But is this just the YouGov “magnifier” working again?
In a poll that will send shock waves throughout British politics a YouGov poll in the Daily Telegraph today puts support for the British National Party up from just above zero to seven per cent.
In echoes of the internet pollster’s surveys in May 2004 showing a huge surge for UKIP ahead of the Euro elections YouGov has the following shares with changes on last month: CON 33%(-3), LAB 35%(-1), LD17%(-1), BNP 7%. Unlike other pollsters YouGov does not usually factor in the likelihood of respondents voting.
When assessing the poll bear in mind that a key weighting calculation that YouGov usually uses is based on the newspapers that those surveyed say they read. In its last published poll this worked out at: SUN/STAR 21.9%: EXPRESS/MAIL 16%: MIRROR/RECORD 16%: FT/TIMES/TELEGRAPH 9.5%: GUARDIAN/INDEPENDENT 4%
Normally YouGov gets many more Guardian and Indy readers taking part so their views are scaled back considerably. Sun and Star participants, on the other hand, are usually in short supply so, for example, their voting intentions last month were magnified by more than a third.
In March 2005 I suggested that those PB.C users who are on the YouGov panel could boost their influence and get invited to take part in more surveys if they told the firm they were Sun regulars. This led to me being banned by the firm’s boss, Peter Kellner.
The poll follows, of course, the great media focus on the BNP in the past week or so with the comments by Margaret Hodge and the report from the Joseph Rowntree Reform Trust which was, in fact, based on surveys carried out two years ago.
These poll findings could not have come at a worst time for the main parties which are in the final fortnight of their campaigning in this year’s local elections because they will give the BNP a bit more credibility.
It is worth recalling how YouGov was first to pick up the move to UKIP two years ago though in the end the pollster produced what were inflated projections. In its final poll which included a large proportion of respondents who had voted by post already YouGov had: CON 26: LAB 24: UKIP 21: LD 13 GRN 6. In fact the
Euro election shares were CON 26.7: LAB 22.6: UKIP 16.1: LD 14.9: GRN 6.3.
So in 2004 YouGov over-stated UKIP by nearly a third – which was not a good performance for the firm.
It was suggested after those elections that the way YouGov carries out its surveys can have the effect of magnifying trends. Is that happening with YouGov’s BNP at the moment?