Chers et chères collègues, je vous prie de bien vouloir trouver ci-dessous la présentation et le programme d'un atelier portant sur les fondements de la randomisation qui aura lieu en ligne les 8 et 9 juin prochains. Le lien de connexion et les résumés seront envoyés en début de semaine prochaine aux personnes qui en auront fait la demande dans un message envoyé à l'adresse suivante : randomization2...@gmail.com. Merci de votre attention, Isabelle Drouet *** Online workshop on the foundations of randomization, June 8th and 9th Randomized controlled trials have been much discussed in the last 20 years for the status they are granted within evidence-based medicine, usually at the top of hierarchies of evidence. Their importance has also been repeatedly underlined in the context of the pandemic. These discussions, in philosophy, in science and in the public debate, often presuppose, more or less explicitly, that the epistemic reasons to randomize are clearly identified and well-known. But this is not the case. The foundations of randomization fail to be consensual even in the one context where it may seem obvious that randomization is the best way to go - that is, to determine whether a given medical intervention has a causal effect on a disease. The most common view, targeted by most criticisms of randomization, is that randomization is a means to balance confounders, known and unknown, between parallel groups, and thereby to ensure that any difference between groups can be interpreted causally. Howe ver, Fisher rather introduced randomization as a device making it possible to calculate the probability of the different possible observations and to determine whether the observed difference is statistically significant. What are the different justifications of randomization and how do they compare to each other? In which situations do they hold and, for that matter, should we randomize at all? The workshop will investigate these questions, which are particularly urgent in the current, pandemic context, where we need to make informed methodological choices as regards the assessment of preventive or curative treatments. How questions about randomization relate to other topics in the philosophy of statistics - primarily the opposition between frequentist and Bayesian approaches - will also be explored. The workshop will bring together philosophers and practitioners to think about these issues. Organization: Isabelle Drouet (SND, Sorbonne Université) Scientific committee: Isabelle Drouet and Anouk Barberousse (SND, Sorbonne Université) Email for information and registration: randomization2...@gmail.com June 8th 1.30 pm - 2.20. Invited talk by Stephen Senn (consultant statistician, Edinburgh): Fisher’s Gambit. Understanding Randomisation 2.20 - 2.55. Evangelos Koumparoudis (Sofia University): Randomized control trials limitations and epistemological aspects break 3.10 - 4.00. Invited talk by Jonathan Fuller (University of Pittsburgh): Statistics, Balance and Bias: Randomization in Clinical Epidemiology and EBM 4.00 - 4.35. Aydin Mohseni and Daniel Alexander Herrmann (University of California, Irvine): Why randomize, really? break 4.50 - 5.40. Invited talk by Fabienne El Khoury (INSERM / Sorbonne Université): Randomised trials: merits, limitations, and field developments 5.40 - 6.15. Elselijn Kingma & Oliver Galgut (King’s College London): Better than randomisation? A philosophical defense of ‘dynamically allocated controlled trials’ June 9th 2 pm - 2.50. Invited talk by Isabelle Guérin and François Roubaud (Institut de Recherche pour le Développement): RCTs in the field of development: a critical perspective with a focus on microcredit sector 2.50 - 3.05. Julio Michael Stern (University of Sao Paulo): Why and how to randomize and audit in legal sortition and clinical trials break 3.20 - 4.10. Invited talk by Maximilian Kasy (Oxford University): Statistical decision theory cannot justify randomization or pre-registration for experiments. 4.10 - 4.45. Michel Shamy (Ottawa Hospital & Department of Medicine, University of Ottawa & Ottawa Hospital Research Institute): Evidentiary standards and the justification of randomized clinical trials: The example of hydroxychloroquine trials for COVID-19 4.45 - 5.20. Konstantin Genin (University of Tubingen) and Conor Mayo-Wilson (University of Washington, Seattle): Randomization, identifiability, and estimation of causal effects -- https://www.vidal-rosset.net/mailing_list_educasupphilo.html
