The final round of Ipsos MORI’s marginal polling for Reuters has been published. These are polls conducted in Labour held marginal seats where the Conservatives need a swing of between 5% and 9% to win – to win on a uniform swing the Conservatives need a swing of about 6.9%, so this sample excludes around about 60+ of the closest Lab-v-Con marginals, and goes up to seats that would grant the Conservatives a pretty substantial majority. To get a majority of 1, the Conservatives would need to win about half these seats, which equates to being neck and neck in voting intention.

The topline figures from MORI today, with changes from last week, are CON 36%(+1), LAB 36%(-2), LDEM 20%(-1). This represents a slight swing to the Conservatives since a week ago, and with a 7 point swing is just about enough for the Conservatives to get an overall majority on a uniform swing. In practice however, these figures would be unlikely to produce a Tory majority – to win on a 7% swing the Conservatives would also need to gain a substantial number of seats from the Liberal Democrats, and this seems unlikely on present national polling.

There are some interesting findings on tactical voting – MORI asked people if the party they said they were voting for was their first choice, or a tactical choice. 9% said they were voting tactically, a figure which has not risen during the election campaign – despite the widespread and intuitive belief that people make such tactical decisions during the campaign, MORI are not picking it up (or at least, not yet). MORI went on to ask them which party would have been their first choice – there were more people who said their first choice would have been the Lib Dems voting Labour tactically than there were voting Conservative tactically, but the sample sizes are so minute I would not put any weight on it.

Finally, unlike previous waves of the study MORI excluded people who said they were not registered to vote. 5% of the sample said they weren’t – I can’t see the question in the full tables so I can’t tell if these were people who gave voting intentions and were excluded, people who said they were unlikely to vote, or whether MORI asked about registration first and didn’t ask unregistered people further voting questions.

UPDATE: Bob Worcester has answered that final question. People were asked if they were registered to vote first, and not asked further questions if they weren’t – so we don’t know what answers they would have given.

NB – If there is heavy traffic tonight and the site gets too slow I’ll be switching comments off during “rush hour” everything is cross-posted to the backup blog at where comments will remain open.

166 Responses to “Ipsos MORI show 7% swing in Lab-v-Con marginals”

1 2 3 4
  1. Hmm not very scientific but I’ve just checked the Tory target seat list on this site. There are 133 seats total that fall to the Tories on a 7% swing. If you strip out the 32 targets that are not held by Labour (on the basis that noone expects gains against SNP/PC and we don’t know about LDs) then that leaves 101 Tory gains. Add that to their 198 seats from last time and you get 299 seats, well short of a majority. So even if you assume some gains from LDs (far from certain) then MORI’s poll doesn’t appear to suggest a Tory majority at all. A Tory government perhaps but not a majority government.

  2. Great Press release that deserves to be laid out in full

    **Ipsos Mori: marginals poll does not indicate a Tory majority**

    Earlier today, an Ipsos-Mori poll of 57 Labour/Conservative marginal seats for Reuters was published with the claim that it showed the Tories on course for a majority of two.

    However, the poll did not include any Conservative/Lib Dem marginals, in which given the improved Lib Dem performance, the Conservatives are expected to perform worse than was indicated at the start of the campaign. The Tories cannot rely on winning seats from the Lib Dems, or even holding their existing position, and it will therefore require a larger swing in Con-Lab marginals to be confident of a majority.

    With this in mind, it is thought that the swing in Lab-Con marginals would need to be at least eight per cent for the Conservatives to be confident of an overall majority.

    Ipsos Mori have issued the following clarification:

    The last of our polls in key marginal constituencies, Ipsos MORI’s new poll for Reuters points suggests that the Conservatives are inching towards a majority with a 7% swing from 2005 in these key battleground constituencies. However, given that national polls are suggesting the Conservatives are more likely to lose seats to the Liberal Democrats than gain from them, this majority is not necessarily guaranteed.

