Abstract
Zachary Schrag would like to put the burden of proof for continuation of research ethics review in the Social Sciences on those who advocate for research ethics committees (RECs), and asks that we take the concerns that he raises seriously. I separate his concerns into a principled issue and a number of pragmatic issues. The principled issue concerns the justification for having research ethics committees; the pragmatic issues concern questions such as the effectiveness of review and the expertise of the committee members. I argue that RECs can be justified by their role in improving ethical practice and in reducing wrongs done to research participants. I propose a model of review for doing this, which I think would also address the pragmatic issues raised. I then offer an account of where the UK ethics review system is now and suggest three steps which could improve social science ethics review in the UK and move it in a perhaps more desirable direction.
Introduction
Zachary Schrag would like to be rid of research ethics committees for social science (Schrag, 2011). He would like to put the burden of proof for continuation of research ethics review on those who advocate for research ethics committees (RECs), and asks that we take the concerns that he raises seriously. He writes persuasively, and the concerns that he raises do warrant attention. His list of criticisms is not short: he thinks the low-risk nature of social research means ethics review lacks a principled reason for existence; that ethics committees set silly requirements; that they lack expertise; that they apply principles in ways that are inappropriate; and that they can be positively harmful. Schrag writes as a social scientist.
I write as a philosopher, but also as someone who works in research governance in a large research intensive (Russell Group) university (though the views expressed here are my own and not those of my employer) and previously worked managing ethics committees in a small university specializing in social science research. I see most of the ethics review applications that go to National Health Service (NHS) RECs from the University where I work and I see the responses that applicants get back. I take Schrag’s concerns seriously, but I don’t think the case for any of them is as compelling as Schrag does and I don’t think that the appropriate response is to scrap social science ethics review.
Of the concerns presented I think one can be treated as an objection in principle, and the rest are pragmatic issues concerning the way that ethics committees operate. The point of principle is important to address: if one cannot make a principled case for ethics review, then Schrag is right and RECs should be done away with. The pragmatic points allow a different response. Even if we grant that there are problems with the way that ethics review works in practice for social science, we can maintain that the appropriate response would be to improve the system of review, rather than scrap the system entirely.
The principled issue in Schrag’s article is whether there is a role or justification for ethics review in the social sciences. And the ‘social sciences’ part of his claim is important: Schrag seems comfortable with there being a role for ethics review in medical research. His claim is that there is something that makes social science research different from other sorts of research.
Schrag identifies two potential factors: the lower risk inherent in social science research and a lack of a social science equivalent to the Beecher report. The Beecher report (Beecher, 1966) was an account of cases of medical research malpractice. Papworth (Papworth, 1967) is broadly Beecher’s equivalent in the UK. The Beecher report was important in showing that there was a problem with research in the US and led to the establishment of the ‘Common Rule’ which requires that all federally funded research in the US has to be reviewed by an Institutional Review Board. Schrag’s claim is that there is no equivalent evidence of systemic wrong-doing in the social sciences.
In what follows I argue that the differences in the risks between social science and medical research are often overstated and that the question of whether there is widespread wrongdoing in the social sciences is largely beside the point in determining whether or not there should be social science ethics review. I argue that ethics review of social science research is best conceived of as a tool for improving ethical practice in research and that social science research presents particular challenges in this area. I respond to the pragmatic points by giving a brief account of what I think a positive and useful ethics review system for social research would look like and then arguing that the United Kingdom in particular is well placed to move towards such a system.
The Principled Claim: Is there a purpose for ethics review of the social sciences?
Any system of ethics review has two basic goals: to prevent unacceptable research practices and to promote good (ethical) research practices. Often those attempting to justify the existence of research ethics focus on the first of these. Indeed, the UK National Research Ethics Service (NRES) states that the purpose of ethics review is to protect research subjects (NRES, 2012) (though it also sees its own role as, at least in part, facilitating research). Schrag unpicks this argument in relation to social science research, making the case that the risks of social research do not justify the existence of ethics review.
