Alexis St-Gelais, M. Sc., chimiste
When discussing scientific topics, it is important to go beyond opinions and hearsay if the objective is to learn something valid or reach a sensible conclusion. As such, the source of the elements submitted to the discussion becomes extremely important. For the best and the worst, we live in an era of informational overflow. The Internet is filled with both excellent and horrible pieces of information, and additional explanations by experts of undefined credibility makes it even more complicated to exactly understand what is going on. One of the most useful and credible source one should turn to when seeking an information (and should be able to cite when making a statement) is scientific literature. But how exactly does that work?
Scientific literature can be categorized by levels. Tertiary literature is the most transformed form: it contains digested information from various other scientific sources, in a form that suits various audiences (including the public), so as to present widely accepted principles and facts. This kind of literature had the advantage of presenting a broader picture. Secondary literature is an expert-conducted review of scientific literature, prepared so as to centralize all information available at a given time about a specific topic, and sometimes put patterns into evidence. It is generally still directed at the scientific community first and foremost. The most abundant literature, however, is the primary literature, that is, scientific articles, thesis, and conference proceedings, mostly, which was peer-reviewed. Figure 1 shows an example of primary literature. It should always aim to present new findings, previously unreported, to the community.

Figure 1. Example of a peer-reviewed article.
Peer-review is a cornerstone of the current scientific practice. It serves two purposes: to ensure that the data presented was collected and interpreted in a way that complies with the state-of-the-art of the field; and to skim overstatements, shortcuts, biases, or otherwise wacky conclusions, or filter work that does not add anything new to the knowledge of the community (or worse, was plagiarized). Here is commonly the steps any scientific work must go through in order to be published in a reputable journal:
- The research itself is conducted, and its results and conclusions are written down in the form of a manuscript, following general guidelines specific to the journal chosen by the authors.
- The manuscript and a presentation letter, along with all relevant supporting information, is sent to the Editor of the journal. It will first be checked by the Editor or his assistants to make sure that it presents sufficient novelty, that is it of general relevance toward to scope of the journal, and that is is decently written (nonsensical language or outrageously weird results will likely be rejected singlehandedly).
- The Editor will seek out reviewers, which are other scientists conducting research in a field closely connected to the topic of the manuscript. Usually, there are two or three reviewers. Their task is to carefully read over the manuscript, and judge whether it is worthy to be published (accepted with or without modifications) or not (rejected). This decision can be weighed by comments and modification requests, which can sometimes be major and involve that the research team gathers more results to make the final result more convincing. This process is entirely anonymous: the reviewers should usually ignore who the authors are, and vice-versa.
- The Editor (also a scientist of the field) makes the final decision based on the opinion of his reviewers, and can add comments and requests of his own. Articles are very seldom accepted without any modification. In case of rejection, the authors can try again with another journal, with a different scope, or enrich their work and prepare a revised manuscript to try again.
- If, over time, an article is found to be problematic (such as this famous case of a study suggesting a link between vaccines and autism, which was filled with flaws, shortcuts and ethical problems), it can be retracted, which means it should not be considered anymore as valid in any regard. It plainly means it should never have been published.
Generally speaking, work that has been peer-reviewed should be considered more reliable than information found on forums, blogs, or by mouth-to-ear. One must be wary even of information from experts themselves (including us) which is strictly based on their scientific authority and not on literature or sufficient data. That is because nobody is entirely sheltered from biases, whether they are conscious or not. The input of fellow scientists goes a step further to improve the quality of the discoveries presented in a scientific paper.
Yet, this is not the whole story, are there are some traps to avoid when looking at primary literature – but that is for another blogpost!