Professional Documents
Culture Documents
Literature on the contribution of peer studied the effect of any aspect of the
review to improving quality peer-review process, such as blinding identi-
ties of reviewers or authors, provision of
In the past two decades, some attempts have submission checklists, and training review-
been made to test the effect of peer review in ers. The majority of the studies identified by
improving the quality of manuscripts. Sev- Jefferson et al. focused on specific editorial
eral studies have compared the initial procedures; only two assessed the effect of
submissions with their published versions and peer review on the quality of manu-
provided evidence that peer review helped scripts.15,21 Both studies revealed beneficial
improve a manuscript in terms of its pre- effects, but they may have only limited
sentation and readability,14–17, reporting of generalizability, due to atypical settings.
statistics,18 accuracy of content,19 and over-
the proportion all quality of the manuscript.20–22 Studies of
There are also problems with reviewers.
of favourable There is ample evidence that editors’
reviewers’ comments revealed that most decisions were closely related to the recom-
reviewers’ comments, even those with a highly critical mendations of reviewers, and the proportion
comments is far tone, showed the reviewers’ commitment to of favourable reviewers’ comments is far
helping authors.4,23,24 Some authors also more influential than the characteristics of
more influential evaluated the helpfulness of the reviewers’ authors and manuscripts.13,22,23,32–37 Ensur-
than the comments that they received.25–28 While ing the quality of reviewers and reviewing is
characteristics the majority of the responding authors, therefore crucial. However, reviewers were
especially those responding to a recent inter-
of authors and national survey,28 favoured the peer-review
frequently criticized for their tardiness in
returning reviews or for returning reviews of
manuscripts system, it should be noted that the voice of poor quality.38–40
unsuccessful authors was absent in these
studies. There was also evidence that many
The disciplinary focus of studies on peer
authors experienced pressure to make revi-
review
sions to conform to editors’ or reviewers’
personal preference, or to make changes that Existing empirical and anecdotal evidence
they felt were wrong.25,29 Even with review- on editorial peer review is patchy in its cov-
ers’ helpful comments, many authors still erage of major disciplines. In the pursuit of a
struggled in revising their manuscripts, systematic review of the literature, the
which implied that reviewers’ contribution researcher developed a database of all rele-
to improving a manuscript was somewhat vant literature. This database contained
limited.4,30,31 Jefferson and colleagues11 con- more than 1,200 items that appeared in jour-
ducted a systematic review of articles that nals or books in all key disciplines from 1960
reported studies on peer review published in to the present, including 385 research arti-
biomedical journals up to June 2000. Of the cles, 650 essays and commentaries that
135 articles that they identified, only 19 were not based on original data but also
Table 1. Distribution of literature on peer review (excluding books and book chapters)
Discipline Research literature Essays, comments, and editorials
n % (n = 385) n % (n =650+199)
Business, economics, management 32 8.3 31 3.7
Sociology, social works, humanities 31 8.1 52 6.1
Education 11 2.9 31 3.7
Science, information, mathematics 15 3.9 51 6.0
Medicine, biology 176 45.7 261 30.7
Physics, chemistry 11 2.9 37 4.4
Psychology 49 12.7 68 8.0
Cross-disciplinary, general 60 15.6 318 37.5
Total 385 100 849 100
contributed to the debate, and 199 editorials Every fourth name in the list (232 individu-
(Table 1). Twenty books were also identified; als in all) was sent a four-page survey (see
13 of these were focused exclusively on peer Appendix). The survey comprised 44 ques-
review and the remainder were about scien- tions, divided into five sections: (1) back-
tific communication and journal publication ground information about the reviewers; (2)
in general, but also covered the topic of peer their opinions of peer review as a method of
review to a substantial extent. Most of the research assessment in their field; (3) their
authors of those works based their discus- experiences as reviewers and authors; (4)
sions or studies on such fields as medicine, their level of success in grant application
psychology, business, and economics; there and journal publication; and (5) an open-
was considerably less coverage from educa- ended question to allow respondents to com-
tion or physics and chemistry. Drawing on ment freely. A total of 84 usable responses
this lack of research in education, physics, from 27 universities were received; this rep-
and chemistry, the researcher decided to resents a response rate of 36.2%, which is
focus her investigation on these three fields. within the common range for mail sur-
veys.41,42 Pearson’s chi-square was calculated
to test for systematic non-response bias by
The questionnaire survey comparing the number of respondents to
This paper reports some of the findings of a that of non-respondents for each discipline.