It should be conceded that some medical research involves more physical risk than social science research. Social science research rarely involves injecting people with ‘irreversibly toxic green stuff’, as Robert Dingwell (Dingwall, 2006) likes to put it. But this difference in levels of risk is not as great as is often portrayed, and critics of social science ethics review often overplay the risks of medical research and downplay the risks of social science research. Much medical research involves little more than the donation of a blood sample or a bit of tissue that was being removed anyway. The risks to subjects in that research are not very different from (and may be less than) the risks involved in telling a researcher you don’t know very well intimate information about your life. Even in the case of the riskiest category of health research, Clinical Trials of Medicinal Products, the risks are balanced with a much more detailed risk assessment and more involved safety precautions than in other forms of research − and of course these risks are balanced by the possibility of greater and more evident rewards. Medics do not inject patients with irreversibly toxic stuff of any colour if they can help it, and the point of much research is to find drugs with better toxicity profiles than the current standard medicine.
It is easy to underestimate the risks that can be inherent in social science research. As Schrag acknowledges, social science research can involve very real risk of harm: interviewing rape survivors in South African slums, illegally trafficked children in Columbia or men who have sex with men in West Bengal (all projects carried out in institutions where I have worked) are not safe things to do for either researcher or researched. And harm in relation to social science can happen. Social researchers very often deal with extremely vulnerable people and potentially fragile personal relationships. Damaging those relationships, or a person’s social standing, or a person’s self-esteem, or a person’s ability to trust, is real harm even if it is not physical.
But social science research does not just involve the risk of physical or emotional harm. The key risks involved are of breaching the trust, exploiting or taking advantage of subjects. When you start researching people directly by accessing data about them, talking to them, and finding out about their lives, you enter into a relationship with them. That relationship engenders responsibilities just as other professional relationships do. To fail to meet those responsibilities is to wrong your research subjects even if it does not harm them in any physical sense and even if they never find out about it.
So even if the point of ethics review is to prevent harm to research participants, the argument that social science research is less risky than medical research is weaker than its purveyors suppose. If the point of ethics review is to prevent the wronging of research participants then there is no reason to exclude social science from ethics review. So (if other research is) ethics review of social science research is justifiable in terms of the prevention of wrongs and possibly in terms of preventing harms.
In the section of Schrag’s article titled ‘Ethics review is a solution in search of a problem’ Schrag argues that medical research ethics review was established to solve a problem: the widespread problems with research to which Beecher and Papworth drew attention. He argues that there is no similar evidence of systemic misconduct in Social Science research, and concludes that there is no valid rationale for social science ethics review.
In this he is assuming a distinction between social research and other sorts of research which is not obvious to those in the review system. What Schrag wants to say is that, rather than treat all research together, social scientists are in a position to say the bad things done are done by those nasty medical researchers, not social scientists. The problem with this sort of dialectic is that it is precisely the claim of those who advocate for ethics review that researchers are too close to their work to be able to assess its risks and benefits − and this line of argument tends to reinforce that view by suggesting that the protesting researchers think their research risk-free.
Even supposing we grant that social science research is somehow less affected by ethical difficulties than any other sort of human subjects research, however, this would not entail that there was no justification of systematic ethics review in social sciences. A widespread pattern of abuse might necessitate such a system of review, but no such pattern is needed to justify it − so long as there are sufficient benefits to be gained from the system in terms of the improvement of ethical practice. I think that there can be such benefits. Ethics review can be justified in terms of quality improvement rather than in terms of regulatory necessity.
People who object to ethics review like to make out that all or most social researchers already build ethics and ethical consideration into their practice. But anyone who has worked with ethics committees for any length of time has reason to believe that while there are many very good and ethical researchers, there are also a number who are just blind to significant ethical questions that their work gives rise to. It takes training and facilitation to learn how to recognize ethical issues in the first place. There is social research suggesting this is the case for clinical ethics committees, which deal with far sharper and more evident ethical issues than research ethics committees (see for instance Pederson et al. (Pedersen et al., 2009) who found in a qualitative study of committee responses to a paper case that the ethical issues were often not uncovered in the committee deliberation, but rather in the researchers’ post deliberation interviews). But the applications submitted by sometimes very experienced academics can make quite clear that there is something that they are missing. As one colleague managing University Research Ethics Committees put it at a conference, ‘If there is no need for ethics review, how is it that the applications we get are so bad?’.
The important thing for social science is to ensure that the ethical conversation has taken place, to ensure that researchers have sat down and thought through whether the way they are going about things is really the best way of answering their research questions − ‘best’ methodologically, but also ‘best’ ethically.
