Problem 2
The Multiple Testing dataset (multiple.txt) is a simulated
dataset which contains 50 variables and 100 observations per
variable. Suppose we know that the first 10 variables have mean
equal to 2 and the rest of them have mean equal to 0. Analyze
the dataset according to the following steps:
1. Perform multiple testing to the population mean vector to
test if it equals to a vector whose elements are all zeros. Set
the significance level at a = 0.1
2. Based on the test results in step 1, calculate the following
quantities: number of type I errors, FWER and FDP.
3. Redo the multiple testing in step 1 with Bonferroni correction
(set a = 0.1). Calculate the FWER of your new test results
4. Redo the multiple testing in step 1 with BH procedure (set
a = 0.1). Calculate the FDP and FWER of your new test results.
How does the results compared with the ones you obtained in
step 1 and step 3?