Given the success of the approach at the 2017 election I expect we’ll see several MRP seat models this time round. The first one to emerge however is one constructed by Focaldata, using data from mixed sources, including YouGov, that Best for Britain have used to drive a tactical voting website. It has caused some controversy – particularly on the comment pages of the Guardian – with people arguing over the validity of its recommendations. I won’t get too far into that (vote for whoever the hell you want), but thought it was probably worth making a few comments about MRP itself, considering it will crop up again through the campaign.

What is MRP?

First, we need to understand what MRP is. It stands for multilevel regression and post-stratification, which almost certainly doesn’t help. There is a academic paper by Ben Lauderdale and his colleagues who run the YouGov MRP that explains it in great detail here, however the short version is that it’s a modelling technique aimed at producing robust estimates for small geographic areas from large national samples. In the context of elections, that means coming up with estimates of vote share in single constituencies based on a big national sample.

Using traditional techniques even very, very large samples don’t contain enough respondents to be a good guide to individual seats. If you had a huge sample of 50000, divided by 632 seats it would still give you less than 100 people a seat – which wouldn’t be enough to produce decent data. I’ve seen this as a naive criticism of the Best for Britain MRP model (there are only 70 people per seat!) but in fact that is exactly the problem that MRP is intended to solve.

MRP works by modelling the relationship between demographic and political variables and voting intention (the multilevel regression part), and then applying that to the demographics and political circumstances in each individual constituency (the post-stratification). So in this case, an MRP model would look at how demographics like age, gender and education relate to vote intention, and how that differs based on political variables (Is there an incumbent MP? Is it a remain or leave area?). That model is then applied to the known characteristics each seat. What that means is the projection in an individual seat is not just based upon how respondents in that seat say they would vote, it’s effectively also based on how respondents with the same demographics in seats with similar political circumstances say they would vote.

How well does it work?

At the last election YouGov had an MRP model that performed very well – correctly predicting the hung Parliament and some of the more unusual election results like Canterbury and Kensington. Clearly, given the accuracy of the YouGov model, it is possible to use MRP successfully to produce decent seat level estimates from a big national sample.

Best for Britain’s defence of their tactical recommendations relies heavily on how well the YouGov MRP model did in 2017. However, not all MRP models are necessarily equal. It isn’t one single model, it’s a technique, and it’s possible to do it well or badly. It is not certainly not a magical guarantee of accuracy. If we look back to 2017 the YouGov MRP model got all the attention, but it wasn’t the only MRP model out there. Lord Ashcroft also commissioned an MRP model, but that wrongly predicted a Tory majority. Just as some polls have been more accurate than others in recent years, some MRPs may be more accurate than others.

The things that drive the quality of a MRP model should be the quality of the data that’s going into it, and the quality of the model itself – have those designing it picked demographics and political factors that allow them to accurately model voting intentions? As an external observer however, it is quite hard to judge that. For the YouGov model there is its track record from 2017. From other MRP models, we’re driving a bit blind. We know it is a technique that can be very successful if done well, but we won’t really know if it is being done well until it’s compared to actual election results.

Are tactical voting recommendations based on an MRP model sensible?

In principle, yes. MRP is obviously not perfect or infallible – nothing is – but it is an established technique for producing estimates of support in small geographical areas from a larger national poll. Certainly it should be better than using a crude uniform swing, or just basing recommendations on what the levels of support were at the previous election and assuming nothing has changed since.

In practice, of course, it depends on the quality of the model and the tactical decisions that people make based upon them – I certainly don’t intend to get into that debate, especially since I expect there will be various rival tactical voting sites with different recommendation, and perhaps different aims and motivations.

543 Responses to “MRP models and tactical voting”

1 2 3 11
  1. First?

  2. “‘I wish we had more time to explain the deal’ – Prime Minister tries to ease concerns on NI visit.”

  3. I have some limited experience of working in a team where statisticians used a multilevel regression and post-stratification model.

