In stark contrast to yesterday's Daily Telegraph poll, which showed a 2:1 majority against the EU constitution, an ICM poll published in the Sunday Telegraph today suggests that the public is "evenly split".
This is an ICM survey which, like the DT's YouGov survey, asked voters the exact question to be posed in the referendum. But this survey shows 39 percent in favour of signing up to the constitution, just two percent behind the "noes" at 41 percent.
The result is clearly anomalous, suggesting a far narrower gap than all other polls published to date which have, almost without exception, shown the “noes” in the lead.
While the "yes" campaigners will take heart from this result, it presents more sanguine observers with a problem as to which poll to believe. Even causal observers of the "poll scene" are aware that there are considerable problems in interpreting polls and here there is also a methodological issue to consider.
YouGov uses online polling and, recently, has come up with good results, making it one of the most respected of the internet polling companies. While there are errors associated with this type of polling, ICM uses telephone interviews, which also have their own errors. Furthermore, as we have found, responses can depend very much on the context in which the questions are asked.
On the basis of previous polls, however, the odds are that ICM is a rogue poll. All the other pollsters, including the EU's own Eurobarometer, are saying differently, and the likelihood is that there is strong opposition to the constitution throughout the country.
However, if a week is a long time in politics, a year is an eternity and anything can happen in the next 12 months. Complacency is not an option. The referendum is not in the bag.