New YouGov Welsh poll

There is a new YouGov poll of Welsh assembly voting intentions for ITV. The topline figures, with changes from the last monthly poll, are as follows:

Constituency: CON 19%(-3), LAB 44%(nc), LDEM 9%(-2), PC 21%(+2)
Regional: CON 18%(-2), LAB 40%(-1), LDEM 9%(-3), PC 23%(+4)

At both regional and constituency level Plaid take second place from the Conservatives. On a uniform swing (and assuming Labour retake Blaenau Gwent now Trish Law is standing down) this would result in the Conservatives losing 4 of their 5 constituency seats to Labour (Monmouth would be the only hold). Add on the regional seats, where the Conservatives would get back some of the seats they lost at a constituency level, and the final result would be Conservatives 10 seats (down 2), Labour 31 seats (up 5), Lib Dems 5 (down 1), Plaid 14 (down 1) – giving Labour an overall majority for the first time.


519 Responses to “New YouGov Welsh poll”

1 9 10 11
  1. @EOIN
    Braveheart or Nad i Ali ? Bollox or Bullets.

  2. @OLD NAT
    Having that self satisfied little twerp against you is a very good start.

  3. Roland,

    I reckon you’d prefer it if your wee man was prancing around in a skirt with a bow and arrow/ sword rather than dodging IEDs in Taleban land. :)

  4. Interesting Labour PPB here. New slogan – “Scotland deserves better.”

    The SNP slogan is “Be part of better”.

    I think Labour have misunderstood. Alex Salmond is a known (and successful) racing tipster, and we probably got funding from Ladbrokes!

    I wonder how many Scots saw the PPB and thought “Scotland deserves better (than Iain Gray)”.

    Has there been any polling (or other research) on the effect of PPBs? Do they influence anyone?

  5. Eoin
    For on eof the samples you done your decimals for the sample size was 2500+ (eg. MoE neglible).

    Actually for a sample of 2500 Margin of Error isn’t “negligible”. It’s +/- 1.95%.

    [Calculated using ComRes’s MoE calculator
    ht tp://www.comres.co.uk/page12574936.aspx ]

    And that’s the “perfect” MoE. YouGov are not taking a random sample from their panel and also are having to adjust the answers so the response is more representative of voters throughout Britain. This means the MoE will be even wider than the 3.90% above.

    And remember that there is still a 5% chance that this is a rogue poll and that the actual figure you are trying to assess is outside that range.

    In this context, looking at figures to two decimal points is just silly.

    Hooded Man
    I see what you did now – though it’s more complicated to do (and explain!) than what I did. However the same objection applies – even if it was meaningful (see my reply to Eoin) the mid-point of the range isn’t much more “likely” than the ends.

    Anthony
    Recent telephone polls give “refused” percentages of ICM 8%; Populus 6%; ComRes 10%; Ipsos-MORI 2% – so it’s not nothing. Refusers tend to be slightly more likely to be female, but that is probably a side-effect of the biggest age group by far being the over-65s. They’re posh no longer though, C1s seem the most reticent.

    Of course the point I was trying to make is that if we get into a “shy coalitioneers” situation this is where they will be hiding.

  6. OldNat,

    Since the days were East Retford was a pocket Borough, money = votes. If Lab spend more that SNP, generally they’ll be more successful. What do they spedn it on? Airtime, media, advertising. You bet propaganda in general is successful. As for PPBs, well they prob help the smaller parties get recognition, and probably save the poor/cash strapped parties some cash. But if your a big fish you would not want to be relying on them.

    I watched Ed M’s speech today. Hmm… a bit negative I’d say. What do reds think of alex salmond? A recent YG poll asked that question, I’ll go find out.

  7. Old Nat,

    this link might interest you if you want to explore it a bit further..

    h ttp://poq.oxfordjournals.org/content/40/2/216.full.pdf

  8. Old Nat

    Not sure about PPBs but the ‘Labour isn’t working’ and ‘Labour’s tax bombshell’ were reckoned to have an impact….

  9. Roger,

    Hooded Man is good with maths. He takes an interest in polls. His % figures to the nearest decimal might not be your cuppa. They might also be silly. I welcome them. They actually inform my poll guesstimates. Which for right or for wrong, tend to leave hooded and I pretty close.

    MoE is theoretical, go back and read Harris’s methodology it was you who posted it. When I questioned Anthony on MoE, he said Harris were right/correct not ot include an MoE with their polls. I have since that moment paid less heed to it.

    Now ‘attack’ me for being ‘silly’ if you choose to do so. this is actually my first post with you on the topic of decimals that Hooded produces, but it seems to have got the juices going in your adjective selection.

  10. Eoin

    Thanks for that.

