Archive for the 'Pollsters/polling' Category


Corbyn would be taking a huge gamble going into an election so out of step with LAB voters on Brexit

Friday, January 11th, 2019

Yesterday in what was billed as his “big Brexit speech” LAB leader Corbyn called for a general election should TMay lose fail to win backing for her Brexit deal in the vote next week. But he’s been far more reluctant to allow Labour to give any backing to the increasing clamours for a specific referendum on the deal.

As is widely known Corbyn has been anti the EU just about all his political career and he’s not going to change now – a position that could be very dangerous at a general election whenever it is held. For the vast majority of those who vote for his party have a very different view of the EU from him.

The chart above is based on data from the mega-poll with a 25k sample from YouGov that was published a few days ago. The question featured is how LAB voters would vote if there was a new referendum.

    I’d suggest that a party leader who is so out of line from what the bulk of his party’s support base wants is treading a very difficult path.

At GE2017 Corbyn was very much helped by the general perception tha LAB didn’t stand an earthly and he came under very little scrutiny. Next time that will be very different.

Mike Smithson


No Leader of the Opposition has rated even nearly as badly as Corbyn and become Prime Minister. An analysis into the satisfaction ratings of leaders of the opposition

Sunday, January 6th, 2019

In the final piece of three, Corporeal looks at the satisfaction ratings of Leaders of the Opposition

The Leader of Her Majesty’s Most Loyal Opposition is a position of great responsibility and impotence. It is traditionally the delicate art of attention grabbing, agenda setting, holding the government to account, and providing an inspiring alternative vision for government on the major issues of the day. Or if all that fails (unkind commentators might suggest that not all the holders of the office have achieved all of those objectives) at least try not to get people to remember to hate you more than the Prime Minister.

Current Situation:

Jeremy Corbyn’s most recent rating was -32, following scores of -31 in October and -42 in September. This is a return to normality for him with about half his ratings being between -20 and -40. The only sustained periods outside this less than ideal range were his first six months, and the second half of 2017 (he started rising in March, peaked in July after the election and slowly slid back down to his current level.

That slight upward shift does have one comforting note, it means for the first time in six months he is not the lowest rating Leader of the Opposition but instead is a comfortable one point ahead of William Hague after the same length of time in office. Here’s a celebratory graph of his time in office:

Historical Context:

Ratings for Leaders of the Opposition have tended to be rather less predictable than Prime Ministers but over a narrower range. They generally don’t get as popular or unpopular but bounce around in a narrower range, fuelled by lower total response rates. Iain Duncan-Smith never had more than 71% expressing an opinion validating his ‘quiet man’ nickname. Corbyn has never rated below 69%, a higher low than anyone but Thatcher (and we don’t have data from the first part of her tenure).

Corbyn has spent most of his tenure battling Michael Foot and William Hague for the bottom spot in the rankings. He and Hague share the dubious distinction of being the only ones never to record a positive rating (Foot was saved by a solitary positive rating of +2 in his first month). Here’s some lines and numbers:


Corbyn’s high ratings mainly show themselves in higher than usual dissatisfaction ratings. He is mostly at the lower end of average in satisfaction ratings, and second to last in the dissatisfaction ratings (behind only Foot). Even during his 2017 peak  he never had lower than a 45% dissatisfaction rating.

On a more positive note for Corbyn is the 2017 election where his ratings spiked to a remarkable degree. To say that it was the greatest improvement in the run-up to an election doesn’t do it justice. He gained 30 points (from -41 to -11) from March to June (and peaked at -1 in July after the election) with roughly equal improvements in his satisfaction and dissatisfaction ratings. The next highest gains over comparable pre-election periods are in the mid-teens (+13 in ’83, +15 in ’92, +16 in ’15). How much this is due to Corbyn, and how much due to Theresa May is an open question.

Here’s some data represented in a chart


No Leader of the Opposition has rated even nearly as badly as Corbyn and become Prime Minister.

(The next lowest is a one-off -22 for Cameron in September 2017, then Thatcher at -15 in March of 1977).

