It’s Monday, so as usual we are spoilt for polls, with new figures from Populus, Ashcroft and ICM (who are now on a weekly rota until the general election), with YouGov to come later on tonight.

Populus have voting intentions of CON 32%, LAB 34%, LDEM 9%, UKIP 15%, GRN 4%. Tabs are here).

ICM have topline figures of CON 34%, LAB 32%, LDEM 10%, UKIP 11%, GRN 5% (tabs). The Conservatives are down 5 points since last week, UKIP up four points – something that is almost certainly a reversion to the mean after the incongruous six point Tory lead last week.

Lord Ashcroft has topline figures of CON 34%, LAB 30%, LDEM 10%, UKIP 13%, GRN 4% (tabs). Ashcroft has also updated his Scottish constituency polls from last week with two new polls in Scotland, both in Edinburgh. Edinburgh North and Leith and Edinburgh South were both Labour -v- Lib Dem marginals at the last election, with the SNP in a poor fourth place. Ashcroft’s latest polls finds the SNP ahead in both, with a 13 point lead in Edinburgh North and Leith and a narrow three point lead in Edinburgh South (tabs, tabs).

YouGov are still to come later on tonight…

UPDATE: The last GB poll of the day – YouGov for the Sun – has topline figures of CON 34%, LAB 35%, LD 7%, UKIP 13%, GRN 5%. So we have two polls giving Labour leads, two giving Conservative leads, and the polls apparently still fluctuating around an underlying picture of Labour and Conservative neck-and-neck.

654 Responses to “Monday polling round up”

1 11 12 13 14
  1. @Charles Stuart

    The main problem I see with phone polls is that they tend to focus their fielwork at weekends only. That could be providing a distorted picture of what will happen on a Thursday election.

    The online polls are more frequent and more up to date.

    Also YG actually were very close to the GE result last time – they had Con 36 Lab 30 LD 25. It ended Con 37 Lab 30 LD 23. I would say that was pretty accurate – wouldn’t you?

  2. @CL45

    Who says Lab is doing badly in Midlands marginals? A single crossbreak?

  3. @ Bill Patrick

    “It’s also why the House of Lords is so great, because it gives such politicians a foot in public life if they want it. The best debates in the House of Lords are like a “greatest hits” album of politics, and actually constitute rational discussions rather than the rather mechanical debates you often find in the Commons, because the Lords are seeking the truth rather than performing for their parties.”

    Didn’t you say you only comment on polling?

  4. RAF

    That would be silly.

    It’s 3 cross breaks….

  5. NorthumbrianScot,

    If we absolutely had to change the House of Lords towards an Americanized system like all the other institutions we have in this country that work well, so it could blow up in our faces like changing to a US-style financial regulatory regime, then your proposal would be preferable.


    Major’s legacy depends a lot on what you focus on, and it’s hard to come to a general judgement. Even his critics tend to say that his biggest problem was his party. The exceptions, of course, are the ultra-Eurosceptics, for whom he’s just as bad as Blair or the Lib Dems.

  6. @ RAF

    I don’t think anyone really knows about the Midlands. Which is why I think the election is so uncertain. I [wouldn’t be confident of Labour winning in the] exit poll comes out unless Labour are at least two points ahead.

  7. RAF:
    Hello to you. I can hear Christchurch Priory Bells ring. (Dissolved by Cromwell in 1535).

    I was replying to Mikey who asked about Labour in the Midlands.


    I think it is important to remember there are still a lot of people not online , and these are mainly elderly , and a lot of elderly people tend to vote Conservative. These people would never be included in an online poll.

  9. MARTIN W.
    I agree with you. Blair would win the Midlands. I think.

  10. @martinw
    “I don’t think anyone really knows about the Midlands.”


    Lord Ashcroft, if you’re reading this, you know what to do. Stop polling the bloody Scots every 5 minutes and head south. The Midlands need you.

  11. MartinW @ 19.00

    If I remember correctly for the 2010 exit poll, they only gave seat numbers and not % vote. I seem to remember frantic cross posting about what the % must have been to get to the exit poll
    seat numbers.

    I mentioned a couple of days ago if we were in ‘ordinary’ times (no SNP/UKIP) I’d take the exit poll as gospel and turn in. This time I won’t believe anything until I see it.

