Last night Labour held Stoke Central and lost Copeland to the Tories. As usual, by-elections don’t tell us a huge amount about the bigger political picture, but are very important in setting the political narrative.

By-elections are very unusual beasts. Because they don’t decide who will form the government for the next five years, only who will be the local MP, people are comparatively free to use them to register a protest. They are much more fiercely contested than your average seat at a general election. The constituency itself will also normally have its own local ideosyncracities that mean it can’t just be read as if it is a microcosm of Britain as a whole. So when people ask me what by-elections tell us, I normally say not much: if the change in the vote is in line with what the national polls are showing then it tells us nothing we didn’t already know, if the change is different to the national polls it’s probably just because by-elections are very different to general elections.

Looking at the two results, Copeland is a marginal seat between Labour and the Conservatives… albeit, one that had been in Labour hands for eighty years. The national polls tend to show the Conservatives about 14 points ahead of Labour, the equivalent of a 3.5% swing from Labour to Conservative since the general election. Therefore if Copeland had behaved exactly in line with the national polls it should have been on a knife edge between Conservative and Labour. In the event the Tories gained it comfortably. We cannot be certain why the Tories did better than the national picture would have predicted, thought the most obvious hypothesis is the unusual nature of the seat: Whitehaven is a town wholly dominated by and dependent on one industry – nuclear power – and the Labour party were perceived as being hostile towards it.

On the face of it Stoke Central was a less interesting result – Copeland is one for for the record books, but Stoke saw hardly any change since the general election (only the Lib Dems really saw a significant increase in their share). However it does perhaps give us a idea of the limits to the UKIP threat to Labour. UKIP were perceived as the main challengers from the beginning and it was a promising seat for them: a somewhat neglected working class Labour seat that voted strongly for Brexit, but with a Labour candidate who was remain. They seem to have thrown all they could at it, but with very little success. Again, we can’t be certain why – Paul Nuttall obviously had a difficult campaign and anecdotally UKIP’s ground game was poor, but there are also wider questions about UKIP’s viability now Brexit has been adopted by the Conservatives and without Farage at their helm. By-elections have often been an important route for smaller parties, getting them publicity and a foothold in Parliament. Whenever there has been a by-election in a northern city in the last five years or so there has been speculation about it being a chance for UKIP, but they never seem quite able to pull it off.

So what will the impact of these by-elections be? Copeland will be a body blow to Labour simply because of how incredibly unusual it is. Governments do not normally gain seats at by-elections. Lots of people will be writing about past examples today – 1982 in Micham and Morden (Lab vote split because of SDP defection, and the government got a surge of support during campaign because of the Falklands); 1961 Bristol South East (Tory gain only because the candidate with the most votes – Tony Benn – was disqualified for being a peer), 1960 Brighouse and Spenborough (ultra marginal to begin with). The fact that one has to go back that far to scrape a few examples that generally have extremely unusual circumstances underlines how freakish this is. The political narrative will go back to how Labour are in crisis…but whether that makes the slightest practical difference, I don’t know. Might it provoke another Parliamentary coup within Labour? Who knows. Might it sow some doubts among Corbyn supporters within the Labour party? Again, who knows. The point is, Labour have had terrible poll ratings for a long time, Jeremy Corbyn has has terrible poll ratings for a long time, but this did not stop him being being relected leader last year. The question of Labour’s leadership is one that seems to be a lot more about the opinion of Labour members than the wider public.


ICM’s regular poll for the Guardian has topline figures of CON 44%(+2), LAB 26%(-1), LDEM 8%(-2), UKIP 13%(+1), GRN 4%(nc). The changes since the previous ICM poll aren’t significant, but it’s worth noting that the 18 point Conservative lead is ICM’s largest for many years (there was a lead of 19 points in an ICM/News of the World poll in 2009 and a 20 point lead in an ICM/Guardian poll in June 2008)

ICM also asked about the position of EU nationals in the Brexit negotiations – 42% think the British government should only guarantee the position of EU nationals in the UK once the EU guarantees the rights of British citizens in the EU; 41% think Britain should do it unilaterally straight away. There is a similarly even split on the fate of John Bercow: 30% think he should stay, 32% think he should resign. Finally they asked about Donald Trump’s visit. 18% think it should be cancelled, 37% think he should be invited but not given a full state visit, 32% think a full state visit should happen. Full tabs are here.

To catch up with some other recent voting intention polls. YouGov’s latest figures came out at the tail end of last week (though fieldwork is now a whole week ago) – topline figures were CON 40%, LAB 24%, LDEM 11%, UKIP 15% (tabs). The lead is similar to that from ICM, but with lower support for the main two parties.

