Peer review

Written evidence submitted by Richard Horton (PR 02)

1. Peer review is a central issue in many scientific controversies and disputes today. Take climate change. In the Times Higher Education , last year, Andrew Montford, author of The Hockey Stick Illusion: Climategate and the Corruption of Science (1), argued that events at the Climatic Research Centre (UK) at the University of East Anglia (CRU) had far-reaching implications for the world of scientific peer review and publishing (2). His charge sheet was sharp and precise: that scientists undermined the peer-review process. Implicit in Montford's argument is that peer review is critical to the process of – and thereby public trust in – science. Writing in The Guardian , George Monbiot put it this way: "science happens to be [a] closed world with one of the most effective forms of self-regulation: the peer review process."(3). But for many of us who do peer review, this "most effective" form of self-regulation is often misunderstood and misrepresented.

Peer review: firewall or the weakest link?

2. For scientific journals, peer review is the (confidential) evaluation of a submitted manuscript by one or more individuals who are experts in an aspect of the work under

scrutiny.

3. Who invented peer review? It's hard to be sure, but possibly the prize goes to Ishaq

bin Ali Al Rahwi (AD 854-931) (4). In his book, Ethics of the Physician , Al Rahwi

apparently encouraged doctors to keep contemporaneous notes on their patients, later

to be reviewed by a jury of fellow physicians. But the serious business of journal peer

review had to wait another 800 years. Henry Oldenburg, editor of Philosophical

Transactions of the Royal Society , was the first modern editor to adopt peer review in

the seventeenth century. He used it to famous effect, provoking often fractious, but

illuminating, debates between scientists across Europe.

4. Today, any scientific journal that lays claim to respectability must have a robust peer review process. At The Lancet , the process goes like this. A research paper is submitted electronically to a secure database and allocated by an editor to a colleague. The first or second editor can reject the manuscript at that early stage if the paper is judged to be scientifically poor, unsuitable for the journal‘s readership, unoriginal, or insufficiently topical. Journals differ here. For The Lancet, around three-quarters of manuscripts are rejected at this point. If a paper survives preliminary editorial review, it is discussed at a pre-review meeting to assess its suitability for external peer review. If judged a potential candidate for publication, the manuscript is sent to three expert advisors, commonly from overseas and representing different methodological dimensions of the research, as well as a statistician. There is always the risk of group-think among experts. That is, there may be an orthodox belief about a particular subject, strongly held, which resists alternative perspectives. Editors try to reduce the risk of group-think by sending papers to different and widely dispersed reviewers, deliberately seeking or even provoking critical reviews (just like Henry Oldenburg). Reviewers are not referees in the sense that they can blow a whistle and call time on the paper. We ask reviewers to provide written comments for authors, confidential comments to editors, and a detailed rating for each section of the paper. Those comments are collected, presented, and discussed at a once-weekly manuscript meeting attended by all the journal‘s editors.

5. At this stage, a paper can be rejected or we can open negotiations with authors. If we proceed, reviewers' questions and concerns are put to the authors, with appropriate

guidance from editors. The authors will reply by answering each question from

reviewers, submitting a revised manuscript that attempts to respond to the points

raised by editors and reviewers alike. The authors may also disagree with or challenge

reviewers with varying degrees of force. The revised paper is discussed again at a

manuscript meeting. The options at this stage are to reject, accept, go back to the

authors with further requests for clarification, or return to reviewers (old or new) for

additional opinions. We proceed with further revisions of the paper until a final

reject/accept decision is made. We know that with such a high rejection rate we may

get it wrong. To limit errors of omission, we have a formal appeals process where

editors promise to look again at a paper, weigh up the authors‘ arguments, and

reconsider our decision.

6. Once the paper is provisionally accepted, the peer review process is not over. The

paper is then passed to a scientifically qualified assistant editor who edits the paper‘s

technical content. Mistakes may still be found at this stage, leading to further editorial

or expert review, even (though rarely) rejection. A lesson learned from sometimes

bitter experience is that a paper is not fully accepted until it is published.

7. Here are some of the commonest questions asked about the peer review process (5).

8. Do reviewers make mistakes in their judgments? Of course, and so do editors.

Sadly, the scientific literature is littered with retractions of papers that once passed the

test of peer review.

