comment on the authors' reply to 'Tobacco consumption and advertising restrictions: a critique of laugesen and meads 1991'

M. J. Stewart
Abbey Management Services

THE SCIENTIFIC METHOD

The difference between an editorial in which someone asserts their opinion about a scientific truth, and a refereed paper in a learned journal, is that the conclusions in the latter should be based on an experiment, or analysis, which can be replicated by other scientists. This system is designed to discourage authors from publishing false claims because of the likelihood of their being rapidly refuted; and to ensure that they are refuted, if an author, by error or corruption, does publish a false claim. In the case of an analysis, both data and method must be available to sceptics so that they can repeat the analysis and see if it yields the results claimed. Laugesen and Meads (1991) did not meet this minimum standard, since they failed to describe their method in a way which would enable anyone else to repeat their analysis.