Archive for the 'Pollsters/polling' Category


Two new polls this afternoon with very different shares for LAB and the Greens

Friday, July 12th, 2019

If LAB had been doing better at recent elections then you might say that Survation has got this right. But Corbyn’s party had a miserable set of locals in May followed up by dreadful Euro elections three weeks later when they slipped down to just 13.6% and could not even maintain position as top party in London. They did, of course, hold Peterborough in June but with a share down 17% on GE2017.

At the Euro election YouGov got the Greens almost dead on and were within a point for the LDs and LAB. Survation had LAB on 23% compared with the UK actual of 13.6%. The firm put LDs on 12% compared with a UK actual 19.6%.

Survation was, of course, the most accurate pollster at GE2017 but, as we’ve seen, being top at one is not always a good guide to the future.

Mike Smithson


If Boris wins then last month’s ComRes poll suggests a huge CON recovery is in prospect

Saturday, July 6th, 2019

Telegraph June 11 2019

The possiblity of a 140 majority would surely tempt the new PM

There is little doubt that the Telegraph’s ComRes poll that appeared on June 11th played a major part in the Conservative leadership contest. It came out at a crucial time just as the MP rounds of voting were due to start.

As the headline above showed it pointed to a Johnson led Conservative Party winning a majority of 140 seats over LAB in a general election. The paper noted at the time:-

“Asked how they would vote if Mr Johnson was Prime Minister compared to his main rivals, 37 per cent said they would vote for the Conservatives, 22 per cent said they would vote for Labour, 20 per cent said they would vote for the Liberal Democrats and 14 per cent said they would vote for the Brexit Party.

Using these figures and taking out those who ‘didn’t know’ and ‘wouldn’t vote’, Electoral Calculus projects that the Conservatives would win 395 seats compared to 151 for Labour, 26 for the Liberal Democrats and no seats for the Brexit Party – suggesting that the former foreign secretary would have the ability to banish Nigel Farage’s party from Westminster for good.

None of the other contenders got anywhere near that sort of success in poll and the unequivocal message from the numbers and the way it was presented in the newspaper was that Boris was the man to lead the Tories to recovery and a substantial victory.

Now that the leadership contest looks a certainty then we’ll get an idea whether the polling got it right.

Certainly the experience of new PMs taking over during a parliament  is that their parties  had substantial polling boosts. We saw that with Major in 1990, Brown in 2007  and, of course TMay in 2016.

If ComRes was right then the new PM effect with Johnson should dwarf them all.  It could have an impact on the Tory performance in the Brecon and Radnorshire by-election that takes place only a week after the new leader will be in place.

Not long to wait.

Mike Smithson




The best test of a pollster is not how they’re currently doing against other firms but what happened last time they were tested

Sunday, June 30th, 2019

I am afraid that I have to disagree with David Herdson on his latest Saturday thread about YouGov understating Labour. Firstly you cannot judge pollsters’ based on their current surveys when less than 5 weeks ago they were tested against a real election involving real voters.

In the two charts above I compare LAB and LD vote shares for the May Euros in their final published polls.  Just two of them can claim to have come out of the election well with the rest trailing some way back.

Just examine some of the exaggerated figures that some pollsters were record reporting for LAB where we had a range from 13% to 25%. The actual GB figures was 14%.

Now look at the second chart showing the final LD shares. These range from 12% to 20%. The actual GB share was 20,4%.

Apart from Ipsos MORI and YouGov the rest really did rather badly.

Because of the low turnout, the 37% that actually happened was broadly anticipated, this was always going to be a challenging election for polling because turnout was everything. If one party’s supporters were less likely to vote  then that presents the pollsters with serious challenges .

The other challenge, of course, was tactical voting generally by remain backing LAB voters to the parties they saw as being most likely to succeed in their region and so the vote could produce the maximum number of MEPs. This helped the LDs and, of course, the Greens to achieve the success that they did. Whatever mechanisms YouGov and Ipsos Mori use they were able to detect better what was the big characteristic of this election.

So when I look at the current polls I regard Survation and Opinium, of the recent ones, as LAB over-staters.

Mike Smithson


The YouGov discrepancy: just how badly is LAB doing?

Saturday, June 29th, 2019


A clear lead or struggling to be neck-and-neck?

Three parties have dominated the coverage of opinion polling and major elections over the last three months. On one side, the Tories have clearly suffered a catastrophic loss, shedding more than half the support they had at the start of the year, losing more than 1300 councillors and then nearly all their MEPs in May. Against which, the Brexit Party has exploded out of nowhere to win the EP elections and to vie for the lead in Westminster voting intention, while the Lib Dems have dramatically recovered from their near-decade-long slump to gain over 700 councillors, more than double their poll share and win their biggest election since 1910 (London, in the EP vote).

