论文部分内容阅读
[目的 ]认识和区分各种率的性质并应用合适的统计分析方法。 [方法 ]利用 Poisson分布原理及其正态近似性解决一些不具有比例的性质的率的统计分析问题。 [结果 ]对那些分子在分母之中 ,但分子有重复计数的率 ,以及分子不在分母之中的率 ,不能用一般适用于对比例作统计分析的方法来分析。率的标准误的计算 ,对样本率与已知总体率的差别作统计检验 ,对两个样本率的差别作统计检验 ,都可以利用 Poisson分布原理及其正态近似性得以解决。[结论 ]应该区别比例与率的统计分析方法 ,“率的标准误”等提法应该与“比例的标准误”等提法区别。
[Objective] To recognize and distinguish the nature of various rates and apply appropriate statistical analysis methods. [Methods] The Poisson distribution principle and its normality approximation are used to solve the statistical analysis problem of some non-proportional properties. [Results] For those molecules that are in the denominator, but the rate at which the molecules are counted repeatedly, and the rate at which the molecules are not in the denominator, can not be analyzed using statistical methods that are generally applicable to the comparative example. The calculation of the standard error of the rate makes a statistical test of the difference between the sample rate and the known overall rate. Statistical tests of the difference between the two sample rates can be solved using the Poisson distribution principle and its normality approximation. [Conclusion] Statistical analysis methods for ratios and rates should be differentiated. References such as “standard errors in rates” should be distinguished from references to “standard errors in proportions”.