What ethics review should be
Schrag notes what I would call the irenic response to ethics review. This says that ethics review itself may be useful, even necessary, but that it needs to be reformed. I have myself written (Jennings SLM, 2010) advocating some of the sorts of changes that he mentions and in my previous job was able to implement some of them (e.g. the short ethics form which basically asks ‘what are you doing?’ and ‘what are the ethics issues and how are you dealing with them?’).
The ethics review I advocate is best seen as a way of ensuring that ethical issues have been considered, and as such the most important part of the review is what happens before the application is submitted. The important part is where the researchers think through the research for themselves and identify the ethical issues in their research. The bureaucratic function of the ethics review on such a model is to support this process of reflection, to make sure it happens, to challenge assumptions and to point out things that might have been missed. The independence of the ethics reviewers from the project is important in this respect for two reasons. One is that an enthusiasm for a research project can lead one to underestimate the real risks that it poses and to overestimate the benefits. The other is that in thinking through the ethical issues in the project the knowledge that you will need to justify your position to a panel of sympathetic but independent people focuses the mind. It means that you have to justify your assumptions and your conclusions. The interests of people being researched do not always align neatly with the interests of the people doing the research, and identifying where those tensions lie can be a valuable exercise which an independent panel can help with.
To be useful the review panel must be made up of (or at least include) lay members and social scientists, preferably with a mix of experience, and committees must be appropriately trained. The administration of the committee is also important: the person organizing the committee can be very influential with respect to where discussion is focused in reviews that take place, who gets selected for membership of the committee and on the way that committee decisions are communicated to researchers.
The outcomes of a review also need to be detailed. The reasoning for requirements needs to be provided. Ethics review will not work as a way of promoting reflection unless its outputs prompt reflection and the process of review is reflexive. This means that committees need to be willing to enter into a dialogue with applicants and be willing to listen to what applicants say. It also means that where committees do set silly conditions on an approval that applicants have to be willing and able to challenge them. All this, of course, requires training for the ethics committees and for researchers.
This is still a prospective ethics review, and the point is well taken that it is in the practice of social research that the real ethical issues arise. Many researchers will have encountered unexpected situations which either raised ethical issues at the time, or which turned out to be ethically problematic upon reflection. But the best way to deal with the saturation of social research with ethical issues is to start considering them early and continue considering them all the way through. Ethics committees should be available to advise on emerging issues (where there is time for this) and should get feedback on issues that do arise so that they are able to better advise in future.
For real benefits to accrue from ethics review, it needs to be useful to researchers. It needs to be something that they engage with in a positive way. There is a difference between engaging with ethics review as a way of improving your practice and just writing the things you need to write in order to get your research proposal passed by a REC. It is only when ethics committees are set up to be venues for ethical reflection, and researchers own them as such, that we will reap the full potential benefits of ethics review.
What ethics review is
There is a literature on problems with ethics review, in medical sciences (Glasziou and Chalmers, 2004; Jamrozik, 2004) (though this has decreased in the UK since NRES took over RECs) just as there is for social science. To some extent this is to be expected − no one really enjoys being regulated, and generating literature is one of the things that academics do. One should be careful of inferring from the existence of such a literature that all academics are unhappy with ethics review. People are often much more motivated to write about bad experiences than good ones. Those who rail against ethics review have often encountered problems which meant their research was delayed; those who have had no such problems are busy getting on with their research and are much less likely to want to stop and write about what a great time they had at ethics review.
Schrag’s case, however, is not just that there are problems with ethics review. It is that there are systemic and very common problems with ethics review. As he says, he collects them. One understands the impulse to do this, and doing so serves an important political purpose in that you cannot fix a system without any evidence that there are problems with it. As a way of generating valid data about a system it is unlikely to give you a balanced view, however. It is therefore difficult to draw conclusions on the incidence of problems with ethics review on the basis of such examples.
There is reason to think that the UK may be in a better position than other parts of the world in respect of many of these problems. Not least because a goodly chunk of the ethics review system in the UK is centralized and has systems in place to deal with the sorts of issues that Schrag raises. The US system of ethics review has also been developed in a fairly bureaucratic and heavy-handed way which means that its requirements are at once unclear and have the weight of law. That said, I think that anyone working with ethics committees, particularly in the university system in the UK, will recognize some or all of the problems that Schrag highlights.