What of the other party in the apparent four-way tie for the lead, Labour? Perhaps because their slump in support began earlier (early March, following the TIG defections), and hasn’t been quite so precipitous, it’s not been so well-observed either.

But just how big has that drop been? Here we arrive at a polling quandary. Present-day polling gives a disproportionate prominence to YouGov, who release far more polls than anyone else. This is unfortunate since their results are quite out of line with other pollsters.

For example, these are the average poll scores from YouGov so far this month:

Brx 24.3
LD 20.5
Lab 19.8
Con 19.3
Grn 9.0

Whereas the average shares from all other pollsters combined (and the relative difference against YouGov) is:

Lab 26.0 (+6.2)
Con 23.8 (+4.5)
Bxt 19.0 (-5.3)
LD 18.0 (-2.5)
Grn 6.2 (-2.8)

These are, clearly, starkly different splits. At the extreme, YouGov shows Labour 4.5% behind the Brexit Party, while other firms have them 7% ahead. Before anyone gets too excited about the implication for seats won implied in any individual poll, let’s remember that someone’s methodology for producing the underlying polling data is very wrong (never mind the methodology for translating the votes into seats – though that’s a discussion for another day).

In as far as we have anything to go on, the chances are that it’s YouGov which is out. At the EP elections, their final poll underreported Con and Lab against the actual results by 2% and 1% respectively, and substantially overstated the Brexit Party (by some 6%), which is very much in line with their Westminster findings compared to other companies (with the possible exception of Opinium).

For Labour, it may be cold comfort that they’re probably polling in first place with a mid-twenties share rather than in third, behind the Lib Dems and perhaps sub-20. After all, for the main opposition party to be in the mid-20s in any circumstance is extremely poor but it’s still not quite the existential crisis that not being the largest left-of-centre party is. Note also the much larger share that YouGov give to the Greens: almost half the Labour share, rather than less than a quarter of it that the other companies find.

Of course, these figures were before Labour casually revived media attention of their antisemitism problem by readmitting an unrepentant Chris Williamson into their party while the EHRC inquiry into it is still ongoing. That, plus the internal Labour criticism to it, might trigger a decline in Labour’s vote share after a month of relative stability, as might well the election of the new Tory and Lib Dem leaders – and goodness knows what might happen after the summer, when the Brexit drama reaches a new climax. Even if Labour is ahead now, there’s no guarantee it’ll last.

David Herdson


The three post Euros polls have had three different parties in the lead

Sunday, June 2nd, 2019


How the final polls did against the actual results – party by party

Monday, May 27th, 2019

As can be seen from the charts above the biggest problem the pollsters had was with LAB which had a huge range in the final polls. Partly, I suspect, this was that Labour supporters had not finalised their decision to vote tactically when they were being questioned.

It is well know that a significant proportion of voters really don’t make their minds up until the last moment. This is why Ipsos MORI generally try to ensure that its final poll before an election involves fieldwork carrying on until latish on the Wednesday evening.

Matt Singh of NCP explained in last week’s PB/Polling Matters Podcast how his final numbers had been impacted by a large proportion of his sample doing the online questionnaire on the the Saturday.

Mike Smithson


Remember that at the 2014 Euro elections YouGov, by some margin, was the most accurate pollster

Monday, May 20th, 2019

2014 Euros polling – Wikipedia

The others overstated UKIP lead by upto 7%

With polls coming thick and fast at the moment the one big trend is that YouGov has been showing markedly better numbers for TBP and the LDs than just about all the others. At times like this it is useful to look at the record and what happened last time.

The table above shows how well YouGov did in 2014 compared with the other firms and overstated Farage’s then party the least.

Clearly that was all five years ago but it is worth highlighting. The key to polling low turnout elections is to ensure that as far as possible your numbers are based on the views of those who have or will actually vote. It is here that YouGov, who first got into online polling nearly 20 years ago, has probably got an edge if only because of the data it has on its polling panel.

But who knows? GE2015 was a shock followed by the Brexit referendum in 2016 and of course GE2017. Might we see something like that when the Euro results start coming out a 1opm on Sunday night?

Mike Smithson


The big post Euro election question is whether the Westminster polls will revert to normal?

Sunday, May 19th, 2019

Above is the latest YouGov polling with the most extraordinary Westminster voting figures that I can recall in recent times. For four parties to be within six points of each other is extraordinary and for the CON+LAB aggregate not to exceed 50% is completely unprecedented.

There is some historical precedents for Westminster voting polls to go slightly haywire before the Euro elections and in the past things have settled down quite quickly afterwards.

The Brexit party rise was largely predicted but what’s worrying both Labour and the Conservatives is that the Lib Dems have made a remarkable recovery with the highest poll share since the coalition was formed in May 2010.

There’s said to be a worry in LAB circles about the party’s defence of the London Mayoralty in eleven months time.

The voting data is the 9k+ sample YouGov poll that was mainly focus on the Euro elections that was published yesterday.

Mike Smithson