Ph.D. study undertaken by the researcher. The resultant chi-square χ2(1) =0.72, P =
The study consisted of six phases: a system- 0.337 indicated no evidence of non-response
atic review of the literature; a mapping of bias.
editorial policies of all existing English- The full questionnaire is given in the
language scholarly journals in education; a Appendix.
pre-interview survey; interviews; a post-
interview survey; and an analysis of review- Results
ers’ comments. This paper draws on the
findings of the pre-interview survey. Whilst
Demographic information
whilst peer
peer review is something that involves all review is
members of the research community, the Table 2 summarizes information on the
exercise of power is often restricted to a demographic and academic background of
something that
dominant few. Academic élites who, in the the reviewers. Most of the respondents were involves all
current context, are defined as long-standing experienced researchers and reviewers. All members of the
researchers and reviewers, are predominant of the 84 respondents had a Ph.D. degree; 37 research
shapers of the peer-review process. It is for (44%) had achieved their doctorate more
this reason that the study’s target sample than 20 years ago. There were 26 professors, community, the
was professors, associate professors, and 30 associate professors, and 28 senior lectur- exercise of
senior lecturers who had served as journal ers; 37 (44%) were from the ‘Group of Eight’ power is often
reviewers in education, physics, and chemis- – Australia’s leading universities. No respon-
try in 37 Australian universities (excluding dent was younger than 35 years old; 28 restricted to a
the author’s own university, the University (33.3%) were between 35 and 49, and 56 dominant few
of Newcastle, to avoid bias). (66.7%) were 50 or older. In the field of edu-
By browsing university websites, a list of cation, 25 males and 22 females responded
930 academics was generated. Physics and and, in physical sciences, 34 males and 3
chemistry were treated as one group – physi- females. Thirty-nine (46.4%) of the respon-
cal sciences – for three reasons: (i) both are dents had served as editor for peer-reviewed
laboratory-based; (ii) physics journals often journals; all had been journal reviewers for
embrace chemistry in their scope, and vice at least 4 years and 61 (72.6%) had reviewed
versa; and (iii) the total population of senior for more than 10 years. They reviewed an
academics in physics and chemistry is average of 14.2 manuscripts per year. Seven
approximately the same as that in education, respondents (8.3%) had served on the
and combining physics and chemistry there- assessment panels of the Australian Research
fore facilitates comparison with education. Council (ARC), the major national research
Table 3. Reviewers’ expectations and perspectives of peer review: means and correlations
Six-point Correlation Compare means
Likert
scale Age Years being a Sex Discipline
reviewer
Statistical test Mean (SD) Kendall’s τ t-test (2-tailed)
What is your impression of the 5.00 τ = 0.212* t(82)=1.79 t(82) = –2.23
effectiveness of peer review? (0.826) P = 0.027 P = 0.077 P = 0.029
To what extent should peer review 5.45 t(82)=2.70
aim to improve manuscript quality? (0.916) P = 0.008
To what extent can peer review 4.88
improve manuscript quality? (1.087)
How often do editors follow 4.98 τ = 0.165 τ = 0.228* t(81) = 2.15 t(81) = –2.27
reviewers’ recommendations? (0.604) P = 0.078 P = 0.019 P = 0.035 P = 0.026
Do reviewers bring different 4.93
expectations to grant and journal (1.045)
article review?