9. Are reviewers objective in their judgments ? Pure objectivity is impossible. For some subjects, an editor can predict the judgment of the reviewer based on past experience with that reviewer. But this misses the point of what an editor is seeking. It is not simply the judgment of reject/accept that an editor wants from a reviewer. That

decision is the responsibility of the editor and the editor alone. What an editor really

seeks is a powerful critique of the manuscript – testing each assumption, probing

every method, questioning all results, and sceptically challenging interpretations and

conclusions. Armed with that critique, the editors decide – and take full responsibility

for deciding.

10. Are reviewers willing to accept new ideas ? Certainly, they are, although they might question those ideas to destruction. The vast majority of reviewers take their

responsibility as advisors very seriously indeed. They themselves are often on the

receiving end of peer review. Most try to be as open as possible to new findings,

although we encourage them to ask difficult and awkward questions.

11. Despite peer review, are authors able to get away with dishonest or dubious

research? Yes, they are. Peer review does not replicate and so validate research. Peer

review does not prove that a piece of research is true. The best it can do is say that, on

the basis of a written account of what was done and some interrogation of the authors,

the research seems on the face of it to be acceptable for publication. This claim for

peer review is much softer than often portrayed to the general public. Experience

shows, for example, that peer review is an extremely unreliable way to detect research

misconduct.

12. Are peer reviewers accountable for what they do? Yes, to the editor. But in a broader sense, to the scientific community and to the public as well. To a large extent, the trust society places in science depends on the scientific process, including peer review and publication, getting it right most of the time.

13. Does peer review improve the quality of published research? In our everyday

practice, we see that it does. And research suggests that it does too (6). Peer review

improves discussion of the limitations of research. It emphasises uncertainty. It invites

justification of generalisability. As one study of peer review concluded, "peer review

is a negotiation between authors and journal about the scope of the knowledge claims

that will ultimately appear in print" (9).

14. Is there still a need for peer review, given the extraordinary ability of the Internet to enable continuous open criticism of research once published (that is, surely a thousand readers as reviewers after publication are better than 4 reviewers selected by editors before publication)? There is no right answer to this question. Certainly, post-publication peer review adds greatly to the understanding of a piece of research. But watching pre-publication peer review in action - both at the macro level of external expert review and the micro level of technical editing - and seeing the extent to which research papers change (mostly for the better) after peer review, I think that pre-publication review still has an important part to play in science. At its best, pre-publication peer review clarifies, introduces uncertainty, insists on placing new work in the context of the totality of available evidence, demands a careful explanation of limitations, and prevents flights of fanciful over-interpretation.

15. Peer review has changed considerably during the past two decades. First, the stakes are higher. Individual and institutional success depends on getting papers published in high-impact journals. Citation data are now a standard metric for measuring research performance. This trend has increased competition and rivalry for places in the best journals. Second, the globalisation of science has expanded the geographic range of papers submitted to journals. Research originating from China, for example, is now far more common than even five years ago. The

internationalisation of science has further intensified competition for publication. Third, research papers are increasingly multi-disciplinary, requiring a much broader

range of expertise during peer review. Fourth, science is a stronger part of our public

culture now than it once was. What scientists used to write only for other scientists is

today available to – and sometimes read by – non-scientists, policy makers, and the

media. Fifth, the importance of statistics has grown substantially. Whereas twenty

years ago The Lancet had no separate statistical peer review process, every paper we

now publish has been carefully scrutinised by an independent statistical advisor.

Editors are now far more aware of analytic errors in research. Sixth, to address the

often conflicting results of individual research studies that are trying to answer the

same (or a similar) question, a new type of research method has been devised – the

systematic, as opposed to the narrative, review. Systematic reviews aim to search for

particular types of study (eg, the randomised trial), then select only the best according

to pre-specified criteria, and, if possible, to combine those findings in a statistically

meaningful way (which is called meta-analysis). Examples include the risk of cervical

cancer among women taking hormonal contraceptives (7) and the effects of a class

of medicines on heart disease (8). In biomedicine, the Cochrane Collaboration is the

most mature example of an effort to create a database of systematic reviews on

treatments. Finally, editors have had to face an upsurge in the discovery of episodes of research misconduct (fabrication, falsification, and plagiarism). The increasing awareness of research fraud had led not only to greater vigilance (hopefully not suspicion) among editors but also to the birth of institutional mechanisms to set standards and advise on research practice (eg, the Committee on Publication Ethics).

