the General Election were tomorrow, which party would you vote for? [IF already voted by post] Which candidate did you vote for?" and poll respondents were then given the candidate names followed by their party affiliations.

See: cdn.survation.com/wp-content/uploads/2024/06/25152633/Islington_North_2024-06-25_Tables.xlsx

Logically that should be true but I am not sure it is in reality.

I recall polls conducted a few day s before the Libcoln by election on March 1st 1973 which had Dick Taverne and the Labour candidate neck and neck  yet on Polling Day Taverne won by 58% to 23%.

[kevinf]

Having spoken to people who have campaigned for Corbyn and others for Labour in Islington North, the feeling is it is very close. Much closer than this poll.

[batman]

yes that has been the mood music. I am actually a little surprised: I have been thinking that Corbyn would win fairly easily but perhaps the generally good campaign by Labour has put them in contention. I'll continue to forecast a Corbyn win for now but would be pleased to be proved wrong

[Pete Whitehead]

Jun 25, 2024 23:21:31 GMT batman said:

yes that has been the mood music. I am actually a little surprised: I have been thinking that Corbyn would win fairly easily but perhaps the generally good campaign by Labour has put them in contention. I'll continue to forecast a Corbyn win for now but would be pleased to be proved wrong

That's exactly my position. Conversely, in Rochdale I had assumed Galloway would lose fairly easily but the chatter is that he is in contention (I remain to be convinced). I would also be happy to see Galloway win whereas it would give me great pleasure to see Corbyn humiliated.

[birkinabe]

The poll has Corbyn and the Labour candidate pretty much tied with eachother amongst those who voted Labour in 2019, with Labour's lead relying mostly on squeezing 2019 LD voters (and Green voters to a lesser extent), and I'm wondering if the perception that the race is closer is consistent with this. On a slightly different note, there are certain findings in the poll beyond the voting intention that Corbyn and his campaign will no doubt be pleased with* (and that I have already seen some of his supporters cite in order to argue he could still win and therefore boost his vote in July)

*these include Corbyn winning in a head-on-head race with the Labour candidate by about 48-40 (55-45 excluding DKs), Corbyn having higher net approval than Starmer (about +22, vs basically +0), and net disapproval of the decision to block him from standing for Labour (about -15).

[YL]

Looking at the tables for the Omnisis poll of Bicester & Woodstock (Lab 31, LD 31, Con 30, Reform 3, Green 3) released today, I notice that the sampling method is described as follows

In street convenience sampling- all residents of Biscester constituency, screened via postcode lookup. Quota cells based on constituency demographics values. Multiple locations within the constituency and times of day used to help eliminate bias

(Yeah, that's not how you spell it.)

This is followed by

We use rake weighting to make our sample representative of the constituency. In particular we weight our data to match ONS targets on the following variables: age, gender, highest qualification, plus 2019 general election vote and likelihood to vote in this upcoming general election.

And there is indeed very strong weighting, particularly on highest qualification. (Degree up, GCSE and A level down.) The Lib Dem vote is very heavily upweighted; the unweighted figures have Labour ahead with the Lib Dems in third. Of course there's no reason to think that the unweighted figures are better, but it does mean that the effective sample size is even smaller than the quoted 458.

The same sampling method was used in their polls of Gillingham & Rainham and Hartlepool, but not the two they did for the Greens, which were phone polls.

[andrewp]

Bristol Central Constituency Voting Intention:

GRN: 49% (+23)
LAB: 40% (-19)
RFM: 3% (+2)
CON: 6% (-8)
LDM: 3% (New)

Via @wethinkpolling, 25 Jun - 1 Jul.
Changes w/ GE2019 Notional.

[Tony Otim]

Given the track record of constituency polling, that's still too close to call.

[arnieg]

So we had the wrong winner in Godalming, Portsmouth North, Islington North and Bicester.

Once again a less than stellar performance.

[eastmidlandsright]

Jul 6, 2024 13:39:39 GMT arnieg said:

So we had the wrong winner in Godalming, Portsmouth North, Islington North and Bicester.

Once again a less than stellar performance.

I am somewhat unconvinced that constituency polling in this country will ever be remotely reliable. You are simply polling too small an electorate to get a good sample.  

[greenchristian]

Jul 6, 2024 13:39:39 GMT arnieg said:

So we had the wrong winner in Godalming, Portsmouth North, Islington North and Bicester.

Once again a less than stellar performance.

But the right winner in North Herefordshire, Waveney Valley, Bristol Central, Richmond & Northallerton, Holborn & St Pancras, Gillingham & Rainham.

A 60% "success" rate suggests that they might be getting better.

[batman]

Well that's true but in some of those cases the percentages were way out from the actual result. Gillingham & Rainham particularly so.

[greenchristian]

Jul 6, 2024 21:04:55 GMT batman said:

Well that's true but in some of those cases the percentages were way out from the actual result. Gillingham & Rainham particularly so.

Which raises the question of the margin of error (which is always higher with these polls due to the smaller sample size) and how much of the discrepancy was due to polling error vs late swing.

Yes, there are problems with constituency polling. I just don't think they're as bad as they've been made out to be here.

[manchesterman]

I wonder how soon post-election we will get the first poll of the new Parliament?

[cathyc]

Jul 9, 2024 17:43:34 GMT manchesterman said:

I wonder how soon post-election we will get the first poll of the new Parliament?

I was polled by Survation this afternoon.

[manchesterman]

Jul 9, 2024 18:49:07 GMT cathyc said:

Jul 9, 2024 17:43:34 GMT manchesterman said:

I wonder how soon post-election we will get the first poll of the new Parliament?

I was polled by Survation this afternoon.

Please tell me the question wasnt " which party do you intend to vote for at the next GE!?" 

[Jack]

I've been asked by YouGov who I voted for and my opinions on all the different parties already.

[cathyc]

Jul 9, 2024 19:09:58 GMT Jack said:

I've been asked by YouGov who I voted for and my opinions on all the different parties already.

Maybe they think that Starmer's going to call a snap election.

[peterl]

Jul 9, 2024 19:09:58 GMT Jack said:

I've been asked by YouGov who I voted for and my opinions on all the different parties already.

I had a similar survey today, but from Survation. It also asked about what forms of ID one has.