So what would it take ?
Opinion polls are a useful tool, but they are not perfect. In 1992 they got the result wrong: in 1997 and 2001 they predicted that Labour's percentage lead in votes would be much larger than it actually was. Pollsters in the US have similar problems. (Remember when we all went to bed under the impression that Kerry was winning and woke up to find that Dubya had the last laugh.) All the pollsters are trying to correct for the bias which understated Conservative (and Lib/Dem) support in the last three elections; they are using different methods to do this, and that's one of the reasons different polls are all over the place. We will only find out whether any of them have got it more or less right on election night. I very much doubt that those polls suggesting a Labour landslide of the same order as 1997 and 2001 can be right. For one thing, in those two years Tony Blair was still popular and most people accepted his self-evaluation as a "pretty straight kind of guy". I...