    While I was happy to accept their conclusions, I understood it less from their description than I now do from Anthony’s!

  4. Excellent write-up from AW as always but just a note on YG’s model.

    Look at 95% confidence level range from GE’17

    CON: 269-334
    LAB: 238-302
    LDEM: 7-19
    SNP: 30-53
    PC: 0-3

    I just want to make sure that when folks state a specific number then they are stating a “central”[1] case which is the sum of all the individual ranges of outcomes (with some fancy stats stuff to aggregate out all the “marginals”).

    PS The model will also sum up the %s and again check those! Overall the “central” numbers for YG model were 30ish out (biased in favour of LAB seats) but well within the 95% confidence limits (apart from PC who won 4 seats)

    PPS For sure their individual seat numbers were very useful in highlighting the “uni-town quake” that led many of accept that Dr.Mibbles was onto something (and make a nice bit of 25-1 cash from likes of Paddy Power). The B4B analysis we have for GE’19 does highlight the “posh M25 area” seats where LDEM will likely do much better than UNS (but bookies are aware of that)

    [1] Central doesn’t necessarily mean “centred” – you can have a “skew” with a “fat” tail on one side but not the other. You obviously also hit “ceilings” (eg SNP can’t go above 59, so you might have a 95% confidence range 35-59 with 50 as the central forecast)

  5. @AW
    Clearly, given the accuracy of the YouGov model, it is possible to use MRP successfully to produce decent seat level estimates from a big national sample.
    Make it so.

  6. Gedling constituency poll is up on Survation’s site (headline VI posted by others on last thread)

    Great example of a seat where Farage will cost CON a gain from LAB.

    I’d also suggest look at:

    Q3: “prefered Brexit outcome” (Remain 29%, Leave Deal+No Deal 56%)

    Q4: “best PM” (Boris 45%, Corbyn 19%)

    That’s the harsh brutality of FPTP and an example of how Farage might well lead to Remain and PM Corbyn :(

  7. One thing I don’t understand is if the basic LTV assumptions are wrong whether an MRP or a conventional poll the result will be wrong.

    Did the YG MRP poll use a different LTV for the various sub groups to the regular YG poll; or did they use the same inaccurate LTV adjustments for young people etc and just get lucky with the MRP as the various inaccuracies just happened to net out abut right?

  8. “The things that drive the quality of a MRP model should be the quality of the data that’s going into it, and the quality of the model itself …”

    Well, precisely. The garbage in garbage out principle. GIGO for short. If the MRP model is bad too, that exacerbates the problem, but the key is what data is going into the model in the first place. If the data is dodgy, the quality of the model is irrelevant.

    There’s a bit of boasting here from Anthony about the accuracy of YouGov’s MRP model in 2017, but how far out from polling day did it become accurate? My memory of the polling for that election was, virtually until the very end, that all pollsters, including YouGov, were some way out in terms of the final result. Survation got it pretty spot on with a virtual eve of election poll, but I don’t recall YouGov being very close to it. Maybe near the end, but my recollection is that they too were showing quite big Tory leads for most of the campaign.

    Were YouGov using just YouGov data in 2017, or was it their model using an average poll of polls?

    P.S. Anthony is very loyal to his employer, I think – and I don’t blame him at all!


  9. I might as well copy it over from the end of the last topic.

    Basically, the method is: you attach an odd (rather than probability) to a person with various attributes (it is much broader, judging by YouGov’s experiment in 2015, including things like newspapers, TV programmes values, and so on, than the usual demographic stuff in the “normal” polls) to vote for a political party, and then you run it against a sample.

    The result of that is used for correcting the a priori odds, and now you run it against another sample, and so on. Soon (it depends on the principles of the attribute), so let’s say after 15 runs the a priori odds are close to the outcomes of the experiment. Mathematically it means, that an extraordinary outcome is needed to substantially (I avoid the word significantly) change the a priori odds (Look up Bayes factor).

    Obviously you need a large population (people you can ask) to run the samples. The MoE would be all over the place, but in this method.

