Sample size determination for single-case experiments based on a hybrid Bayesian approach

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K03224
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 10020:Educational psychology-related
• Research Institution: Shiga University (2021-2022); Joetsu University of Education (2019-2020)
• Principal Investigator: Okumura Taichi 滋賀大学, データサイエンス学系, 准教授 (90547035)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 単一事例実験

Outline of Final Research Achievements

We proposed a method to generate a data set from the posterior predictive distribution based on data obtained by a multiple baseline design and to estimate the confidence interval width and power of the test when data were collected in that design. We were able to flexibly perform accuracy and power analyses even for designs with different intervention starting points and number of time points depending on the subject. On the other hand, it may not be robust to mis-specification of the assumed model, and tended to estimate a smaller sample size than required when performed without considering serial correlation of errors.

Free Research Field

教育心理学

Academic Significance and Societal Importance of the Research Achievements

単一事例実験は心理的介入の個人レベルの効果とその個人差を検証する上で広く用いられている方法である。一方、対象者数、時点数、介入開始ポイント、変化の非直線性、系列相関構造、結果変数の取りうる範囲など選択肢の多様さや複雑さが統計的アプローチを導入する上で大きな障壁となってきた。本研究で示したアプローチは標本サイズ決定に注目し、ハイブリッド型ベイズアプローチの採用により多様なデザインやモデルの扱いについて一定の解決策を示したものと考えている。