Data : same 10 people measured 1 hour after Type A and Type B
A <- c(15,26,13,28,17,20,7,36,12,18)
B <- c(13,20,10,21,17,22,5,30,7,11)
d <- A - B
c(diffs = I(d), mean_d = mean(d), sd_d = sd(d), n = length(d))
## diffs1 diffs2 diffs3 diffs4 diffs5 diffs6 diffs7 diffs8
## 2.000000 6.000000 3.000000 7.000000 0.000000 -2.000000 2.000000 6.000000
## diffs9 diffs10 mean_d sd_d n
## 5.000000 7.000000 3.600000 3.098387 10.000000
a : hypothesises H0: mean(A − B) = 0 (no difference in mean concentration) Ha: mean(A − B) ≠ 0 (two-sided)
b : Paired t-test (α = 0.05), report p-value and conclusion
t.test(A, B, paired = TRUE)
##
## Paired t-test
##
## data: A and B
## t = 3.6742, df = 9, p-value = 0.005121
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
## 1.383548 5.816452
## sample estimates:
## mean difference
## 3.6
conclusion : the average within person difference is about 3.60 mg%. The paired t-test p value is about 0.005 (df = 9). Since p < 0.05, I reject H0. In this sample A is higher on average.
c : two-sample t-test instead of paired, whats p value?
t.test(A, B, var.equal = TRUE)
##
## Two Sample t-test
##
## data: A and B
## t = 0.9802, df = 18, p-value = 0.34
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -4.116103 11.316103
## sample estimates:
## mean of x mean of y
## 19.2 15.6
t.test(A, B)
##
## Welch Two Sample t-test
##
## data: A and B
## t = 0.9802, df = 17.811, p-value = 0.3401
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -4.12199 11.32199
## sample estimates:
## mean of x mean of y
## 19.2 15.6
Data:
Active <- c(9.50, 10.00, 9.75, 9.75, 9.00, 13.00)
NoExercise <- c(11.50, 12.00, 13.25, 11.50, 13.00, 9.00)
a : hypothesises H0: the two groups have the same distribution Ha: Active < NoExercise (one-sided; expect earlier walking with Active)
b : why non-parametric small samples (6 vs 6), ties, and normality is doubtful. The rank based Mann Whitney test is safer here
c : Mann-Whitney
wilcox.test(Active, NoExercise, alternative = "less")
##
## Wilcoxon rank sum test with continuity correction
##
## data: Active and NoExercise
## W = 9, p-value = 0.08523
## alternative hypothesis: true location shift is less than 0
The one sided p value is about 0.09. Since p ≥ 0.05, I do not reject H0. The active group looks a little lower, but the evidence isn’t strong with n=6 per group.