The increasing use of the lay public as researchers in studies on issues that affect them could be producing unreliable research findings, a conference heard this week.
Social scientists at the meeting warned that the surge in interest in so-called "participatory research", where the public - including children - undertake research, was following a flawed political agenda to involve the views of patients in research projects.
But researchers pioneering the involvement of non-academics in social and health research argued that the new method led to a better understanding of the issues under investigation, and a less patronising approach to research.
The issue emerged as the key talking point at the annual conference of the Social Research Association held at the School of Oriental and African Studies in London last week.
Martin Bulmer, professor of sociology at Surrey University, said: "This argument in favour of using volunteers in the research process is untenable. You don't need to be affected by an issue to understand it - you don't need to have been Caesar to understand Caesar.
"It serves a political agenda, providing patients or those affected with another platform for their views - without actually advancing the research process."
He added: "Disability research provides an example of wherethe boundaries blur. It's certainly true that researchers have often treated disabled people as if they are not there but it's getting to a point where they are pushing the argument too far."
But Paul Burton, head of the School for Policy Studies at Bristol University, argued that involving people in research allowed them to play a central role in shaping public policy.
He said: "Before the use of participatory methods, research about mental health services was done in a patronising and paternalistic way with little or no reference to the views of those who use these services, resulting in poor quality services."
Sheila Peace, of the Open University, meanwhile said that participatory research should be part of a multi-method approach.
"Often we end up with data that isn't very good and we as researchers have to decide if we can or can't use this - a sort of quality assurance on the data. It could be argued 'why do it in the first place' but this approach has enabled us to have a much greater breadth of research over a much greater time."
Sue Duncan, Chief Government Social Researcher, said: "We encourage researchers to provide the best available evidence - which can take a variety of forms - to inform government decision-making. I encourage productive ways of involving stakeholders in the research process, be they customers, policy-makers or practitioners, but that need not mean they actually conduct the research."