TNS show narrow NO lead in AV race

The Independent this morning carries a new TNS-BMRB poll on the AV referendum. It shows a much tighter race compared to the recent polling by ICM and YouGov, with YES on 32%, NO on 34%, 21% don't kmow and 13% won't vote.

Leaving aside their Scottish polling, I think this is the first TNS GB polling we've seen since the election, though notably this poll was conducted online, rather than the face-to-face methodology TNS was using for political polls before the election.

There are no tables available yet, so I can't really speculate on the difference between what ICM and YouGov are finding and the TNS findings. Note that the fieldwork was conducted between the 14th and 18th April, so covering the same period as ICM (15th-17th), and slightly earlier than YouGov's fieldwork (18th-19th), so the differences are likely to be methodological or due to the question, rather than timing.

The Independent's headline for the poll - "Voting referendum neck-and-neck as Yes campaign gains support", incidentally, wins our coveted "crap media reporting of polls" award by claiming this shows the YES campaign gaining support. It does not - there is no earlier TNS poll to compare it too, and it predates one of the polls showing a bigger lead (Andrew Grice's actual article does not make the same error!).

UPDATE: Tables for the poll are here, all appears to be above board and shipshape. The topline figures aren't adjusted by likehood to vote, but that doesn't make a massive difference anyway, so no obvious explanation as to why it is showing a different picture.