Comparer des méthodes

	Recherche de cohorte bayésienne	Recherche par sondage bayésien
Domaine	Conception de la recherche	Conception de la recherche
Famille	Process / pipeline	Process / pipeline
Année d'origine	Formalised in health research from the 1990s onward	1980s–2000s (modern applied development)
Auteur d'origine	Synthesis of cohort epidemiology (Doll & Hill, 1950s) with Bayesian inference (Bayes, Laplace, Jeffreys)	Thomas Bayes (theorem, 1763); applied to survey methodology by Donald Rubin, Andrew Gelman, and others (1980s–2000s)
Type	Quantitative longitudinal observational design	Quantitative observational research design with Bayesian inference
Source fondatrice	Ibrahim, J. G., & Chen, M. H. (2000). Power prior distributions for regression models. Statistical Science, 15(1), 46–60.	Gelman, A., & Carlin, J. B. (2007). Some issues on the foundations of statistics. In A. Gelman & J. B. Carlin (Eds.), Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press. ISBN: 978-0521686891
Alias	Bayesian cohort study, Bayesian prospective cohort, Bayesian longitudinal cohort analysis, Bayesian follow-up study	Bayesian survey analysis, Bayesian survey methodology, Bayesian polling, Bayesian questionnaire analysis
Apparentées	4	4
Résumé	Bayesian cohort research follows a defined group of individuals over time to track outcomes, and uses Bayesian statistical inference to update beliefs about risk, incidence, or causal effects as follow-up data accumulate. Prior knowledge — from earlier studies, registries, or expert judgment — is formalised into a prior distribution and combined with the cohort's likelihood to yield a posterior distribution that quantifies uncertainty in a directly interpretable way.	Bayesian survey research applies Bayesian statistical inference to survey data, combining prior knowledge or beliefs about population parameters with observed questionnaire responses to produce posterior probability distributions. Unlike null-hypothesis significance testing, this approach quantifies uncertainty directly, incorporates prior evidence, and yields probabilistic statements about parameters of interest — making it especially powerful for small samples, sequential data collection, and contexts where substantive prior knowledge exists.