Pollwatch – State of the parties (summer 2010): Reasons to keep calm and carry on

By Stephen Tall | Sat 21st August 2010 - 5:57 pm

There have been a rather astonishing 36 opinion polls in the six weeks since LDV’s last polls round-up at the beginning of July. Thirty of those 36 originate from just one polling company, YouGov.

So let’s bring you up-to-date with July and August’s polls in chronological order of publication:

July average: Con 42%, Lab 36%, Lib Dem 15%

August average (to date): Con 41%, Lab 36%, Lib Dem 14%

Anthony Wells’ UK Polling Report blog’s current polling average shows the following scores on the doors:

Lib Dems 15%, Lab 36%, Con 40%

As Mark Pack rightly noted recently here, “with the next general election years away, national voting intention questions don’t mean very much this far out.” But, regardless, the polls will be closely scrutinised by politicians, the media and supporters alike to try and discern how voters are reacting to the ‘new politics’. So let’s take a look at the figures from each of the main parties’ perspectives …

Conservatives:

the Tories will be reasonably cheered by their current polling, with the party enjoying a narrow but clear lead over Labour in most polls. Compare the current polls with 1979, the last time they came to power, when Labour had moved into a narrow lead by August.

However, it’s noticeable that in August only YouGov is showing the party polling above the psychologically important 40% mark, with ICM, ComRes and Harris all showing them below. Perhaps more crucially for the stability of the Coalition, no poll has yet shown the Tories in a commanding enough lead to win an outright majority.

Labour:

there’s no doubt Labour is buoyed by the opinion polls showing them within touching distance of the Tories, and level-pegging according to ICM’s August poll. Three factors suggest they should put the champagne on ice.

First, as I pointed out this past week, Labour is losing the economic argument, with the majority of the public signed-up to the Coalition’s austerity measures. Assuming the economy moves out of recession before the next election, therefore, Labour risks being blamed for the slump and receiving none of the credit for recovery. Given most elections are determined on the basis of the Government’s economic credibility, this should worry Labour: all their leadership contenders appear to be playing a very short-term game.

Secondly, as the 1979 polls show compared with the 1983 result, it’s quite possible for an opposition to be in the lead at this point in the election cycle and still crash to a heavy defeat at the ballot box next time round.

And thirdly, Labour has yet to elect its leader, and so is currently a safe repository for most anti-Coalition/Government dissatisfaction. Once a leader is chosen, they will be subject to scrutiny, the media will challenge them on their response to the economy, and they will start having to make choices, not all of which can be popular.

Lib Dems:

Clearly at the moment the Lib Dems have most reason to be troubled by the polls. Compared with our general election performance of 24%, we have shed a significant number of votes. That there has not been greater angst to date is a product of two factors.

First, the party knew full well what it was getting itself into when it signed up for the Coalition: a rocky ride, with ups and downs – and that applies to the polls as much as it does to policies.

And secondly, we are a long way out even from next May’s crucial polls in Scotland, Wales and English councils. As we all recall from the last parliament, the news media was at pains to write off the party every time a bad poll came out – and then came the general election campaign, and the party’s traditional polling resurrection as the news broadcasters were forced by law to do their jobs properly, and report politics fairly for a change.

ICM v YouGov: you pays your money…

There is a further factor. As I mentioned in the opening paragraph, these peacetime polls are utterly dominated by YouGov. Only two other pollsters in the past six weeks have published more than one poll to enable us to show how different methodologies can produce significantly varying averages in the Lib Dem share of the vote:

The difference between YouGov and ICM (which most poll-watchers regard as the industry’s gold standard given their consistent accuracy in predicting vote shares over successive general elections) is more than should be expected.

But it’s not hard to see why if you look at the weightings used by each pollster in order to achieve a balanced sample of responses. As Mike Smithson’s PoliticalBetting.com recently pointed out, YouGov is alone among the main pollsters in using a party ID weighting system which pegs the Lib Dems at 12%, with Labour and the Tories on 32.5% and 28.5% respectively. This is a stark contrast with ICM which uses past vote recall to peg the Lib Dems at 16%, with Labour and the Tories on 21% and 25% respectively.

Of course, such subtleties are not reflected in the news media reporting of opinion polls which simply pounces on the ups/downs* (however marginal) and produces an automated headline with the words boost/slump* (*delete as appropriate).

So let LDV offer readers something that the news media usually forgets when it comes to polls: context. Let’s look at ICM – the mainstream pollster which produced the most accurate general election prediction – and its Lib Dem polling figures in August over the years:

ICM’s Lib Dem scores:

Notice anything there? Well, what sticks out to me is the consistency of recent years’ Lib Dem poll ratings according to the most reliable opinion polling company in the UK.

So there you go… you have your choice. If you’re a glass half-empty Lib Dem or a Coalition-sceptic/hater then read the YouGov runes and enjoy. If you’re a glass half-full Lib Dem and a Coalition well-wisher, take heart from the fact that ICM suggests Lib Dem support holding steady.

