I've been having a look at the datasets for YouGov's Labour leadership poll showing Jeremy Corbyn in the lead, and the most interesting thing is that the results have been weighted by recalled vote from the 2010 leadership election. That's important, because last time around YouGov conducted a poll that wrongly showed party members backing Ed Miliband - in the end, it was David Miliband who won comfortably in the members' section of the electoral college. Sure enough, the raw unweighted sample of today's poll has Ed Miliband ahead in the recalled vote, so it's pretty clear that a significant adjustment was required. What remains to be seen is whether the crude act of reweighting to the actual 2010 result is sufficient, because the two Milibands don't have direct counterparts in this year's election. Could there be a particular 'type' of Ed Miliband supporter who is still over-represented?
One thing that does add to the credibility of the poll is the fact that the vast bulk of the weighted sample are actual Labour members, and even those people put Corbyn well ahead on first preferences - although they do produce a dead-heat between Corbyn and Burnham in the final run-off.
On the other hand, a possible flaw is that there are roughly equal numbers of men and women in the weighted sample, which may not reflect the real shape of Labour's internal electorate. Corbyn is doing significantly better among women (he's 51-49 behind Burnham among men in the run-off), so if female members and registered supporters are over-represented in the sample, the result should be a little closer.
As I speculated last night, it's actually fairly close between Cooper and Burnham in the race to avoid elimination after Liz Kendall's votes are redistributed, with Burnham just surviving by a 29% to 26% margin - well within the poll's claimed margin of error. As it turns out, though, Corbyn only does slightly better on lower preferences from Burnham supporters than he does from Cooper supporters, so the identity of his final opponent may not make much difference to his chances of winning, unless the race is ultra-tight.
There's been speculation today that the impact of this poll may in itself change the final outcome, in the same way that the famous YouGov poll during the independence referendum completely changed the dynamics of the campaign, and in an unhelpful way for Yes. We're already seeing the equivalent of the "shock and awe" campaign being directed at Corbyn over the last few hours, so that only leaves one question to be answered. When can we expect "The Vow"?