Archive for the 'Pollsters/polling' Category


Today Corbyn’s Labour lead the ABC1s by 12%

Friday, July 7th, 2017

Corbyn’s Labour now have a 12% lead with the upper and middle classes, let that sink in.

I think whilst this might be an outlier it is probably a result of Mrs May and her close advisers decision to abdicate from focusing much/any time on the economy during the general election campaign lest Philip Hammond had a good campaign and made it impossible for Mrs May to sack Philip Hammond on June 9th.

The Tories must hope whomever succeeds Mrs May won’t make such a stupid and petty decision like that at the next general election, I suspect once the upper and middle classes realise what Corbyn and McDonnell’s economic and tax policies are, they’ll flee back to the Tories, even in 1997 (and on June 8th this year) the Tories won the ABC1s quite comfortably.



Toxic Theresa’s terrible polling continues as YouGov gives Corbyn’s Labour an 8 point lead

Thursday, July 6th, 2017

She must be facing her Leo Amery moment very soon if Labour leads like this become the norm or larger.

Surely the Tory party must realise that Mrs May is re-toxifying the Tory party and making Corbyn as Prime Minister inevitable, she needs to go now if there is going to be a Tory party worth salvaging.

Whilst Mr Corbyn must be feeling very happy, the Tories have portrayed him as an economic illiterate, terrorist sympathising, Trot, and yet he leads by 8%.



I hope this is not the first step in the state regulation of polling

Wednesday, July 5th, 2017

The House of Lords announced last week they would looking into opinion polls, The Lords say

The results of political opinion polls have become an increasingly prominent feature of British politics, caused by parties reacting to changing economic and social trends, the media seeking more frequent measures of parties’ standing, and technological innovation and falling costs in the polling industry. In the period between the 2010 and 2015 general elections, over 2,000 voter intention polls were conducted by polling companies—on average more than one poll per day for five years. Meanwhile, as the use of social media and online sources have increased, their influence on political discourse has also grown.

There are a variety of ‘official’ ways in which the general public is encouraged to get online to influence political discussions in Parliament. In recent years, there has also been a growth in the number of independent websites and digital publications which encourage political dialogue and support online campaigns. While both political polling and social/digital media undoubtedly have an effect on political discourse, relatively little work has been carried out to assess the combined effects of both. An ad hoc committee might therefore wish to consider how the two factors interrelate, as well as how each issue affects politics on its own.

One of Lords involved in this is Lord Foulkes who has in the past ‘has tabled a private member’s bill bringing the multi-million-pound industry under the control of a state-appointed regulator rather than the current self-regulatory body, the British Polling Council. The state regulator would be empowered to consider whether polls should be banned in an election period.’

State regulation of polling is a huge concern for me, I fear it will lead to pollsters being unable to present their findings that turn out to be accurate, for example general election polling by Survation or YouGov’s model would be impermissible with this idea.

As for banning polls during an election period, that would likely lead to inaccurate and ill informed speculation and hints which would help nobody, least of all the voters or the political parties.




The scale of LAB’s lead in the parliament’s first polls is unprecedented

Monday, June 26th, 2017


Never before has main opposition party had such margins after an election

We have now had three voting polls since the general election and all of them, as can be seen in the table above, have shown clear leads for Labour.

This is highly unusual and almost unprecedented. Almost always the first polls after a general election see the winner doing better than it did in the voting on the day.

Thanks to Mark Pack’s excellent Pollbase place we can ascertain that there has only ever been one case before of the main opposition party beating the election winner in the early polls of a new parliament.

The exception could bring some cheer to the Tories because Margaret Thatcher’s Conservatives struggled in the early days of her Parliament after her 1979 General Election win. Then Gallup and NOP recorded LAB leads of up to 1.5% in the first surveys and we all know that Mrs T went on to win a landslide four years later.

But 1.5% is nothing like the scale of the first three voting intention polls of this parliament which have seen gaps of 3-6%.

My reading is that Corbyn is still benefiting from the sheer shock of the June 8th result partly because most of the pollsters got it so wrong. If all of them had been producing numbers like Survation then my guess that the impact could have been less.

