Would Blackpool have been different if a poll had not been suppressed?

November 1st, the day Labour was planning for the general election, is probably a good moment to reflect on the amazing events of the past six weeks. And one element that nobody’s really focussed on is the impact of Observer decision not to publish on September 29th an Ipsos-Mori poll showing the Tories 13% behind.

For if it had been the splash lead rather than what did appear (above) then would the conference, which started that day, have gone so easily for the leadership and could Labour’s hoped for Conservative implosion actually have happened?

Let’s recall the sequence of polls during that frenzied week after Gordon had made his first conference speech as Labour leader and Prime Minister. On the the Saturday afterwards the Times and the Telegraph both published polls reporting double digit leads for Labour. Populus had Labour 10% ahead while YouGov in the Telegraph reporting an 11% Labour lead.

The next poll that was due to have gone out was the main Ipsos-Mori’s survey for the month putting Labour on 44% to 31% for the Tories where fieldwork had finished the previous Wednesday. So the three main surveys that weekend would all have had double digit leads with the last to be published putting the gap at 13%. The move to Labour and Brown would have seemed unstoppable. What an awful start for the Tory conference?

But something happened at the Observer four days earlier to make them think twice about publishing it. Instead they commissioned a second survey from Ipsos-Mori, this time a phone poll, which showed a 7 point Labour lead – the same as a BPIX survey in the Mail on Sunday.

This had a remarkable affect on the mood in Blackpool. Somehow being seven points adrift felt manageable in a way that a 13% deficit would not have. And although you should only compare surveys from the same firm there was little doubt that the polls published that Sunday morning looked slightly less daunting as delegates enjoyed the Blackpool sunshine.

In years to come political nerds will produce lots of counter-factuals about the November 1st general election that wasn’t.

I’m convinced that it was the Observer that did it for Dave.

Site traffic reaches record levels. The total number of page downloads during October was 1,560,751 – a figure that surpassed the previous high in September by more than half a million.

Mike Smithson