(reviewed for 10 years or longer) were less whether they had ever refused to act as a
supportive of the idea that it was up to peer reviewer, either for some papers or for some
review to improve quality. To clarify this journals. Those answering ‘yes’ were asked
point, a reviewer wrote that ‘too many to give two reasons for their refusal. Fifty- although
authors used peer review to improve their four (64.3%) responded ‘yes’, and gave a
papers . . and often submitted papers that total of 100 reasons (Table 4). In 37 cases reviewers
were not up to the standard of submission’. reviewers chose not to review some papers generally
The third question ‘To what extent can because they were outside their field of expected peer
peer review improve manuscript quality?’ expertise; in 23 cases they did not have
was rated somewhat lower than the second enough time; and 4 indicated conflict of review to
one, which suggests that, although reviewers interest because they knew the authors. Sim- improve
generally expected peer review to improve ilarly, in 14 cases the reviewers declined to manuscript
manuscript quality, they were less certain of review for some journals due to lack of
its ability to achieve this. Whether reviewers expertise; in 10 cases because of time pres-
quality, they
felt that editors followed their recommenda- sure; and in 12 because they thought the were less
tions (question 4) was significantly related quality of the journal was poor or they had certain of its
to reviewers’ level of seniority, in terms of been treated badly by the journal in the past. ability to
age and years of reviewing; physicists and The second question asked reviewers to
chemists rated it significantly higher than give two reasons why they did agree to achieve this
education academics (P < 0.05). Sixty-one
(72.6%) of the journal reviewers also had Table 4. Reasons for declining to review
the experience of assessing research projects
for grant application. They generally brought Reasons Frequency (%)
different expectations to the review of grants For paper For journal Total
and journal articles (mean = 4.93), but no
significant difference was found by disci- Lack of 37 (57.8%) 14 (38.9%) 51 (51%)
expertise
pline, seniority, or sex.
Lack of 23 (35.9%) 10 (27.8%) 33 (33%)
time
Reviewers’ experiences of peer review Other 4 (6.3%) 12 (33.3%) 16 (16%)
Questions in section 3 sought responses on Total 64 (100%) 36 (100%) 100 (100%)
aspects of the reviewers’ experiences of peer answers
review. The first question asked reviewers offered
review. A total of 134 reasons was given; no one said ‘yes’. The last question of Sec-
these were grouped into five categories (Table tion 3 asked reviewers whether grant or
5). Each reason was coded once per individ- journal reviewing was the more important
ual, so the total number of reasons exceeded activity for them. Among the 58 reviewers
the total number of respondents. The who responded to this question, 35 review-
reviewers regarded their contribution firstly ers thought journal peer review was more
reviewers did as part of their professional responsibility, important, as opposed to 7 who thought
not generally which was mentioned 48 times (35.8%). grant peer review was more important. Six-
regard Some reviewers regarded peer review as a teen felt the two were equally important
good opportunity for professional develop- activities.
reviewing as ment, in the sense of ‘keeping abreast of
a means to current developments in the field’, ‘getting a
better feeling of what papers are accepted’, Respondents’ overall level of success as
improve the researchers
and gaining ‘something to put on CV’. Some
quality of reviewers agreed to contribute their time The next section of the questionnaire looked
manuscripts or and expertise with the aim of keeping the at respondents’ own level of success in terms
to maintain the peer-review system going and to ‘share of gaining grants and publishing their
the load with other academics’. While the research outcomes (Table 6). For the period
quality of reviewers deemed reviewing as ‘an expecta- 2001–5, 67 (79.8%) of the respondents won
journal tion of my academic job’, they did not one or more competitive grants from the
publications generally regard it as a means to improve the ARC or the Australian National Health
quality of manuscripts or to maintain the and Medical Research Council (NHMRC);
quality of journal publications, reasons 40 (47.6%) won competitive grants from
which were mentioned by only 12 of them other national funding bodies; 45 (53.6%)
(9%). received institutional research grants; and
The next three questions solicited infor- 83 (98.8%) published articles in peer-
mation about guidelines for reviewers. reviewed journals. The success rates self-
Nearly 80% of the journals for which respon- reported by individual respondents (e.g. pub-
dents reviewed almost always provided them lications against submissions) were high for
with guidelines. All but two reviewers most, ranging from 38 to 100%. As shown in
almost always followed the guidelines, and Table 6, on average, academics from physics
65 (77.4%) felt the guidelines helped them and chemistry reported higher rates of suc-
generate better reviews, with the relation- cess in publication (93.6%) than those in
ship between their following the guidelines education (90.9%), and in winning grants
provided and ultimately generating better from the ARC and NHMRC (64.6% com-
reviews significantly positive, r = 0.35, P = pared to 44.7% – statistically significant, P
0.02. There was no significant difference in < 0.05) and other national funding bodies
reviewers’ perspectives by discipline. (70.8% compared to 67.1%). In contrast,
Next, reviewers were asked whether they education academics won significantly more
had been paid by any journal for reviewing; grants from their own institutions (95.1%
compared to 80.3%, P < 0.05). In addition, Another reviewer described his experience
Pearson coefficient correlations were calcu- as an author, feeling disadvantaged by the
lated to test for the impact of personal anonymous reviewing system:
success on their impression of the peer-
review system. Respondents’ overall impres- On a number of occasions, I have submit-
sion of the effectiveness of peer review was ted a paper to a journal, had it rejected,
significantly related to their success rates in submitted it again to a different journal
gaining grants from the national funding and had it accepted. This is not explained
bodies of ARC and NHMRC (r = 0.45, P < by the rejecting journals having extremely
0.001) and successful publication (r = 0.24, high standards, but by unprofessional ref-
P < 0.05). ereeing. The anonymous review, which I
guess was imported from science, has
major flaws in my field, which is badly
Reviewers’ comments affected by ideological wars. Current
Lastly, respondents were invited to make
practice of anonymous reviewers and some reviewers
non-anonymous authors is inappropriate,
any additional comments or suggestions
and encourages reviewers to hide behind
maintained
about peer review; 46 of them shared their that subjectivity
their anonymity.
