我用一种重复的方法进行了一种方差分析:
aov <- aov(score ~ group*time + Error(subject/time), data=valueAest)
但是,关于组间差异是来自治疗前,治疗后还是两者均难以解释的结果?
Error: subject
Df Sum Sq Mean Sq F value Pr(>F)
group 1 12.01 12.013 4.424 0.0421 *
Residuals 38 103.17 2.715
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Error: subject:time
Df Sum Sq Mean Sq F value Pr(>F)
time 1 0.012 0.0125 0.029 0.866
group:time 1 0.013 0.0125 0.029 0.866
Residuals 38 16.475 0.4336
所以我决定进行事后分析。在查看其他人的答案时,我使用了multcomp库:
lme_score = lme(score ~ group, data=valueAest, random = ~1|subject)
anova(lme_score)
require(multcomp)
summary(glht(lme_score, linfct=mcp(group = "Tukey")), test = adjusted(type = "bonferroni"))
但这给了我“ ncol(linfct)”的错误不等于“ length(coef(model))”的错误。
我也尝试过:
Mixed_Fitted_Interaction<-emmeans(aov, ~group|time)
Mixed_Fitted_Interaction
# pairwise comparison
pairs(Mixed_Fitted_Interaction)
这给了我以下结果:
time = recreation1:
contrast estimate SE df t.ratio p.value
map - VR -0.75 0.397 49.8 -1.890 0.0646
time = recreation2:
contrast estimate SE df t.ratio p.value
map - VR -0.80 0.397 49.8 -2.016 0.0492
在我幼稚的头脑中,这看起来很有希望,我是否可以得出结论,其重要性实际上并非来自治疗,而是来自人口差异?