论文部分内容阅读
本文介绍一种缺失数据配伍组设计方差分析的精确方法——回归分析方法,该方法不同以往使用的估项方法,其实质是把方差分析模型视为一般线性模型,采用边界条件法处理,也适用于配伍组设计方差分析非缺项情形。我们将回归方法和估项方法同原来未缺项的方差分析结果对比,认为回归方法比估项方法准确,尤其在处理组结果上,回归方法比估项方法更准确。
This paper describes an exact method of analysis of variance for the design of variance analysis for the missing data combination group—regression analysis method. This method is different from the estimation method used in the past. The essence of the method is to regard the variance analysis model as a general linear model and adopt the boundary condition method. Applicable to the compatibility group design variance analysis is not the absence of the situation. We compare the regression method and the estimation method with the original analysis of the variance analysis of the missing items. We believe that the regression method is more accurate than the estimation method, especially in the treatment group results, the regression method is more accurate than the estimation method.