Brian Glancey (Letters, 12 November) states that "48 per cent of New Statesman readers believe that the CIA should be licensed to kill". He is presumably referring to your "online" poll (result published, 5 November). Such polls tell us precisely nothing about NS readers. They exclude those readers who are not "online" and the even larger group who have better things to do with their time. They have no safeguards against multiple voting, or even whole offices (Millbank, or perhaps the CIA) being instructed to "vote". The NS scrupulously refrains from telling us how many votes were cast. Is Glancey so benightedly stupid that he thinks the magic label "online" confers some scientific status on such results, as opposed to the old-fashioned method of actually speaking to people?

Ian Birchall
London N9

This article first appeared in the 19 November 2001 issue of the New Statesman, And now the trouble really begins