We move on this week to the vital vote on TMay’s Queen’s speech and what has actually been agreed with the DUP.

Mike Smithson


A fortnight on from the eve of GE2017 and a look back at those final polls

Wednesday, June 21st, 2017

They all understated LAB

Mike Smithson



The polling numbers that should really scare the Tories – the oldies are abandoning Mrs May

Tuesday, June 20th, 2017

The narrative of what drove the shock result in the general election is becoming well established. Those in the younger age segments turned out to vote on a scale that hadn’t been anticipated and they were much more pro LAB than CON.

The result was that instead of losing seats to CON Corbyn’s LAB made gains off the Tories, the SNP and the LDs on a scale that alongside a handful of LD gains from CON caused Mrs May to lose her overall majority. This was in spite of the blues making gains off the SNP north of the border.

There’s one group of voters that traditionally the Tories have been able to regard as bankers – the oldies, those in the growing 65+ age segment.

In its only published post GE17 polling YouGov asked its favourability questions on parties and party leaders. These found TMay in serious negative territory almost across the board and almost on par with Corbyn at his worst.

My analysis of that, reflected in the chart above, finds that the PM is struggling to hold onto to the oldies. The chart numbers are based on subtracting those who have an unfavourable view of TMay from those who have a favourable one.

As is shown she had been doing extremely well with the 65+ group with huge favourability margins. In the latest polling she’s still in positive territory but only just by just 4 points. It used to be 55%

This doesn’t bode well for the blue team if there is a new general election which the party’s precarious parliamentary position might well lead to.

She used to be an electoral asset. Now it is looking like she is an electoral liability.

Mike Smithson


Those who rubbished Survation and the YouGov model look pretty stupid now

Saturday, June 10th, 2017

The finalisation of the count in Kensington means that we now can compute the final actual voting share percentages for GE17 the ones against which the pollsters should be judged. These are for GB only excluding Northern Ireland.

What is very striking is how understated the Labour share was with some huge variations in the wrong direction. Those pollsters that had developed turnout models following GE2015 got it most wrong.

In the final few days it was very noticeable how polls that didn’t fit the prevailing CON landslide narrative came under ultra criticism from those who did not like the numbers.

Remember that Survation poll that was attacked for the numbers sampled who had said that they had watched the Question Time special with Corbyn and TMay.

This and other polls that recorded highish turnout rates from the younger generation were fiercely criticised. I received hosts of attacking Tweets even for having the temerity to circulate poll numbers from Survation and YouGov.

The prevailing CON big majority narrative was overwhelming and some leading pundits now have egg on their faces for following it.

Congratulations to Survation and those behind the YouGov model who are the polling heroes.

What happened on Thursday means that LAB has been understated in two of the past three elections.

Mike Smithson


At least one of the final polls, surely, will have got GE2017 right?

Thursday, June 8th, 2017

No one can accuse the pollsters of herding this time

With just one firm still to publish, Ipsos-MORI for the Standard, the above Wikipedia list looks like the almost final polling table of 2017.

The variation between the firms is simply amazing and unprecedented in any previous general election. One thing is for certain some reputations will be made tonight and some will be trashed.

In many ways I admire the bravery of those pollsters who have not felt the need to follow the main pattern. ICM’s Martin Boon has remained very strong in his defence of his methodology as has Damian Lyons Lowe of Survation – the major firms at opposite ends of the divide.

Reputations are set to be made or broken as the results come in and I admire the strength of character of those not prepared to follow the trend.

This is a sharp contrast with two years ago when Survation refrained from publishing what would have been the most accurate poll of the campaign. They did this for fear that they were so out of line. As it was they had got it almost completely right.

At the other end of the scale is ICM which over the decades had built up a reputation as the gold standard of polling. That was when they were a telephone pollster. Their GE2017 surveys have all been online.

Survation has used both online and phone methodologies at this election producing pretty similar figures.

Given the outcome of the election, a continuation of TMay’s CON government in some form, appears not to be in doubt for election geeks GE2017 is an intriguing battle of what is the best way of surveying political opinion.

Well done to ICM and Survation. You might both be wrong and the answer lies somewhere in the middle but you were both brave enough to stick by your guns.

Mike Smithson