    Now you need to match these odds with the demographics (wwell, the attributes you used) with constituency information, and you can predict constituency level outcomes with a very strong odd.

    It is unclear if YouGov tests the outcomes by grouping constituencies (it is actually dependent on the level of information about the constituency).

    So, the logic is: individuals vote. Individuals vote according to certain attributes or values or beliefs or habits but there is an error rate partly because we cannot ask everyone and also because there is a level of uncertainty and also statistical fluctuation. The method filters this out to a good degree so person A having attributed B,C,D,E, ..N is likely to vote (or doesn’t vote) to a political party. Now we check the distribution of such an lifestyle person A in various constituency. This gives us the likely proportion of such a person voting for party x.

    It has been around for a few hundred years, just didn’t get much attention because of the intensive calculation needs, algorithm and a relatively large population (which is a sample) to be used for subsamples. The first app in the US Apple store that could do the calculations appeared about 4 years ago.

    It is also disliked by some statisticians because you have a priori odds (i.e. it conflates deductive and inductive differential reasoning – when you have your standard poll, you cannot go beyond the data. If you start from.assumptions about the chance Labour has, you cannot go beyond the assumption, only accepting or rejecting t on the basis of the data, so you cannot go beyond the hypothesis. In the MRP method you try to combine the two, which is quite problematic, but the problem is managed through the Markov Chain (but of course the problem is not eliminated).

  10. Differential – inferential

  11. Quite a timely report, when we are thinking about care costs and the public sector/private delivery models –

  12. YG article (with polling info) on “pacts” and “coalitions” (incl issue of Indy if SNP MP voters are needed for LAB to form/function govt)

    Boris still has a week to think about it before candidate lists close, after that he’ll have to rely on “paper” candidates and Tactical Voting. Squeaky bum time for sure…

  13. Thanks Anthony for the new thread and the lucid explanation of MRP.

  14. @ Crossbat

    Yougov had a plus 3 and plus 4 for Tories in the last few days but plus 7 on their final poll. Two were pretty close to the result- just not the final one. So I guess it is possible that MRP would have come from all those polls and been fairly compatible? Maybe?

  15. It`s a good description of the MRP technique from Anthony, but he doesn`t say clearly if YouGov are using it this time. Maybe their 2017 model has been tweaked.

    I found the daily predictions for individual seats in Scotland useful in 2017, and seeing the day-on-day changes gave some idea of how strong the predictions were. The actual narrow gains of Labour from SNP looked likely in the two-week run-up, narrow because for 2 days they were Labour gains, then 1 day SNP hold, then the same again.

  16. @Shevii

    Thanks. My sense is that even though there can be last minute vote switching in elections, by the time you get within a week of polling, barring seismic events, things are pretty much settled. TV debates over, manifestos launched, voters registered,leaflets issued, doorstepping and canvassing completed. Voting determinants have probably exhausted themselves too.

    On the flip side of that, with all the factors I’ve listed above still open, I think we’re very premature using current polls to determine election outcomes.

    Premature bordering on silly perhaps.


    “Clearly, given the accuracy of the YouGov model, it is possible to use MRP successfully to produce decent seat level estimates from a big national sample.”

    I am afraid not, that’s a dangerous assumption.

    The fact that last elections YouGov MRP got close to the final result doesn’t meant ; ” successfully to produce decent seat level estimates” because it could have been in the words of Professor McGonagall; “Pure Dumb Luck!”

    I tend to think MRP works but I am not prepared to back it as accurate on the basis of one poll from one Pollster in one Election.

    I need more than that!


  18. @crossbat11
    In fairness to AW and YouGov and all polling companies, I think that the result of this GE might be difficult to predict two days after polling day let alone two days before.
    Given the two main parties both want to borrow to spend on much the same things, NHS, police, the North etc, if there is no majority can we rule out a Con-Lab coalition (to keep the Union)?
    So far the most unlikely thing I have seen in politics is Iain Paisley as First Minister and Martin McGuinness as his Deputy in Stormont. But Boris and Corbyn, Javid and McDonnell as the new Quad would set the bar higher.