16. Because of the faith journal editors have in peer review, together with the empirical evidence we believe exists to support peer review, we take it very seriously indeed (9). That said, editors are well aware that peer review is anything but

uncontroversial. Scientific discoveries that later turn out to be flagrant episodes of

dishonesty – from Woo-Suk Hwang's fabricated claims in Science about cloning

embryonic stem cells, to Andrew Wakefield's falsifications in The Lancet – are not

uncommon. They raise troubling questions about the robustness of peer review.

Editors are only too well aware of the limitations of the peer-review system. Authors,

for example, can be deeply resistant to responding to questions from anonymous

critics (this fact at least partly drives the argument for fully transparent peer review,

where reviewers have to disclose their names to authors). The reluctance of some

authors – and some very famous authors, at that – to take the comments of their peers

seriously stems from the fact that they believe they have no peers. As one historian of

peer review put it, somewhat poetically, "anyone who possessed the MD degree had

no reason to defer to any colleague as an expert greater than he or she" (10).

17. So what is peer review in today's scientific culture? Various views have been more or less vividly expressed. Peer review is a sacred academic cow, according to one editor (11). Everyone – scientists, the public, policymakers, politicians – would like to believe that peer review is a firewall between truth and error (or dishonesty) (12). But as the editor of one leading specialist medical journal has rightly pointed out, "There is no question that, when it comes to peer review, the reviewers themselves are the weakest (or strongest) links" (13). This frustration among editors and scientists that peer review cannot always live up to the claims sometimes made for it produces frequent expressions of dismay. Is peer review a castle built on sand or the bedrock of scientific publishing (14)? Is peer review a landmark, landmine, or landfill (15)? Or, put bluntly, is peer review simply in crisis? (16). Is it "a flawed process at the heart of science and journals" (20)?

18. Unfortunately, there is evidence of a lack of evidence for peer review‘s efficacy. In 2002, Tom Jefferson and colleagues published a startling systematic review of all the evidence about editorial peer review in biomedical journals. Their exhaustive search yielded only a handful of studies. The conclusion? "Editorial peer review, although widely used, is largely untested and its effects are uncertain" (18). They went on, "Given the widespread use of peer review and its importance, it is surprising that so little is known of its effects." Jefferson and his colleagues have confirmed their

observations more recently (19). Their findings have been replicated by others (20).

To be fair, there is some evidence that micro peer review – technical editing – can

improve papers in biomedical journals (21). But, once again, this evidence is not as

robust as one would either like or have expected.

19. Jefferson extended his investigation of peer review by arguing that the objectives of the review process were also unclear (22). Without clear objectives, proving the value of peer review (or not) would be impossible. After almost 350 years of journal peer review, our zeal for and confidence in the peer review process seem inversely

proportional to our knowledge about what it actually does for science. Those who

make big claims for peer review need to face up to this disturbing absence of

evidence.

20. Worse still, what evidence is slowly accumulating should perhaps make scientists,

policymakers, and the public pause. Many who place great weight on the reliability of

the peer-reviewed scientific literature believe that it reflects the judgment of the

scientific community about the quality of research. But evidence suggests that

acceptance of research for publication may well depend on factors other than

scientific quality alone (23). Furthermore, peer reviewers will disagree greatly in their

recommendations to editors about a particular research paper. Yet editors seem to be

significantly influenced by reviewers who, when the quality of their advice is

measured independently, turn out to be extremely unreliable in their overall

judgments (24). Editors, some critics could reasonably argue, need to pay less, not

more, attention to the recommendations of their peer reviewers.

