The tables for last night's YouGov/Times poll, which showed Labour taking a four-point lead, are now out. As I pointed out last night, it’s always risky to get carried away with a single poll – moves only matter if they are real and sustained.
Very helpfully, YouGov asked respondents if they’d (a) watched the debates in full, (b) seen portions or clips, or (c) hadn’t watched the debates. If we take just those that said they watched in full, we can see that they make up about 19% of the sample, equivalent to about 9 million voters. If we include the some/clips viewers, we get up to 40%, or about 18 million. The actual viewing figures were more like 3 million. And as Alberto Nardelli spotted, 51% (23 million) are “likely” or “very likely” to watch the seven-way debate this week. If I were ITN, I’d settle for those figures…
So it appears that the sample contains a response bias, with people that watched the interviews far more likely than the eletorate as a whole to be in this poll. As this group was a lot more Labour than the broader electorate, both on current and past voting, it seems that there were simply too many in the sample. If we were to weight down those in the poll that watched (just the ones that watched the whole lot) to the actual viewing figures, the two largest parties are essentially neck-and-neck.
Among those that didn't watch the debates at all (N=1,079), CON 36 LAB 30… http://t.co/ffuWvtrriS #GE2015 pic.twitter.com/4ZrvwQVdfZ
— NumbrCrunchrPolitics (@NCPoliticsUK) March 29, 2015
Another approach is to separate out those that watched all or part of the interviews, and compare their current voting intention with their 2010 vote. (Due to rounding in parts of the YouGov tables, these numbers won’t be exact, but they’ll be very close). The current voting intention of this subset gives Labour a 14 or 15 point lead. But their 2010 vote showed a 7-point Labour lead, so the swing since 2010 among these voters is a shade under 4%, a swing equivalent to Labour and the Conservatives being virtually neck-and-neck among all voters – basically where YouGov’s polls have tended to be recently.
So as usual, it’s best not to get too carried away until we get a few more polls.
And finally, some hilarious archive footage: Hague, Robinson, Harman, Salmond, Darling, Blair, Kinnock, Galloway, Gove, Farage and Clegg when young!
Hilarious! Footage of politicians when younger (incl Blair, Hague, Clegg and Farage) http://t.co/3DIFWlrMgz #GE2015 pic.twitter.com/Lc21QIxGSE
— NumbrCrunchrPolitics (@NCPoliticsUK) March 29, 2015
Tomorrow we get at least three national polls, with Populus in the morning, Ashcroft in the afternoon and YouGov/Sun in the evening, along with the traditional but symbolic visit of the Prime Minister to the palace to seek a dissolution of parliament for the election.
Update (22:45): Well it’s fair to say that some people didn’t really like the above… And then as if on cue (I honestly didn’t know it was coming!) out came ComRes for the Daily Mail, showing a swing in the opposite direction and a four-point Conservative lead, the biggest for the Tories in a ComRes phone poll since 5th September 2010. Some people need to understand that caution is a virtue…
ComRes/Mail: CON 36 (+1) LAB 32 (-3) LIB 9 (+1) UKIP 12 (+2) GRN 5 (-2) Dates 28th-29th N=1,005 Writeup http://t.co/5kavgv59d4 #GE2015
— NumbrCrunchrPolitics (@NCPoliticsUK) March 29, 2015
It goes without saying that all the caveats about sudden moves now apply in the opposite direction. It could be a genuine move, or it could be random variation. I haven’t seen the tables yet, so can’t analyse the results in any great detail. Let’s wait for a few more polls…