几种常见的显著性检验方法.pdf

《几种常见的显著性检验方法.pdf》由会员分享，可在线阅读，更多相关《几种常见的显著性检验方法.pdf（2页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、几几种种常常见见的的显显著著性性检检验验方方法法-CAL-FENGHAI.-(YICAI)-Company One1(John Wilder Tukey)test(John Wilder Tukey)test最着名的有 2 个：（1 1）Tukey test for multiple comparisonsTukey test for multiple comparisons主要应用于 3 组或以上的多重比较。比如说一共有 4 组数据，两两比较产生 6 个统计值，Tukey-test 用于生成一个 critical value 来控制总体误差（Familywise error rate，FER

2、），与 Tukey test 相类似的是 Dunnett test，它是控制多对一比较（即 3 组同时和一个参照组比较）的 FER。（2 2）Tukey trend testTukey trend test主要用于检验同一药物不同剂量下和参照药物的线性关系。Tukey trendtest 简单但及其高效,是生物统计学常用的方法。T 检验，这是 1905 年氏首先提出的，当时他以“Student”为笔名发表，故至今有的书籍仍称之为“学生氏检验”。t 可能是倍数的意思(times)，t 就是样本均数 SX(x)与总体均数(“)间相距几倍标准误(sx)。t 检验是用于比较两均数间相差是否显着的。t

3、检验过程：是对两样本均数(mean)差别的显着性进行检验。唯 t 检验须知道两个总体的方差(Variances)是否相等；t 检验值的计算会因方差是否相等而有所不同。也就是说，t 检验须视乎方差齐性(Equality of Variances)结果。所以，SPSS 在进行 t-test for Equality of Means 的同时，也要做 Levenes Test for Equalityof Variances。3.3.Dunns multiple comparison testDunns multiple comparison test Dunns test calculates a

4、 P value for each pair of columns.These P values answerthis question:If the data were sampled from populations with the same median,what is the chance that one or more pairs of columns would have medians as farapart as observed here If the P value is low,youll conclude that the difference isstatisti

5、cally significant.The calculation of the P value takes into account the numberof comparisons you are making.If the null hypothesis is true(all data are sampledfrom populations with identical distributions,so all differences between groups aredue to random sampling),then there is a 5%chance that at l

6、east one of the posttests will have P,or .The calculation ofthe P value takes into account the number of comparisons you are making.If the nullhypothesis is true(all data are sampled from populations with identical distributions,so all differences between groups are due to random sampling),then there is a 5%chance that at least one of the post tests will have P.The 5%chance does not applyto EACH comparison but rather to the ENTIRE family of comparisons.2