其他
随机区组设计，R语言轻松实现统计分析
> block <- c(rep(1, 3), rep(2, 3), rep(3, 3), rep(4, 3)) ##生成区组
> treat <- c(rep(c("A", "B", "C"), 4)) ##生成组别
> alt <- c(76, 86, 115, 12, 38, 85, 40, 81, 103, 12, 33, 57)
> alt <- data.frame(block= block, treat= treat, alt= alt) ##建立数据框
> alt$block <- as.factor(alt$block) ##因子化
> alt$treat <- as.factor(alt$treat) ##因子化
> model <- aov(alt ~ treat+ block, data= alt) ##随机区组方差分析
> summary(model) ##查看模型
Df Sum Sq Mean Sq F value Pr(>F)
treat 2 6074 3037.0 33.54 0.000554 ***
block 3 6458 2152.6 23.77 0.000992 ***
Residuals 6 543 90.6
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
> TukeyHSD(model) ##两两比较
Tukey multiple comparisons of means
95% family-wise confidence level
Fit: aov(formula = alt ~ treat + block, data = alt)
$treat
diff lwr upr p adj
B-A 24.5 3.853961 45.14604 0.0251803
C-A 55.0 34.353961 75.64604 0.0004428
C-B 30.5 9.853961 51.14604 0.0094278
$block
diff lwr upr p adj
2-1 -47.33333 -74.230267 -20.436399 0.0036152
3-1 -17.66667 -44.563601 9.230267 0.2062995
4-1 -58.33333 -85.230267 -31.436399 0.0011919
3-2 29.66667 2.769733 56.563601 0.0333717
4-2 -11.00000 -37.896934 15.896934 0.5343125
4-3 -40.66667 -67.563601 -13.769733 0.0077913
> res <- residuals(model) ##生成残差
> shapiro.test(res) ##评估残差正态性
Shapiro-Wilk normality test
data: res
W = 0.88283, p-value = 0.09529