thoughts. The reviewers generally valued
peer review as ‘the key to maintaining qual- Several reviewers expressed their concern was inevitable
ity’ and supported its function in providing about the lack of induction into the review in peer
an imprimatur. Conversely, some reviewers process – academics should ‘have the neces- judgement
maintained that subjectivity was inevitable sary skills to peer review but these skills are
in peer judgement; the process sometimes rarely (if ever) discussed’. There was also
took an unnecessarily long time and sub- criticism of publishers in that they ‘made use
stantially slowed the publication process; of academics without paying them but yet
some noted that, as authors, they had charged academics and libraries extortionate
encountered ‘very irresponsible reviewers’ fees to access works whose quality was
and some reviews were ‘disappointingly brief’ ensured by academics’.
and ‘surprisingly unhelpful’. One reviewer Many respondents pointed out that it was
commented: increasingly difficult to find time to review
promptly and thoroughly. Some stated they
The peer-review process works well when always tried to give timely and useful feed-
reviewers are concerned to promote schol- back because they understood ‘the review
arship and knowledge in an area of study. process can take too long’. Again, some
The peer-review process does not work reviewers emphasized the heavy workload
well when it is used by reviewers to ex- and time pressure they faced and appealed
clude promotion of alternative views other for greater recognition of their contribution
than their own or when reviewers are not by employing institutions, as ‘an important
sensitive to the feelings of authors and aspect of academics’ work’. A reviewer sug-
write inappropriate and sometimes offen- gested a possible solution to the shortage of
sive reviews. time:
There is a lot of duplication and time were reported as being dissatisfied with the
wasted through submissions to different process because they thought the comments
journals, each conducting its own peer re- they received from reviewers or editors were
view. A central agency conducting peer unhelpful.29,46 However, from the outset
review that would make it available to journals have placed the ultimate responsi-
journals would improve this situation. bility for the quality of the manuscript
squarely with the author. This divergence in
expectation can lead to considerable frustra-
Discussion tion for authors, reviewers, and editors.
This study has its limitations. For example, The reviewers’ reasons for agreeing or
the sample was small and entirely limited to declining to review conformed to the find-
successful Australian universities, and the response ings of a recent study in medicine. In a
researchers and rate was not high enough for safe generaliza- survey of 551 medical journal reviewers, Tite
experienced tion. This study did not attempt to study less and Schroter47 found that the most frequent
experienced and less powerful members of reasons for agreeing to review were: contri-
reviewers were bution of the paper to the field; relevance of
the research community or those from other
generally disciplines. Yet the responses still conveyed the topic to own field of interest; opportu-
satisfied important messages. While the respondents’ nity to learn from the paper; and sense of
overall attitude towards the effectiveness of professional duty, while the main reasons for
with the declining to review were all related to lack of
peer review differed significantly by disci-
effectiveness of pline, the expectations and understandings time: conflict with other workload; having
the current that they brought to reviewing did not. to review too many manuscripts for other
system of Their approach to reviewing was primarily journals; and tight deadlines for completing
related to their level of seniority as academ- the review.
