Does this poll show the impact of Farage/UKIP?
CON 40 (-2) LAB 27 (-1) LD 18 (+1)
Tories down to 40% with YouGov
The online pollster, YouGov, which has taken such a dominant position in UK polling, has another survey out this morning which might cause some concern at Cameron Central. For although the party is still in the 40s it is only just there and the lead over Labour is down to its lowest since June.
We have not got details of the fieldwork period or a break-down of “others” which is up two points. My guess is that it took place while Nigel Farage was making the headlines and that UKIP is up. Also at the end of last week there was the Green Party conference.
YouGov polls have been particularly sensitive to UKIP and in the past two Euro elections the firm has over-estimated the anti-EU party. I am surprised that neither the Telegraph’s Andrew Porter nor Ben Brogan pick up the UKIP impact.
Even so the changes are all within the margin of error and we very much need to if this trend is being picked up by those firms that do it differently. I understand that there’ll be at least one phone poll out tonight.
YouGov restricts its samples to those on its polling panel on whom it has got a mass of data going back several years. These people get repeatedly questioned and, unlike the telephone pollsters, the firm takes no account of certainty to vote.
With the phone pollsters any voter in the country, theoretically, who has a land-line can be called irrespective of whether their number is listed or not. A key element in working out all their shares is how likely it is that those who take part will actually go out and vote. Thus in last week’s MORI poll the Tories had a 17 point lead amongst those certain to vote but only an 8 point one from all those who named a party.
UPDATE 0845: Peter Kellner of YouGov has emailed me to suggest that I was a bit unfair to his firm in this piece. He points out that YouGov came out best of all the pollsters in the June Euro elections and that its UKIP share was out by only one point. This followed a change on their methodology from 2004. Fair enough Peter I am very ready to acknowledge that.