  19. J S-B

    There have been lots of unlikely things in politics. Boris Johnson is one of them.

    Trusting someone who considers the best way to cook is to “whack it in the microwave, gas mark 4” would appear unlikely, but seemingly a plurality of folk in England wish to see their books cooked in that way.

  20. @JSB

    Given the two main parties both want to borrow to spend on much the same things, NHS, police, the North etc, if there is no majority can we rule out a Con-Lab coalition

    I think this could happen before long, yes. Not under either of the present leaderships, but maybe not far away, particularly if the LibDems and SNP make the gains predicted in this election.

    If outright majorities become a fanciful idea, then the idea of a German-style Grand Coalition between LAB and CON could become a real thing. Especially as for a while at least it would likely have the kind of numbers where both parties could easily focus on their common areas and ignore their fringe MPs, rather than being entirely beholden to them (and others) if they try to cobble together a standard multiparty majority.

  21. EOR

    What I think you are saying is that English MPs could happily coalesce around a domestic policy approach that isn’t dissimilar to what has been common ground in Scotland for the last 20 years.

    In the actual “North” of GB (as opposed to the more geographically centred bit referred to) there has been a concentration on building infrastructure, avoiding privatisation of NHS provision and maintaining police numbers.

    You don’t need a “Grand Coalition” to do that, you just need to dump the obsession with holding elections on an FPTP basis.

    It’s quite easy.

  22. @OLDNAT
    J S-B

    “but seemingly a plurality of folk in England wish to see their books cooked in that way.”

    No, we just believe in allowing those who work to keep more of their money.

  23. those who work


  24. Crossbat11,
    “. My sense is that even though there can be last minute vote switching in elections, by the time you get within a week of polling, barring seismic events, things are pretty much settled. ”

    This is not my recollection of the last election. My perception was that each successive poll after the election was called came closer to the actual result. In many ways the final result, the most up to date poll of all, was consistent with extrapolating a steady movement towards remain.

    However I didnt see this as typical of elections. My interpretation is that for many key voters the biggest issue was leave or remain, that this crossed people’s traditional party loyalty, and so they had a conflict who to vote for. They did not resolve this until they faced the final decision, maybe pen in hand. Similarly, the great importance placed on leave/remain caused many to turn out who would not normally and this invalidated turnout models and their input to result estimates.

    The election was unusually not decided by people in the middle of the traditional lab-con divide. It was not a push against people in the middle of lab-con moving them one way or the other, though no doubt there was that movement going on too. It was about pushing leave or remain into one of the big two parties, and a little into the minor leave/remain parties, where different other issues applied.

    The take I get from the AW writeup on MRP is that it is simply a model like any other. It is subject to errors due to bad data and alternatively due to bad modelling assumptions.

    Right now we know the data is wrong. It is stated to be wrong if you read the terms of the polls. It is a snapshot of views on the day the data was collected, not a snapshot of views on polling day.

    A properly conducted MRP model predicting results today has the same problems of a uniform swing model, that the data is likely to change markedly before polling day so any prediction today is wrong.

    If this election follows the pattern of the last – and not the pattern of earlier recent ones – there will be a big swing as time goes by. The MRP prediction by the end will be very different to what it is now.

    Now, one could imagine building into the model a correction for campaign swing, but how would you do that? Asume there will be 20% swing to labour by the end, as happened in 2017? If so, I suspect that would make a huge difference to what any website at the moment is stating as voters best choice to achieve a particular aim.

  25. Jonathan Stuart-Brown,
    “Given the two main parties both want to borrow to spend on much the same things, ”

    The change of policy of conservaives abandoning austerity is entirely consistent with a party that believes Brexit will cause a major hit to the economy. If they are elected there will have to be increased borrowing and spending in a keynsian stimulus to prevent a Brexit recession. So their policy statements are entirely consistent with a party which has not changed its mind on the desireability of cutting the size of government, but simply with weathering that expected recession.