Read more by Stephen Tall or more about featured, icm, Polls or yougov.
This entry was posted in Op-eds and Polls.

Advert

18 Comments

J 21st Aug '10 - 6:02pm

Excellent Post. Well researched and reasonable.
[I am a member of the Labour party]
Andrew Suffield 21st Aug '10 - 7:10pm

And I’ll repeat what I’ve said in other places: yougov’s method appears to be statistically bogus. They’re asking people “what party do you identify yourself as?” and weighting the answers based on that – but self-identification is so strongly tied to current voting intention that their weighting probably dominates the results. It seems unlikely that the yougov polls can have any meaningful accuracy here – and if it’s not accurate to within 5%, all the variation we’re seeing has been swallowed.

What can we tell from the yougov polls? Well, it’s basically ignoring everybody who votes for the party they identify as; those people are replaced with the weighting figures. So it’s just considering people whose vote differs from their self-identification. There will be only a handful of people who identify with one party and intend to vote for another (probably tactical transfers), so the bulk of the results will be people who have “no party identification” – floating voters.

That suggests the Tories are currently dominating opinion of the floaters (being around 12%-14% above their weighting), Labour have some moderate support (around 7%-8% above their weighting), and the Lib Dems have a little bit (around 2%-3%). That translates to around half the unaffiliated population supporting the Tories, about a third supporting Labour, and about a tenth supporting the Lib Dems. Accuracy of extrapolating this way is appalling, so make of that what you will. The Tories are definitely the favourite right now, though.
Andrew Boff 21st Aug '10 - 7:40pm

You really don’t need polls to tell you whether or not you’re doing the right thing.
Anthony Aloysius St 21st Aug '10 - 8:07pm

“And I’ll repeat what I’ve said in other places: yougov’s method appears to be statistically bogus. They’re asking people “what party do you identify yourself as?” and weighting the answers based on that – but self-identification is so strongly tied to current voting intention that their weighting probably dominates the results.”

Surely if that were the case it would tend to suppress, not magnify, changes in party popularity.

You might try posting your comments on Anthony Wells’s blog and asking him what he thinks.
mpg 21st Aug '10 - 8:30pm

Really interesting post. Agree its far too early for thoughts of suicide from Lib Dem supporters. However, polls do tells us how we are perceived now, and we should be mindful of this to say the least.

Agree that the real measure of the Lib Dems position with the electorate will be the local elections next year. So until then, its wait and see.
Rob Sheffield 21st Aug '10 - 8:37pm

@Suffield

And I’ll repeat what I’ve said in other places: yougov’s method appears to be statistically bogus.

I suggest you read UKPR- all your bogus qualms are dealt with in full!
Maria 21st Aug '10 - 10:01pm

Great post.

But how much sense does it make to do such an August-on-August comparison? I am not sure what it can tell us (good or bad). The LibDems are now in such a different position! For once, the LibDems – especially several leading politicians and LibDem policies – have a very high profile in political reporting at the moment. How much coverage was there last summer and in those years before? Clearly, this has been very mixed coverage, hence no gain in the polls from it.

Yet, it’s unlikely that the similar numbers should be read as a sign that there hasn’t been a lot of change behind the scenes. It’s difficult to say, but the similar level of support may hide significant voter movement, with some traditional supporters currently contemplating to move elsewhere, while new support has also been gained. We have yet to see how stable both the defections and the gains will be.

I think that now more than ever, there is little to go on at the moment in order to work out how these figures will relate to votes next May, let alone in 2015! I would guess, however, that a higher profile should be a good thing in the long run….
Andrew Suffield 21st Aug '10 - 10:54pm

I suggest you read UKPR- all your bogus qualms are dealt with in full!

No, they aren’t. What UKPR actually says is that political weighting is controversial and nothing more than guesswork on the part of the pollsters.
jayu 21st Aug '10 - 11:21pm

Assuming the economy moves out of recession before the next election,……..

Would someone like to inform Stephen that we are not in a recession at the present moment. Probably a Freudian slip, in anticipation of the results of the coalition’s policies. He is right, in one sense. In the words of Bill Clinton,

It’s the economy, stupid.

No matter what the LibDems may think that they achieve from this coalition, the fact that they fully signed up to the the Tory economic position, will always outweigh whatever good they think that is.
Dave Page 22nd Aug '10 - 12:20am

I’m sure you’d be happy if that were true, jayu, but there are clear Lib Dem influences in the Budget, not least restoring the pensions link, raising Capital Gains Tax and increasing the bottom Income Tax threshold to take hundreds of thousands of low earners out of tax altogether.

But hey, don’t let the facts get in the way of your mindless bile.
Anthony Aloysius St 22nd Aug '10 - 12:39am

“What UKPR actually says is that political weighting is controversial and nothing more than guesswork on the part of the pollsters.”

