I am performing a simulation study where different values of an exposure X are repeatedly assigned to the same individual (200x in a sample of >10000 people). The way that assignment is done is performed in two fundamentally different ways: process 1 and process 2.
We then calculate the association between measure X and an outcome Y. Thus, we get 200 different values of association for process 1 and 200 different values of association for process 2.
The purpose of the analysis is to compare the means and variances of the association scores that result from process 1 and process 2. The default way to test for this would be the T-test and F-test (for difference in the mean and the variance, respectively). However, the two samples are dependent. Not only that, they consist of values derived from the same individuals. Which statistical approach would be best to formally compare the means and the variances of processes 1 vs. 2?
