Communicate’s May poll for the Independent has topline figures with changes from last month of CON 35%(-1), LAB 31%(+4), LDEM 19%(-3). This largely reflects the trend seen in the other May polls, which have seen a Labour recovery, most often at the expense of the Liberal Democrats – although it is worth adding that in this case of Communicate Labour are recovering from an exceptionally low figure last month.

UPDATE: Several more things – firstly I still haven’t seen the details of how Communicate weight their polls by past vote, but looking at their tables they seem to weight people who voted for “other” parties in 2005 very highly. In reality 8% of people in Great Britain voted for a party other than the main three. Because of false recall ICM and Populus weight their samples so that amongst voters the share is actually lower than that at around 7%. Amongst those who voted in Communicate’s sample 12% of people reporting voting for an “other” party in 2005. Perhaps this explains the rather low shares of the vote for the main parties.

Secondly, Labour voters seem to be much firmer than usual in their intention to vote. The normal pattern is that Conservative and Lib Dem voters say they are more likely to vote than Labour voters do. In this poll the Tories, as usual, had the most committed voters, with 64% of Tory voters saying they were 10/10 certain to vote, but Labour were very close behind with 60% of their voters 10/10 certain. To some extent the increase in Labour support since Blair’s resignation does seem to be down to a re-invigoration of existing Labour supporters. It is the same if you look at the detailed tables of ICM’s last two polls, on ICM’s figures without turnout weighting they would have been up just one point, with turnout weighting they go up two.

Finally a lot of newspaper coverage has taken up the question on whether Brown or Cameron would be more likely to keep their party united, where Brown led by 40% to Cameron’s 37%. This is being widely contrasting to a question from a Communicate poll at the end of last month where 64% of respondents said they thought Labour were divided and 36% thought the Tories were.

For a whole barrel load of reasons it isn’t comparable in the slightest. The question last month asked people to say which parties were divided, they could have picked all three if they wanted. Today’s question asked people to pick either Brown or Cameron, you couldn’t pick both. The question last month asked about the parties as they are now, the questions published today ask people what they think will happen in the future. The questions last month asked about the parties and how united they were, the questions published today asked about the abilities of Brown and Cameron to deal with and contain such division.

In the defence of the media, I haven’t spotted anyone drawing a direct comparison between the two questions, but I’m certain that is meaning many people who hadn’t read either of the questions would take away. If you want to draw a trend from one month to the next, the questions have to be comparable.

Comments are closed.