That’s pretty ludicrous, even by your standards. I mean, you do realise that Anthony Wells works for YouGov, don’t you? But I suppose anything’s possible.
jayu 22nd Aug '10 - 12:59am

But hey, don’t let the facts get in the way of your mindless bile.

Spoken like a true Cleggeron. Or should that be Cleggeroon?
MatGB 22nd Aug '10 - 3:38am

Anthony, we do know that Wells works for YouGov, I had a brief correspondence with him on this very topic before the election wherein I expressed my (serious) concerns about YouGov’s weighting, Stephen’s post is in part based around analysis and research I did and am still doing.

It’s my contention that YouGov’s methodology of pre-weighting the panel is interesting and could be said to be controversial. But that’s their’ choice. However, I am utterly convinced that their respective numbers for party ID are wrong, mostly in that they over weight for the ‘big two’ parties, which should probably be down weighted to the mid 20s, and partially that they slightly under weight for the Lib Dems. But they don’t say how or why they pick those weightings (unless I missed something, if you can link me somewhere where it’s explicitly stated why they use those numbers), and I belive them to be innaccurate, and this therefore distorts their results.

YouGov has always historically given the LDs the lowest poll ratings outside an election campaign, and those “results” do not carry through to actual results. In both local govt elections and general elections, the party is gettting results in the low to mid 20s, yet even when those results are coming through, YouGov puts the party on the low to mid teens. ICM puts the party on high teens to low 20s.

I chose to believe ICMs methodology is sound, YouGov’s remains unproven for voter prediction surveys. If and when they start getting consistently accurate results (and I hope they can at some point sort their methodology out so that they do), then I’ll start taking YouGov polls for voting intentions seriously. But it hasn’t happened yet.

But how much sense does it make to do such an August-on-August comparison? I am not sure what it can tell us (good or bad). The LibDems are now in such a different position!

Yes, we are in a very different position, however the reason it’s pertinent is that elements of the media are putting about that the Coalition, as a specific, has caused a slump in LD poll ratings, whereas the reality is that the poll ratings are about normal.

The positive coverage the party or policies are getting is good, but there’s a lot of negative coverage, especially in the newspapers that are nomrally the most supportive.

I was actually expecting, and predicted, that the coalition would hurt the party in the polls and we’d get squeezed more than normally. That, currently, doesn’t appear to be the case, at least from the ICM polling. I expect it will happen though, but it may recover as well.
Anthony Aloysius St 22nd Aug '10 - 8:51am

“Anthony, we do know that Wells works for YouGov …”

My point is that Andrew Suffield _doesn’t_ appear to realise that, as he would have us believe that Wells himself has described YouGov’s methodology as “nothing more than guesswork”!
Grammar Police 22nd Aug '10 - 9:38am

@ Andrea – “How much coverage was there last summer and in those years before? Clearly, this has been very mixed coverage, hence no gain in the polls from it. ”

I think one thing that people forget is that a lot of coverage from Lib Dems comes from the work that local parties do. During an election, that massively increases, and is part of what raises our profile (as well as the media) during elections. So although we now have ministers, local parties will not be sustaining the same level of activity as during the election, just the same as every August . . . .
MatGB 22nd Aug '10 - 3:52pm

Well yes, but it’s an accurate paraphrase.
John Leston 24th Aug '10 - 10:07am

Really useful analysis. Thank you. But it seems to betray an interesting National Liberal mindset – look guys, following an election where in many ways we were the big story and now with Libs in government for the first time in more than two generations, we have great news – we are no weaker in the polls at this time of year than we have been since 2006! Come on, joining with the Tories was a dreadful and unnecessary mistake; we all know it was a mistake. I regret that I expect LibDem poll figures to decline further since we no longer have any convincing answer to the question – “So, remind me, just what are the LibDems for? Why would I want one?”
Charlotte 26th Aug '10 - 5:19pm

with the majority of the public signed-up to the Coalition’s austerity measures.

Signed up that is, until it actually affects them directly.

And who can blame them. I suspect Lib Dem MPs will be just as provincial: in favour of cuts until it directly affects their own constituency and chances of re-election.

See, for example, the survey conducted before the election by TheyWorkForYou.com. Note how many MPs always ‘strongly agreed’ that more public money should be spent on a service in their constituency. ie, from someone else to their area:

http://www.guardian.co.uk/politics/person/1667/tim-farron

Turkey’s don’t vote for Christmas, obviously, but what this does suggest is that the wider good of the public purse is still a secondary importance to whichever particular local bandwagon they want to jump on. ” ‘Cuts are fine, just not in our area,’ says local hard working MP,” etc.

@JohnLeston

“So, remind me, just what are the LibDems for? Why would I want one?”

Hopefully being in Government gives all LDs a chance to express this properly, and not have to rely on bar charts and tactical voting which accounts (despite the OP’s optimism) for a lot of the flakiness in LD support since Cleggmania, which I’m happy to say is not mirrored by membership levels which have gone up in that time.