I still feel unconfortable using "majority" in this kind of sense, more especially without even a "relative" added. And that isn't actually because of the way the figures worked out. I'd object to it if it had come out the other way. Honestly.

In this case 21% of people declined to express an opinion.   But what if it had been 90%? If that had happened I suspect that Wolfgang would agree that, even if virtually all the 10% had said "Yes" (or virtually all had said "No"), it would be a distortion of the result if anyone were to say that this was the view of "the majority". So what is the magic figure between 90% and 21%?   I don't think there is one. I think the only responsible thing is to take a little more time and trouble and be more precise.

People reading the Guardian story could read down a couple of paragraphs and see the actual figures - but, for most people, the way it's reached them has been through the TV news - and it's seemed to me that every news bulletin which has referred to this poll has just gone with the headline, and failed to clarify the results. We've been repeatedly told that the poll shows that most people welcome Bush's visit - and that just is not a true summary of the result.