论文部分内容阅读
目的 :探讨几种非独立试验资料的内部相关性。方法 :对非独立数据建立随机效应模型 ,利用广义估计方程之残差估计组内相关系数。结果 :组内相关反映了非独立试验资料间的共变关系或同趋性 ,由广义估计方程导出的组内相关是以方差齐性为前提的。结论 :对非独立试验 ,不仅要关心处理组间的效应 ,亦应关心非独立数据间的相关程度。用随机效应模型分析非独立资料 ,可得到更多的信息。
Objectives: To explore the internal correlation of several non-independent test data. Methods: A stochastic effect model was established for non-independent data, and the intra-group correlation coefficient was estimated by using the residuals of the generalized estimation equation. Results: The cohort correlations reflect the covariance or covariance among the independent test data. The intragroup correlations derived from the generalized estimating equation are based on the homogeneity of the variance. CONCLUSIONS: For non-independent trials, care should be taken not only to deal with the effects between groups, but also to relate the extent of non-independent data. Use random effects model analysis of non-independent data, you can get more information.