论文部分内容阅读
本文讨论了非独立计量资料中因变量的协方差结构和内部相关性。在忽略类间相关的情况下，分别给出了不考虑协变量和考虑协变量时内部相关的计算方法和步骤。实例分析表明，混杂因素对内部相关有较大的影响，只有合理地构造协方差结构，才能得到合理的、不含混杂的内部相关。
This paper discusses the covariance structure and internal correlation of dependent variables in non-independent measurement data. In the case of ignoring inter-class correlations, the calculation methods and steps for internal correlations without considering covariates and considering covariates are given. Case studies show that confounding factors have a greater impact on internal correlations. Only reasonable construction of covariance structures can produce reasonable internal correlation without confounding.