Abstract: Split-plot design is one of the most popular experiment designs due to its unique advantages, such as the flexibility of adding treatments and controlling experimental errors with various levels. However, its statistical analysis of data is complicated, and the relevant and effective statistical software is rare. In order to establish a statistical analysis approach with simple operation, strong practicability, and correct calculation results for split-plot data, the program module general linear mixed model (GLIMMIX) of SAS was used to perform statistical analysis of the split-plot design data, and the advantages and disadvantages of GLIMMIX compared with the traditional program module general linear model (GLM) were illustrated. The results showed that compared with GLM, GLIMMIX could automatically select the correct error variance and the degree of freedom to calculate the statistics, which avoided the issue of not being able to calculate the statistics easily when using GLM in some cases. The shortcomings of GLM in statistical analysis of crack design and other advantages of GLIMMIX analysis were illustrated. Compared with the MIXED module, GLIMMIX was considered as the preferred module for statistical analysis of split-plot design data.

Key words: split plot, statistical analysis, GLMMIX, SAS, analysis of variance, multiple comparison

摘要： 裂区设计能够灵活地增加试验处理和进行误差分级控制，在农业试验中应用广泛。但数据的统计分析较复杂，目前有效的相关统计软件十分缺乏。为了建立操作简单、实用性强、计算结果无误的统计分析手段，采用SAS广义线性混合模型（GLIMMIX）程序模块进行裂区设计数据的统计和分析，并通过实例分析说明GLIMMIX相较于传统一般线性模型（general linear model，GLM）程序模块的优缺点。结果发现，与GLM相比，GLIMMIX能够自动选用正确的误差项方差和自由度进行统计量计算，克服了GLM在某些情况下难于计算所需统计量的问题。实例验证说明，采用GLM进行裂区设计统计分析的不足及GLIMMIX分析的其他优点，并与混合程序模块（MIXED）进行了比较，认为GLIMMIX是裂区设计数据统计分析的首选模块。

关键词: 裂区设计, 统计分析, 广义线性混合模型（GLIMMIX）, SAS, 方差分析, 多重比较