In science, peer review means the collective activity of researchers who consider critically the work of other researchers (their “peers”). These assessments can be for a specific search submitted for publication in a scientific journal or to be presented at a conference but can also cover all the work of the researcher or group of researchers evaluated, especially when recruiting a candidate to a position or in the evaluation of research projects by public institutions and private organization (such as a foundation). In scientific journals, the peer review is conducted by reading committees that decide if the report of a research paper submitted for publication and is acceptable or not.
The peer review is a fundamental principle of the scientific research, whether in “exact” science or the humanities and social sciences. It is used both as regards the publication of articles in journals and for the recruitment and promotion of teacher-researchers and funding for their research projects.
The reading committees are an integral part of the scientific process. The normal route of a new result from its discovery to its recognition by the scientific community goes into effect by publishing works that have led them in scientific journals as they undergo working to criticism of researchers in the same field. Traditionally, the dissemination of scientific work is mainly done through conferences and written reports which, to be accepted, must first be subject to careful review by a small number of experts appointed by the conference organizer or the editorial board of the journal. The same evaluation takes place in relation to scientific journals of humanities and social sciences. However, a larger share of the work in these areas escapes journals, since it can use other media, especially books.
Some journals have pushed the system of peer to invite a large number if not all researchers in the field to criticize the articles they publish: the open peer review. Only critical deemed most interesting are finally published with the original article and often a response to criticisms of the authors. This is for example the case of the journal Behavioral and Brain Sciences.
The emergence of open access publishing systems has slightly changed the situation in scientific dissemination. Indeed, according to the terms and conditions associated with this system, authors can deposit in a free access access directory their preprints, that is to say, the articles they submit manuscripts to journals. So they become available, at least in their original form, long before their eventual publication. An example of this type of directory is Arxiv, home of preprints in physics, mathematics and computer science, among others. It should be noted however that there are some rules slightly restricting deposition in Arxiv: an article must be submitted for an email address hosted by a scientific institution and an author publishing for the first time may have to endorse a bid by a person who has submitted successful article.
In the case of a publication in a scientific journal, the manuscript proposed by one or more researchers is received by the director of the magazine or the editor. This is usually a leading researcher in his field, who voluntarily assume this function. His tasks are mainly to choose, in collaboration with the Editorial Board, members of the reading committee, and ensure communication between reviewers and authors of the article.
Upon receipt of the manuscript, the Director shall decide, after a quick read, if the item is potentially publishable in the review. The relevance of the referred question and interest of outcomes for readers of the journal are evaluated based on publishing criteria specific to the magazine. The proportion of manuscripts rejected in this first phase is highly variable depending on the journal; this percentage is 90% for multidisciplinary journals such as Nature or Science, according to them.
In cases where the Director decides to continue the process, he contacts several (usually two) specialists whose profile matches the themes and techniques discussed in the article. These specialists then have a few weeks to read the manuscript thoroughly and report indulging their overall impression of the article, as well as their specific comments on errors or inaccuracies.
The director takes the reports of members of the reading committee and send to the authors its decision to publish or not publish the article. The publisher often follows the advice of reviewers, but his decision remains independent. If in doubt, for example if the reports are clearly contradictory, the Director may appeal to a third or fourth assessment. In any case, it sends a letter to the authors with explanations of its decision and a copy of the reports. In addition to the acceptance and rejection without conditions, the decision may be more nuanced (acceptance subject to corrections or invitation to submit a new version).
Traditionally, the reports are anonymous, only the name of the director being known to authors. For its part, the authors name can be known or not to the reviewers; it is called single or double-blind assessment, respectively. However, more and more journals encourage transparent management of the process and allow or require the disclosure of the names of members of the reading committee. Some people even make the reports of the reviewers, anonymous or not.
One of the most common criticisms about the evaluation process by peers is that it is slow, and it usually takes several months or even years in some areas, for an article to be published eventually. In fact, much of the communications, in areas such as astronomy and economics, are made before the peer review, through preprints submitted on directories such arXiv.
Other concerns relate to ability of peer review to ensure the quality of the articles.