In an otherwise thoughtful article, John Harris neglects one important virtue of pre-election polls (I’ve seen all the ‘landslide’ polls – but they can’t tell us what’s really going on in this election, 23 June). I have spent most of my adult life in constituencies where, in retrospect, voting for the government I wanted would have been best served by voting locally for another party. I do not need help in deciding which issues are important to me or which government is more likely to deliver the outcomes I want, but I do need help in deciding where my vote would best be placed to secure the national outcome I favour. Well-structured polls are a help with this. To give an example, it is clear from an overview of the six MRP polls I have examined that the party I’m inclined to favour has little chance of winning in my constituency. But if I want to rid myself of the worst government in my lifetime, armed with MRP data, the logical thing for me to do is to vote not for my favoured party but for a third party that has a chance of beating the Conservatives in this seat. I shall know on 5 July whether I have made the right choice, but my chance of doing so is much enhanced by the existence of constituency-level polls. Adrian Carter Penselwood, Somerset John Harris is right that much election coverage is focused on polls that are intent on telling us what the election outcome will be, even though no votes will be counted until after 10pm on 4 July. It is not, however, a new issue. The late socialist historian EP Thompson writing on the 1959 election in the New Reasoner noted that: “A psephologist is a man employed by the mass media to research into what people think the mass media has told them to think.” Of course there are exceptions, such as John Curtice, who provide genuine and critical analysis, but Thompson’s point still stands 65 years on. Keith Flett Tottenham, London