editorial peer ics and as reviewers, rather than to their The reviewers’ responses to the question
review discipline. of why they declined to review raised a ques-
The successful researchers and experi- tion that was not anticipated when the
enced reviewers in this study were generally survey was developed. A majority of the
satisfied with the effectiveness of the current reviewers indicated that they had received
system of editorial peer review, irrespective some manuscripts for which the subject was
of their discipline, sex, or academic status. outside their field of expertise (51%, see
Most respondents rated ‘should improve Table 4). This would seem to be a common
manuscript quality’ highly as the main func- problem; author responses as reported in
tion of peer review (though there was a lack several previous studies suggested that edit-
of consistency), but this was not their pri- ors might not necessarily know who would
mary reason for undertaking the job of be a competent reviewer in a particular
reviewing (Table 5). The system of peer area,25,27,29,48,49 which, according to Calla-
review was put in place by journals in the ham and Tercier,50 is a ‘crucial limitation in
17th century in order to assist editors in the the peer review process’.
selection of manuscripts for publication.1,43 The reviewers in this study took the task
Numerous editorials in various journals indi- of reviewing seriously. Their contribution in
cate that journal editors regard peer review reviewing is all the more precious in view of
as the best tool for ensuring the scientific the fact that they were seldom rewarded
quality of publications – a tool for both explicitly, either by their institutions or by
gatekeeping and quality control. This study the journals. However, when time pressure is
confirmed the idea that reviewers think so predominant, as suggested by Tite and
too. Schroter,46 what could the author expect
Do academic authors expect peer review from the reviewer other than a tick in the
to improve the quality of their own manu- box of ‘Accept’ or ‘Reject’ and probably a
scripts? It would seem that many do. Several line or two pointing out the most obvious
studies have shown that authors generally flaws of the article?
expect their submissions to be improved in At present, peer review is supposed to
the peer review process.8,28,44,45 Authors identify high-quality research. However, so
far no agreement has been reached on objec- 14. Biddle, C. and Aker, J. 1996. How does the peer
review process influence AANA Journal article read-
tive ways to measure improvement in ability? Journal of the American Association of Nurse
manuscript quality, nor is it possible to dis- Anesthetists, 64: 65–8.
tinguish between the quality improvement 15. Pierie, J.-P.E.N., Walvoort, H.C. and Overbeke,
brought about by peer review and that A.J.P.M. 1996. Readers’ evaluation of effect of peer
review and editing on quality of articles in the
resulting from subsequent editing. As a Nederlands Tijdschrift voor Geneeskunde. The Lancet,
result, it is not possible objectively to deter- 348: 1480–3.
mine the effect of peer review on manuscript http://dx.doi.org/10.1016/S0140-6736(96)05016-7
quality. Peer review still has to be regarded 16. Metoyer-Duran, C. 1993. The readibility of published,
accepted, and rejected papers appearing in College &
it is not possible
as an under-studied process with uncertain Research Libraries. College & Research Libraries, 54: objectively to
outcomes. 517–25.
17. Roberts, J.C., Fletcher, R.H. and Fletcher, S.W. 1994. determine the
References Effects of peer review and editing on the readability of
articles published in Annals of Internal Medicine. Jour-
effect of peer
1. Zuckerman, H. and Merton, R.K. 1971. Patterns of nal of the American Medical Association, 272: 119–21. review on
evaluation in science: institutionalization, structure http://dx.doi.org/10.1001/jama.272.2.119
and functions of the referee systems. Minerva, 9: 18. Cobo, E., Selva-O’Callagham, A., Ribera, J.-M., manuscript
66–100.
http://dx.doi.org/10.1007/BF01553188
Cardellach, F., Dominguez, R., and Vilardell, M. 2007.
Statistical reviewers improve reporting in biomedical
quality
2. Jefferson, T., Wager, E. and Davidoff, F. 2002. Measur- articles: a randomized trial. PLoS One, 3: 1–8.
ing the quality of editorial peer review. Journal of the http://dx.doi.org/10.1371/journal.pone.0000332
American Medical Association, 287: 2786–90. 19. Purcell, G.P., Donovan, S.L. and Davidoff, F. 1998.
http://dx.doi.org/10.1001/jama.287.21.2786 Changes to manuscripts during the editorial process:
3. Faupel, C. 2002. Editor’s comments. Sociological characterizing the evolution of a clinical paper. Jour-
Inquiry, 72: 1–3. nal of the American Medical Association, 280: 227–8.
http://dx.doi.org/10.1111/1475–682X.00001 http://dx.doi.org/10.1001/jama.280.3.227
4. Gosden, H. 2003. ‘Why not give us the full story?’ 20. Campanario, J.M. 1996. Have referees rejected some
Functions of referees’ comments in peer reviews of sci- of the most-cited articles of all times? Journal of the
entific research papers. Journal of English for Academic American Society for Information Science, 47: 302–10.
