Bob Worcester on the MORI 28% Tory lead
September 19th, 2008(This is the second guest slot that Bob Worcester, the founder of MORI (now Ipsos-MORI) has produced for PB in which he reviews the polls from his experience of nearly 40 years conducting and observing polls and how they are reported.)
“Brown doomed as Tories head for 330 majority”
Have I got news for you, as they say, especially directed at the sub-editor on the Express who used the latest Ipsos MORI poll showing a 28% Conservative lead with the Tories at 52% share to write the ludicrous headline above over the story on 18 September by their political editor, Macer Hall. In his otherwise excellent copy, he talked about the ‘forecast’ represented by our poll. Have I got news for him, and many other journalists, print and broadcast, who so mislead their readers, listeners and viewers?
Polls don’t ‘predict’, ‘forecast’, or show anybody about to ‘head for 330 majority’ at the next general election, probably 20 months hence. Polls measure, within statistical reliability of c. plus or minus 3% how each party stood at the time the fieldwork was conducted.
Pollsters don’t read tea leaves, gaze into crystal balls, read politicians palm or the like, but they do read polls pretty carefully, and pay close attention to what people say in focus groups and other means by which they gain their information, and if they are any good, they rely on objective and systematic, hopefully reliable, information to inform them what the mood is at the moment, and if they take it upon themselves to forecast, they do so with good humour and the knowledge that inevitably they’ll be shown up at some point, and with the knowledge that the media will jump on it and not let them forget it.
I’ll never forget blowing a by-election years ago, and when confronted by the inevitable demand by the media to fess up to my mistake, said at the first question from the presenter, Peter Jay it was, “I blew it”. After the filming, he shared with me the list of half a dozen follow up questions he’d prepared to get me to admit just that, and laughed that I’d left him with not a lot to question, especially since his next question was ‘why?’, to which I replied that it had only happened the day before, and we didn’t have a clue, but would turn every stone until we knew exactly why we’d failed to reflect the likely outcome from a poll published the day before the by-election (not 20 months hence).
This month’s findings, was taken the weekend that the Labour Party was reeling from the first of its series of defections and resignations, calls for a leadership contest, and general malaise with Gordon Brown’s leadership, at the end of the lacklustre Liberal Democrat conference.
With political news ‘hot’, and before the worldwide financial crisis took over the front pages and lead items on the broadcast media, the poll received widespread coverage and comment. Much was made of the fact that the Tory lead cracked the 50% ceiling for the first time in 20 years, when in August 1988 the Tories reached 51%. But what happened then? Just over a year later, Mrs Thatcher lost her leadership contest, and John Major’s early lead slumped and he barely won the 1992 election. In fact, if one person in two hundred across the country who voted Conservative had voted for the second party in their constituency, it would have been a hung parliament instead of the 21-seat majority he received.
The Daily Record’s sub-editor who wrote the headline saying ‘Poll leaves Labour facing a wipe-out’ should also be listed in the “Sinners List”, as it is no more true necessarily now than it was in November 1981 when Mrs Thatcher was the least popular Prime Minister and the Tories tied a 27% share with Labour’s 27% and the Liberal/SDP Alliance went to 44% in the MORI poll (and to 51% in Gallup’s), yet she went on to win her famous 1983 general election victory when the Conservatives got the 44%, Labour was unchanged at 28%, and the Alliance came third, with 26%, a 9% swing in under two years (of course the Falklands War intervened.
There’s much more in the poll to interest, and indeed surprise, the careful reader. That for instance a majority of the public agree that Gordon Brown is doing a reasonable job in difficult circumstances, that by two to one people say they would prefer a Prime Minister who mainly acts on the views and opinions of the general public to make decisions (64%) to a Prime Minister who mainly trusts his own judgment and experience to make decisions (32%).
Finally, one journalist seemed to be puzzled by the fact that the bid jump in the Tory lead was mainly at the expense of the Liberal Democrats rather than Labour. Have I got news for him? Yes, it’s because the 16% lead among the 81% who might vote does soar to 28% among the 55% who say they are ‘absolutely certain’ to vote. This illustrates how many ‘natural’ Labour supporters are saying now they’ll not turn out to support Brown when the election does come. Labour’s 24% standing in the impossibly high 81% rather than at the 29% of certain voters at the 55% turnout represents my “30%/30%/20% rule”, that the core vote for Labour is about 30%, the core vote for the Tories is about 30%, and the core vote for the t’others is about 20%, leaving the 20% of ‘floaters’ to decide the outcome and size of majority of every general election.
Keep taking the polls…
Bob Worcester
Ipsos MORI interviewed 1,017 British adults 18+ by telephone 12-14 September 2008. Data were weighted to reflect the adult British population.
MessageSpace Advertising
Thank you Bob.
Perhaps you’d give us some thoughts on the US election before polling day.
…………………………
Latest Gallup Tracker :
McCain 44% .. Obama 49%
http://www.gallup.com/home.aspx
- “Ipsos MORI interviewed 1,017 British adults 18+ by telephone 12-14 September 2008. Data were weighted to reflect the adult British population.”
It would be more accurate to say:
- Ipsos MORI interviewed 1,017 adults in Great Britain 18+ by telephone 12-14 September 2008. Data were weighted to reflect the adult population of Great Britain.
1. NI was not included (it never is)
2. Not everyone interviewed was necessarily British
3. I assume that the weighting was done to GB demographics, not UK demographics
Thanx Bob. But with a survey of just above a thousand and on the phone at that how accurate would it really compare with an Internet survey?