    The summary “Exposure was moderately correlated with political knowledge and interest. Highly exposed voters were somewhat more likely to attach higher agenda priorities to issues and candidate attributes emphasized in the commercials. Personal affect toward each candidate was mildly associated with advertising exposure frequency. ” isn’t exactly encouraging for parties’ hopes that their PPBs will be hugely influential!

    However, in a close election every edge will count.

  11. If on Monday reds range was 38.2-39.19
    If on Tuesday reds range was 38.3-39.21
    If on wednesday reds range was 38.4-39.3
    If on thrusday reds range was 38.55- 39.49

    All four polls may have shown a 39 but it would, for me at least, tell the full picture. A consistent trend such as this would tell me that reds have a slight ‘very’ upward trend.

    1% of the vote is 270,000 people. Sometimes I think we lose sight of that.

  12. *not tell the full picture.

  13. This post [below] is Anthony’s rebuking me for chastising Harris for not including MoE. Since this post, I give much less credence to the concept of an MoE. That might make me ‘silly’ but that is my reasoning [see below]
    _____________

    Eoin –

    Only Harris use “propensity weighting” – YouGov’s attitudinal weighting is party ID and newspaper readership. Essentially it is just weighting by attitudinal things drawn from parallel face-to-face surveys.

    Don’t mock them for that lack of margin of error, they are entirely correct it in. The 3% margin of error we normally quote for polls is indeed based on some theoretical perfectly random poll of 1000 with a 100% response rate, which never happens in real life. When ICM, Populus, YouGov, Ipsos MORI and so on quote margins of error, strictly speaking they are pretty spurious (you can actually factor in things like response rate, stratification of the sample, weighting factors and so on… but nobody ever does for media polls).

    The real picture is as Harris puts it – there is a margin of error of some sort, but it cannot be theoretically calculated. I still think it’s better to give the 3% or so figure, even though it’s precision cannot be justified, because it reminds people the margin of error is there, whether or not we can quantify it.

  14. @Roland Haines I think the last few days have shown the coalition plans unravelling, this will feed into the polls over the next few weeks and months.

  15. Roger Mexico,

    I see Eoin has articulately saved me the trouble of responding to most of your points.
    It’s not ‘complicated’. Takes less than a minute.
    Not trying to sell it to anyone else anyway, Eoin and I are interested whether LDs were nearer 10 or 12 if YG polls 11.
    Stats. Some people love em, others hate em. No harm done.

    Dispute the ‘meaning’ by all means but not the ‘logic’!
    It sounded like you were close to a concession that you were wrong ? ;-)

  16. OldNat,

    Your closing sentence is the tenet of my political philosophy. In a Gordon Brown Control freakish kinda way, I don’t think you can ever let your eye of the ball once. Take an early night, take a day of campaigning, inconsistency in policy, all breads failure. Thomas Edison’s work ethic still rings true to this day.

  17. @GARY GATTER
    I suppose its a matter of timing. I think if they were going to fall through the floor it would have already happened. The big thing for me is the fact that Labour take the wrap for many of our problems, conversation in the dog and duck, “bloody cuts, yer but Labour did go mad with the money”. They, Labour, will have to come up with something to damage the coalition at the level you believe and I used to believe.

  18. @ROLAND
    When it comes to who to blame for the crisis, Labour won’t have to do anything except wait. Unfortunately for the Coalition the cuts and blame for them will come to rest more and more on their shoulders.
    In the US, the economic crisis is now well and truly Obama’s.
    Voters’ memories are extremely short.
    Unfair I know, but that’s democracy for you.

  19. Hooded Man

    Well I wasn’t wrong in my calculations, but I’m afraid I was confusing people (including myself) a bit by looking at two different problems at the same time. One was what you were doing, which was a sort of mathematical puzzle – given these figures how do I work out the total number of people who expressed a voting preference (#VP)?

    From something Eoin wrote I assumed you’d been doing it one way (working out a range for those who didn’t express a preference and then applying that to the total sample). If you do that, I wrote, then you have to calculate separate ranges for DKs and NVs rather than adding the percentages together and working a range out for that. I was also worried that Eoin was using the range mid-point as the answer rather than realising that #VP was equally likely to be anywhere in the range.

    As it turns out you were doing it differently -calculating a range for #VP from the each three party percentages. This also meant you could then narrow the range down to the overlap between the three ranges and get closer to the result. ;) Of course you could also use the range I got to get closer still and, with more effort, work out other ranges from other figures shown to eventually get to the right result for #VP.

    As I said it’s a mathematical puzzle with a correct answer, but it was the second problem – that Eoin was using your figures as if they meant something – that really worried me. I’ll reply to him in the next thread as the caravan has moved on.

1 9 10 11