    Whereas labour if elected would either see a softer brexit with a smaller recession, or indeed no brexit and a remain boost to the economy. A very different background for its planned spending.

    Labour would be facing a pickup in the economy, whereas con would be facing recession.

  26. “Boris Johnson was seen last night contradicting his own government’s position on how his Brexit Deal will work between Northern Ireland and GB. The video was quickly shared and widely retweeted on social media.”

  27. Oldnat

    ” seemingly a plurality of folk in England wish to see their books cooked in that way.”

    as long as its spoken with an Etonian accent then they vote for any rubbish!

  28. Sam thanks for sharing that video.

    Mr Johnson says there will be no “customs checks” for goods from NI to GB. That’s semantics. There will still be the export form filling needed – but he claims that such form filling is not a check its an “administrative process”. He didn’t rule out there being such “admin processes” on goods from GB to NI — and it seems the NI Conservatives with him were reassured by his words but they don’t tell the whole truth at all.

  29. Prof/Sam,

    Correct me if I am wrong but the administrative process is a post activity declaration, quarterly perhaps, so is semantically different to customs checks.

    However, there would have to be some spot checks, as self declaration has to be policed; and, this is where he is being economical with the truth.

    I hope he gets asked the simple question ”Will there be spot checks and if not how can ensure compliance? ”

  30. ProfHoward/Sam,

    It might just be his typical style, but Johnson does come across as a little drunk.

    Only problem is, the video seems to have a number of discontinuous cuts in it. Playing it on the internet, these maybe come across as interruptions in data delivery, but if you download the whole thing and play it, they are part of the video.

  31. Its also astonishing how bland and uncritical the BBC report is, going with what must be Mr Johnson’s preferred headine of how he’s a ” passionate unionist” instead of him apparently not knowing his own deal and being misleading:

  32. Jim Jam – I don’t know about spot checks. Certainly Stewart Wood (the labour peer) has established that there will be export forms needed to be filled out. It seems one of the people at that meeting (Irwin Armstrong) is claiming that Johnson said there’d be no extra form filling needed – which if true would clearly be in contradiction of his own policy.

  33. @ON

    Comment from last thread noted.

    I really should know better than to reference Scottish politics!

    But I think you understood my broader point.


    I think this may be helpful.

    “Between Northern Ireland (NI) and GB:
    On sales from GB to NI: EU tariffs will be applied if the good is at risk of being moved into the EU.
    The definition of ‘at risk’ is to be determined by the joint committee
    Where the EU tariff has been levied, and if it can be subsequently proven that the good has not entered the EU, customs duties can be reimbursed, but subject to EU state aid limits. At present, it is not clear what would constitute proof.
    While for goods which are not ‘at risk’ there will be no tariffs to pay – this does not mean there will be no checks or controls at the border as the EU will want to be assured that the correct declarations are being made.
    On sales from NI to GB: The WA calls for “unfettered access” so in principle, there will be no tariffs or customs checks although export declaration forms may be necessary.[1] With regard to regulatory checks, the Government’s impact assessment states that the “Protocol contains no requirement for additional regulatory checks on goods moving from Northern Ireland to Great Britain.”[2]”

  35. @ CB11 12:00


    Thanks. My sense is that even though there can be last minute vote switching in elections, by the time you get within a week of polling, barring seismic events, things are pretty much settled. TV debates over, manifestos launched, voters registered,leaflets issued, doorstepping and canvassing completed. Voting determinants have probably exhausted themselves too.”

    This sounds very sensible and is likely to be often the case.

    I do recall one apparent exception though – I think it was the US presidential election of 2000, Bush v Gore.

    Although I cannot remember all the details now I think it was to the effect that a last minute piece of news had appeared in newspapers or other media and this was thought to have produced a late swing to one candidate.

    The reason this was attributed to the news media was that the swing became slightly more pronounced as the results or exit polls covered the later time zones of USA, i.e. the early voters could not have seen this news.