Oh Jeremy Corbyn: greatest campaigner in history.


Here it is Thatcher’s term where we have limited and patchy records.

The mean change in score for February-May in the years before and after elections was -1.5


In the event of an election it’s certainly plausible that Corbyn will show another surge, but it seems likely that he has a ceiling in terms of people he can attract. If there is an election in 2019 we may see another hung parliament if he can’t enthuse more voters, if there isn’t then you start to wonder how long the Labour party will feel without improvement in the polls before they start to get restive.



Theresa May is more popular through the first thirty months of her tenure than Thatcher and Cameron. An analysis into Prime Ministerial satisfaction ratings

Sunday, January 6th, 2019

In part two of three, Corporeal looks at Prime Ministerial satisfaction ratings

Prime Ministers are, of course, towering figures in public life. Pillars of UK society that are respected and loved in equal measure and enter government with the goodwill of the nation behind them. Then with fairly predictable regularity they slip from the hearts of the public and in some cases end up getting burnt in effigy.

Current Situation:

The most recent result for Theresa May came in mid-December and landed at -22, which is both a large up-tick and a surprisingly good result. It’s a ~10 point jump from where she’d been polling for the three months prior to that, if it turns out to be more than a blip the it slots her into (a distant) third place rankings-wise. If she falls back to the negative thirties she’d still be sitting at a similar level to Thatcher, Major, Brown, and Cameron at similar points of their tenures. Here are some lines and numbers.

Her fall really started in the month before the 2017 election. A month before the election she was +20 (with similar scores in the months before that), a week before she was -7, a month after she was -25 and has been bouncing around that level ever since.

Historical Context:

The traditional British maxim is that all political careers end in failure, and there is a certain brutal familiarity to the ratings trends. Most PMs peak very early on with good ratings based on fewer people providing negative responses. Through their first year these negative responses return with predictable (inevitable?) regularity. The main exception to that is one Margaret Thatcher, who has by far the worst scores for the first six months but made it back into positive figures three times later on (end of the Falklands war in ’82, and around the ’83 and ’87 elections). Blair (essentially all of ’01) is the only other Prime Minister to have a sustained positive period after three years in office. So a resurgence is pretty unlikely, on the brighter side for May, PMs have been this unpopular and continued on for many years.

As with the government ratings, I looked for an election time bump and while there is some sign of it (mean gain of 2.5) only half were ultimately positive and there’s a lot of variation going on. May’s performance in 2017 is by far the worst of any recorded here. For an election held on the 8th June her scores were:

March 14th: +13

April 25th: +19

May 17th: +20

June 1st: -7

Giving an overall rating change of -20, with the next worst being John Major in 1992 with a change of -8.


The ‘high’ line is almost entirely Blair (with a single appearance from James Callaghan), and his early popularity is generally a level above everyone else. Theresa May is actually 5 points above the median score but 5 points below the mean and the main reason is Blair pulling the mean up by about 6-10 points. His later ratings drag him down until he still ends up with negative averages across his entire tenure (as does everyone but Callaghan). He was really really popular until he really really wasn’t.

High response rates almost always result in poor ratings, the only real outlier in this case was Margaret Thatcher who managed multiple positive scores with huge respondent rates (around 90 is a normal settling point, above 95 is unusual) including possibly her definitive score of 50% satisfied, 49% dissatisfied in February of 1982. Everyone knew where they stood on the most divisive PM of recent history.

At the other end of the chart low total response rates are usually driven by low negative responses and point to great net scores (like Blair’s 65% satisfied 5% dissatisfied rating of May 1997). Gordon Brown is our standout here in somehow managing to churn out repeated negative ratings. There were 31 results with 87% or less giving an opinion, 25 positive results, and Brown with 6 negative ones. He was disliked at normal levels, but had the least enthusiasm behind him of any Prime Minister here. At no point were more than 44% satisfied with him (the lowest by far). No flash, but not much to smile about either.


