Shy Liberal Democrats

As regular readers and watchers of polling methodology will know, one of the questions facing pollsters is what to do with people who say don’t know or refuse to answer voting intention questions. A purist approach is to ignore them, to base reported figures only upon people who actually say who they would vote for or, for companies using a squeeze question, who they are most likely to vote for – this is the approach used by most pollsters in the UK. The alternative approach is to estimate how they would vote based on one of their other answers.

This re-allocation of don’t knows was pioneered by ICM after the 1992 debacle when the pollsters all vastly overestimated Labour. They noticed that there were a lot of former Conservatives now saying don’t know, and theorised that many of these were people who would still vote Tory, but were reluctant to admit to pollsters that they were supporting an unfashionable party. This was supported by recontact surveys after elections, when people who had said don’t know to pollsters prior to the election did tend to end up voting for the party they had done at the previous election.

As a result, ICM started reallocating 50% of don’t knows and refusals to the party they said they supported at the previous election. During the 1990s this invariably helped the Conservatives, and became known as the “Shy Tory adjustment”. The name stuck, so it was often still known as the shy Tories through the last two terms of the Labour government, when it actually tended to help Labour instead. Andrew Cooper of Populus called them “Bashful Blairites”.

In recent months it has changed again, polls are now showing a large proportion of people who voted Lib Dem in 2010 saying they don’t know how they would vote in an election tomorrow, and ICM’s reallocation of don’t knows is now favouring them. In ICM’s last three polls the re-allocation of don’t knows has bumped up the level of Liberal Democrat support by 2 points – yesterday’s topline figures of CON 36%, LAB 38%, LDEM 14% were CON 37%, LAB 40%, LDEM 12% without the don’t knows.

Naturally this leads us to the question of whether this is a sound thing to do. Certainly there is solid evidence to back up what ICM and Populus do. Re-contact polls after past elections always supported it, and ICM’s re-contact survey this time round found that about 50% of people who said don’t know in the pre-election polls did indeed end up voting for the same party they did in 2005. Equally, at past elections the adjustment has tended to make ICM’s figures more accurate. None of this guarantees it will still work in the future – the current political situation is rather unusual and these former Lib Dems may behave differently – but it’s a sound starting point.

The difference between re-allocating don’t knows and not doing so is more one of principle. Should polls report just what people say, or should we estimate what the people who refuse to answer think? There are good arguments for both, and in terms of voting intention I think it’s a positive that both are produced. We can see what people are actually saying, but also take account of the fact that there are lots of former Liberal Democrats out there saying don’t know, who may or may not filter back come election day. It’s just important to know which pollsters already factor in those don’t knows, and which ones don’t.

284 Responses to “Shy Liberal Democrats”

1 4 5 6


    The point is if you arent voting for a party based on their manisfestos what are you voting for. Just becuase you dont win a majority does not give you the right to back away from your promises.

    November 24th, 2010 at 10:51 pm”

    Just looking at this thread this a.m. Epochery – your response exactly sums up what Far Easterner was talking about regarding posters as far as I can see!

    Maybe it will just have to wait until Labour are faced with the same scenario before the dose of realism suddenly strikes home.

  2. @Roger, Amber.

    I understand why the students would oppose a change of policy that is to their financial detriment. I’m just making the point that there are issues that (perhaps) justify outporings of rage and violence of the streets. The idea that if your salary climbs above 21k you will be asked to pay back part of the costs of your higher education doesn’t seem to fall into that category to me.

    If the policy was, say “Preventing illegal immigrants from receiving emergency lifesaving medical treatment” or “Abolishing courts and allowing the police to decide who is guilty or innocent” I could understand this kind of demonstration. I’d probably be on the streets breaking something myself. But the mechanism by which higher education is funded doesn’t seem to me to be the life and death issue that some are trying to portray it as.

  3. outporings = outpourings

  4. The march in London seemed to be mostly children truanting from school.
    I’d agree with the comments that the polls aren’t saying anything new. Compared to say Ireland things are pretty quiet here.

  5. Neil A,

    If the policy was, say “Preventing illegal immigrants from receiving emergency lifesaving medical treatment” or “Abolishing courts and allowing the police to decide who is guilty or innocent” I could understand this kind of demonstration. I’d probably be on the streets breaking something myself. But the mechanism by which higher education is funded doesn’t seem to me to be the life and death issue that some are trying to portray it as.

    Not meaning to be controversial here on a difficult subject, isn’t this nearly what is being proposed for the pilot for dealing with domestic violence? If Radio 4 reported it correctly, if Police attend a reported domestic violence incident, even if there is not enough evidence to charge anyone, they could ban someone from the address for 48 hours, followed up by Court Order to extend it?

