The Telegraph tomorrow has a new poll of marginal seats carried out by Crosby/Textor, conducted in the 140 Labour seats where the Conservatives need the smallest swing to win, and the 20 Lib Dem seats where the Conservatives need the smallest swing to win. Full details of the polling are here.
The poll found a swing of 7.5% from Labour to the Conservatives in Lab-v-Con marginals, enough for the Conservatives to win around 103 seats on a uniform swing. In the LD-v-Con seats though, the poll found a swing of 3 points from the Conservatives to Lib Dems, so clearly the Conservatives would be unlikely to win any LD seats, and if the same swing occurred in Con-v-LD seats, they’d expect to lose some.
Looking at the technical details of the poll, Crosby/Textor essentially used a methodology very similar to that of MORI. It was a telephone poll using quotas, it was weighted by demographics but not by past vote, there was a squeeze question for don’t knows and it was filtered by only those saying they were 10/10 certain to vote. Perhaps not surprisingly, it also found a similar result to MORI’s poll of marginals today, which found a 7% swing to the Conservatives in marginal seats.
Two interesting things to point out. Crosby/Textor asked two voting intention questions: a standard one, and one prompting using candidate name and party. Prompting by candidate name increased the support for the incumbent party, reflecting the effect of the personal vote. The swings reported are from the candidate prompted questions.
Secondly, while it’s nice to have some LD-v-Con marginals polling, I’d treat it with some scepticism. I can’t find a break down in the tables for how many people were interviewed in Lab-v-Con seats and how many in LD-v-Con seats, but if it was in proportion to the number of seats being polled, the sample size for LD-v-Cons would be absurdly small. Secondly, the figures themselves look odd – CON 41%, LAB 3%, LDEM 53%. Now, I’m sure Labour’s vote is going to get squeezed in those seats… but 3%?