Thomas Elaine
Primary Care Sciences Research Centre, Keele University, North Staffordshire, UK.
Musculoskeletal Care. 2005;3(4):201-12. doi: 10.1002/msc.11.
This article, the third and final article in the series, aims to give health care professionals (HCPs) a sound and helpful introduction to medical statistics (Thomas, 2004, 2005). A brief summary of the content of the previous articles is given in Table 1. The current article will cover the area of basic statistical tests with the aim of guiding HCPs to the correct test for a particular research question and dataset. The article will not go into great depth of the formal methods of calculation required for all the tests covered but I would suggest that the reader refer to standard textbooks (Jordan et al., 1998; Swinscow, 1998; Altman, 1994; Bland, 2000), the help sections of statistical packages (SPSS or Stata), or consult a statistician. For ease of reference within the article the tests have been grouped by the data type, i.e. numerical or categorical. Further separation within each data type has been carried out depending on the number of groups being compared, whether the groups are independent, the size of the sample, and, in the case of numerical data, the distribution of the variable. For quick reference, two other tables are also presented which summarize which analysis methods should be used in each situation.
本文是该系列的第三篇也是最后一篇文章,旨在为医疗保健专业人员(HCPs)提供扎实且实用的医学统计学入门知识(托马斯,2004年,2005年)。表1简要总结了前几篇文章的内容。当前文章将涵盖基本统计检验领域,目的是指导医疗保健专业人员针对特定研究问题和数据集选择正确的检验方法。本文不会深入探讨所涵盖的所有检验所需的正式计算方法,但我建议读者参考标准教科书(乔丹等人,1998年;斯温斯科,1998年;奥尔特曼,1994年;布兰德,2000年)、统计软件包(SPSS或Stata)的帮助文档,或咨询统计学家。为便于在文章中参考,这些检验按数据类型进行了分组,即数值型或分类型。在每种数据类型中,还根据要比较的组数、组是否独立、样本大小,以及对于数值数据而言变量的分布情况进行了进一步划分。为方便快速参考,还给出了另外两张表,总结了在每种情况下应使用的分析方法。