Archive for the 'Pollsters/polling' Category


Sadiq Khan 20% ahead of Goldsmith according to Survation phone poll

Friday, April 29th, 2016


So on the face of it a great poll for Sadiq Khan, Labour, Corbyn and those PBers who got on the Labour man at 33/1 in March 2103 when Henry G Manson tipped him.

But the survey took place before the blow up involving the last LAB mayor of London, Ken Livingston and phone polls have not had a good record on London Mayoral contests.

Of the 2012 final polls the least accurate was the only phone survey.

There’s also an issue that Labour voters might be less likely to turnout given it looks like a foregone conclusion.

Just in case I’ve been laying off large parts of my 2013 Sadiq bet at odds of up to 17/1 on Betfair so I’m certain of a substantial pay day next Friday whatever happens.

Mike Smithson


Next Thursday could start to restore our confidence in the polls

Thursday, April 28th, 2016

Exit poll

Alastair Meeks on the importance of the London, Scottish & Welsh surveys

The 2015 general election was a disaster for the polling companies. On the eve of the election, all the pollsters were predicting a hung Parliament with the Conservatives and Labour neck and neck. In the event, the Conservatives were 6% ahead of Labour and got an overall majority.

Since then, the pollsters have flagellated themselves, put on hair shirts and sought to uncover what exactly went wrong. They have conducted investigations, issued reports and held symposia on the subject. They have put in place corrective measures. But we don’t yet know whether the time for remorse is over. There remains a gnawing anxiety that the pollsters might still be getting it wrong.

The general election was not an isolated failure. The Scottish independence referendum polling was fairly uniformly 3% off, the same margin as the standard general election error. This wasn’t much noted at the time but if the error had been 3% the other way from the published polls, Scotland would now be independent. In a close referendum, accuracy to plus or minus 3% is not much use.

The EU referendum betting reflects that. The betting markets are apparently moving independently of any polling. Opinion polls are being treated as having junk status. This seems excessive. We can at least expect them to be giving us a sense of which way opinion is moving, even if their absolute accuracy is suspect.

In any case, we have an upcoming opportunity to calibrate their accuracy. The 5 May round of elections will allow us to see the accuracy of polling in Scotland, Wales and London on those elections. There has been plenty of polling of all three of these elections. So watch them carefully: they will be invaluable in helping us determine how effectively the pollsters have got to grips with their problems.

If the polls perform reasonably well against the actual outcome, take note. So if Mayor Khan has won a comfortable victory, Labour are left running a minority government against a Plaid Cymru opposition and the SNP increase their overall majority, perhaps it’s time to start taking the polls a bit more seriously again when placing your bets. After all, it would be a shame to be completely discounting a potential source of information, wouldn’t it?

Alastair Meeks


New ComRes phone poll has REMAIN retaining its 7% lead

Sunday, April 17th, 2016

EU Ref polling   Google Sheets

But there’s a sharp increase in don’t knows

Almost one of the constants of this campaign has been that the Inners are doing a fair bit better with phone polls than online and so the pattern continues tonight.

The first phone poll to be carried out wholly in April, by ComRes for the Sun, has REMAIN maintaining its 7% lead – a gap which is very much in line with the other phone polls that we saw at the start of the month.

The main difference with the latest poll is that there’s been a sharp increase in the number of DKs – up from 11% at the end of March to 17%.

While the phone surveys are showing margins of this scale it is hard to call this election for other than REMAIN though things can change.

Mike Smithson


The perceptions on the Tories and Labour

Sunday, April 17th, 2016

YouGov have published some polling, conducted within the last week on which groups the voters identify the Tory Party and the Labour Party with. The findings aren’t that surprising. The Tories are perceived to be really close to the rich, businessmen/The City, and voters in the south. Whilst Labour are seen as being really close to trade unions, the working class, and benefit claimants.

The most interesting finding from this polling was that the Tories are seen as being not close to older people, and that Labour have better net rating with how close they are to older people. Now whilst not every older person is a pensioner, you can make a strong case that the segment of society that the governments of David Cameron have looked after the most is older voters, particularly pensioners. Recently the work and pensions select committee announced that they would investigate claims that baby boomers get more out of the state than they put in while younger generations lose out.

I suspect the reason for this particular polling finding is that Labour is seen as being close to benefit claimants, as they have been for a while, and older people such as pensioners because as recipients of the state pension are seen as benefit claimants, whilst with the benefit cap, the Tories are seen as no friends of benefit claimants.

This shows that once again in politics sometimes perceptions matter more than the facts, especially when you consider that a month ago, Iain Duncan Smith resigned as Secretary of State for Work and Pensions ‘because he was frustrated that Downing Street and the Treasury refused to consider controversial cuts to universal pensioner benefits…Friends of the former work and pensions secretary said he was fed up of being asked “again and again” for cuts to working age benefits and those for disabled people, while the money spent on older voters remained untouched.’



