It would be wise to view a new poll with cautionby Peter Kellner / January 8, 2018 / Leave a comment
Just before Christmas, I warned Remainers not to get too excited about a poll showing a ten point lead for staying in the EU. The opposite now applies. In his latest column for the Sunday Times, Dominic Lawson reports a YouGov poll showing a Leave lead widening from two to nine points in the past three months. Once again, I doubt whether there has been any real shift in public opinion.
The poll Lawson cites was one of YouGov’s Eurotracker series. It is conducted monthly, but this particular question is asked less frequently. This is how it is worded:
At this point would you prefer that Britain stays in or leaves the European Union?
In early October, it found a 46-44 per cent margin in favour of leaving; in the latest survey, conducted just before Christmas, the Leave advantage had widened to 48-39 per cent.
It is just possible that the shift is the result of sampling fluctuation. However, that is unlikely to be the whole explanation: a five point drop in the Remain percentage looks too large.
There is, though, another explanation. It concerns polling methods rather than shifts in public attitudes. A new question was added to the December poll that was not asked in October:
Martin Schulz, party leader of the Social Democrats in Germany (SPD) announced at the SPD party congress his vision to transform the European Union into the United States of Europe by 2025 with a common constitutional treaty. EU members who do not agree with this federal constitution would then automatically have to leave the EU. Generally speaking, do you support or oppose the vision of the United States of Europe by 2025?
Only 10 per cent of British voters back this idea.
The Remain-Leave question was asked immediately afterwards. Respondents were asked about Brexit with the threat (or, for 10 per cent, dream) of a “United States of Europe” fresh in their minds. By inserting this extra question at this place in the survey, YouGov may well have affected the result of its Remain-Leave tracker question.
For years, survey researchers have known of the “order effect”: the words, thoughts and propositions of particular questions may affect the way respondents…