    Yes, I know it is a Labour proposal being enacted.

    Neil, how would the average Police Officer feel about this, do you think? I think it put quite an erroneous responsibility on them. What if they see no evidence, but don’t want to risk a further attack that could be blamed on the Officer for not acting in the right way?

  6. Old nat

    Plaid/Labour, SNP/Labour?
    The differences IMO include the fact that Plaid is no longer (if it ever was) a party focussing on independence but is more devolutionist. This also means it is more clearly on the left (much) than the SNP. I think the current poll is very similar to the last two?

  7. @ Wolf
    “The march in London seemed to be mostly children truanting from school.”


  8. @Garry K,

    Fortunately I’ve been in specialised departments so long I’ve never had to deal with any of this new “Order based policing” culture.

    The plan seems to be that rather than investigating and attempting to prove specific crimes to a criminal standard (beyond reasonable doubt) you gather less rigorous evidence and apply for an order telling the person to stop. Their failure to acquiesce is then a criminal offence.

    I’ve always thought this approach was a bit underhand and what we really need to do is work out why we’re so c**p at investigating the actual crimes in the first place.

    I’ve not read anything on the latest proposals, but from what you describe they follow in the footsteps of a series of attempts by the powers that be to curtail police discretion in dealing with “domestic” incidents. Some strong-willed police officers have and probably still will resist this but the majority will go with the path of least resistance and cover their own backs.

    The problem with domestic violence is straightforward. For whatever reason, most victims of it don’t actually want to end their relationship or for any action to be taken against their abuser. If they call the police at all, it is usually only because they want the immediate incident to be resolved (generally because they’re afraid of being seriously hurt) not because they want their partner to be moved away. There is a host of good explanations for why people behave this way, but so long as they are a consenting adult who is not sectionable, it is their right to make choices that may be bad for them and put them at a risk of harm.

    My personal view is that the victim culture, by making the authorities shoulder the blame for the decisions of the victims, has led to a situation where we are basically being told to treat adult victims like children.

    I’ve never dealt with domestic violence in isolation, as I’ve mostly been involved with the children. That adds an entirely different dynamic as the moral hazards of not intervening in a violent household with children are rightly borne by the authorities. Sadly, the most effective way to ensure that domestic abusers are kept away from homes with children has always been to make it clear that the children may be removed otherwise.

    Ultimately I think domestic violence is one of those problems that will always be with us, and efforts to totally eradicate it are likely to stray over into the serious infringement of civil liberties. I am all for finding ways to encourage the victims not to put up with it, and I think the police deserve praise for the way cooperative victims are supported these days.

  9. The students, many of whom were school children.
    Some that I saw had walked out of school in their shirts and (Asda) V-necks.

    I would like to see the police show more consideration, and set a good example to these young people.

    This would mean making a distinction between those who engage in lawful protest, and those who commit unlawful acts. Subjecting them to the collective punishment of ‘kettling’ for upwards of 4 hours in freezing conditions is not proportionate. (Caroline Lucas made this point in parliament.)
    I know this is asking a lot of the police in difficult circumstances… so what’s new?

    Nick Clegg made a point of being seen with students at every opportity during the election campaign… NUS is making the point that “No more broken promises” was the slogan to one of the LD PPbroadcasts.

  10. Billy Bob,

    I have always wondered about the effectiveness of Kettling.

    I used to watch Leeds United, and the police tactics used be having small groups of Officers all over the place, and if any groups were getting ‘giddy’ they would deal with it when the group was small and break it up. During this, the ring leaders were dragged off to the awaiting police vans. This meant that a large group never formed, and the people most likely to cause bother were lifted out of the situation.

    From years of watching English fans abroad, you could plainly see that letting a large group gather and keeping them together increases tempers and the perceived strength in numbers leads to an increased risk of violence.

    I have to say when it comes to policing football crowds our Boys in Blue are the best I’ve seen in the world. The way it is done on the continent, large groups of police in riot gear, was always more provocative and more likely to lead to disorder and injuries.

  11. Silghtly off topic but has there been a poll in Oldham East and Saddleworth? The by-election won’t be until the New Year but a poll now and one just before might be quite interesting.

  12. @Garry K

    My experience of watching a *lot* of football in the 80s was that policing in some areas seemed almost designed to provoke. Treat people like animals and you can’t be surprised if they react like them. Kettling is no different. Sometime soon, kettling is going to result in a major riot or a major tragedy.

  13. Colin – posts actually about polls are hardly ever off-topic!

    But the answer is no, by-election polls are relatively rare creatures these days.


