LETTER: Boundaries of political science

Sir: David Gardner(Letters, 16 March), is wrong to assert that the study undertaken for the BBC and ITN constitutes "the only serious independent academic study" of the impact of changes to constituency boundaries. In a paper to be published shortly in the British Journal of Political Science, we employ an alternative approach which produces a figure very similar to that of Dr Mawhinney.

Interestingly, in the light of your earlier report on the battle for the Labour nomination in Swindon North, that seat provides the best example of the scope for disagreement. Using local election results (the BBC/ITN method) suggests a Labour lead of almost 1,000; basing the estimate on changes to the socio-economic composition of the seat (our method) indicates a Conservative majority of over 7,000.

Estimating the outcome of hypothetical contests is an inexact science.

Professor Ron Johnston

Dr Dave Rossiter

Department of Geography, University of Bristol

Dr Charles Pattie

Department of Geography, University of Sheffield