Analyzing Effect Size Indices for Single : Case Research Data

DOI Web Site

Bibliographic Information

Other Title
  • 一事例実験データの処遇効果検討のための記述統計的指標について : 行動分析学研究の一事例実験データの分析に基づいて
  • イチ ジレイ ジッケン データ ノ ショグウ コウカ ケントウ ノ タメノ キジュツ トウケイテキ シヒョウ ニ ツイテ コウドウ ブンセキガク ケンキュウ ノ イチ ジレイ ジッケン データ ノ ブンセキ ニ モトズイテ

Search this article

Description

Researchers have suggested various effect-size indices for assessing the effects of treatments in single-case designs or for synthesizing the outcomes of single-case experiments. But there are no criteria for these measures yet. The purpose of the present study was to examine the distributions of effect sizes and the relations between the measures that have been proposed. Single-case data were selected from 63 articles in the Japanese Journal of Behavior Analysis. The following measures of effect size were calculated for each : (1) percentage of non-overlapping data(PND), (2) and (3) Busk and Serlin's (1992) effect size (ES_BS1, ES_BS2), (4) mean baseline reduction (MBLR), (5) Hershberger, Wallace, Green, and Marquis's (1999) effect size (ES_H), (6) last treatment day (LTD), (7) Center, Skiba, and Casey's (1985-1986) effect size (ES_C), and (8) Allison and German's (1994) effect size (ES_AG) .The results of this study were compared with the results in Parker and Brossart (2003) and Campbell (2004). Using cluster analysis, we classified the 8 effect sizes into 3 groups (group 1 : ES_H, MBLR, ES_C ; group2 : ES_BS2, ES_AG, LTD, ES_BS1 ; group 3 : PND).

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top