These crazy stats that show Theresa May is more popular through the first thirty months of her tenure than Margaret Thatcher and David Cameron.

Theresa May: Worst election campaigner ever.

(Maybe this needs work, I’d click on them though).


A standard here, Callaghan’s patchy results make him hard to measure against (although he seems to hold up well from what we have, and if his first year was even averagely popular then he’d do even better).

I looked at the ratings movement in years outside elections from February to May and it came out with a mean change of -3.5 which would put the relative over-performance in election year scores at around the same level as with government ratings. But it still feels noisy.

Most total response rates settle into the low 90s after about the first year, my suspicion is that early satisfaction ratings are more important than early net ratings but I haven’t (yet) done the work to see how predictive they are since even I have a limited desire for spreadsheeting.


Theresa May is not popular (all this work for such great insights) but compared to her predecessors she’s rating pretty well. I’m sure this will be of great comfort to her in the times ahead.



The May government’s net satisfaction ratings are on par with Thatcher’s and Cameron’s

Sunday, January 6th, 2019

Only the Blair government’s net average ratings were better

Governments are pretty much always unpopular. The anonymous grinding mills of policy and administration rarely inspire enthusiasm, much less devotion. Democratic Rep. Emanuel Cleaver once joked that the US Congress was less popular than Satan, a claim that as far as I can find hasn’t been tested by any pollster (perhaps something for Lord Ashcroft to consider during his American travels).

In 2013 during a government shutdown the great American public earnestly declared to Public Policy Polling they preferred root canals, head lice, and venereal disease to their legislative representatives.

The UK government is not a direct comparison to Congress, and we don’t know how they would do in a direct popularity face-off with cockroaches and colonoscopies, but it has a long history of low ratings. The mean net rating of the results was -29.6, paired with a nicely matching median of -30 and a beautifully close to flat overall trend-line (that was very slightly positive).

Current Situation:

The most recent government satisfaction net rating is a less than spectacular -45, which is a solid uptick from the -52 recorded in September which was the lowest it’s been under Theresa May’s premiership (having fallen from -4 when she started). It’s in what we could see as the third distinct section of the net rating (these divisions seem arbitrarily neat but by signal or noise I think they can be seen).

In her first year of office it bounced between -1 and -20. In her second year it bounced between -27 and -41. In the four results since we have scores of -47, -52, -48, and now -45. Here’s a chart to look at:

Historical Context:

As bad as it looks this is not that unusual for government satisfaction ratings. The ratings show precisely one sustained period of net positive approval for the government, Tony Blair entered Downing Street on a wave of popular approval with a +37 net rating that slowly wore down into negative figures at the end of 1999 and remained almost completely (a brief post-9/11 period excepted) negative after that.

Here’s a second chart to enjoy:

As you can see the satisfaction ratings under May are pretty much par for the historical course. As much as the daily news sends headline writers looking for new ways to say omni-shambles (maybe Armando Ianucci has a spare moment between films to create a follow-up) the public has what we might call historically normal levels of contempt for their government.


The key to high net ratings is mainly avoiding too many people hating you. Blair’s +37 debut net rating came on a 46-9 split (his record-setting high of +38 on a 53-15). Thatcher, Callaghan, and May all managed 46 satisfied ratings, with net scores of 1, 0, and -1 respectively. Positive net ratings cluster around lower numbers of people expressing an opinion. Once you get to 90% response rates everything is very negative, with a few exceptions that can be mostly explained by exceptional circumstance (Thatcher and the Falklands, Blair after 9/11).

The highest response rates are dominated by the lowest scores, familiarity may breed contempt but contempt clearly breeds certainty. Which means plenty of representation for Brown and Major, the two lowest scoring PMs by far, neither of whom benefited from the clean slate start enjoyed by Cameron and Blair (but not, notably, by Thatcher). Measured against the other replacement leaders (is there a better phrase?) Theresa May’s government ratings are holding up well.

