R语言实现常用多重比较方法

在单因素方差分析ANOVA中，如果该因素影响比较显著，那么需要进一步利用多重比较方法比较该因素不同水平的影响，确定不同水平下该因素的影响是否显著。常见的多重比较方法主要有两种，LSD法和Tukey HSD法。下面对R语言中，这两种多重比较方法的实现进行举例。

前期数据如下，影响因素为group，指标为value：

> head(tarD)                   value  group      sample   timeA0522W11NC1 0.0002053745 normal A0522W11NC1 11weekA0522W11NC2 0.0031773712 normal A0522W11NC2 11weekA0522W11NC3 0.0060378288 normal A0522W11NC3 11weekA0522W11NC4 0.0017626931 normal A0522W11NC4 11weekA0522W11NC5 0.0018035261 normal A0522W11NC5 11weekA0522W11NC6 0.0036690067 normal A0522W11NC6 11week> tmp <- aov(value ~ group, tarD)

最小显著差数检验法(LSD法)

> res <- LSD.test(tmp, 'group', p.adj = 'bonferroni')> print(res$groups)       trt        means M1   normal 2.576910e-03 a2    drug3 7.552555e-04 b3    drug2 7.269247e-05 b4 high_fat 6.220610e-05 b5    drug1 2.954733e-05 b

Tukey氏固定差距检验法(Tukey HSD)

> TukeyHSD(tmp)  Tukey multiple comparisons of means    95% family-wise confidence levelFit: aov(formula = value ~ group, data = tarD)$group                         diff           lwr          upr     p adjdrug2-drug1      4.314514e-05 -0.0015468705 0.0016331608 0.9999916drug3-drug1      7.257082e-04 -0.0008643074 0.0023157239 0.6929965high_fat-drug1   3.265877e-05 -0.0015149488 0.0015802664 0.9999969normal-drug1     2.547362e-03  0.0009997549 0.0040949700 0.0002613drug3-drug2      6.825631e-04 -0.0009487586 0.0023138847 0.7563196high_fat-drug2  -1.048637e-05 -0.0016005020 0.0015795293 1.0000000normal-drug2     2.504217e-03  0.0009142017 0.0040942330 0.0004945high_fat-drug3  -6.930494e-04 -0.0022830651 0.0008969662 0.7277757normal-drug3     1.821654e-03  0.0002316386 0.0034116699 0.0175538normal-high_fat  2.514704e-03  0.0009670961 0.0040623113 0.0003161> TukeyHSD(tmp)$group                         diff           lwr          upr        p adjdrug2-drug1      4.314514e-05 -0.0015468705 0.0016331608 0.9999915820drug3-drug1      7.257082e-04 -0.0008643074 0.0023157239 0.6929965170high_fat-drug1   3.265877e-05 -0.0015149488 0.0015802664 0.9999969171normal-drug1     2.547362e-03  0.0009997549 0.0040949700 0.0002612744drug3-drug2      6.825631e-04 -0.0009487586 0.0023138847 0.7563195891high_fat-drug2  -1.048637e-05 -0.0016005020 0.0015795293 0.9999999705normal-drug2     2.504217e-03  0.0009142017 0.0040942330 0.0004944674high_fat-drug3  -6.930494e-04 -0.0022830651 0.0008969662 0.7277757202normal-drug3     1.821654e-03  0.0002316386 0.0034116699 0.0175537862normal-high_fat  2.514704e-03  0.0009670961 0.0040623113 0.0003161003

参考

1. https://wenku.baidu.com/view/0a54c8d3a58da0116c174981.html###