- In this pandemic period, scientists are working harder to understand how the Covid-19 works.
- They publish the results of their work in studies.
- To find their way, 20 Minutes asked a researcher for advice on how to calmly approach a scientific study.
There would be inequalities between the blood groups facing Covid-19, the virus would survive several hours on certain surfaces, certain treatments would fight against the disease ... These hypotheses, widely reported in the media, are from studies. Faced with an unknown disease a few months ago, scientists are redoubling their efforts to understand this new coronavirus and its modes of transmission. However, not all of these studies have the same strength of evidence.
Faced with this flow of information, what are the right questions to ask in order to fully understand a scientific study? Publications in professional journals are primarily intended for researchers. It is possible to find your bearings by asking yourself a few questions, which Bastien Castagneyrol, a researcher at Inrae and author of articles on the subject for The Conversation site, specifies for 20 Minutes .
"In times of emergency, the first question to ask is to know where the study comes from," says the researcher. He recommends checking whether it has been reread by other experts. “What is happening right now is that the articles are put online before this proofreading stage. These are what are known in the jargon as " pre-prints ". The authors make work available to the public that they consider successful, but it is work that has not yet been evaluated by the scientific community. It's like the beta version of software: there are certainly some interesting things in it, but it's not the final version that you can trust. "
The analogy of chocolate cake
How is this proofreading process going? "When researchers submit an article to a scientific journal, the publisher will propose the article to experts in the field," says Bastien Castagneyrol. The objective? “Check that the reasoning is correct, that the way it is presented is correct and that there are no major flaws in the methodology. The article can be reworked if the publishers believe that there are inaccuracies or ambiguities. As a result, it can take several months before an article is published.
This work is essential because, if the methodology is too imprecise, other teams will not be able to use it. "What makes the weight of a discovery is the fact that other studies confirm what has been done," says the scientist. To be able to confirm this, it is necessary to be able to use the same methodology. By analogy, if you want to make a chocolate cake, you go to a recipe site. If you only have the photo, you will not be able to redo the same cake. You need the detailed recipe. "
In the study, Bastien Castagneyrol advises to look at the part devoted to the statistical processing of the data. If it is too short, watch out! “When, in an article, a paragraph on the statistical methodology for processing data makes two lines, it appeals to me. These data may be valid, but we have to say a little more so that we can trust them or not. "
This description of the data is found in all the disciplines, underlines the researcher. “This is what makes the research process which is shared by all colleagues from all disciplines, on an international scale. Methodological consensus is reached on this. "
Not all journals have the same rigor
Be careful, however, not all newspapers have the same rigor. Some are well established: Nature , Science or The Lancet . However, having a “slap in English name”, as David Louapre points out on his “Amazing Science” channel, is not necessarily a sign of quality. Another source of confusion is that some unreliable journals may have names very close to well-established journals. "Since the advent of the Internet, newspapers have developed an economic model that requires authors to pay to be published," says Bastien Castagneyrol. Some newspapers do this transparently and very cleanly. For others, their motivation is simply to bring in items to bring in money. In these newspapers, which are generally called predatory journals, the quality of the article can be poor. »How to spot these bad apples? "Is it a well-established newspaper for ten or twenty years, or is it a newspaper that was created last week? What fees are asked of the authors? This is what should put the chip in the ear. "
Once all these verifications have been carried out, does that mean that we can trust the work presented? Not entirely, because there are what scientists call "levels of evidence", which Bastien Castagneyrol describes: "A study is interesting. Several studies, even with a small sampling, which say the same thing, are all the more interesting. It is a first level of evidence. The second level are large double-blind * randomized studies. Finally, once we have accumulated enough elementary bricks, we can bring everything together in a meta analysis, this is the ultimate level of proof. It takes time, because you need these elementary bricks ”.
He added: "The time of the crisis, the time of the media, the time of research, these are different times. "
* In a double-blind study, the study objects, for example patients, are separated into two groups. One group receives the experimental treatment, the other a placebo or the usual treatment. Neither the patients nor the doctors know in which groups the patients are.
Do you want the Fake off team to verify an info, photo or video? Fill out the form below or write to us on Twitter: https://twitter.com/20minFakeOff
20 Minutes is a partner of Facebook to fight against fake news. Thanks to this device, users of the social network can report information that they believe is false.
By the Web
Coronavirus: Four Tips for Investigating Intoxication
Culture
Coronavirus: "The physical visit is only part of the museum public experience today," says Roei Amit
- Covid 19
- Coronavirus
- Fake Off
- Science
- study