20 janvier 2006

Publication scientifique : qui veut me relire ?

(Nouvel Obs) Le scandale qui a éclaté en Corée du Sud avec le Dr Hwang a jeté le doute sur l’efficacité des méthodes de relecture et de vérifications mises en place par les journaux scientifiques primaires. Ces journaux s’appuient sur des comités de lecture composés d’autres chercheurs (peer review) pour valider les résultats soumis par les auteurs d’un article. Dans la foulée de Hwang, un autre chercheur vient d’avouer sa fraude : le Norvégien Jon Sudbø avait inventé ses données pour une étude sur le cancer parue en octobre dernier dans The Lancet.

Ce chercheur d’Oslo a créé les 900 personnes de la cohorte, leur histoire et les effets de médicaments anti-inflammatoires sur la survenue d’un cancer de la bouche. C’est une épidémiologiste de l’institut national de santé public de Norvège qui a découvert le pot aux roses. Comme les éditeurs de Science, contraints de rétracter les deux papiers de Hwang, les responsables du Lancet rappellent que le système de relecture est efficace pour écarter des travaux mal faits mais pas pour détecter une fraude intentionnelle et bien troussée.

Les revues primaires ne peuvent pas pour autant faire l’économie d’une révision de leurs méthodes. Dans son édition publiée aujourd’hui, Nature propose des modifications pour évaluer les travaux scientifiques, notamment dans le domaine du clonage. Dans le cas de Hwang, la revue a demandé à un scientifique indépendant de mener les analyses permettant de vérifier que le chien Snuppy avait bien été cloné. Faut-il étendre ce type de procédure ? Nature craint que les contraintes pèsent lourd sur la recherche pour un nombre de cas de fraude finalement peu important. La revue conclut son éditorial en appelant de ses vœux suggestions et discussions.

Cecile Dumas

Mobile phone use not linked to increased risk of glioma brain tumours

Mobile phones are not associated with an increased risk of the most common type of brain tumour, finds the first UK study of the relationship between mobile phone use and risk of glioma. The results are published online by the BMJ today.
The four year study by the Universities of Leeds, Nottingham and Manchester and the Institute of Cancer Research, London found those who had regularly used a mobile phone were not at a greater overall risk of developing this type of tumour.
There was no relationship for risk of glioma and time since first use of a mobile phone, lifetime years of use and cumulative number of calls and hours of use. Risk was not associated with phone use in rural areas which was found to be associated with an increased risk in an earlier Swedish study.
A significantly increased risk was found for tumours which developed on the same side of the head as the phone was reported to have been held but this was mirrored by a decrease in the risk on the opposite side of the head making it difficult to interpret as a real effect.
This finding may be due to people with glioma brain tumours linking mobile phone use to the side of the tumour and therefore over reporting the use of a phone on the same side as their tumour. This results in under reporting use on the opposite side of the head, say the authors.
Mobile phones have been available in the UK since 1985, but widespread use did not begin until the late 1990s making the number of long term users (over 10 years) quite small. This study had limited numbers for estimating the risk of using a phone over a long period.
Early mobile phones were designed to use analogue signals and emitted higher power than current digital phones but the study showed no increased risk of glioma brain tumours with the use of analogue phones.
Notes to Editors: There are over 4,000 new cases of brain tumours per year of which glioma is the most common type. Early symptoms include headaches and feelings of nausea. The causes of these tumours are currently unknown.
The study was conducted between 1 December 2000 and 29 February 2004 and included people living in the Thames region, southern Scotland, Trent, the West Midlands and West Yorkshire.
966 people with glioma brain tumours (cases) and 1716 healthy volunteers (controls) were interviewed about their previous mobile phone use history including how long they had used mobile phones, the number and duration of the calls they made and what make and model of phone they had used.

=> L'extrait de l'étude du British Medical Journal

Higher Education Fuels Stronger Belief in Ghosts

LiveScience Managing Editor

Believe it or not, higher education is linked to a greater tendency to believe in ghosts and other paranormal phenomena, according to a new study.
Contrary to researchers' expectations, a poll of 439 college students found seniors and grad students were more likely than freshmen to believe in haunted houses, psychics, telepathy, channeling and a host of other questionable ideas.
The results are detailed in the January-February issue of the Skeptical Inquirer magazine.
'Not Sure'
The survey was modeled after a nationwide Gallup Poll in 2001 that found younger Americans far more likely to believe in the paranormal than older respondents.
The new study was done by Bryan Farha at Oklahoma City University and Gary Steward Jr. of the University of Central Oklahoma.
In general college students checked the "Believe" box less than the general population surveyed by Gallup. But the lack of "Don't Believe" responses among college students was lower for six of the 13 categories: psychic or spiritual healing, haunted houses, demonic possession, ghosts, clairvoyance and witches. That means a higher percentage of college students put themselves in the "Not Sure" column on these topics.
Less skeptical
More significantly, the new survey reveals college is not necessarily a path to skepticism in these realms.
While 23 percent of college freshmen expressed a general belief in paranormal concepts—from astrology to communicating with the dead—31 percent of seniors did so and the figure jumped to 34 percent among graduate students.
"As people attain higher college-education levels, the likelihood of believing in paranormal dimensions increases," Farha and Steward write.
The media are likely responsible for some people's beliefs in alien abductions and other paranormal concepts, the scientists write, based on their survey of existing studies. And some people tend to selectively confirm whatever ideas might be in their heads. Even smart people might believe in something offbeat because, in part, they're good at defending whatever they believe.

The Poll Results How the college students responded in the new poll on a percentage basis (in bold) compared to the 2001 Gallup Poll (parenthesis). Figures are rounded.
Paranormal conceptBelieveNot Sure

Don't
Believe

Psychic or spiritual healing or the power of the human mind to heal the body

56 (54)26 (19)18 (26)
ESP or extrasensory perception28 (50)39 (20)33 (27)
That houses can be haunted40 (42)25 (16)34 (41)
That people on this Earth are sometimes possessed by the devil40 (41)28 (16)31 (41)
Ghosts or that spirits of dead people can come back in certain places and situations39 (38)27 (17)34 (44)
Telepathy, or communication between minds without using the traditional five senses24 (36)34 (26)42 (35)
That extraterrestrial beings have visited earth at some time in the past17 (33)34 (27)48 (38)
Clairvoyance, or the power of the mind to know the past and predict the future24 (32)33 (23)42 (45)
That people can hear from or communicate mentally with someone who has died16 (28)29 (26)54 (46)
Astrology, or that the position of the stars and planets can affect people's lives17 (28)26 (18)57 (52)
Witches26 (26)19 (15)55 (59)
Reincarnation, that is, the rebirth of the soul in a new body after death14 (25)28 (20)57 (54)
Channeling, or allowing a "spirit being" to temporarily assume control of a human body during a trance10 (15)29 (21)61 (62)

SOURCE: Bryan Farha and Gary Steward Jr. via the Skeptical Inquirer


Il est dommage que les étudiants en sciences n'aient pas été considérés séparément des autres, comme cela a été fait lors d'études précédentes.