    Of course we have only one time zone at the moment for December 12th. We could compare morning v evening voters I suppose?

  36. Judy Curry has noted the Thompson and Smith paper and added her own comments.

  37. @JAMESB

    Indeed, though on the flip side, not having endless options to stymie a minority government might have forced a vonc and GNU rather than MPs maintaining a kind of puppet government.

    Possibly but the idea of the government doing the will of parliament is surely justified even if there is not a GNU. As I understand it and I have been pointing out There was not a remain alliance: The Tory Rebels wanted brexit but not a no deal brexit and a GNU would not have provisioned this and they did not have common ground but they did have enough common ground to rule out no deal. To my mind the reality was that the Government refused scrutiny and withdrew the WAB which tend to suggest that either they did not want the bill passed since any amendment could be withdrawn after an election anyway

  38. Lab held council seat in Croydon comfortably yesterday-tho their vote share was down 10 per cent,Tories down 4,Lib dems up.10.

    Corbyn needs to attack Libdems on Revoke at some point .

  39. @DANNY

    Then the question is, does society need to change so that young people do not say what they believe starting from birth, or does society change to become more accepting of all sentiments diverging from a very narrow political correctness.

    The problem with the internet is your whole life is available to view. yet people act as if they have the privacy that they have relinquished. Some of the comment people are being pulled up for however are to my mind complete b~ll#cks. Some warrant an apology and some warrant dismissal, currently I reckon they have it right.

    What it does do however is mean that our musing on the internet should be seen as public statements and why I hatte Twitter is that trying to make a sensible argument in 256 characters is impossible and so you get very crude debate and comment.

  40. DAVWEL

    Re yesterdays question to me:

    My source was h ttps://

  41. Does anyone know what the LD policy is on which Brexit option should be vs remain in a 2nd red? Given that it is impossible for them to win a majority, this is a very pertinent issue.

  42. A cautionary note: Yesterday’s GFS 12z run showed massive polar high pressure blocking into December. If verified by subsequent runs, this would point to a repeat of December 2010. About the worst possible backdrop for an election.

    Too early to worry about that now, and other runs are not in agreement, but there are more general signs of things turning colder later in December.

  43. Looking at alternative polling sites corresponding with EC, I see they are predicting a Tory win in Ynys Môn. The Alliance could be crucial there.

    They prediction for Cardiff Central is a Lab hold with c. 35% of the vote.

  44. If the first day was bad for the Tories, yesterday was a disaster for Labour IMO!.

    What happened to the LibDems in Cornwally? I don’t usually comment on Council elections but what about:

    IND (Moorcroft): 38.5% (+38.5)
    CON: 34.5% (-2.2)
    LDEM: 17.5% (-31.6)
    GRN: 8.6% (+0.8)
    IND (Harris): 0.9% (+0.9)

    Independent GAIN from Liberal Democrat.

    I thought the Lib Dems were targeting the South West. Were there any special reasons for that result?

  45. @CARFREW

    Whether it allows “lots” is something else entirely.

    A developed economy has to do a lot of investing in new stuff to get growth, because the pre-existing has already been exploited.
    I do not agree. Some developing countries have been spend R&D on items and have been able to take advantage of this to the point that they are outrunning developed nations so for example whilst China is a massive polluter it also has the largest carbon neutral energy generation by far, It has planted more new trees than anyone else, part of that is due to their size but R&D spending is no guarantee of anything. Indeed we forget all the failures of our R&D spending and only look at successes

    Even Germany only averages about, what, roughly 1.5% growth over the last couple decades? If we want to do significantly better than that we need a lot more R&D.

    US growth is not predicated on R&D at the moment it is predicated on debt so it is difficult to argue that it is R&D that is driving or that will drive markets. R&D in itself is very hit and miss.