The worst scores can be found neatly in a single result. December of 1994, John Major registered an 8-86 split for the lowest satisfaction rating, the highest dissatisfaction rating, and the worst net rating (-78). He owns the lowest 39 ratings, with no-one else falling below -63.

Here’s a more colourful way of looking at it:


It’s government satisfaction ratings, short of printing it on a model’s underwear it’s never going to be eye-catching to anyone who doesn’t have an unhealthy relationship with spreadsheets and Joe Twyman’s twitter page.


In one of my many fun-filled evenings spent with my spreadsheets I had a brief peak of excitement when I noticed that May’s current ratings were approximately as a low than the final ratings of all the Prime Ministers  rated apart from Thatcher. But after a celebratory cup of tea I dug a little deeper and found that was the result of repeated bumps in popularity for the government in the run-up to elections (mean gain of 7.4 in the four months prior to an election and only one negative score in ’01). I took a look at the government performance in the run up to the normal election months (May/June in the years before and after each election and found very little average change (mean gain of 1.1, and that’s largely driven by 1982 and the impact of the Falklands War). But take the election effect with its own caveats.

The intermittent polling and lack of early scores on Callaghan (his first one comes at 12 months into his premiership) makes his scores a little enigmatic.


Theresa May’s government is about averagely awful in the eyes of the public. Either the last 40 years of British governments have been almost entirely awful, or the British public loves to hate its governments. Or both. Thank goodness they have the money, power, and status, to help them cope with it.



Since the end of October only one survey from a pollster other than YouGov has recorded a CON lead

Tuesday, January 1st, 2019

All the others have had LAB leads or were tied

There’s a new poll reported in the Times today from YouGov which has a Conservative lead albeit a reduced one of 2%.

It has not yet been added to the Wikipedia table featured above of every published poll. When looking in detail at the list one thing is very striking and that is that the Conservatives leads are almost totally from YouGov. All the other polls bar one in early November from Kantar are either showing ties or LAB leads..

Because YouGov publishes more polls than anyone else this might be giving a distorted view of public opinion during what is quite a critical time politically in the UK.

At one stage I used to keep a table showing just the latest poll from each pollster and it might be a case for doing that to get a better and more rounded impression of what is going on.

Nate Silver in the US always talks about the “house effects” of different pollsters and records how particular firms can be out line with the others.

We saw before GE2015 how some pollsters would always record larger UKIP shares than others and there are many examples of what can be termed house effects.

Mike Smithson


Looking back over 2018: Alastair Meeks reviews his predictions

Wednesday, December 26th, 2018

It’s that time of year when everyone makes predictions about what is coming up.  Oddly, you get rather fewer articles looking back at the previous year’s predictions. Funny that.

I make predictions each year to set out my considered thoughts and to see where they went wrong, in an effort to see where I can get better when assessing scenarios for my betting. So how did I do twelve months ago?

For once, not too badly at all. I made four broad predictions, as follows.

1) There will be a Brexit deal, substantially on the EU’s terms

This is what I wrote:

“Theresa May is using tactics without strategy. Politically, that serves her quite well, even if it isn’t good for the country. Despite caving in on more or less everything, the media coverage of her initial deal was excellent and her opponents on all sides were discomfited. Aside from a few rumblings from those Leavers whose preferred version of Brexit would be tectonic, she carried all before her at home.

We can expect to see the same trick repeated. Since the government has no strategy and no deal is worse than a bad deal, a bad deal will be done, substantially on the EU’s terms. This time the risk of hardliners opposing the deal will be much greater. There seems, however, to be a majority in the House of Commons for a bad deal. So I expect that a bad deal will be done and Theresa May will again look like a winner.

Much of the year will be taken up with alarums, excursions and brouhahas on the Brexit negotiations. We should ignore them all.  We won’t.”

That stands up quite well. A deal was negotiated. It is a bad deal and substantially on the EU’s terms. OK, it hasn’t been agreed and it is already unravelling. You can’t have everything.

