Exercise 2
Due date : 2/4 before 21:00, via email. Send a short pdf with
your code, results, and discussion.
Can be done individually or in pairs
Small data set 1 non paired
Using the files a1.csv and b1.csv from Exercise 1, run a Monte-Carlo
permutation test (5000 repetitions) and report the p-value. Compare with the ones
reported by the t-test and Wilcoxon sum rank (from exercise 1).
paired data set 1
Using the paired data set paired.csv
from exercise 1, compute the Monte Carlo permutation p-value.
this discussion explains the paired permutation test. Compare
with the results of the paired tests in Exercise 1.
Confidence Intervals
File
single.csv contain one set data. Compute the confidence interval for the mean using
- the reverse t test
- the reverse Wilcoxon test (it gives the interval for the
median not the mean)
- bootstrap (using the BCA method) with 5000 repetitions.
Significance test for categoric data
File
two.csv contains a paired binary data
(coded as 0/1). Run the following precedures and compare the
p-values. Discuss if it is appropriate or not to run the test (I am
not sure the is a correct answer to this question!!)
- McNemar test
- (non paired) Chi-square test
- paired t test
- Wilcoxon signed rank test