27 November, 2019| Analysis, GE2019|Ben Walker Ben Walker

The Brexit Party’s pullout demonstrates a problem for pollsters

The Brexit Party’s decision to stand down may have boosted the Conservatives in the polls, but it’s difficult to predict how many of those votes will actually materialise on polling day.

Photo: Getty Images

The announcement earlier this month that the Brexit Party will not stand in any of the 317 constituencies won by the Conservatives in 2017 was welcomed by Tory activists, who saw it as improving their chances of taking a number of marginal seats from Labour. Since then, national polling has seen a boost in Tory fortunes to the detriment of the Brexit Party. But it remains to be seen if this will materialise on the ground. Will Brexit Party voters will turn out as enthusiastically for Boris Johnson as they did for Nigel Farage in the European elections in May? How many of those voters will vote differently in a general election – and how many, in this election, will turn out at all?

To answer some of these questions, it is worth knowing how polling companies tackled the issue of the Brexit Party standing in less than half of the country.

Here’s how it works: when there isn’t an election campaign happening, pollsters typically include all of the popular parties active in the country among the options available in the voting intention questions. So, before it is known which constituencies the Green Party will stand in, the Greens are offered as an option to all respondents. Since the Brexit Party was formed in January, pollsters included the party in their initial list.

Once an election has been called and the candidates are known, pollsters then change their methodology: the voting intention question broadly goes from “if there was a general election tomorrow, which party would you vote for?” to “there is a general election happening on [X date], which of these candidates, who are standing in your constituency, would you vote for?”

YouGov, the most prolific of all pollsters, offers only the candidates standing in the constituency where the respondent lives. Survation and ICM have adopted the same method, and ComRes and BMG have indicated they will be doing similar in future polling.

This shift in methodology affects how smaller parties are seen in the polls, as their share of national voting intention is reduced by the reduction in the number of seats in which they’re standing.

The Brexit Party is an especially interesting example of this. By halving the number of its candidates, it went from being a national party to a small party overnight, and the share of its voting intention in the polls has plunged. But this does not tell the whole story. While the Brexit Party now polls at less than 5 per cent nationally, its share in the seats where it is standing may be two or three times higher. The same is true of the Greens.

Other pollsters, including Ipsos MORI, Kantar, and Panelbase, address this to a certain extent by asking respondents for their second preferences. Those who select the Brexit Party as their initial voting intention, but live in a seat where the party is not standing, have their second preference used instead.

However, there is no guarantee that voters who are unaware that their first choice isn’t standing will commit to their second preference. It may be more likely, when they discover their chosen party isn’t standing, that they decide not to turn out. The Brexit Party’s pull-out may have given the Conservatives a boost in the polls, but predicting how many of those votes materialise on polling day is a harder problem.

Ben Walker

Ben Walker is the Founder and Editor of Britain Elects.