Assignment week 4.2 : Analysis of continuous data three or more group comparison
Statistic summary of amyalase level among the three groups of patients
setwd("C:/TMHG531/assignment")
ind.dat3 <- read.csv("ind_t_test3.csv")
#Group 1
summary(ind.dat3[ind.dat3$group==1,2])
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.0 39.0 47.0 66.0 94.5 135.0
#Group 2
summary(ind.dat3[ind.dat3$group==2,2])
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 12.00 48.00 100.50 94.55 134.00 200.00
#Group 3
summary(ind.dat3[ind.dat3$group==3,2])
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 78.0 97.5 129.5 161.5 227.5 344.0
Boxplot of Amylase level among the three groups of patients
boxplot(amylase~group, data=ind.dat3, main="Amylase level among the three group of patients",
xlab="group", ylab="amylase level", col="pink")
** Analysis of Variance of data**
res.aov <- aov(amylase~group,data=ind.dat3)
summary(res.aov)
## Df Sum Sq Mean Sq F value Pr(>F)
## group 1 58698 58698 17.26 0.000137 ***
## Residuals 47 159856 3401
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Find the pair with diffrence mean of amylase level
pairwise.t.test(ind.dat3$amylase,ind.dat3$group, p.adjust.method="bonferroni")
##
## Pairwise comparisons using t tests with pooled SD
##
## data: ind.dat3$amylase and ind.dat3$group
##
## 1 2
## 2 0.44781 -
## 3 0.00032 0.00727
##
## P value adjustment method: bonferroni
Checking normality of the distribution of data
plot(res.aov)
Homogeneity of the data
library(car)
## Loading required package: carData
bartlett.test(amylase~group, data=ind.dat3)
##
## Bartlett test of homogeneity of variances
##
## data: amylase by group
## Bartlett's K-squared = 8.4709, df = 2, p-value = 0.01447
RESULT
Descriptives
| Group | N | Mean | Min | Max |
|---|---|---|---|---|
| 1 | 15 | 66.0 | 29 | 135 |
| 2 | 22 | 94.55 | 12 | 200 |
| 3 | 12 | 161.5 | 78 | 344 |
Hypothesis
| H0 | μ1 = μ2 = μ3 |
|---|---|
| H1 | Not all means are equal |
Analysis of Variance
The P-value is 0.000137 which less than significant level of 0.05. The null hypothesis is rejected. Therefore, there are significant difference between the groups.
| Df | Sum of Squares | Mean Squares | F-value | P-value | |
|---|---|---|---|---|---|
| Between Groups | 1 | 58,698 | 58,698 | 17.26 | 0.000137 |
| Within Groups | 47 | 159,856 | 3,401 | ||
| Total | 48 | 218,554 | 62,099 |
Normality of the distribution
According to the plot of the fitted againts residuals and the normal QQ, the data are random and independant and are normally distributed.
Multiple comparisons T-test
The P-value of pairing group 1 againts group 3 (0.00032) and group 2 againts group 3 (0.00032) is less than the significant level of 0.05. Therefore, mean of group 1 and group 2 and mean of group 2 and group 3 are significantly different.There are no evidence that mean of group 1 and group 3 are difference asthe P-value is more than the significant level of 0.05.
| Pair | P-value | |
|---|---|---|
| Group 1 | Group 2 | 0.44781 |
| Group 3 | 0.00032 | |
| Group 2 | Group 3 | 0.00727 |
Bartlett test of homogeneity of variances
The P-value is 0.01447 which is less than the significant level of 0.05. This is the evidence of the variance across groups are statistically significantly difference. Therefore, we can assume the heterogeneity of varience in the diffrence groups of patients.
| K-squared | Df | P-value |
|---|---|---|
| 8.4709 | 2 | 0.01447 |