What if the fieldwork had taken place yesterday?

As has been noted often here a key element when assessing a poll is the timing of the field-work. In fast-moving political situations like we’ve seen this week the “when” can play a critical role.

So the overnight YouGov poll in today’s Telegraph has to be looked at in terms of when it took place. The figures, as discussed on the previous thread, were with changes on the previous survey from the pollster – CON 43% (+2): LAB 32% (-2): LD 14% (nc)

There’s no doubt that Labour has had a terrible week in the media – but it has been a week that has got incrementally worse almost with each news bulletin. Last night’s move by Harriet Harman to implicate the Brown camp, the splash in the Guardian this morning, could take this to a new level.

As Anthony Wells in UK Polling Report notes the fieldwork for this latest YouGov poll took place from Monday to Wednesday – but the experience is that half the respondents complete and submit their forms on the first day with probably only a small minority, in this case, waiting until the end.

On Monday the proxy donations story had nothing like the potency that it reached last night when the Harman split emerged and news of the two police investigations came out.

Fortunately for Brown when the news is less bad and less relenting the chances are that some of the lost ground in the polls might be recovered.

In the betting there’s been a move against Labour and to the Tories on the commons spread markets – where the number of seats the parties will get at the general election are traded like stocks and shares. The Tory “buy” price is now about 306 seats. I increased by Tory buy position last night and now have £150 a seat at risk across three bookmakers at an average of 293 seats. With spread-betting you don’t start making money until the price has moved to cover the spread which is usually six seats.

UPDATED 0905 “The boss of YouGov, Peter Kellner has emailed me to say “the full details of our Telegraph poll will be up on our site later today; but in the light of your comments, I thought you should know that the voting intentions were the combination of two separate polls, one Monday-Wednesday (the monthly tracker poll), the other Tuesday-Thursday (when we also asked the questions about perceptions of party sleaze, which the Telegraph commissioned after Monday evening’s disclosures, by which time the tracker poll was already under way). As it happens, there was no material difference between the two surveys – both showed an 11% Con lead, 43-32.”

Mike Smithson