I disagree.

First of all, the voters selected were from London and the surrounding areas. Whilst it's true that London is a large part of the English electorate, vote share outside London and the South is quite different. Had the audience been selected from, say, Liverpool then the Conservatives would have rightly been angry at such a biased sample.

Secondly, you're implying that polling doesn't effect the electoral outcome, when it clearly does. Motivating voters - not to mention motivating your own activists - is a lot easier when the media coverage is positive. There is an effect that George Soros called "reflexivity" in relation to financial markets, where it is impossible for individuals to make neutral predictions, because their judgements affect the outcome. This effect also applies in politics.

If you're suggesting that voters voted "exactly" as your focus group suggested, then you're saying that the election was decided based on some very narrow criteria - particularly views about the individual leaders. Therefore, a 'scientific' survey which 'proves' that David Cameron is popular must be influential in persuading other voters to support him, right? Because leader's personalities are what you claim people are deciding on. The more people think that a person is popular, the more likely they are to vote for him (or his party). I think David Davis might know a thing or two about this effect!

I also note that you don't choose to deny bias, only claiming that your results were later reflected in the election.

It's not wrong for pollsters to have opinions. It's not wrong for pollsters to do surveys which show the strengths of one party, or the weaknesses of another. But I think it is wrong to conceal any personal bias from the television audience. And if you read the original pieces posted here, some questions have been raised about just which opinions were given prominence when the piece was aired on TV.

Mr. Luntz, you're an excellent pollster and everyone says so. You're one of the best and that's precisely why people take you seriously. This means that you know that a skilful pollster can generate whatever results he likes, and you are most certainly a skilful pollster. Did you - even unconsciously - give the Conservative party a boost with your focus group? I think that you did.

Finally, it appears that you have predicted not just the result, but the Conservative response to it.

Now, I'll admit that I found the election results disappointing in some cases, though certainly not all. I'm not going to deny that many of the opinions expressed in your focus group are true. But I do question the prominence given to certain opinions, and the conclusions drawn from them. This is a particular problem when emphasis is placed on the skill and reputation of the pollster to interpret these results, or to guide the group in their own interpretation.

Anyhow, thanks for the response. We won't be apologising for raising the question of bias, but we're happy to debate the issue with you.

Liberal Democrats. Whining Here.