No. That poll didn’t put Ed ahead in the Prime Minister stakes.

An hour and a half watching Ed Miliband debate four people who are not going to be Prime Minister.

That is the ordeal you had to go through in order to be qualified to answer Survation’s post-debate poll, which included the ‘sensational’ result that respondents preferred Ed Miliband to David Cameron by 45% to 40%. The figure set some even seasoned commentators agog at Ed’s miraculous turnaround on the preferred Prime Minister stakes, following years of languishing twenty or so points behind the Conservative leader.

Everyone should hold their horses. People who watch debates are, at the best of times, the electorally aware and highly partisan, largely tuning in to have their prejudices confirmed. This is going to be especially true of people who knew ahead of time (in order to be asked by Survation to take part) that they were going to don their political anoraks for Thursday evening’s performance.

So when looking at respondents’ Prime Ministerial preference in the post-debate poll, we therefore have to consider that these voters will be far more wedded to their party than the average. By way of illustration, a Survation poll published just over a week ago showed that while 90 per cent of people planning on voting Conservative thought David Cameron would make the best Prime Minister, 63 per cent of Labour voters thought the same of Ed Miliband. When you look at yesterday’s poll, those numbers are 98 per cent and 92 per cent respectively. Of course they are – while less engaged Labour voters may not think the party’s leader should actually enter Number 10, these partisans are unwavering.

Moreover, for the 8 per cent of voters calling themselves ‘undecided’, amongst whom the poll found Ed with a 21 point lead, ninety minutes of Labour-led Cameron-bashing is going to act as an extremely skewed primer. Rather like sitting someone in front of Jaws for an evening and then asking them if they fancy going scuba diving.

In short, the poll is no doubt reflective of those who consciously took time out of their day to watch the debate (doubtless a smaller number than the 4 million who actually tuned in), but these are neither reflective of the public at large nor those who will decide the election.

Ed’s numbers may well be improving, but these figures tell us very little about that.

Mark Gettleson is an elections and polling analyst and Director of Portobello Communications.