Skip to Main content Skip to Navigation

Recherche biomédicale et journalisme en situation d'incertitude : validité des résultats de la recherche biomédicale et couverture médiatique

Abstract : Many academic publications are devoted to the « reproducibility crisis » in biomedical sciences. Their authors distinguish this lack of reproducibility from fraud or plagiarism. This “crisis” deals with a much larger phenomenon encompassing many scientific disciplines: a large amount of scientific results are disconfirmed by subsequent studies.This lack of reproducibility is to be expected: knowledge production is an incremental process where early, promising yet tentative findings are validated through replication. Indeed, scientific results are uncertain per se. The problem, however, is that this uncertainty does not seem to be taken into consideration when science “meets” the public, especially through the media.In this dissertation we studied how the media presented this uncertainty when dealing with biomedical findings. To do so we first created a large, original database of scientific studies investigating the association between risk factors (genetic, biochemical, environmental) and pathologies from three biomedical domains; psychiatry, neurology and a set of four somatic diseases. We evaluated the validity of each initial study by comparing their results to the result of meta-analyses on the same subject. The replication validity is low: 65% of initial studies are disconfirmed by corresponding meta-analysis even when they were published in high-ranking journals. We then identified which studies were selected by the press: initial studies published in prestigious journals and relevant to the readers were preferentially covered. Their validity was nonetheless poor with more than 50% being subsequently invalidated. The press rarely mentioned these frequent invalidations. Analysing the newspaper article contents, we found that journalists and their editors do not deal with scientific uncertainty. Indeed, the majority of newspaper articles referred to the study as being an initial study but only 21% indicated that the results needed to be replicated. Moreover those statements were made by scientists and have become scarce in most recent articles. A survey of 21 science journalists confirmed that journalists still consider high-ranking scientific journals to be reliable sources of information. However, these journalists were not familiar with the incremental process of knowledge production: two-thirds did not know that early findings were uncertain, or confused uncertainty with fraud. The other third knew about the uncertainty of initial results but found it hard to take it into account in their articles because of their respective hierarchy.More generally, the dissertation discusses the influence of extra-scientific factors upon the production of scientific knowledge. We conclude that the scientific assessment process based on the number of papers published in high impact factor journals, combined with the scientific institutions’ orientation towards the media, might undermine the reliability of scientific results, and this in academic publications as well as in the media. Indeed, journalists’ working conditions are deteriorating and most do not seem to properly grasp how scientific facts are produced. This might be damaging for public trust in biomedical research and public debate about health-related issues.
Document type :
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Friday, January 12, 2018 - 4:09:08 PM
Last modification on : Friday, October 23, 2020 - 5:03:16 PM
Long-term archiving on: : Sunday, May 6, 2018 - 3:17:52 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01682946, version 1



Estelle Dumas. Recherche biomédicale et journalisme en situation d'incertitude : validité des résultats de la recherche biomédicale et couverture médiatique. Science politique. Université de Bordeaux, 2017. Français. ⟨NNT : 2017BORD0737⟩. ⟨tel-01682946⟩



Record views


Files downloads