    ICM 36 38 14


  15. Wandering Welshman,

    Based on a unified national swing, under the current system and boundaries, that equates to a Labour majority of 12. (Calculator on this site.)

    What does it mean with AV and a reduction to 600 seats? Best ask Mystic Meg!

  16. (From my post yesterday):
    “Yesterday’s poll (Con 42 Lab 40 LD 10) seemed to be quite significant because behind the drop in the Con lead (+3% to +2%) the poll included a much higher proportion of 2010 GE Conservative voters than in the previous week.

    That is, the ratio of 2010 Con to Labour voters yesterday was 643/497= 1.294. In the previous (3% Con lead) poll the ratio was 1.210 and trend towards an increasing Con lead over the past week had been closely correlated with an increase in the ratio from only 1.072 when Lab was 5% ahead.

    Put another way, you would have expected yesterday’s change in the sample to have increased the Con lead if the previous correlation had continued. Instead the Con lead diminished.

    So yesterday’s poll may have marked a more significant movement in voting intention back towards Labour than appears to be the case. Tonight’s poll might just reflect that movement rather than be something new. If it is a repeat, we will see the ratio falling back a bit from 1.294 when YouGov publishes its detailed tables tomorrow.”

    To update this, the YouGov detailed tables do indeed show that the 2010 Con/Lab GE vote ratio fell back a bit to 1.229 in last night’s 40/40/9 poll taken on 23/24 Nov.

    This ratio is still a tiny bit higher than the equivalent ratio of 1.210 that accompanied the poll of 41/38/11 taken on 21/22 Nov. So the erosion of the 3% Con lead since 21/22 Nov seems to point to a definite shift in voting intentions since that poll. But I think the bulk of the shift occurred between 21/22 and 22/23 Nov rather than between the 22/23 and 23/24 Nov polls.

  17. he also said AV would favour the conservatives …..which is strange as they will be voting against it

    his comment was based on the current system on that poll ..his words being

    ” based on the poll, the coalition would still edge it ”

    hes going to of course try to wriggle as its a sporting bet


  18. The latest YouGov tables also reveal that, of the LibDem 2010 GE voters, 64% of those who have made up their minds would for a different party to the LibDems (and more would vote for Labour than the LibDems).

    So going back to Anthony’s original thread about the assumptions that ICM make, it seems entirely wrong IMO for any reallocation of some of the past LD “dont knows” to be 100% back to their original party, when only 36% of past LD “do knows” are now sticking with that party.

  19. Another typo, sorry.

    last post should read
    “64% of those who have made up their minds would vote a different party to the LibDems”

  20. or even
    “64% of those who have made up their minds would vote for a different party to the LibDems”

  21. Wandering Welshman

    i said that if there was an election I reckon the coalition would edge it. You have to consider many things which these polls do not consider.

    these polls attempt to normalise voter behaviour in each constituency as being uniform, it is not.

    the margin of error is +-3%, far too great to gain any meaningful information. The BBc have been tracking polls since 1983. since that date they have recorded 1708 polls, here are hte results

    Party / vote to gain majority / Poll wins / %age
    Lab / 38% / 746 / 43.7%
    Con / 42% / 152 / 8.9%
    Coalition / 50% / 810 / 47.4%

    so according to the polls labour beat the tories to a majority five times since records began yet have been in power for 13 years compared to the tories 14 years. (with three terms and four respectively).

    it appears that mid term polls need to be treated with extreme caution, certainly when the tories are embarking on a programme of austerity measures to rebalance the books.

    furthermore, in all of the above polls, the result predicted when compared to the next election was correct only 36% of the time, giving an incorrect reslt 64%. this means that these polls predict the wrong answer more than they predict it correctly.

    so if you look at what the poll predicts, then reverse it, you are more likely to get the result of the next election, a bit like the monty hall problem, but for politics

  22. Phil,

    I have been looking at the data you provided over the last few days.

    I have looked at it as if it were a control chart, and the data was statistically in control, ie the variation present is normal process variation, not special variation.

    I want to get more data and try it via a CUSUM analysis. I have found patterns in data previously using this tool that a control chart failed to see.

    I have also tried some correlation tests between the Con lead from the Con/Lab figure. I got a result of R-Sq(adjusted) of 83.7%. Removing one outlier from the data It improved to 99.2%.

    I need to do more analysis, but I need to get the raw data from home and use Minitab at work to complete it.

    You have got me thinking here though…

  23. @GarryK

    Thanks for that, Garry. If you’re looking for strong correlations, it might be best to break the data up into snapshots of a few days. I think you may find a reasonably strong correlation over any short term period while voting intentions are stable, but as soon as there is an underlying change in voting intentions the correlation will reduce as you’ll be into a new series of data.

