Poorly planned research liable to produce misleading results is unethical.
Local Research Ethics Committees in the U.K. have a duty to assess the scie
ntific merit of studies, of which statistical aspects are an essential comp
onent, yet few have access to a biostatistician. I present personal experie
nce of the workload involved and general issues arising in submissions to a
large teaching NHS Hospitals Trust, with emphasis on sample size. These ge
neral issues should apply to analogous committees in countries other than t
he U.K. The opportunity to influence the quality of all local medical resea
rch is argued to be reasonable incentive for more voluntary involvement by
statisticians. (C) 1998 John Wiley & Sons, Ltd.