21. Scepticism about peer review is healthy. But every editor knows that peer review can be an indispensable aid to his or her work. Peer review can rescue science from

embarrassment and error. An extreme example goes some way to showing why. Peter

Duesberg is a well-known molecular virologist who believes that HIV is not the cause

of AIDS. In 2009, the journal Medical Hypotheses published a paper by Duesberg

arguing that the deaths attributed to AIDS in South Africa were false. The editor of

Medical Hypotheses operated an editorial policy of no external peer review. The

justification was that peer review might suppress creative thinking. In the case of the

Duesberg paper, the idea that HIV does not cause AIDS was not new. More

importantly, South Africa is only now reversing its disastrous denialist policies on

HIV-AIDS. To consider Duesberg's old (and discredited) idea at a critical moment for

the country he was writing about would, most reasonable editors might conclude,

require some kind of external peer review to assist decision-making. The editor did

not seek expert reviews. He accepted the paper within a few days of its submission.

Many scientists in the AIDS community were appalled. They wrote to the publishers

(Elsevier, also the publishers of The Lancet ) to complain. Elsevier removed the paper

from its online database pending the results of an independent investigation. The

Lancet was asked to review the paper. We did so and the reviews were uniformly and

deeply critical. No journal could have conceivably published the Duesberg paper

based on these reviews. The Duesberg paper remains retracted, excised from the

scientific literature. Here is an example of what can happen when peer review is

excluded from a journal's processes, and why peer review can bring important

information to bear on judgments about the suitability of research for publication.

Thanks to these events, this particular journal will now implement peer review. Meanwhile, the publishers have found a new editor (25).

Peer review under pressure

22. It is common for editors to have multiple, intense, and sometimes sharp interactions with authors and reviewers. Publication matters. Authors and reviewers are frequently passionate in their intellectual combat over a piece of research. The tone of their exchanges and communications with editors can be attacking, accusatory, aggressive, and even personal. If a research paper is especially controversial and word of it is circulating in a particular scientific community, third-party scientists or critics with an interest in the work may get to hear of it and decide to contact the journal. They might wish to warn or encourage editors. This kind of intervention is entirely normal. It is the task of editors to weigh up the passionate opinions of authors and reviewers, and to reflect on the comments (and motivations) of third parties. To an onlooker, these debates may appear as if improper pressure is being exerted on an editor. In fact, this is the ordinary to and fro of scientific debate going on behind the public screen of science. Occasionally, a line might be crossed. We experienced such a border crossing recently, where several reviewers and third parties encouraged us to delay publication of a paper for non-scientific reasons (26). Defining that line is a crucial task for editors.

23. One issue that is important to solve for the peer review process to work effectively is the full disclosure of all financial and relevant non-financial conflicts of interest. If a

research paper about drug A for disease Y is sent to a reviewer who has shares in a

company that makes drug B, also for disease Y, there is a potential for the introduction of bias into that reviewer's advice to the journal – favouring drug B over drug A. The editor may still want and value that reviewer's advice, but s/he needs to know about the reviewer's financial conflict to judge the weight s/he gives to the review. Non-financial conflicts may be even more important. If a scientist has devoted a life's work to theory A about disease Y, then clearly s/he might be biased if s/he is sent a manuscript that criticises theory A and proposes an alternative and compelling theory B for that same disease. Again, the editor would expect the reviewer to declare any non-financial academic or intellectual conflicts that might have the potential to influence that reviewer's critique.

24. It would be wrong for editors not to listen to advice about publication even after

acceptance of a paper. A paper is only fully accepted when it is published. New

information that informs the decision to publish a provisionally accepted paper before

publication can be very valuable. The Lancet has rejected papers in this twilight zone

of peer review. After publication, criticism is common and welcome, even lethal

criticism. This is the much vaunted self-regulation of science – except that sometimes

editors and authors are reluctant to act when things go wrong after publication.

25. Much has been made of whether scientists should or should not take public positions on the meaning of their data, especially if those data relate directly to policy or practice. The reality is that they do, all the time. Science does not exist in a political

vacuum. The idea that scientists are neutral observers, bereft of opinions, is naïve. In biomedical and public health research, scientists are often quick to make statements applying their data to the real world. They will often do so passionately and be well known for those passionate views. Indeed, the current climate of science is such that scientists are encouraged at every stage of their research to consider the impact – economic or human – of what they do, and to trumpet that impact. Research assessments in the future are likely to include a measure of impact when judging the quality of a scientist‘s work. In relation to peer review, the scientific, policy, or political positions an author, reviewer, or editor may hold could intervene to bias a review in one particular direction. There have been many examples of such conflicts in other scientific disciplines – eg, psychology (27) and genetic epidemiology (28). These episodes are troubling, but an almost inevitable consequence of the way peer review is ordinarily done.

