Sheeber L B, Sorensen E D, Howe S R
Oregon Research Institute, Eugene, USA.
J Psychiatr Res. 1996 May-Jun;30(3):185-99. doi: 10.1016/0022-3956(96)00012-X.
This paper discusses several data analytic technique, for examining treatment efficacy in pretest-posttest control group designs. The following approaches are described: ANOVA on post scores, ANOVA on difference scores, split-plot repeated measures ANOVA, profile analysis, and ANOCOVA with prescore as the co-variate. Guidelines for choosing between available techniques are provided; the primary focus here is on the nature of the null hypothesis, the assumptions underlying the approach, and the power of the procedure. The importance of examining the characteristics of the data set in selecting an analytic technique is illustrated.
本文讨论了几种数据分析技术，用于检验前测-后测控制组设计中的治疗效果。描述了以下方法：对后测分数进行方差分析、对差异分数进行方差分析、裂区重复测量方差分析、轮廓分析以及以预测试分数作为协变量的协方差分析。提供了在现有技术之间进行选择的指导原则；这里的主要重点是原假设的性质、该方法所基于的假设以及该程序的功效。说明了在选择分析技术时检查数据集特征的重要性。