    There must be a relationship as around 90% of 2010 Con and Lab voters have not changed their past allegiance.

    I’ve tried a time series linking points on an Excel scattergraph plotting the ratio (x) against the Con lead (y). Movement towards the bottom right or top left highlights a shift in underlying voting intention in favour of Lab and Con respectively.

  24. you said the coalition would cling to power based on the 36 38 14 poll………those figures …..

    thats a 36 38 14 vote share on election day

    it wouldnt

  25. I said in my opinion in an election the coaliion would pip it. I never said it was based on the poll. you are twisting what i said. the poll figures give a labour government using flawed methods (ut the best we have).

    we’d need a poll in every constituency to get a true indication. bear in mind from a sample size of 1,000, depending on the poll collator, you could end up with a sample of less than 100 when stratified. not really giving good confidence levels is it.

    so i ask the quesiton, based on the poll results, are you saying this would guarantee a labour government, or simply give a good indication of a labour government. the two are not the same

    (apologies for any typos)

  26. you really do not get it do you. it is a rough approximation only. it is possible that getting 38% of the popular vote the number of MP’s labour have could go down. likewise if voter behaviours changed we could see the libdems rise dramatically in line with the popular vote.

    you seem to say it is cast in stone. i say it is not. its a guideline, a yardstick and nothing more. even the poll collators agree with this. why can’t you

  27. you said in a reply to the poll of 36 38 14 ..which is the only poll mentioned by me or UC in that thread …..

    ” even if a hypothetical election followed the results of the poll , would not give labour majority in the commons , the coalition would edge it ”

    wed 24 nov, 20:41

    clearly the assertion on here and other respected analysis sites that this would give a labour majority is wrong

    and you are right ..and you didnt say that anyway …

  28. yougov

    lab 39

    con 42

    ld 10

  29. labour largest party , short of clear majority by 12 seats

  30. how many times do you need to be told. its what would happen if there was a national swing. suvh a thing does not exist, a swing happens at a constituency level. therefore it is only a best guess rather than an accurate prediction, there are too many factors involved for it to be anything other than that.

    why no comment that of the 1708 polls recorded by the BBC since 1983, only 36% of them have accurately predicted the outcome of the next general election? you wish to use this method, which is wrong 2 out of 3 times, to say that an outcome will happen for sure – madness imho.

    lets face it, the major unknown in all of this is voter behaviour. they may say one thing in a apoll, but when it comes to voting, they may do another. there any many factors involved which are used to create a best guess, such as stratified sampling, or assigning don’t know votes based on previous voting habits.

    I’ll say it again, based on the polls and they way they attempt to extrapolate to a national level, I still think come a flash election, the coalition would edge it. We are only six months in to parliamentary session

  31. @ Amber
    “Students saw Clegg & Co. sign a pledge; now they are exercising their democratic right to protest against a policy that isn’t what they voted for & the best you can do is dig out the old ‘revolting’ pun.”

    Fair point

  32. @ Neil A
    “Whilst I am no fan of tuition fees, I do find the sight of “angry” students vandalising stuff because they want the right to earn big salaries without repaying loans a bit stomach-churning.”

    The web is full of stories that the police van was deliberately parked there and the students kettled so that they would vandalise it for the cameras, and to give the police good press whilst they were waving batons at schoolgirls. At least the press were wearing riot helmets. For instance:

    h ttp://

    @ Wolf (& Billy Bob)

    “The march in London seemed to be mostly children truanting from school.
    I’d agree with the comments that the polls aren’t saying anything new. Compared to say Ireland things are pretty quiet here.”

    I find the tone of your post patronising towards young people attempting to exercise their democratic rights and being surrounded for hours by the police. Let us see if they turn up to the next one on Tuesday and use better tactics to stay ahead of the police wall of shields.

  33. @ Garry K (& Neil A)
    “Not meaning to be controversial here on a difficult subject, isn’t this nearly what is being proposed for the pilot for dealing with domestic violence? If Radio 4 reported it correctly, if Police attend a reported domestic violence incident, even if there is not enough evidence to charge anyone, they could ban someone from the address for 48 hours, followed up by Court Order to extend it?”

    Not meaning to be partisan but one sometimes receives the impression they are making policy up as they go along. It may be a good idea to think through consequences first.

  34. Returning to the original post….

    I am Lib Dem founder member, former parliamentary candidate and an approved candidate for next May’s council elections.

    I am not sure how I am voting in the General Election in 2015.

1 4 5 6