26. The intersection of politics and science in well shown in the field of climate change. The Skeptical Environmentalist , written by Bjørn Lomborg and published by

Cambridge University Press, led to huge pressure on the publishers to withdraw the

book (29). Although the manuscript was reviewed by four experts who all

recommended publication, the scientific backlash was acute. Letters of protest were

written to newspapers. One scientist refused to work with Cambridge University Press

ever again. Lomborg was attacked physically. Chris Harrison, in his thoughtful reflections as the editor at Cambridge University Press who dealt with Lomborg's book, points out that peer review offers no guarantees of always ensuring the truth (32). But in the case of The Skeptical Environmentalist , the concerns were as much political as scientific. The publication of this book by a respected scholarly press might play to a particular political agenda and could be used and abused by vested corporate and political interests. Harrison rejected the idea that he should have applied these kinds of value judgments in the editorial process. He defended the scholarly publishing industry's commitment to pluralism.

27. This commitment to pluralism would be the likely view of many scientific editors,

even when controversy follows. One might conclude that these kinds of extreme debate, although difficult, are part of the normal fabric of scientific discourse. The question to be answered is: where is the line to be drawn between vigorous scientific exchange and improper attempts to close down debate (these two positions can be remarkably close to one another)? But one should also be conscious of what some observers have described as the "chilling" effect of political controversy on science. A survey of US National Institutes of Health scientists revealed that many engaged in self-censorship after they found themselves the subject of political criticism for their work (33). Political disagreement over science can shape not only the behaviour of scientists but also the future of science itself. Increasingly, commercial, as well as political, interests are also intervening to threaten the integrity of peer review (34).

28. Peer review and publication can provoke important questions about access to data.

During the review process, reviewers may seek more information. Except in

allegations of fraud, it would be highly unusual to provide or request raw data (even

then, journals expect institutions to take responsibility for investigating the

authenticity and reliability of original data). But access to data may be sought after

publication. This is a highly contentious and unresolved issue. In the field of medicine, these issues are currently the subject of much disagreement. While many parties might like to see greater sharing of data, this practice remains unusual. The Wellcome Trust is taking an especially strong interest in data access. It has proposed a code of conduct calling for "maximum public access to data of public health importance." The very fact that this proposal is being made illustrates the point that routine access to data is not a settled issue or a universal norm in science, as some

claim.

29. The issue of retention of records and exclusion of data is also a matter relevant to the peer review process and the ordinary working of journals. Journals do expect records to be kept for limited periods (say, 5 years, although journal practices vary). And they are comfortable with the exclusion of data provided that those exclusions – and the reasons for exclusion – are fully described, with appropriate sensitivity analyses being completed where necessary.

30. Two additional dimensions of peer review must be noted. One relates to confidentiality, the other to uncertainty. Editors send manuscripts to reviewers based on a principle of confidentiality. The author expects the editor to maintain a covenant of trust between the two parties. The editor will not misuse the author's work by circulating it outside of the confidential peer review process. The editor expects that covenant of trust to be honoured by the peer reviewer. No manuscript should be passed to a third party by a reviewer without the permission of the editor, usually on the grounds of improving the quality of the critique of the manuscript by involving a colleague in the review process. A disclosure to a third party without the prior permission of the editor would be a serious violation of the peer review process – a breach of confidentiality. It is also of paramount importance to report fully in all published scientific papers both quantitative and qualitative measures of uncertainty. One of the main benefits of peer review is to focus on areas of potential uncertainty and to ensure that those uncertainties are fully acknowledged, measured, and reported.