Purposes, 2, 87–101.
http://dx.doi.org/10.1002/(SICI)1097–4571(199604)4
http://dx.doi.org/10.1016/S1475–1585(02)00037–1
7:4<302::AID-ASI6>3.0.CO;2-0
5. Miser, H.J. 1998. The easy chair: journal editing as I
21. Goodman, S.N., Berlin, J., Fletcher, S.W. and Fletcher,
see it. Interfaces, 28: 115–23.
R.H. 1994. Manuscript quality before and after peer
6. Abby, M., Massey, M.D., Galandiuk, S. and Polk,
review and editing at Annals of Internal Medicine.
H.C. 1994. Peer review is an effective screening pro-
Annals of Internal Medicine, 121: 11–21.
cess to evaluate medical manuscripts. Journal of the
http://www.annals.org/cgi/content/full/121/1/11
American Medical Association, 272: 105–7.
22. Smigel, E.O. and Ross, H.L. 1970. Factors in the edi-
http://dx.doi.org/10.1001/jama.272.2.105
7. Horrobin, D. F. 1990. The philosophical basis of peer torial decision. The American Sociologist, 5: 19–21.
review and the suppression of innovation. Journal of 23. Beyer, J.M., Chanove, R.G. and Fox, W.B. 1995. The
the American Medical Association, 263: 1438–41. review process and the fates of manuscripts submitted
http://dx.doi.org/10.1001/jama.263.10.1438 to AMJ. Academy of Management Journal, 38:
8. Sweitzer, B.J. and Cullen, D.J. 1994. How well does a 1219–60.
journal’s peer review process function? A survey of 24. Jauch, L.R. and Wall, J.L. 1989. What they do when
authors’ opinion. Journal of the American Medical Asso- they get your manuscript: a survey of Academy of
ciation, 272: 152–3. Management reviewer practices. Academy of Manage-
http://dx.doi.org/10.1001/jama.272.2.152 ment Journal, 32: 157–73.
9. Chubin, D.E. and Hackett, E.J. Peerless Science: Peer 25. Bedeian, A.G. 2003. The manuscript review process:
Review and US Science Policy. Albany, State University the proper roles of authors, referees, and editors. Jour-
of New York Press, 1990, p. 4. nal of Management Inquiry, 12: 331–8.
10. Fletcher, R. and Fletcher, S. 1997. Evidence for the http://dx.doi.org/10.1177/1056492603258974
effectiveness of peer review. Science and Engineering 26. Rowlands, I. and Nicholas, D. 2006. The changing
Ethics, 3: 35–50. scholarly communication landscape: an international
http://dx.doi.org/10.1007/s11948-997-0015-5 survey of senior researchers. Learned Publishing, 19:
11. Jefferson, T., Wager, E. and Davidoff, F. 2002. Effects 31–55.
of editorial peer review: A systematic review. Journal http://dx.doi.org/10.1087/095315106775122493
of the American Medical Association, 287: 2784–6. 27. Thyer, B.A. and Myers, L.L. 2003. An empirical eval-
URL: http://jama.ama-assn.org/cgi/content/full/287/ uation of the editorial practices of social work jour-
21/2784 nals. Journal of Social Work Education, 39: 125–40.
12. Cole, S., Simon, G. and Cole, J. R. 1988. Do journal Available at http://www.accessmylibrary.com/coms2/
rejection rates index consensus? American Sociological summary_0286–23466319_ITM
Review, 53: 152–6. 28. Mark Ware Consulting. Peer Review in Scholarly
13. Hargens, L.L. 1988. Scholarly consensus & journal Journals: Perspective of the scholarly community – An
rejection rates. American Sociological Review, 53: International Study. London, Publishing Research
139–51. Consortium, 2007. Available at http://www.publishing