  12. @ MartinW


    I remember in 2010 the exit poll, which was actually individual seat based rather than a UK exit poll given as %. it was near enough bang on the result I think in terms of seats won?

  13. “I don’t think anyone really knows about the Midlands.”

    There be Dragons!


  14. Omnishambles @ 19.04

    Agree about the Ashcroft Scotland obsession. It’s all over bar the counting up there.

  15. @ Omnishambles

    “Lord Ashcroft, if you’re reading this, you know what to do. Stop polling the bloody Scots every 5 minutes and head south. The Midlands need you.”

    Yes, I wonder what the fixation with Scotland is this election? LOL!

  16. @ Tony Dean

    “I remember in 2010 the exit poll, which was actually individual seat based rather than a UK exit poll given as %. it was near enough bang on the result I think in terms of seats won?”

    You are right! I remembered after posting that the exit poll is actually a prediction of seats rather than percentages. But in my defence, the essence of my post is still the same – in that I believe that there’s so many unknowns that the Tories could get back in with 2% less vote share than Labour.

  17. Chrislane I may be wrong, but I remember Major’s efforts on the peace process were a bit hampered by his small majority and the need to keep the Rev Paisley happy. He certainly laid foundations that a large majority Govt could build on.

  18. @ Rayfromthenorth

    same reply as to Tony Dean!

  19. @ MartinW

    I agree! But sincerely hope you are wrong!

  20. RAF

    Previous polls show Lab doing well in London and the North West but not so in the Midlands. The latest cross breaks seem to reinforce what is going on. Hence my question.

  21. The BBC’s exit poll is designed to predict the swing in marginal seats and, therefore, the total seat numbers, not the overall GB shares of the vote. Hence the figures they announce at 10pm on election day will probably be projected seat numbers, you won’t get projected national shares of the vote until much later in the day.

    (Indeed, behind the scenes the data we got from John Curtice from the exit poll at 10pm was seat numbers and IIRC changes in the vote in different types of seat – there weren’t any national shares of the vote until later)

  22. Latest forecast changes; two improve for Labour, one improves for the conservatives. Giving us one tie, four Labour leads and two Conservative leads.

    Tables here (with larger graph):!496536&authkey=!AL86mCleBMIBeaU&ithint=file%2cpdf

  23. Anthony @ 19.26

    I asked a question the other day about why the graphics never tell us if an actual declaration result jives with the exit poll forecast.Can this not be done for individual seats then?

  24. Anthony.

    The BBC prediction at 10pm in 2010 was bang on.
    However I remember 1987 when they announced it was too close to call and Thatcher won relatively comfortably.

  25. @ Shandy 5.25

    My point exactly. Without Clegg the LDs will be too busy looking inward to do a deal with anyone. They could probably agree amongst themselves to abstain on the confidence/no-confidence motions likely to be needed under the FTPA and let the numbers Lab or Con can muster from the other parties determine the outcome.

  26. Listening to John Major, I’m reminded that his speaking style is underrated. He’s not charismatic, or rousing, or funny, but he sounds very genuine and as if unscripted in a way that very few politicians can actually achieve, even though they all want to be able to pull it off.

  27. This is interesting- about 2m have registered online in last few weeks – (lots of nice graphs and data) – dominated by younger age groups. There is also a late surge- 485k on Aoril 20th (last day) alone!

    Will this impact the polls? Will it help Labour and the Greens?

  28. I think the midlands will be the last region to succumb to the Milihunk charm but ssuccumb they surely will!

  29. Tom Newton Dunn (Twitter)
    Could fears of a Labour-SNP power sharing deal be the big #GE2015 game changer? We have an interesting @YouGov poll later.

    I guess tonight’s YG poll will be out earlier than usual :(

  30. @ northumbrianscot

    “I agree with Norman…..” (and Michael)

    Me too. It’s all a bit weird.

  31. RAF

    What were you saying about twitter comments last night…….

    Then it will really be LOL in the political sense.

    JOHN TT. I agree that Major was hampered by his small majority.

    I remember the 1987 Election party we held that night; going very quiet!

    Great Links, thank you

  33. So it looks like a Con lead tonight in YouGov. Well it did have to happen.

  34. Mikey – after the exit poll also got it wrong in 1992 they completely changed their approach. 1992 and before it was aimed a getting a national picture.