Opinium also had voting intention figures in the Observer at the weekend. Over recent months Opinium have tended to be something of an outlier, showing Labour leads of seven or eight points rather than the double digit leads consistently reported by other companies. This fortnight they showed a shift towards the Conservatives, putting their figures more in line with other companies: CON 40%, LAB 27%, LDEM 8%, UKIP 14% (tabs here.)


-->

Ipsos MORI’s monthly political monitor for the Evening Standard came out earlier today. Topline voting intention figures are CON 40%(-3), LAB 29%(-2), LDEM 13%(+2), UKIP 9%(+3). The Tory lead remains pretty steady (note that the increase in the UKIP vote is probably largely a reversion to the mean following an anomolous 6% last month).

Satisfaction ratings with the party leaders are plus 17 for Theresa May (53% are satisfied, 36% are disatisfied) and minus 38 for Jeremy Corbyn (24% satisfied and 62% disatisfied). That includes 22% of Tory voters who say that are “satisfied” with Corbyn’s leadership… I suspect they don’t mean that in a complementary way.

Nothing else has been published yet (MORI normally ask a few other questions, but I expect they’ve held them back to give the Standard another story), all the details so far are over here.


“But the sheer size of the survey […] makes it of interest…”

One of the most common errors in interpreting polls and surveys is the presumption that because something has a really huge sample size it is more meaningful. Or indeed, meaningful at all. Size isn’t what makes a poll meaningful, it is how representative the sample is. Picture it this way, if you’d done an EU referendum poll of only over 60s you’d have got a result that was overwhelmingly LEAVE… even if you polled millions of them. If you did a poll and only included people under 30 you’d have got a result that was overwhelmingly REMAIN… even if you polled millions of them. What matters is that the sample accurately reflects the wider population you want them to represent, that you have the correct proportions of both young and old (and male & female, rich & poor, etc, etc). Size alone does not guarantee that.

The classic real world example of this is the 1936 Presidential Election in the USA. I’ve referred to this many times but I thought it worth reciting the story in full, if only so people can direct others to it in future.

Back in 1936 the most respected barometers of public opinion was the survey conducted by the Literary Digest, a weekly news magazine with a hefty circulation. At each Presidential election the Digest carried out a survey by mail, sending surveys to its million-plus subscriber base and to a huge list of other people, gathered from phone directories, membership organisations, subscriber lists and so on. There was no attempt at weighting or sampling, just a pure numbers grab, with literally millions of replies. This method had correctly called the winner for the 1920, 1924, 1928 and 1932 Presidential elections.

In 1936 the Digest sent out more than ten million ballots. The sample size for their final results was 2,376,523. This was, obviously, huge. One can imagine how the today’s papers would write up a poll of that size and, indeed, the Digest wrote up their results with not a little hubris. If anything, they wrote it up with huge, steaming, shovel loads of hubris. They bought all the hubris in the shop, spread it across the newsroom floor and rolled about in it cackling. Quotes included:

  • “We make no claim to infallibility. We did not coin the phrase “uncanny accuracy” which has been so freely applied to our Polls”
  • “Any sane person can not escape the implication of such a gigantic sampling of popular opinion as is embraced in THE LITERARY DIGEST straw vote.”
  • “The Poll represents the most extensive straw ballot in the field—the most experienced in view of its twenty-five years of perfecting—the most unbiased in view of its prestige—a Poll that has always previously been correct.”

digestpoll

You can presumably guess what is going to happen here. The final vote shares in the 1936 Literary Digest poll were 57% for Alf Landon (Republican) and 43% for Roosevelt (Democrat). This worked out as 151 electoral votes for Roosevelt and 380 for Landon. The actual result was 62% Roosevelt, 38% for Landon. Roosevelt received 523 in the electoral college, Landon received 8, one of the largest landslide victories in US history. Wrong does not nearly begin to describe how badly off the Literary Digest was.

At the same time George Gallup was promoting his new business, carrying out what would become proper opinion polls and using them for a syndicated newspaper column called “America Speaks”. His methods were quite far removed from modern methods – he used a mixed mode method, mail-out survey for richer respondents and face-to-face for poorer, harder to reach respondents. The sample size was also still huge by modern standards, about 40,000*. The important different from the Literary Digest poll however was that Gallup attempted to get a representative sample – the mail out surveys and sampling points for face-to-face interviews had quotas on geography and on urban and rural areas, interviewers had quotas for age, gender and socio-economic status.

pic2097election

Gallup set out to challenge and defeat the Literary Digest – a battle between a monstrously huge sample and Gallup’s smaller but more representative sample. Gallup won. His final poll predicted Roosevelt 55.7%, Landon 44.3%.* Again, by modern standards it wasn’t that accurate (the poll by his rival Elmo Roper, who was setting quotas based on the census rather than his turnout estimates was actually better, predicting Roosevelt on 61%… but he wasn’t as media savvy). Nevertheless, Gallup got the story right, the Literary Digest hideously wrong. George Gallup’s reputation was made and the Gallup organisation became the best known polling company in the US. The Literary Digest’s reputation was shattered and the magazine folded a couple of years later. The story has remained a cautionary tale of why a representative poll with a relatively small sample is more use than a large poll that makes no effort to be representative, even if it is absolutely massive.

