From Ted Lovesey
Ten years ago, Ioannidis identified wide-ranging problems with scientific research, but this is not a new phenomenon. Studies of aircrew fatigue during the second world war had little success in quantifying the origins of poor performance, until cognitive psychologist Frederick Bartlett realised that piecemeal studies were inappropriate and all relevant factors are needed to produce sensible results.
Somewhere along the way Bartlett’s message has been lost. Most researchers have been trained to limit the variables in an experiment so that statistics are manageable and correlations can be made. This makes life easier but means that some important factors are omitted that could have had a major influence upon the results.
Perhaps we should be critical of the way in which experimental design is taught.
Stoke Gabriel, Devon, UK