  3. Graeme on the previous thread has quoted opinium as showing

    33- 28-27

    If this presumably new poll is correctly reported that would be a change from one week ago of
    Lib -1

  4. first

  5. Opinium/Daily Express Poll- CON 33 LAB 28 LIB DEM 27

    BBC Seat Projection

    Lab 283 Con 255 LD 83 Others 29

  6. Wow Opinium showing Labour largest party now like the 2 polls yesterday. And Conservative stuck at 2005 defeat level in 3rd poll now in last 24hours. Opinium has shown better Conservative leads. And 1/3 undecided in Mori marginal

  7. sorry, fourth :-)

  8. Opinium for the Express:

    C – 33% (-1%)
    Lab – 28% (+3%)
    LD – 27% (-1%)

  9. Opinium good news for reds, bad for blues.

  10. Opinium for the Express:

    C – 33% (-1%)
    Lab – 28% (+3%)
    LD – 27% (-1%)

    yet another lib-lab poll: james murdoch must be planning to plant another microphone…..

    Interesting to see if CoNres and YG match this poll and the two yesterday that clearly showed the ‘big Mo’ (what a ‘larf) screeching to a halt :-)

  11. I think cameron did ok at the citizens uk debate

    hes certainly not an ignorant fellow and I feel he may well be close to a one nation conservative like clarke

    but both he and clegg were second stage to brown

    a superb speech from him

  12. @ Anthony,

    great assessment.

    “In practice however, these figures would be unlikely to produce a Tory majority – to win on a 7% swing the Conservatives would also need to gain a substantial number of seats from the Liberal Democrats, and this seems unlikely on present national polling.”

    I wonder how Reuters, a news agency I have enormous respect for did not pick up on this, and went with their Tory close to majority headline. I have no doubt that they simply missed the point as they focused solely on Lab-Con marginals forgetting the LibDem, rather than any bias on their party. They too above the fray for that.

    I was working on Con-Lib marginals before I commented, to make sure I had all the facts before I spoke, but seeing as you already reached that conclusion, I guess I can comfortably say that I agree with that assessment 100%.

    I will still release my Con-Lib marginal workings later on tonight.

  13. The Con/Lab marginals include about 5 3-way marginals, which presumably would not definitely go to the Tories. It’s not enough to overtake Labour, they have to stay ahead of the 3rd party in the race.


    Thanks – as always – for picking up & posting the ‘myth busting’ stuff. You are 8-)

  15. I’m sure some blues will writing this off as another rogue bank holiday poll.

    Whilst at the same time taking heart from the Mori marginal poll.

    And Reds will do vice versa.

    Guess we’ll know on election day.

  16. Much more sensitive for Lib Dems to be seen officially requesting tactical voting than it is for Labour to do so:

    1/ Their new, younger supporters may not even get the concept.

    2/ It would play very badly with some Lib Dem/Tory waverers in their marginal seats.

    I repeat what I wrote yesterday: achieving the right balance in tactical voting will be as easy as herding cats. It will also make for just as interesting viewing on Thursday night. It’s going to be very odd I think.


    Re your kind offer to trawl through old posts etc: that’s fine by me – not sure why you made it sound like a threat by the way, I couldn’t really give a toss.

    I’ve never claimed any of my predictions were cast-iron and happily amended them as events themselves altered.

    What I’ve consistently felt was that the Tories were not sufficiently popular to gain an overall majority – but, of course, it’s obviously possible – and also, that given the different circumstance of this election, the Lib dem vote would not collapse as you have maintained.

    There were no “trust me’s” in my posts.

  17. But what do we think of Opinium? Its Polls on 19 & 26 April differed from others at those dates (in as much as they identified larger Conservative leads) And what are the field work dates?

  18. Are there any field dates for Opinium?

    I am please to see they are now converging with YG etc.

  19. Opinium is not worth the paper its printed on. Its a pit if your red leaning but since the company is completely untested in elections we cannot take it seriously. Anyone could come up with figures that seemed plausible.