Why hasn’t it been agreed? What I got wrong was the intransigence of Leavers. I thought that they’d decided to compromise when they accepted the preliminary agreement last December. It turns out they simply hadn’t understood it. This gave Remainers the cover they needed to oppose it too. Brexit now looks in serious peril in a way I could never have imagined a year ago.

2) The party leaders will stay the same

This was a braver prediction than it looked at the time and I knew it. Theresa May spent most of the year fending off a challenge. Lib Dems have been predicting Vince Cable’s retirement ever since he took the job on. Even Jeremy Corbyn has had a more fraught year than he would have hoped for as some of his past activities have left him with a lot of explaining to do, explaining that he has been notably ineffectual at.

I will not be making the same prediction this year.

3) There will be more Cabinet departures

There were. Aside from the three who left in the January reshuffle (one of whom, James Brokenshire rapidly returned), five more fell by the wayside, four of whom did so over Brexit.

The long continuity of ministers under David Cameron is a thing of the distant past. The Cabinet is held together by short dowels rather than screws and nails.

4) Labour and the Conservatives will remain roughly neck and neck in the polls

I give myself full marks on this one too. The most recent poll from Opinium had both main parties on 39%. You can’t get much more neck and neck than that.

However, this is what I said about the referendum polling:

“I’m also not expecting to see too much change in the polling on the referendum decision. If a deal is concluded, as I now expect, and the economy does OK in 2018, Leave might well start to pull ahead a bit – maybe as far as 55:45. But the only concessions that the government has made to Remain supporters have been extracted by the EU.

If the government wants to start converting Remain supporters in numbers it is going to need to show that it can include their values in its vision of Brexit. Since it isn’t even trying, we can expect a hardcore group of Remainers for the indefinite future. Christmas 2018 is likely to have just as many family arguments about Brexit as Christmas 2017 and Christmas 2016. Happy New Year!”

Contrary to my expectations, Leave have been unable to convert their institutional advantages into a lead in the polls over Remain and as I write Remain’s lead is heading towards 55:45. However, I was right about the continuing lack of harmony between Leavers and Remainers. Positions are getting steadily more entrenched and becoming a part of many people’s personalities. That in turn will have a big part to play in how 2019 turns out. I shall turn to that next.

Alastair Meeks


Tonight’ big Brexit polling news – LAB could slip to third place if it helped CON pass Brexit

Saturday, December 15th, 2018

The S Times is reporting a YouGov survey of 5,000 voters, commissioned by the People’s Vote campaign, showing that support for LAB could fall from 36% to 22% if it helped the Tories to pass a compromise deal with Brussels like the one advocated by Theresa May.

In these circumstances, the LD would move from 10% to 26% — their highest rating in any poll since GE2010.

    I should say that I am generally sceptical about polls of this nature especially when they are commissioned by a campaign group. When you ask voting questions based on “what if” scenarios it is asking a lot of the process to come up with precise measurements as we see here.

But there’s little doubt where the vast majority of LAB voters stand on Brexit and that is some distance from what Mr. Corbyn believes. Because everything has been focussed on the Tories LAB has got away with relatively little scrutiny.

By 68% to 11% by voters generally didn’t believe that that Corbyn could get a better Brexit deal if he were PM. The split amongst LAB voters was by 47% to 30%.

In the betting the Betfair exchange price on a second referendum before the end of next year is 40%

Mike Smithson


Latest YouGov tracker finds the Brexit “wrong” lead over “right” in double figures at record level

Wednesday, December 5th, 2018

With the political process of the UK leaving the EU completely dominating the headlines the latest YouGov brexit tracker has unwelcome news for those who want to follow the referendum result.

The figures are in the chart above. 49% believe that in hindsight it was wrong to vote to leave the EU with 38% saying it was right.

The question has been asked several times a month by YouGov in exactly the same form since the referendum in 2016 and this, by far, is the biggest lead lead for “wrong” and it’s the first time that those believing that Brexit was right is below 40%.

This trend chart from the Times has just been published.

This will give heart to those MPs at Westminster who are pressing for a second referendum.

Mike Smithson