Political

Can you trust Parliamentary by-election polls? The Week in Polls, 5 June

We’ve already had two focus groups published for the June Parliamentary by-elections in Tiverton & Honiton and Wakefield (short version: double grim news for the Conservatives), but this week has brought the first opinion poll.

Carried out by JL Partners for the Sunday Times, it shows Labour well ahead in Wakefield. It’s showing a big swing to Labour, one that would result in a headline-grabbing comfortable win, though also a swing still well short of the sort Labour was getting in the run-up to the last time it ousted a Conservative government (in 1997).

Wakefield by-election poll

However, the poll is only one poll, taken several weeks before polling day. Constituency polling, moreover, has a pretty mixed record in British politics. Add to that the greater volatility of voters in the closing stages of a by-election rather than a general election, and it would be easy to present a superficially smart argument about how little to trust the poll.

The evidence, though, suggests otherwise. Let’s take a simple measure: did the last poll carried out in a Parliamentary by-election before polling day get the winner right or not? The score is 9 correct to 2 wrong in the 11 by-elections polled over the last decade. One of the two exceptions, moreover, was Richmond Park, where a poll taken early in the campaign – more than a month before polling day – was taken too early to pick up the big late surge to the Liberal Democrats. Even that, though, had got right that the Lib Dem vote was up sharply. Which leaves the one other error – Batley and Spen – where Survation’s poll seems to have been mostly undone by failing to pick up the support for George Galloway (6% in the poll, 22% in the result). Note that the Wakefield poll didn’t prompt for the Yorkshire Party, who are standing in Wakefield.

Also in the poll’s favour is that it fits with what we know from other polling. National polls have the Conservatives down over ten points and Labour up approaching ten points since 2019. Add in the usual exaggerations of by-elections, and the JL Partners figures look plausible for a by-election like this.

All in all, therefore, while it is just one poll, it’s a promising pointer, and more robust evidence than you often get from a journalist turning up in a constituency for a few hours for some vox popping. (The Guardian‘s piece about Chesham and Amersham that managed to omit any references to people voting Lib Dem just a few weeks before Sarah Green’s victory is a classic of the genre, though in really skilled hands, such vox popping can be illuminating, just as a day of spot canvassing around a constituency by a skilled canvasser can be.)

JL Partners pollster James Johnson helpfully dug into the details of the poll beyond what the Sunday Times wrote-up in a Twitter thread. He concluded:

The main hesitations about voting Conservative: trust, Boris, and a sense the Tories are out of touch & only care about the rich. All signs are that partygate has crystallised historic concerns about the Tories and turned the people of Wakefield decidedly against them.

Boris Johnson being an electoral liability matches the continuing finding in national polls that large majorities want him to resign and the net negative ratings for the government overall.

Elsewhere in polling, this week’s national voting intention polls continued the picture of the Conservatives in the low 30s, bumping around the Jeremy Corbyn Labour Party’s level of support in 2019 (33%), Labour consistently ahead but only rarely in the 40s, the Lib Dems up into double figures such that a 10% score is now a disappointment rather than a cause for celebration, and neither the Greens nor Reform prospering from the struggles of Labour and the Conservatives.

Other polling told us that:

Starting next week, these round-ups will move over to an email newsletter, powered by Revue. Sign up to receive the Week In Polls into your email inbox here.

Polling UnPacked book cover and Sunday Times review quote

3 responses to “Can you trust Parliamentary by-election polls? The Week in Polls, 5 June”

  1. Unsolicited testimonial. Mark’s book is bloody good. Wonderful blend of evidence and anecdote. And a great relief to me as I had been toying with writing something similar (though doubtless less good) – no need now.
    David Lipsey
    Chair, Lords Select Committee on Opinion Polls 2018

    • Polling relies heavily on the media commissioning it, and I think the possibility of Labour finally making a good gain from the Conservatives is seen as more newsworthy – and so more worthy of polling – than the possibility of the Lib Dems making a third gain from the Conservatives. That would be a more sustained trend, and so ironically both more notable and less newsworthy.

Leave a Reply

Your email address will not be published.

All comments and data you submit with them will be handled in line with the privacy and moderation policies.