I do believe that internet replies are far more revealing than answering on the phone; the questionee has more time to consider the question, and as a result, more accurate. Furthermore, the number interviewed are usually considerably increased.
The above does not mean I support any particular polling company.
Bob , I agree with your comments it is one poll subject to usual M of E , however with respect one of your specific comments is incorrect . The poll was not taken at the end of a lacklostre LibDem conference it was raken at the weekend whilst the conference was getting under way .
How likely do you think a hung parliament is Bob?
I suspect Labour will get over 30%, and the Tories under 40% at the election, and if Brown isn’t leading Labour it could be close still.
I’m quite confident there will be a hung parliament, with perhaps a 5% swing to the Tories. Tory by-election performance is vastly improved, but not consonant with the huge poll leads. Crewe was no Dudley (either 1968 or 1994) for instance…
re 3. Only 70% of adults have internet access which makes it harder to find a valid sample.
Excellent article, Bob. Thank you very much.
Rumours of a Labour wipe out much exaggerated then?
Our ‘Ave It’ will be mightily disappointed, though doubtless he will take issue with you in his own inimitable way.
Yes, many thanks Bob for taking the time and effort to write at such length. Thanks must also go to Mike for arranging this, demonstrating again the considerable influence reputation and influence of PB.com.
Who knows, one day even Nick Clegg might be persuaded actually contribute something worthwhile to the site instead of just going through the motions.
This is a great article, and should be a must-read for all political journalists.
I have one question, which is on the final point, about the extra lead coming from the Lib Dems. Is Bob suggesting that Labour’s vote simply won’t turn out, but that the Lib Dem vote is essentially waverers who have in this case gone from Tory to Lib Dem, or is there a likelihood-to-vote variance within the Lib Dem numbers that is creating the appearance of a LibDem-to-Tory shift, whereas in fact it is indirect movement via the Labour Party exacerbated by likelihood-to-vote?
Great piece.
5. Have we had another English Parliamentery by election since the Crewe by election? Isn’t there a risk that your putting too much emphasis on that one result? What if there was another english by election next week, and the Tories achieved a 21% swing, as opposed to the 17% they got at Crewe? Would that mean Cameron is heading for Downing St? Or would it just mean that in that one constituancy the Tories achieved a 21% swing and thats all?
Thanks to Bob for the article.
4. Yes it was before the Nursery Rhyme Gaff and the Pension Gaff!
With regard to the thread:
It is an interesting article and polls are important, one must question how, given the different polling methodoligies the LD’s can get 12%, then 20% or the Tories 52% and then IIRC 44%! Bit too much volitality there - either way it points to a crushing Tory victory and at least a third of LD MP’s not being returned.
Some LD’s think the 20% poll means they have gone from the mortary to the recovery room, I would not get your hopes up to much.
10. Perhaps, but we’ve had Henley and Glasgow East, which saw nothing spectacular for the Tories…
12. Well, we know the Tories are nowhere in Scotland and probably never will do any good there (certainly not until after seperation)
As far as Henley goes, that was a Conservative heartland that had been won previously by a huge personality/cult politician. The only thing that could happen in that seat was that the Tories vote could go down.
As far as I can see, we simply haven’t had enough by-elections in seats like Crewe, to say where the Tories are standing in actual polls in swing seats. Its may be the 17% they got in Crewe is where they are. It may be they are doing better or worse than what they polled in crewe. We smily haven’t had enough elections to be able to form an opinion. If only more politicians kicked the bucket like they used to years ago…
The only problem is that the Lib Dem conference was by no means “lucklustre”. It was very good indeed.
The Lib Dems are improving their popular standing, as shown first by Luntz (ugh!) and now by You Gov. The most recent Mori poll was seriously flawed, as was discussed on a previous thread.
14. Objectively, the Clegg speech was recieved objectively; to some outlets slightly positive. It would not explain a 8% jump in the polls (Nearly double the 12% reported a few days before). I think it can be put down to different polling methodology. Best just to take an average of the last 3 or 4 polls.
14. Gifford the Mori poll which appropriately for the first time allowed for public/private sector split was not necessarily flawed. Similarly we don’t know if the latest poll is flawed or if the higher LD share is largely due to Margin of Error differences etc. We also should not treat a “one-off” Luntz session with carefully selected participants as representative. In short we need more evidence before drawing conclusions as to whether or not the LDs have improved their so-called “popular standing”.
Thanks for the interesting comments, Bob - the ‘doing a reasonable job under difficult circumstances view (held especially strongly among people certain to vote) sits oddly with the 28% lead (not to mention the ‘everyone hates him’ stuff we get on this site).
If the LibDems can consistently poll 20% this far out from the general election while the public is only just getting to know their new leader then I’ll take odds on them gaining 20 seats.
I can’t believe the Conservatives will be able to keep up the level of their support when they are being consistently outshone on the economy and have yet to publish any policies.
Worse for Cameron is the fact that things can’t get any better for him - he has peaked too early!
That is, unless they can publish a series (not just one) of comprehensive new policies to public acclaim, which also subsequently stand up to the increased level of scrutiny they should expect from any prospective new government. I doubt they can.
So those pesky LibDems must be feeling quietly optimistic for their own fortunes given that the race is only just starting to warm up.
Am I correct in my thoughts that Brown/Darling seem to have played a pretty straight bat over the last few days -certainly the contrast with Major and particularly Lamont`s arrogance during the ERM crisis is quite stark .
A much better article than the previous one by Bob Worcester which hardly told us anything we didn’t already know.
Re 3: Internet polling also allows for people to click through their responses as fast as possible. Is it only me that clicks through their YouGov polls as fast as possible to earn my 50p?