    Are they in the BPC can someone tell me?

  20. : Opinium/Daily Express: Con 33 (-1) Lab 28 (+3) Lib Dem 27 (-1)

    Great Press release that deserves to be laid out in full
    **Ipsos Mori: marginals poll does not indicate a Tory majority**
    -Where can I find the press headline you are referring to?

  22. Have found your posts very informative and the best site around. Well done thanks for all the information on the pollsters results.

  23. “I’m not sure I should say this on camera,” says a tieless Tony Blair, but decides to press on anyway with a good old dig. “The Clegg business is absurd, you know. The policies are nonsense.” Speaking in Tamworth, Staffordshire, he also praises Labour, saying “we should be proud of what we’ve done” and this election “is about what we can do in the future if we’re given the chance”.

    From BBC

  24. Eoin – they are in the BPC, though frankly the information they release on their website is very scant indeed. You can’t even work out the weighting targets they use.

  25. By the way, given a choice between the Lib Dem vote collapsing and a Tory Govt and I’d go for the former anytime.

    It’s quite odd how we’re getting quite close – within just 3% now – of a joke post some time ago, suggesting 30/30/30 as the result.


  26. Do these “marginal” polls have a track record? Were they used in 2005, for example?

  27. @ 33% .. and how many points better is that than 2005 zilch

  28. Rob Sheffield: absolutely right, I agree with you.

    Just as these polls, which only focus on Lab/Con marginals, were not being viewed accurately a few weeks ago when it was likely that the swing in LD/Con and Con/LD marginals would be from LD to Con, they are similarly being inaccurately viewed today when the LD/Con swing is likely to be in the opposite direction.

    The Tory majority of 2 is misleading because the Tories are quite likely to lose seats to the LDs such as Eastbourne. In fact losing Eastbourne would wipe out the Tory majority given by the marginals poll.

  29. @Éoin: I agree. I have always ignored Opinium, even when it looked good for LibDems.

  30. Worth bearing in mind that the last Opinium poll showed a Tory rise on a day when three other polls showed them dropping. I think the changes today are probably an unwind of an unusually good sample for the Tories last time. More or less in line with the other recent polls.

  31. Opinium are a member of the BPC (as stated on their website)

  32. The MORI poll is the best marginal poll the Tories have had in 2010. THree days before an election that can only be a positive thing for them.

  33. @Theresa

    Its pasted in full above but page is here (posted at 17:01)

    h ttp://

    You can always select a passage and paste into google as well.

  34. Interesting on lack of increase in tactical voting reported during campaign. Wonder if explanation is:

    1) it’s more of a phenomenon in Lib-Con seats, where people who start off saying Labour on national grounds, then move to LD once the local LD campaign (as they always do) drum home the local tactical position? There’s less relentless tactical focus in Lab-Con seats, a smaller pool of potential tactical voters, and more ‘consumer resistance’?

    In 2005 many of the rather few Labour increases in vote share were in Lib-Con seats, a perhaps odd reflection of the final Lab campaign message not to let the Cons in by voting LD!

    2) The national rise in LD credibility has inoculated some of their supporters against a tactical squeeze in Con-Lab seats.

  35. I’ve noticed a lot of people still pushing the line that labour is “always” overestimated in the polls and people “always” show a surge for the cons in the polling booth.

    It occurs to me that even if this were true, its hardly helpful if there is a decline in their lead prior to polling day. You’d just end up back where you were, not any further forward.

    An advance following a retreat just takes you back into no-man’s land.

  36. @ RHUCKLE

    I think you are referring to my post about GO to SUE.

    As you say, it was absolutely light-hearted & was either auto-modded because of a word or phrase or Anthony was concerned it would be taken out of context & start a ‘car crash’ of daft posts.

    @ Anthony

    I wouldn’t have posted it except we were at the end of a thread & I thought few people would see it.

    Sorry, it was thoughtless of me when you are trying to keep traffic at a manageable level.

