Departmental Bulletin Paper ベイズ推定による情報仮説の評価 : その理論と各種モデルへの応用について
Recent advances in Bayesian evaluation of informative hypotheses : Its theory and applications to psychometric models

岡田, 謙介

6pp.9 - 17 , 2016-03-15 , 専修大学人間科学学会
Null hypothesis significance testing has been one of the most frequently used statistical techniques in psychological research. However, the hypotheses considered in null hypothesis significant testing are usually either unrealistic or uninformative. In contrast, in the context of Bayesian statistics, researchers can directly evaluate the researchers' informative hypotheses by simple arithmetical calculations. In this paper, first the basic theory and practice of Bayesian evaluation of informative hypotheses are presented. Next, the specification of prior distributions is discussed. The paper concludes with a review of applications, which includes multiple regression, contingency tables, structural equation modeling, and meta-analysis.

Number of accesses :  

Other information