The future of peer review

31. Peer review is a human process and so will always contain flaws, produce errors, and occasionally mislead. Given that journals are the gatekeepers of scientific publication, they have enormous – probably too much – influence over the reputations of scientists, research units, and universities. Many measures of academic success

depend upon journal publication – promotion, tenure, grants, fame, and personal

wealth. It is not surprising that journals, and the main decision aid used by journals

(peer review), are the subject of constant tension and occasionally explosive

controversy. At such moments, it is not only essential to be clear (and modest) about

what peer review can do, but also to look for opportunities to do better. Journal articles are highly stylized reports of research. The linear and logical style of the research they report rarely presents a true or accurate picture of how a piece of

research was done. As the Nobel laureate, Peter Medawar, put it (32) in his essay, Is the Scientific Paper a Fraud? (to which he answered that it was),

"[the scientific paper] misrepresents the processes

of thought that accompanied or give rise to the work

that is described in the paper...The scientific paper in

its orthodox form does embody a totally mistaken

conception, even a travesty, of the nature of scientific

thought."

32. Medawar‘s point was that, "There is no such thing as unprejudiced observations." To add insult to injury, research papers may not even fully represent the views of the

authors who completed the work (33), and when faults are found after publication

those faults may be completely ignored in the subsequent use of that research (34).

There are actions that the scientific community could take to improve this far from

happy state of affairs surrounding one its foundational processes. First, there are new

opportunities and techniques available to search out, identify, and eliminate (or at

least reduce) unwanted bias in the peer review process (35, 36). Second, all young

scientists should receive formal training – which they currently do not – in the

standards and ethics expected in the peer-review process (37). It is scandalous that

peer review is simply not taken as seriously as it should be in the training of scientists.

The result is that peer review is often idiosyncratic and sometimes unreliable, fueling

scientific controversies, such as that over climate science, rather than defusing those

controversies. Strengthening the training, standards, and expectations around peer

review would do much to make the quality of peer reviewing part of the formal

appraisal of a scientist‘s contribution to his or her subject. There is a demand for

training in peer review (38). And the ethical dimensions of the review process are

now sufficiently concerning to scientists that they merit training as much as the more

formal methodological aspects of reviewing (39). Disappointingly, existing training

packages in peer review deliver little benefit to the quality of the peer review process

(40-42). Third, the peer review process is enormously inefficient. Individual journals will undertake peer review and reject manuscripts that will then cycle around other

journals until either the paper is accepted or the authors are sufficiently exhausted that

they abandon attempts at publication. In the face of such gross inefficiencies, some

scientific communities have tried to bring journals together to cooperate and make the

review process not only more efficient, but also less costly on the time and energy of

reviewers, authors, and editors (43). Alternatively, there may be intra-journal

procedures that can be introduced to deliver more efficient peer review (44). Fourth, journal editors should adopt more effective methods to resolve disputes between authors, reviewers, and readers. Within the journal, an ombudsperson operating independently of the editors can be one useful way to resolve intractable disagreements about journal processes (45). If a dispute remains impossible to resolve, journal editors can take their concerns to the Committee on Publication

Ethics, a charity that aims to set standards for journal practices, including peer review.

Journal editors should consider using this facility more often than they currently do –

in some ways, it represents the collective wisdom of a wide range of journal editors, a

collective wisdom that any scientific editor can draw upon in times of crisis. Lastly, peer review should be a subject for research in its own right. Although there is

a small group of scientists who study peer review (a biomedical peer review congress

is held every 4 years), that community is extraordinarily fragile when measured

against the size and importance of the contribution peer review makes to science (46). Historically, science funding bodies have been reluctant to invest in research on peer

review. This reluctance is partly responsible for the present vacuum in our knowledge

about the way scientific knowledge is constructed, reported, and discussed. One

positive result of the debate over the role of CRU scientists in peer review might be to

encourage funding bodies – such as the Medical Research Council and the National

Institute for Health Research – to take the science of peer review far more seriously.

33. Journals have inevitable limitations. When a paper with important policy implications is considered, editors can ask authors to balance their conclusions by putting the work in the context of existing evidence. Or we can commission an editorial that does the same. But a journal cannot adjudicate a public debate, and neither can conventional peer review. For those occasions when science meets (or clashes with) policy, there may be a case for referring that area of controversy to an independent body for a public inquiry - a National Agency for Science and Health, for example.