  37. “The year of the broken swing” (Jon Snow blog)

    h t tp://

  38. I don’t recall any analysis of local by election results on here, which I find slightly surprising.

    My crude analysis indicates that in the four months ending March 2010, percentage share of the vote was:
    CON 33
    LAB 22
    LD 26
    Others 19 (with, in particular, UKIP and BNP doing well)

    Now, one might say, that such results bear little relation to the GE outcome. Maybe true, but compared with the corresonding period before the 2005 election, the changes are:
    CON n/c
    LAB -2
    LD -5
    Others +7.

    My interpretation has always been that CON really should have been doing better than the low to mid 30s to achived a comfortable working majority.

  39. @Gattino,

    I don’t think it’s either or. The suggestion is that the Tories outpoll their final opinion poll figures. Whether that’s going to be true or not this time does not depend on what happens to their opinion poll figures in the last week…

  40. Strange how the left of centre posters suddenly putting faith in an opinium poll. Personally I’d lay more store in Mori – a very good marginals poll for the Conservatives.

  41. Opinium Research are registered with the BPC according to their website

  42. He’s done it again! Gordon Brown has just given the best speech of the whole campaign to citizen UK. Head and shoulders about Clegg & Cameron. Not seen such an inspiring speech for a long time. He received a rapturous reception. People actually chanting his name! :-)

    If any of this stuff makes the news, I reckon it will persuade a lot of the 1/3rd who have not made up their minds in that marginal poll.

    Then again ….I must get real ….they will probably only show the heckler! ;-)

  43. @JohnG,

    Local authority by-election results are not weighted in any way and therefore don’t give a good indication of national popularity. What the scores are is very dependent on exactly which seats come up for grabs.

  44. This also doesn’t take into account three way races where lib dems can gain seats.

    The 23 lib dem tory targets are showing no gains in polls so this poll would leave them well short of a majority.

  45. Does anyone have any stats on whether the weather has any impact on voting?

    The weather in middle England marginals on Thurdays is 13C with the occasional rain shower.

    This may appear an odd question. But for those that are not certain to vote, if the weather is grim, they might not bother. This might affect Labour more than the Tories ?

  46. @Paul Bristol,

    My impression of the Citizens UK “debate” is that the audience were fairly partisan anti-Tory. Brown may well have done well, but he and Clegg received rousing cheers before saying a word. If it gets coverage it could still help him though.

  47. Maybe the overall result will have to await the outcome of the Thirsk and Malton poll. There have been boundary changes, but it looks like a nailed on CON win.

  48. Do we know what MORI showed as the national vote share on these same dates? It’s hard to know how much extra swing there is if we don’t have a national baseline. The last MORI national poll was a rogue, so we need to know whether they’ve ironed that out.

  49. @Gattino

    “I’ve noticed a lot of people still pushing the line that labour is “always” overestimated in the polls and people “always” show a surge for the cons in the polling booth.”

    I have posted here in the past that this could be the first election with a ‘shy labour (read Brown) voter”.

    The shy voter has always been used to describe why there seemed to be an unwillingness for people to admit to voting for the nasty, mean, selfish Tories…until they got to the voting booth. hence the Labour vote being overstated in polling- though this has decreased markedly over the last 4 elections.

    But that theory basically IMHO does not work today: when the embarrassing party to admit you support (in our ‘image obsessed society’) is actually LABOUR (and Brown) this time and its very *fashionable* to support trendy and young Dave or ‘the new-Nick’.

    I actually think there is little if any ‘shy voter’ skewing going on in the polls: I think the average of the ‘big four’ YG/ ICM/ COMRES/ MORI is pretty accurate this year.

    But IF there is a ‘shy voter’ pattern going on I suspect it is more likely people saying they are undecided who will then actually vote Labour in the privacy of the polling booth.

  50. I’m hoping for a very complex configuration of weather conditions on Thursday.

1 2 3 4