Ethical scrutiny of university research is patchy despite the fact that more institutions than ever have committees to vet projects, according to a new study.
The survey, funded by the Nuffield Foundation, found that a "significant number" of university ethics committees do not scrutinise all of their research.
Anthea Tinker and Vera Coomber, researchers from King's College London, had responses from 78 institutions but found they had sufficient information to fairly assess procedures at 57.
Of these, just five met the complete criteria - such as whether student research is scrutinised and whether committee members come from a range of disciplines - scoring ten out of ten, 25 scored between eight and nine and 21 met between five and seven. Only four scored between one and four and two met none of the criteria.
Some universities had one institution-wide committee that vetted all research. Others devolved responsibility to committees at school and departmental levels, which then referred projects to the main institutional committee for approval.
Of the 78 universities that responded to the survey, half indicated that between one and 50 applications were made to each institution-wide committee a year.
Sharon Witherspoon, deputy director of the Nuffield Foundation, said that the research showed that there have been many constructive developments in British universities over the past two years but that there "is still a way to go".
Ms Witherspoon writes in the study that while many institutions have developed more effective ethical guidelines on the sort of research they carry out, "they have not always worked out a plan to ensure appropriate scrutiny for all the relevant non-medical research".
The report also discovered that while most committees covered research carried out by staff and postgraduate students, not all considered projects carried out by undergraduates.
The study also found that some ethics committees met infrequently. In half the universities that had one overall committee, meetings took place between one and three times a year. The frequency of meetings in institutions with devolved systems was similar.
Membership of committees is also crucial, the study found. "Fellow researchers do not necessarily know best and a small circle of people can be open to criticism of cronyism," it says.
Lay members from outside the university were found on 80 per cent of institution-wide committees. A third had one lay member and another third had two.
Peter Cotgreave, director of Save British Science, warned that expecting all university research to be considered by an ethics committee would be "bureaucracy gone mad". "If it's not relevant, then you do not need approval," he said.