卡方检验误用辨析  

Analysis on the misuse of chi-square test

作　　者：张平原 王开发[1] Zhang Pingyuan;Wang Kaifa(School of Mathematics and Statistics,Southwest University,Chongqing 400715,China)

机构地区：[1]西南大学数学与统计学院,重庆400715

出　　处：《中华烧伤与创面修复杂志》2024年第5期482-488,共7页Chinese Journal of Burns And Wounds

摘　　要：计数资料在生物医学研究中十分常见,此类资料常常被整理为列联表的形式。对计数资料,常见的研究目的是检验2个因素之间是否独立,因此常用χ^(2)检验来进行统计分析,却容易忽略χ^(2)检验的适用条件和不同条件下对结果的校正。此外,对计数资料,还有检验2个分类变量之间是否存在线性变化趋势、结果是否一致等研究目的,因此如何根据不同的研究目的选择合适的检验方法,也是实际应用时值得关注的问题。该文通过系列示例比较系统地总结归纳了各类计数资料的常用统计分析方法,并对常见的误用情况进行了辨析。Count data are very common in biomedical research,and such kind of data is often organized in the form of contingency table.For count data,the common research purpose is to test whether two factors are independent,therefore chi-square test is often used for statistical analysis,but it is easy to overlook the applicable conditions of chi-square test and the correction of results under different conditions.In addition,for count data,there are also other research purposes,such as testing whether there is a linear trend between two categorical variables,whether the results are consistent,and so on.Therefore,how to choose appropriate testing methods based on different research purposes is also a problem worth paying attention to in practical applications.In this study,the commonly used statistical methods for various count data are systematically summarized through a series of examples,and the common misuses of the chi-square test are analyzed.

关 键 词：生物医学研究 数据说明 统计 计数资料 列联表 χ^(2)检验 

分 类 号：O212.1[理学—概率论与数理统计] R195.1[理学—数学]