How did the pollsters do in the local elections? Part 2

Welcome to this week’s edition which takes a second look at how the pollsters did in the local elections. (See part 1 here.) This time its vote share polls that I’m looking at.

In addition, paid-for subscribers will find 10 insights from the last week’s polling at the end, along with all the footnotes.

Before we get to that, this week’s disapproving stare is for the Daily Telegraph which reported of a new Blue Wall poll: “The poll numbers suggest that the Tories are now on an upward trajectory”. Here’s the zoomed in part of the pollster’s graph showing the current and previous Conservative ratings from its Blue Wall polls:

Upward trajectory, hey?

As ever, if you have any feedback or questions prompted by what follows, or spotted some other recent polling you’d like to see covered, just hit reply. I personally read every response.

---

Been forwarded this email by someone else? Sign up to get your own copy here.

If you’d like more news about the Lib Dems specifically, sign up for my monthly Liberal Democrat Newswire.

---

My privacy policy and related legal information is available here. Links to purchase books online are usually affiliate links which pay a commission for each sale. Please note that if you are subscribed to other email lists of mine, unsubscribing from this list will not automatically remove you from the other lists. If you wish to be removed from all lists, simply hit reply and let me know.

---

How the vote share polls did this May

Shortly after the local elections, I took a look at how two MRP studies did. Short version: YouGov’s was pretty good at picking out the right overall stories of political movement, though struggled more at picking up how far the movement was going in specific places, especially where the movement was coming from a party bucking national trends. With Electoral Calculus, however, the basic story was wrong for the Greens, the Lib Dems and the Conservatives.

But there were also two pollsters who did voting intention polls for the local elections. How did those fare? This is trickier than it may appear to answer because the polls were measuring something that doesn’t exist.¹

Those local election voting intention polls were measuring how people would vote if every party stood in every ward up for election. Therefore they were not measuring how people all across the country would vote. Rather, they were were only measuring how people in the places with local elections would vote.

Fair enough, you might think - but that means the polls can’t simply be compared with the national vote share projections either from John Curtice and the BBC team or from Michael Thrasher and Colin Rallings as those projections are projections across the whole country. Therefore you can’t simply compare the local election voting intention polls with those projections as, for example, Wales - without any local elections - is omitted from the former but included in the latter. So too London.

Ah, but you might think that all you need to do therefore is compare the polls with the actual vote shares in the places that did have elections. That too, however, is problematic because, unlike a general election, local elections don’t see all the main parties stand nearly everywhere. This year, for example, only 77% of vacancies were covered by a Labour Party candidate, significantly less than the number of vacancies with Conservative candidates.²

Therefore, what the local elections voting intention polls are capturing is neither what the actual vote totals tell us nor what the projected vote shares tell us either. Those are, though, the yardsticks we have so let’s look at how the polls did against those:

Sources: Omnisis poll, Survation poll, Britain Elects results spreadsheet.

Overall, that’s pretty good. Both polls got the broad picture right, including of the Lib Dems doing much better than the national voting intention picture and of the Greens having a healthy share of the vote, though clearly behind the Lib Dems.

Both polls also ‘exaggerated’ the Labour lead - 11 points and 10 points compared with the 2 point reality. Curiously, the polls were closer to the national projections that included areas the polls didn’t sample.

An extra rough yardstick to try therefore is to look only at wards which had both Conservative and Labour candidates.³ In those wards, Labour had a 5 point lead - higher than the 2 point lead across all wards - and close to, though still less than, the 10 and 11 point leads in the local elections polls.

Before rushing to conclude that this therefore suggests the polls in general, including national voting intention polls, are over-estimating Labour’s lead, a few pieces of caution are in order.

First, we only have one poll each from two different pollsters. That’s a much smaller data set than we have for national vote share polls.

Second, turnout is a major challenge for voting intention polls. Much more work has gone into trying to handle this properly for national voting intention polls than for local election polls.

Third, the overall evidence is that the lower turnout is in an election, the less accurate polls are - and in these local elections turnout was around half of what we’d expect at the next general election.

Fourth, the national vote share projections from these local elections are about in line with what we’d expect to see compared with past local elections if the general election voting intention polls are right.

Fifth, remember where we started: none of the yardsticks are actually measuring what the polls tried to measure.

All of which means that, absent other clues suggesting the same thing, the local elections voting polls aren’t a reason to think the national polls are wrong.⁴

The main conclusion to draw from them, rather is that for local elections, it’s the questions other than vote share that are the most interesting and useful. Survation’s finding that potholes mattered more to people than Brexit, for example, helped explain what was going on in these elections.

Though let’s hope the reaction to all this is to have more polls like this in the future - as well as more MRPs like the YouGov one - as that would start to provide a bigger bank of vote share changes between polls at different rounds of local elections to look at too.⁵

---

National voting intention polls

Here’s the latest from each currently active pollster. 28% is a good number to have in mind when looking at the Conservative figures as that’s Labour’s notoriously disastrous vote share in 1983 under Michael Foot. Sunak’s party continues to flirt with that sort of support. (Not something you’d spot from, ahem, certain political punditry corners, and see also The Sun in the 10 insights section below).

For more details and updates through the week, see my daily updated table here and for all the historic figures, see PollBase.

Last week’s edition

Coalition of chaos: Project Fear or Project Flop?

Know other people interested in political polling?

Conservative voters trust The Guardian more than the Daily Mail, and other insights from this week’s polling…