YouGov: Tory voters most ready to change EURef vote when asked what they’d do if BREXIT would cost them £100 a year

Friday, April 15th, 2016

Datawrapper    UgHzg    Visualize

There’s a new YouGov poll out which has REMAIN back with in lead from the level pegging. Actual figures are 40% to 39% so all within margin of errot.

At the end of the survey YouGov posed this question.

“Imagine that if the UK left the European Union the standard of living would be lower and people would on average be £100 a year worse off. In those circumstances, how would you vote in the referendum: Should the United Kingdom remain a member of the European Union or leave the European Union?”

The chart shows the net change in the LEAVE position when the new information is added.

There was similar polling during the 2014 Scottish IndyRef which I recall had been prompted by Professor John Curtice of exit poll fame. In this case it would have been helpful to have the reverse question – what if REMAIN was ging to cost you £100 a year.

Mike Smithson


Why LEAVE needs to neutralise PROJECT FEAR if it is to win

Thursday, April 14th, 2016

Datawrapper    QtFA7    Publish

Worry about the unknown is a great vote driver

In the 2010-2015 period I repeatedly suggested that a good guide to the general election outcome was YouGov’s “who is responsible for the cuts” tracker. Throughout the entire five year period Labourwas blamed more than others and so it turned out to be.

For the referendum on June 23rd I have been looking around for another possible tracker and believe I have found one. It is featured in the chart above and is about the risk of leaving the EU.

As we get nearer to the day the IN campaign is going to focus on the risk of Britain taking such a step. We call it Project Fear which of course is what it is.

What’s striking in the subgroup splits in the chart above are the figures for Conservative voters. They split on voting intention to leave by a 3% margin yet on the risk/safe question
there is a 10% lead amongst those who think BREXIT will be risky.

No doubt the newly designated official body for the OUT campaign, Vote Leave, already has advanced plans on how it is going to deal with it. On this, I’d suggest, the referendum outcome will depend.

Mike Smithson


If there is a “Bradley effect” in the Mayoral race it’ll have a lower impact in London than elsewhere

Thursday, April 7th, 2016


There has been a fair bit of talked about a possible “Bradley effect in the London mayoral election on May 5th. This refers to the well observed effect of people telling pollsters that they will support a non-white candidate when in fact they don’t end up doing so.

There are two reasons to believe why this might not be significant in the London mayoral election even though the Labour candidate, Sadiq Khan,is a Muslim whose parents emigrated to Britain from Pakistan.

Firstly this is generally seen in phone calls where the interviewer effect can come into play. Well almost all the surveys so far who is London election have been online and so one would not expect the same impact.

The second reason why it might not be important is the nature of London electorate. It has the highest proportion of people who were not born in the UK of anywhere in the country and significantly large Asian and Muslim communities.

Interesting the latest ComRes poll, featured in the chart above broke down the sample by ethnic origin and the results are there to see. Zac Goldsmith is clearly winning the white vote but this is more than compensated for by Sadiq’s support from London BME voters.

A major uncertainty about the election and the polling is whether turnout is being properly reflected. This is of course the old polling problem and one that we are only too aware off because of what happened in May last year.

Another feature of the Com poll is that Zac Goldsmith’s background as an environmental campaigner does not appear to be winning Green party second preference votes. Of the 19 GRN first choice responders in the poll not one placed Zac as 2nd preference,

Mike Smithson


Trying to work out who will turn out in the referendum of June 23rd

Wednesday, March 30th, 2016

Datawrapper    QQeoI    Publish

New study tries to explain why phone and online polls are giving different EUref results

As we get closer to the referendum there’s a lot of effort going on to try look at the polling more closely so we don’t end with another GE2015.

The results of a Populus/Number Cruncher Politics study on the difference between the online and phone surveys was looked at on Newsnight last night and I’m hoping it will be possible to link to the actual document during the day. One of the conclusions relates to turnout and its link to demographics which is why I’ve featured the Ipsos-MORI GE2015 data in the chart above.

The notion that Leave voters are more determined to vote in the referendum, based on what Leavers have been telling pollsters, has been little questioned. But this conclusion is heavily dependent on sampling people representatively in terms of relative political interest. Turnout self reporting is known to be misleading this far out. If instead we consider high and low-turnout demographics, older people are more likely to vote to Leave, but more affluent people are more likely to vote to Remain than less affluent people. In fact if we compare General Election turnout between constituencies, we find that turnout is not higher in more Eurosceptic areas – in fact it is slightly lower.

Another feature identified in the report is that those who respond to phone polls are more likely to give socially liberal responses than those online. Maybe that is the interviewer effect. Thus 40% of the phone sample felt that racial equality had not gone far enough compared with 24.9% of the online group. This compares to 31.9% in the huge BES face to face study.

One suggestion the authors make is that polls might be weighted by social attitude with, presumably, the BES being the norm.

Mike Smithson