The figures for ComRes’s rolling poll tonight are CON 35%(nc), LAB 25%(-1), LDEM 27%(+1), with changes from yesterday. The poll is made up of samples from Monday and Tuesday.

In the comments below the last post there was quite a lot of discussion about the maths of the rolling poll, and whether the consistency of this poll supports yesterday’s big jump in the Tory lead that ComRes found, but none of the other pollsters did.

As far as I’m aware ComRes’s rolling poll is interviewing 500 people a day, and each days figures are made up of the last two samples. What appears to have happened this week is that ComRes had figures pretty much in line with other companies in their Monday poll – suggesting pretty normal samples on Saturday and Sunday. Their Tuesday poll dropped the Saturday sample and added the Monday sample, which pushed the Tories up 3 and knocked Labour and the Lib Dems down 2, suggesting a Monday sample with a very large Tory lead. Today’s figures keep that very Conservative Monday sample, but replace the Sunday data with data from Tuesday.

The fact the figures haven’t changed implies that Tuesday’s data is much the same as Sunday’s, and that ComRes’s bigger Tory lead is all down to a strange sample on Monday. ComRes only do their rolling poll between Monday and Wednesday, so we won’t get any figures tomorrow. If we did, then the Monday sample would be dropped and replaced with data from Wednesday, and would probably have shown the Tory lead shrinking again.


564 Responses to “ComRes rolling-poll – 35/25/27”

1 10 11 12
  1. With Thirsk delayed until June – and it would have been counted as a straight Con hold – that’s one more they need for a majority on May 7th.

    Not that I expect it to be that close really.

    If it is then I shall eat me chocolate.

    {oh, hang on: I just did]

  2. Tony D
    Well we had better ask FrankG himself as my reading of his sentence is quite clear. Let me try it. He says that 18% of LDs will vote tacticaly in a Lab /Con marginal. 13.5% will vote Lab. 4.5% will vote Con. So the net effect is a benefit of 9% of LD votes for Labour.

    That’s it isn’t it?

  3. Howard

    Career politicians have a whole lot of ability and talent. For the most part as Cameron candidly admitted when introducing his Big Society Mano, they don’t have a clue and have been conning the public for more than 40 years, so they want to involve us together etc etc

    You know these guys aren’t really motivated by public service, its how they make their living. So different skillsets

  4. The daily YouGov poll is a bit disappointing for LibDems (though it would have been very welcome a week ago), but it’s only one poll & we always need more than one to show a trend.

  5. The daily Yougov poll is within the margin of error. I don’t think the Lib Dems will be worrying.

  6. @ Michael

    Thanks for the info – funnily enough I just realised the same thing listening to PM!

    @ Al J

    Thanks – I’m going to try very hard not to get hooked! :)

  7. @ GILES

    Hello & welcome – I just read the Indie article you mentioned, in which NC comes to his senses & realises he doesn’t get to dictate who leads the Labour party.

    If Nick Clegg cannot resist serving up soundbites then doing a u-turn the day after, his reputation & credibility will be eroded.

  8. @Eoin Clarke

    ‘No chance on Birnmingham Hall Green
    The yellow and blue vote combined does not even exceed Red’

    Just looked it up
    Notional 2005 figures
    Lab 16799
    LD 12608
    Con 7892
    Others 6931

    Maj 4191 (9.5%)

    Swing Lab/LD 4.75% needed

    It would go

  9. i have a request for the perversely obsessed with ze numbers like me. At the behest of FRANKG i was asked to make a comparison with reported tactical voting in the 2005 election. I unfortunately cant find such numbers, so am hoping that a guru out there has a reference for me.

    I reject the premise of Frank that in assessing voting we should compare to 2005. reason being that what is important is the vote swing as it is now. Tactical voting is only a nuance of FPTP, and effects votes cast. a swing in tactical voting has no impact. All we need to know is what is the trend atm, and how will ppl vote tactically this time round and thus effect the trends we are seeing, particularly in the marginals. But surely nothing will be lost from a comparison, and if anything i am sure it would be interesting to see.

    On the point of tactical voting, always look at how the Libs will vote tactically. It is their tactical vote that will effect the Lab/Con marginals. The Con/Lib and Lab/Lib marginals are also interesting but surely we all agree that Libs can never have a majority of seats and their surge will effect both Lab and Con seats, thus cancelling out each other. Those seats will only be relevant in an extremely tight race (Cons only 1-2% from Lab, which we don’t seem to have.

    What i can say is that in 2005 tactical voting gave Labour around 33 seats (I had the reference somewhere but can seem to find it either at the moment – must be my nerves due to the upcoming debate). So definitely tactical voting is important.

    I do have an impression, and it is just an impression that this time round the ABC and ABL numbers are pretty much at par with each other atm, though with a slightly higher ABC (again this is an impression, ie i have not crunched numbers in this regard as such, but i from the numbers i have seen it seems to be so.
    iI can fish out the numbers in this regard if needed).

    However, the numbers are also telling us that those Lib voters that claim that they will vote tactically the break down is 3:1 Lab:Con. This is the significant number. I suspect the attack by the Tory press today will only confirm this trend.

  10. frank G,

    not if these polls keep going the way the are

    48hours ago Libs led by 3%.

    48 hours on they now trail by 6%.

    The only 4.55 I see, is the 4.5% dip they have taken in two days.

  11. @Xiby

    Many thanks for that explanation.

    Very much disagree about not including the Tactical voting last time around. The majorities in for instance in Lab/Con marginals which we consider when deciding if that seat is vulnerable already includes tactical voting from last time. The same people saying they will tactical vote again this year represents a nil effect on that majority, because it already inclues their votes.

    For example if the current lab/con majority in a seat includes say 6% LD tactical voters from last time. This effective cuts down the actual number of new tactical LD voters from its net 9% to 3%. If the amount of LD tactical voters last time was 12% of their vote and only 9% net tactically voting now then this is decrease in support of 3%.

    Without considering the tactical voting in those seats last GE2005, you cannot just attribute tactical voting this year as additional tactical voting to last year.

    Your statement that it gave Labour 33 seats probably means that tactical voting against Con was a lot, lot heavier than your current net amount of 9% of the LD vote in that seat.

  12. @Eoin Clarke

    ‘not if these polls keep going the way the are’

    Naughty! We were discussiing the situation before the latest poll, when the Lab/LD swing was 7%

    Agree with you that if this poll is the GE figure, it would not fall

  13. @Eoin Clarke

    “not if these polls keep going the way the are’

    In addition to my last. neither would
    Blaydon
    Swansea S
    Northampton N
    Glasgow N

  14. @FrankG,

    Agreed. Keep these types of posts up, they are thoroughly thought provoking. :)

1 10 11 12