Research ethics review in the UK is broadly split between research that involves NHS patients recruited through the NHS and research which does not. Research ethics review for projects involving NHS patients is managed by the National Research Ethics Service (NRES) and is run according to the Governance Arrangements for Research Ethics Committees (GAfREC) (DoH, 2010). NRES also run a specialist committee for social science ethics review which allows for review of research in social care contexts where there is no other committee available to review the research and which allows some research to be reviewed where there is a statutory requirement for the research to go through an NRES REC (e.g. research involving participants without the capacity to consent).
NRES has been successful in standardizing applications to different committees, and have committees cross-review studies in order to help encourage consistency of review. My experience of NRES committees bears out Hedgecoe’s research (Hedgecoe, 2008), suggesting that there is no systemic bias against qualitative research in their committees; while I have occasionally come across silly decisions, these have not been particularly confined to social science or qualitative research. My experience is, however, that the complexity and amount of human interaction involved in social research makes it more difficult to explain to an ethics committee than other research, and consequently more difficult to identify and address the issues that are likely to arise in it.
Ethics review in universities is far patchier than in NRES. The last university-wide survey of RECs in universities was the Tinker report in 2004 (Tinker and Coomber, 2004). Since then the Economics and Social Research Council has introduced its Research Ethics Framework (the 2005 Research Ethics Framework has now been superseded by the 2010 Framework for Research Ethics (ESRC FRE) (ESRC, 2010)), which required that all ESRC funded research undergo ethics review, and compliance with the ESRC FRE has become part of the conditions of funding of other research councils funding social research. The result has been a revolution in ethics review in universities, with any university with a substantial research profile now having some sort of ethics review process. The quality of these reviews is likely to be patchy, but more problematically, there is no national system in place for evaluating the quality of consideration generated by such RECs.
It seems to me likely that many of the pragmatic problems with ethics review that Schrag identifies are apparent with at least some university RECs.
How to get from what ethics review is to what it should be
How do we get from where we are to where we want to be? Well I think that the existence of NRES and the relatively flexible nature of its university RECs put the UK in a strong position. Three things need to happen.
First, we need a change in attitude from social science researchers. Rather than complain about the fact of ethics review, they need to focus on what they would like ethics review to achieve. Ethics review can be helpful to researchers, but for that to be the case they need to own it and use it as a tool for improving their practice (similar to peer review), not treat it as a compliance hurdle to be overcome. Despite what researchers may think, RECs in universities and NRES think of themselves as facilitating and supporting research. RECs know how vital high quality research is. NRES has shown itself to be responsive to criticism and it has vastly improved in many of the areas where it has been criticized in the past. If researchers are not getting what they need from ethics review, then this is something that can be reported back to them constructively. Similarly, those running ethics review systems in universities are generally reasonable people. Researchers need to engage with these systems to ensure that they are promoting good research. Conference time would be far better spent on how to make ethics review useful to researchers rather than on bemoaning ethics reviews’ existence.
Second, we need more coordination amongst university RECs. UREC review probably is inconsistent at the moment, and there is no guarantee that if another university has reviewed something that this has been done to an acceptable standard. University ethics coordinators need to be speaking to each other more than they are and seeking to establish national standards for ethics review. These need to include not just the bureaucratic process but also the culture and level of commentary that reviews should provide.
Third, we need a coordinated research programme on research ethics and a systemic way of feeding its results through to researchers and RECs. One of Schrag’s more telling criticisms is that ethics committees lack expertise, not just on the research they review, but also on research ethics. This is, I fear, fairly common and is at base a training issue. NRES committee members are all required to undergo a certain amount of training, and training provision is also a requirement of the ESRC FRE, so in theory training should be taking place. A national consensus on what should be included in such training, what it is important to cover, and what conception of ethics review should be promoted would do a good deal to support consistency of review, but it would also be useful to feed new research results concerning research ethics and participant safety into such training.
No one likes their behaviour to be regulated, and it is only natural that those who are subjected to regulation will object to it. Similarly, it is only natural that those such as myself who are or have been responsible for managing ethics review systems will seek to defend them. What both sides have in common is an interest in doing good research, in uncovering and unpicking the ethical challenges and tensions involved in this and making sure that the research that we conduct is as good as it can be. We can acknowledge that not all ethics review systems are as good as they should be and that even the best can be improved. Researchers and RECs need to work together to this end.
What I know, after years working in this area, is that I would not agree to be a subject in any research that had not been subjected to ethics review by a suitably constituted REC. I wouldn’t advise anyone else to do so either.