手法を比較

	ベイズ型コホート研究	ベイジアン調査研究
分野	研究デザイン	研究デザイン
系統	Process / pipeline	Process / pipeline
提唱年	Formalised in health research from the 1990s onward	1980s–2000s (modern applied development)
提唱者	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)
種類	Quantitative longitudinal observational design	Quantitative observational research design with Bayesian inference
原典	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
別名	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
関連	4	4
概要	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.