The question of why the Digest poll was so wrong is interesting itself. Its huge error is normally explained through where the sample came from – they drew it from things like magazine subscribers, automobile association members and telephone listings. In depression era America many millions of voters didn’t have telephones and couldn’t afford cars or magazine subscriptions, creating an inbuilt bias towards wealthier Republican voters. In fact it appears to be slightly more complicated than that – Republican voters were also far more likely to return their slips than Democrat voters were. All of these factors – a skewed sampling frame, differential response rate and no attempt to combat these – combined to make the Literary Digest’s sample incredibly biased, despite its massive and impressive size.

Ultimately, it’s not the size that matters in determining if a poll is any good. It’s whether it’s representative or not. Of course, a large representative poll is better than a small representative poll (though it is a case of diminishing returns) but the representativeness is a prerequisite for it being of any use at all.

So next time you see some open-access poll shouting about having tens of thousands of responses and are tempted to think “Well, it may not be that representative, but it’s got a squillion billion replies so it must mean something, mustn’t it?” Don’t. If you want something that you can use to draw conclusions about the wider population, it really is whether it reflects that population that counts. Size alone won’t cut it.

=

* You see different sample sizes quoted for Gallup’s 1936 poll – I’ve seen people cite 50,000 as his sample size or just 3,000. The final America Speaks column before the 1936 election doesn’t include the number of responses he got (though does mention he sent out about 300,000 mailout surveys to try and get it). However, the week after (8th Nov 1936) the Boston Globe had an interview with the organisation going through the details of how they did it that says they aimed at 40,000 responses.
** If you are wondering why the headline in that thumbnail says 54% when I’ve said Gallup called the final share as 55.7%, it’s because the polls were sometimes quoted as share of the vote for all candidates, sometimes for share of the vote for just the main two parties. I’ve quoted both polls as “share of the main party vote” to keep things consistent.


ComRes have a poll in Sunday’s Independent and the Sunday Mirror. Most interestingly, it found that people agreed by 45% to 39% that John Bercow was right to refuse to invite Donald Trump to address the Commons, but also that people thought by 47% to 37% that the Queen should meet Donald Trump if he visits the country. As we’ve already seen elsewhere, the British public have little sympathy for Donald Trump’s immigration policy (33% think he was right, 52% think he was wrong) though it’s worth noting that the question wording went considerably wider than Trump’s actual policy (ComRes asked about halting immigration from “Muslim-majority” countries in general, whereas Donald Trump’s policy deals with seven specific countries they claim have an issue with terrorism or vetting).

The poll also had voting intention figures of CON 41%, LAB 26%, LDEM 11%, UKIP 11%, GRN 4%. This is the first ComRes voting intenton poll since way back in June 2016 – after one of the poorer performing polls in the EU referendum (the final ComRes poll had Remain eight points ahead), they paused their voting intention polls while they conducted a review into their methods. They have now recommenced voting intention polls with – as far as I can tell – no changes to their pre-referendum methods. ComRes’s view appears to be that the referendum was an exceptional event, and while the turnout model they adopted after the polling errors of 2015 worked badly there, it worked well at the London mayoral election, so is being retained for Westminster polls. For better or for worse, the ComRes results seem to be very much in line with those from other companies, with a Conservative lead in the mid-teens.

Full tabs for the ComRes poll are here.

While I’m here, I should also mention a BMG Scottish poll that came out at the start of the week (I’ve been laid low with a heavy cold). Voting intention in a second independence referendum stood at YES 49%(+3.5%), NO 49%(-3.5%). This is the lowest lead for NO that any Scottish Indy poll has recorded since the EU referendum. This was interpreted by the Herald as a response to Theresa May’s announcement of her negotiating stance on Brexit. I think that is somewhat premature – so far we’ve had two Scottish polls conducted since May’s speech, a Panelbase poll showing a very small (and not statistically significant) movement towards NO and a BMG poll showing a somewhat larger (but still barely significant) movement towards YES. In short, there is nothing yet that couldn’t be normal sample variation – wait for the next few polls on attitudes towards Scottish independence before concluding whether there is or is not any movement. Full tabs are here