    Germany 1.5% growth is actually starling considering that it had to go through the unification of the East Germany and that they did it all alone they had a decade of stagnation incorporating the East with huge fiscal transfers. They are still amongst the leaders in R&D in the EU so I suspect it is not fair to single them out

    My experience of saubsidies/grants is actually mixed. When I worked In Northern Ireland, I had to deal with INI and the EU the EU had some funds for start ups and local employees and the INI were absolutely terrible around a third of monies were returned to central government because the criterion was impossible for many companies to meet

    We don’t even have proper full employment (as opposed to zero hours etc., so there is still room for lots more investment.

    You have constantly argued for service industries being the new grail our expansion in service industries has been in those roles where low skills and insecure employment is the norm. Entrepreneurs do the investing in such endeavours states can only subsidise what there is, They can start new industries granted but very few government do this or if like in terms of the US (and the EU) they do they hand them over to private industries

    According to Alec, investment in batteries might fall foul of the EU. But just batteries alone is a big deal. How much more might fall foul of such rules?

    ALEC has already responded that you have misrepresented him and in my view on whether batteries can be subsidised are known to you

  46. Morning all. Further to the issue of ever more stuff being sneaked into trade deals, the SNP are trying a counter. It’s also acting on the democratic legitimacy thing:

    ”The SNP is to launch its election campaign by promising to bring forward legislation to protect the NHS from privatisation and future trade deals.

    The NHS Protection Bill would block any UK government from using the NHS as a “bargaining chip” in trade talks.

    If passed it would also give devolved parliaments in Scotland, Wales and Northern Ireland a veto on these deals.“

  47. Howard, while you are on.

    Can you confirm that your MP is Chris Graythng?

    A much maligned cabinet minister would be accurate.

  48. @Alister1948

    “I do recall one apparent exception though – I think it was the US presidential election of 2000, Bush v Gore.”

    There probably have been incidences of last minute vote switching due to an event very late on in an election campaign, and you may well be right that the Bush v Gore presidential election was one such example, but I think they must be very few and far between.

    I thought it was interesting in 2017 how that eve of poll Daily Mail 14 page splash on Corbyn’s links to terrorist organisations had so little apparent impact. Maybe this says something about the fading influence of the printed press, and mainstream media generally, but I suspect the Mail saved it up to have what they hoped would be a significant effect on undecided voters. They attempted to reprise the Sun’s famous “Will the last person to leave please switch the lights off” headline in 1992, with Kinnock’s head made into a lightbulb. It came out on polling day. It was the Sun wot won it then, but the Mail obviously failed in a similar bid 25 years later. Different world now, I think, and strident tabloid voices tend to fade on the wind nowadays. That’s frustrating for them, but better for our democracy that right wing proprietors and editors don’t feel they own election outcomes anymore.

    I must admit, I’ve always been fairly sceptical of the oft touted notion that many voters enter a polling booth and their pencil hovers over the ballot paper while they make their final decision. Undecided to the very end. Maybe a handful, but many that undecided won’t vote at all and my sense is that the vast majority of voters march down to polling stations totally sure about how they’re going to vote.

    Nearly all having made up their minds days or weeks before, probably years in some cases!


    Gedling constituency poll is up on Survation’s site (headline VI posted by others on last thread)

    Great example of a seat where Farage will cost CON a gain from LAB.

    I’d also suggest look at:

    Q3: “prefered Brexit outcome” (Remain 29%, Leave Deal+No Deal 56%)

    Q4: “best PM” (Boris 45%, Corbyn 19%)

    That’s the harsh brutality of FPTP and an example of how Farage might well lead to Remain and PM Corbyn :(

    Bit as I remember you wanted this election, you’v been crowing for it , you have argued that this was the real people’s vote……

    I accept that much of what you have been saying is partisan but isn’t this why it is a peoples vote it is an election using a flawed system to allow a minority to impose something on a majority.

    FPTP make this not one election but actually 650 different elections
    That is the beauty and the horror of all this. The majority of votes don’t have any effect and that is why it is going to be difficult to call and more over a majority of the electorate will not be happy with the result.

1 2 3 11