34. The best one might hope for the future of peer review is to be able to foster an

environment of continuous critique of research papers before and after publication.

Many writers on peer review have made such a proposal, yet no journal has been able

to create the motivation or incentives among scientists to engage in permanent peer

review (47-49). Some observers might worry that extending opportunities for

criticism will only sustain maverick points-of-view. However, experience suggests

that the best science would survive such intensified peer review, while the worst

would find its deserved place at the margins of knowledge. This process of weeding out weak research from the scientific literature can be accelerated through more formal mechanisms, such as the systematic review. A systematic approach to selecting evidence focuses on the quality of scientific methods rather than the reputations of scientists and their institutions. This more rigorous approach to gathering, appraising, and summing up the totality of available evidence has been profoundly valuable to clinical medicine.

35. More importantly, intensified post as well as pre publication review would put uncertainty – its extent and boundaries – at the centre of the peer review and publication process. This new emphasis on uncertainty would limit the rhetorical power of the scientific paper (50), and offer an opportunity to make continuous but constructive public criticism of research a new norm of science

1. Montford A. The Hockey Stick Illusion: Climategate and the corruption of

science (Stacey International, 2010)

2 . Montford A. Heated discussions. Times Higher Education March 25, 2010:

43-44.

3. Monbiot G. Our narrow, antiquated school system is at the root of the climate email fiasco. The Guardian April 6, 2010: 25.

4. Spier R. The history of the peer review process. Trends in Biotechnology

2002; 20: 357-58.

5. Hernon P, Schwartz C. Peer review revisited. Library and Information Science

Research 2006; 28: 1-3.

6. Goodman SN, Berlin J, Fletcher SW, Fletcher RH. Manuscript quality before

and after peer review and editing at Annals of Internal Medicine . Ann Intern

Med 1994; 121: 11-21.

7. Smith JS, Green J, de Gonzalez AB, et al. Cervical cancer and use of

hormonal contraceptives: a systematic review. Lancet 2003; 361 : 1159-67.

8. Jun M, Foote C, Lv J, et al. Effects of fibrates on cardiovascular outcomes: a

systematic review and meta-analysis. Lancet 2010

9. Green SM, Callaham ML. Current status of peer review at Annals of

Emergency Medicine . Ann Emerg Med 2006; 48: 304-08.

10. Burnham JC. The evolution of editorial peer review. JAMA 1990; 263: 1323-9.

11. Fitzpatrick JJ. Peer review: a 2008 report on the sacred academic cow.

Applied Nursing Research 2008; 21: 53.

12. Harms M. Peer review: the firewall of science. Physiotherapy 2006; 92: 193-

94.

13. DeMaria AN. Peer review: the weakest link. JACC 2010; 55: 1161-62.

14. Berger E. Peer review: a castle built on sand or the bedrock of scientific

publishing? Ann Emerg Med 2006; 47: 157-59.

15. Balistreri WF. Landmark, landmine, or landfill? The role of peer review in

assessing manuscripts. J Pediatr 2007; 151: 107-08.

16. Mulligan A. Is peer review in crisis? Oral Oncology 2005; 41: 135-41.

17. Smith R. Peer review: a flawed process at the heart of science and journals. J R Soc Med 2006; 99 : 178-82.

18. Jefferson T, Alderson P, Wager E, Davidoff F. The effects of editorial peer

review. JAMA 2002; 287: 2784-86.

19. Jefferson T, Rudin M, Brodney Folse S, Davidoff F. Editorial peer review for

improving the quality of reports of biomedical studies. Cochrane Database

Syst Rev 2007; 2 : MR16.

20. Richards D. Little evidence to support the use of editorial peer review to

ensure quality of published research. Evid Based Dent 2007; 8 : 88-89.

21. Wager E, Middleton P. Technical editing of research reports in biomedical

journals. Cochrane Database Syst Rev 2008; 4: MR2.

22. Jefferson T, Wager E, Davidoff F. Measuring the quality of editorial peer

review. JAMA 2002; 287: 2786-90.

