データ
| Sequence |
SID |
Period1 |
Period2 |
| RT |
1 |
74.675 |
73.675 |
| RT |
4 |
96.4 |
93.25 |
| RT |
5 |
101.95 |
102.125 |
| RT |
6 |
79.05 |
69.45 |
| RT |
11 |
79.05 |
69.025 |
| RT |
12 |
85.95 |
68.7 |
| RT |
15 |
69.725 |
59.425 |
| RT |
16 |
86.275 |
76.125 |
| RT |
19 |
112.675 |
114.875 |
| RT |
20 |
99.525 |
116.25 |
| RT |
23 |
89.425 |
64.175 |
| RT |
24 |
55.175 |
74.575 |
| TR |
2 |
74.825 |
37.35 |
| TR |
3 |
86.875 |
51.925 |
| TR |
7 |
81.675 |
72.175 |
| TR |
8 |
92.7 |
77.5 |
| TR |
9 |
50.45 |
71.875 |
| TR |
10 |
66.125 |
94.025 |
| TR |
13 |
122.45 |
124.975 |
| TR |
14 |
99.075 |
85.225 |
| TR |
17 |
86.35 |
95.925 |
| TR |
18 |
49.925 |
67.1 |
| TR |
21 |
42.7 |
59.425 |
| TR |
22 |
91.725 |
114.05 |
プログラムと結果
変量効果モデル
library(nlme)
tab361 <- read.table("table361.csv", sep=",", header=T)
nrow1 <- nrow(tab361)
sequence <- rep(tab361$Sequence, rep(2, nrow1))
SID <- rep(tab361$SID, rep(2, nrow1))
Period <- rep(1:2, nrow1)
Drug <- rep(" ", nrow1*2)
y <- double(nrow1*2)
for (i in seq(nrow1)) {
if (tab361$Sequence[i] == "RT") {
Drug[i*2-1] = "R"
Drug[i*2] = "T"
} else {
Drug[i*2-1] = "T"
Drug[i*2] = "R"
}
y[i*2-1] = tab361$Period1[i]
y[i*2] = tab361$Period2[i]
}
tab361.2 <- data.frame(Sequence=sequence, SID, Period=as.factor(Period), Drug, y)
fm1.lme <- lme(y ~ Sequence+Period+Drug, data=tab361.2, random=~1 | SID)
anova(fm1.lme)
numDF denDF F-value p-value
(Intercept) 1 22 431.7741 <.0001
Sequence 1 22 0.3745 0.5468
Period 1 22 0.2151 0.6474
Drug 1 22 0.3754 0.5463
Split-plot モデル
fm2.aov <- aov(y ~ Sequence+Period+Drug+Error(factor(SID)), tab361.2)
summary(fm2.aov)
Error: factor(SID)
Df Sum Sq Mean Sq F value Pr(>F)
Sequence 1 276.0 276.0 0.3745 0.5468
Residuals 22 16211.5 736.9
Error: Within
Df Sum Sq Mean Sq F value Pr(>F)
Period 1 36.0 36.0 0.2151 0.6474
Drug 1 62.8 62.8 0.3754 0.5463
Residuals 22 3679.4 167.2
参照
