LETTER : Truth in polls

Click to follow
The Independent Online
Sir: Conrad Jameson's article about the opinion polls (24 April) makes no contribution whatsoever to aiding understanding why poll results taken at the same time sometimes diverge, or why the polls failed to predict the result of the 1992 election.

His thesis appears to be that so many respondents lie to pollsters as to make accurate polling impossible, and that this was the cause of the polls' error in 1992. The theory of lying respondents was investigated and specifically rejected by an exhaustive report following a two-year inquiry sponsored by the Market Research Society; it simply does not explain the facts. The explanations the MRS report endorses, with 160 pages of evidence - late swing, differential refusal and failures in the quota sampling - Jameson dismisses or ignores.

It is Jameson's explanation that is unsupported by the evidence, not that unanimously reached by the MRS Inquiry team of experts, which included independent academics and market researchers as well as pollsters.


Chairman, Market and Opinion Research International

London SW1