A new YouGov poll in today’s Telegraph is a prime example of the difference sample error can make – the topline figures are CON 38%, LAB 31%, LDEM 18%. The poll was begun before the YouGov poll in the Sunday Times, but there was a substantial overlap in the fieldwork, so the simple explanation for why one YouGov poll shows the Conservative lead static and one show it falling is normal sample error. For all the attention we pay to small changes in parties’ support in voting intention polls, movement in a single poll can always be just the normal variation between different samples. What we should be looking for, is trends that are supported across a number of different polls, only then can we be confident that real change is occuring.

YouGov other questions report similar findings to the myriad of other polls on the Labour leadership we’ve seen in recent weeks. Gordon Brown’s ratings are continuing to fall – in February YouGov found that 36% thought Brown would be a good Prime Minister, 33% that he wouldn’t (a net figure of +3), by the beginning of September the net figure had fallen to minus 8, now only 27% think he will be good, with 44% disagreeing, a net figure of minus 17. Amongst Labour voters 56% think Brown will be a good Prime Minister, 19% think he won’t. As in ICM’s recent poll when asked who they would prefer as Prime Minister, Brown trails behind Blair (27% to 32%).

A list of questions about how people see Gordon Brown reveals the usual pattern. Brown has positive figures on competence, decisiveness and being effective. Everywhere else he is regarded negatively, being seen as concerned only for himself and his party, being uncaring, untrustworthy, unlikable and unable to unite the nation.

The same opinions about Brown came out in two ‘focus groups’ for the Guardian and the Times/Newsnight. Both reported that people thought that Brown was strong on substance, but weak on likeability. The more interesting finding amongst both groups was that John Reid, whose star had seemed to be fading since his boost in popularity after the terrorist arrests, was viewed more positively than Alan Johnson, who recently has seemed to be the man the media are gathering around to promote as the “anti-Brown”. Reid was seen as strong and tough, but willing to listen. Johnson was seen as pleasant enough, but not really a heavyweight, and a bit boring.

The headline reports on the two groups differed – the ICM focus group was reported as finding that, despite these faults, Brown still came out top. Frank Luntz’ reported that his group didn’t put Brown as first choice. At the end of the day it doesn’t matter – these aren’t quantative groups, you shouldn’t try and use them to work out who comes top or bottom or whatever – that isn’t he point of qualitative polling. These studies don’t tell you what proportion of people think positively or negatively about people, and shouldn’ tell you who is favourite. They tell you why people think positively or negatively about the contenders.


Comments are closed.