DECONSTRUCTING STATISTICAL QUESTIONS

Too much current statistical work takes a superficial view of the clie nt's research question, adopting techniques which have a solid history , a sound mathematical basis or readily available software, but withou t considering in depth whether the questions being answered are in fac t those which should be asked. Examples, some familiar and others less so, are given to illustrate this assertion. It is clear that establis hing the mapping from the client's domain to a statistical question is one of the most difficult parts of a statistical analysis. It is a pa rt in which the responsibility is shared by both client and statistici an. A plea is made for more research effort to go in this direction an d some suggestions are made for ways to tackle the problem.