23. Aarssen LW, Lortie CJ, Budden AE, et al. Does publication in top-tier

journals affect reviewer behaviour? PLoS ONE 2009; 4 : e6283.

24. Kravitz RL, Franks P, Feldman MD, et al. Editorial peer reviewers‘

recommendations at a general medical journal: are they reliable and do editors

care? PLoS ONE 2010; 5 : e10072.

25. Enserink M. Elsevier to Editor: change controversial journal or resign. Science

2010; 327 : 1316.

26. Horton R. Maternal mortality: surprise, hope, and urgent action. Lancet 2010;

27. McCarty R. Science, politics, and peer review: an editor‘s dilemma. American

Psychologist 2002; 57: 198-201.

28. Calnan M, Smith GD, Sterne JA. The publication process itself was the major

cause of publication bias in genetic epidemiology. J Clin Epidemiol 2006; 59:

1312-18.

29. Harrison C. Peer review, politics, and pluralism. Environmental Science and

Policy 2004; 7: 357-68.

30. Kempner J. The chilling effect: how do researchers react to controversy. PLoS

Medicine 2008; 5 : e222.

31. Curfman GD, Morrissey S, Annas GJ, Drazen JM. Peer review in the balance.

N Engl J Med 2008; 358 : 2276-77.

32. Medawar PB. The Strange Case of the Spotted Mice (Oxford, 1996).

33. Horton R. The hidden research paper. JAMA 2002; 287 : 2775-78.

34. Horton R. Postpublication criticism and the shaping of knowledge. JAMA

2002; 287 : 2843-47.

35. Bornmann L, Mutz R, Daniel H-D. How to detect indications of potential

sources of bias in peer review. J Informatics 2008; 2 : 280-87.

36. Ross JS, Gross CP, Desai MM, et al. Effect of blinded peer review on abstract acceptance. JAMA 2006; 295 : 1675-80.

37. Walbot V. Are we training pit bulls to review our manuscripts? J Biol 2009; 8 :

24.38. Snell L, Spencer J. Reviewers‘ perceptions of the peer review process for a medical education journal. Med Educ 2005; 39 : 90-97.

39. Resnik DB, Gutierrez-Ford C, Peddada S. Perceptions of ethical problems

with scientific journal peer review. Sci Eng Ethics 2008; 14 : 305-10.

40. Schroter S, Black N, Evans S, et al. Effects of training on quality of peer

review. BMJ 2004 doi: 10.11136/bmj.38023.700775.AE.

41. Callaham ML, Tercier J. The relationship of previous training and experience

of journal peer reviewers to subsequent review quality. PLoS Medicine 2007;

4 : e40.

42. Schroter S, Black N, Evans S, et al. What errors do peer reviewers detect, and

does training improve their ability to detect them? J R Soc Med 2008: 101 :

507-14.

43. Saper CB, Maunsell JHR. The Neuroscience Peer Review Consortium. Brain

Res 2009; 1272 : 1-2.

44. Johnston SC, Lowenstein DH, Ferriero DM, et al. Early editorial manuscript

screening versus obligate peer review. Ann Neurol 2007; 61: A10-12.

45. Horton R. The journal ombudsperson: a step toward scientific press oversight.

JAMA 1998; 280 : 298-99.

46. Linkov F, Lovalekar M, LaPorte R. Scientific journals are faith based‖: is

there science behind peer review? J R Soc Med 2006; 99 : 596-98.

47. von Segesser LK. Peer review versus public review – new possibilities of online publishing! Interactive Cardiovascular and Thoracic Surgery 2002; 1:

61-62.

48. Mandviwalla M, Patnayakuni R, Schuff D. Improving the peer review process

with information technology. Decision Support Systems 2008; 46: 29-40.

49. Liesegang TJ. Peer review should continue after publication. Am J Ophthalmol 2010; March; 149 : 359-60.

50. Horton R. The rhetoric of research. BMJ 1995; 310 : 985-88.

Note: A longer version of this paper was submitted to the Muir Russell inquiry into the events that took place at the Climatic Research Unit of the University of East Anglia.

Declaration of Interest: I edit a medical journal, The Lancet.

Richard Horton

Editor

The Lancet

9 February 2011