    Ray – the exit poll model itself does put out predictions for individual seats which are added up to make the projection, but the BBC don’t use the individual projections in the broadcast coverage.

  35. NorthumbrianScot & Bill Patrick
    I sometimes wonder if the best policy for the Lords might be election for an indefinite term but with some kind of recall – strikes a compromise between the undemocratic state of the Lords at the moment and the indignities we might see if the Lords were to be more concerned with getting re-elected than doing the right thing for the country.

    Of course I also see a few problems with this idea!

  36. RAF

    The rumour is Tories up to 38% in tonight’s poll.

    No figure for Labour mentioned yet.

  37. Anthony @ 20.06

    Thanks for that, I do remember vaguely Kellner at a big screen and occasionally saying wtte “higher/lower swing than our exit poll forecast” which immediately had me saying out loud “show it on the bl##dy graphics then”. I don’t understand why they don’t.

  38. @Anthony

    Why was the exit poll wrong in 1992? Shy Tories, or wrong social class weightings, or something else?

  39. @Gary Gatter

    One thing thwt is clear from your plots is that either ElectionsEtc or YG Nowcast will take the wooden spoon in the competition for best protection over the period you cover.

    If we end up with the Tories having most seats, then YGN will be out in the cold. Ditto for EE If Labour come out ahead.

    Given that it doesn’t offer any individual seat projections, there is no way of benchmarking EE prior to May 7. But based on evidence that YouGov has not so far done a good job in predicting Ashcroft results, my inclination would be to start burning the initials YGN into the trophy.

  40. Unicorn @ 20.14

    The YG forecast (not nowcast) has Con ahead

  41. SLAM

    If its another Tory lead, it’ll just be another outlier

  42. TND at the sun as a habit of taking a single poll as a game changer only to go a little quiet the next day

  43. You can do your own focus group on how people decide to vote just by looking at twitter

    Finding #1
    Candidates have their addresses printed on the ballot paper. If the address is not local, some people may change who they intended to vote for

  44. @Adam B

    It’s an outlier if it significantly varies from the mean only to revert to the mean tomorrow.


    This is not a twitter rumour. It’s a tweet from someone who has seen the figures.

  45. Gary Gatter

    Thanks for your link, showing your sheet. I really like your layout and style.

    I liked the layout style of May2015’s seat calculator, but they don’t seem to update the other pollsters regularly enough. So I started to paste their numbers into Excel and manually add the newer numbers from the other pollsters plus Electoral Calculus and produced an average column.

    PS. One day last week Ladbrokes’ seats added up to a total of 658 !!!

  46. RAF

    I know what an outlier is, my uni degree was maths! I was being flippant in that it seems that whenever a Tory lead of more than 1% comes up, people say its an outlier. I’m not saying that the Tories are 6% ahead but there do seem to be an increasing amount of larger Tory leads

  47. ALAN [re UNICORN]
    That line of code was from the HTML generated by the script. To get it you need something to run the script (normally your browser). It makes automation difficult to say the least, (as well as, as Barbazenzero found, where the numbers were being pulled from seems a bit archaic).

    Alan, the penny has just dropped that you’re talking about the HTML physically written to the screen rather than the source code itself.

    What software are you using to view code such as the “<span style="left: 43.96%; width: 12.56%; border-color: rgb(0, 135, 220)" in your 4:31 pm post on p17 of yesterday's thread?

    It would involve some minor reverse engineering as you describe in that post, but if the output HTML code can be captured [as opposed to the HMTL source] it can do the job UNICORN is looking for.

    I’m not aware of that option being available in MSIE or Chrome, but am no longer familiar with other current browsers.

  48. @RayfromtheNorth

    The YG forecast (not nowcast) has Con ahead

    True, but as Anthony has pointed out, the forecast is not based on a formal model. It is Peter Kellner’s assessment of what is going to happen.

    It is tests after the event that will matter. No doubt númerous individual pundits and even UKPR posters will produce better predictions than some of the models. But there isn’t agreat deal anyone can learn from that.

    In contrast, the models that do well will probably still be working away in April 2020. Others will either be forgotten or revised to correct for their errors.

  49. New thread

  50. AdamB

    If its another Tory lead, it’ll just be another outlier


    Like Labours outliers, if that is what you mean, then yes.

1 11 12 13 14