The constituency opinion polls you really should ignore

Polling station sign in Watford 2019

A round of constituency “polling” figures have been appearing in local newspaper titles from both the Northcliffe and Trinity Mirror groups. The reason for the quote marks? All the stories I’ve seen so far have been for online surveys carried out with self-selecting samples of people who are not weighted in any way to be representative of the public.

No surprise then that where voting figures are given, some of them are highly implausible. Do you really think, for example, that in Truro and Falmouth the Conservative vote has collapsed from 44% to 18%, or even, if we recalculate the figures to exclude don’t knows (something the newspaper, the Cornish Guardian, didn’t do), to 26%?

A similar survey conducted for the Stoke-on-Trent Central by-election┬áhad Labour and Ukip neck and neck on 26% when published two days before polling day by the Stoke Sentinel. The actual result? A clear Labour win by 37% – 25%. The Conservatives secured 24% compared with the 12% in the survey.

Tempting though the polling figures are to quote when they happen to give a result that looks nice, the only real lesson to draw from these stories is how poor local newspaper coverage of politics can be. When you’re reading such stories on pages drowning in ads (the story linked to above had 20 ads and 123 – yes, 123 – different ad tracking cookies – when I checked it), it’s not a happy picture of the state of local political coverage. We need vibrant local newspapers to cover local politics and hold local politicians to account. These stories aren’t it.

P.S. Also to be ignored is the “constituency poll” publicised by Clive Lewis.

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Your email address will not be published.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.