Urban75 Home About Offline BrixtonBuzz Contact

Political polling

I am keeping my eye on Survation, the only company that got it basically bang on in all three of the last GEs, they have the Labour lead up to 21% now.

1718819657120.png
 
Labour have been polling 15-20% ahead of the tories for about a year now - and the bi-elections and local election results are in line with the polling. There is nothing to suggest that labours vote share is being overstated. Sure - people are genrally "meh" about starmer but that is vastly negated by the overwhelming "please just fuck off and keep fucking off" that something like 75% of voters feel about the tories. Even their base are clearly voting out of loyalty and/or inbuilt fear of the evil of the labour party rather than any positive feelings about the tories. Not sure about complacency - agasint that their is the motivation of hurting the tories and ever better organsied tactical voting.
There are over a dozen polling companies doing several polls a week using a wide range of methodology and all of them are predicting a tory meltdown - its just a question of weather its as bad as 1997 or even worse - sorry - L meant even better.

I'm talking about polling day vote drop which is not something you can spot in the polls or even the bi-elections/local elections really. I question how much people are motivated to flog the Tory dead horse especially in constituencies where it really doesn't matter. The rest is in line with what I have said!!!

Fwiw the polls are predicting much worse for the Tories than 1997. I think it might be closer to 1997 in terms of seats than the polls suggest.
 
I am keeping my eye on Survation, the only company that got it basically bang on in all three of the last GEs, they have the Labour lead up to 21% now.

View attachment 429659
Yes, Survation interests me as well.

Looking at the latest L-C leads across the main pollsters, it's clear that, amongst the "nowcasters", (who largely set aside the DKs and generally produce larger L leads), Survation has the lowest L-C lead at 18%:-

1718821521729.png 1718821554808.png
 
Yes, Survation interests me as well.

Looking at the latest L-C leads across the main pollsters, it's clear that, amongst the "nowcasters", (who largely set aside the DKs and generally produce larger L leads), Survation has the lowest L-C lead at 18%:-

View attachment 429663 View attachment 429664

That 18% was Survation's last poll, with the one published today, it's now 21%.
 
Amongst the "LP friendly" 'nowcasters' there's a lead range of 9% points; 27 (Deltapoll) to 18 (Survation)
Amongst the "middling" 'squeeze questioners' there's a lead range of 3% points; 22 (Focaldata) to 19 (Savanta)
Amongst the "Tory friendly" 're-weighters' there's a lead range of 3% points; 19 (YouGov) to 16 (More in Common)

Obviously the bulk of the variation in polling leads relates to these 3 methodological areas relating to the DKs. What makes this GE a little more 'squeaky-bum-time' for the pollsters is that DKs are running at about 2X the comparable figures from this point in the 2019 GE.

Interesting times.
 
I'm talking about polling day vote drop which is not something you can spot in the polls or even the bi-elections/local elections really. I question how much people are motivated to flog the Tory dead horse especially in constituencies where it really doesn't matter. The rest is in line with what I have said!!!

Fwiw the polls are predicting much worse for the Tories than 1997. I think it might be closer to 1997 in terms of seats than the polls suggest.
well - i was contesting your idea that the polls were overstating the labour voteshare - when there is nothing to suggest that. I think some polls may be understating the tory votes - based on that fact that a polling level of sub 25% (sub 20% in some polls!) is so uprecedented I think a bit of caution might be applicable. A labour vote share in the low 40s is - however - not particaulrly extrodianry and - given how hated the tories are - you might expect to be higher. So the "meh" factor WRT to labour seems to be represented in the polling.
 
well - i was contesting your idea that the polls were overstating the labour voteshare - when there is nothing to suggest that. I think some polls may be understating the tory votes - based on that fact that a polling level of sub 25% (sub 20% in some polls!) is so uprecedented I think a bit of caution might be applicable. A labour vote share in the low 40s is - however - not particaulrly extrodianry and - given how hated the tories are - you might expect to be higher. So the "meh" factor WRT to labour seems to be represented in the polling.

No there isn't any direct evidence to back up my claim, it's just inference. Getting the vote out is an important factor on the day, I'm just giving that factor some thought. The fun thing is we'll find out if I'm right in a couple of weeks.
 
well - i was contesting your idea that the polls were overstating the labour voteshare - when there is nothing to suggest that. I think some polls may be understating the tory votes - based on that fact that a polling level of sub 25% (sub 20% in some polls!) is so uprecedented I think a bit of caution might be applicable. A labour vote share in the low 40s is - however - not particaulrly extrodianry and - given how hated the tories are - you might expect to be higher. So the "meh" factor WRT to labour seems to be represented in the polling.

I am not sure the Tory voting level needs much caution, TBH.

Here Labour canvassers are working the area hard, for the first time ever in a GE, they even flooded my village, which really isn't Labour territory at all, the one that knocked my door said they had started in the south of the village, with the seriously posh houses and long drives, told me around 60% of people had told them that despite being life-long Tory voters, they wouldn't be voting this time.

I suspect this will be duplicated all over the country, shedloads of Tory voters staying at home, and the head-banging ones voting for Reform, is going to seriously hit them hard.
 
My mum let me stay up to watch 1992 and I've watched every single results programme ever since (except Theresa May's 2017 experiment and I wanted to weep when I realised that I was falling asleep because that's a great sign of getting older)
 
This poll is from right academic Matthew Goodwin's polling outfit who unsurprisingly don't appear to have a great reputation, methodologically. Complete outlier and perhaps more of an attempt to seed a narrative
 
The only media outlet using 'pollingpeople' in GB News, it's a total nonsense, they need investigating by the British Polling Council.
Goodwin's law: "As GB news polling grows on longer, the probability of artificially promoting the Nazis rises"
 
I've always enjoyed the thrill of a new poll, but this election it's all started to feel chaotic (even putting Goodwin aside). Beyond the basic fact of a Labour landslide, nothing seems to agree. Too many polls, too many methodologies, too hard to interpret likely trends.

The MRPs seem particularly misleading. I've just read a Twitter post suggesting that an MRP has the Greens ahead by a whopping 47% in Brighton, yet I'm sure another MRP had them losing a few days ago?
 
Oh, so they are, I searched the BPC list for PollingPeople, not PeoplePolling. :facepalm:

I'll put a complaint in, in the morning.
I can't see the raw data, but someone who claims to have done has said on twitter that the panel cohorts were hugely skewed with 60 18-24 year olds and about 300 who are 75+; that would do it.
 
Back
Top Bottom