We did it for the Crump Test, the Randomization Test, and the \(t\)-test… We make fake data, we simulate it, we compute the sample statistic we are interested in, then we see how it behaves over many ...We did it for the Crump Test, the Randomization Test, and the \(t\)-test… We make fake data, we simulate it, we compute the sample statistic we are interested in, then we see how it behaves over many replications or simulations. Notice, if I told you I ran an experiment with three groups, testing whether some manipulation changes the behavior of the groups, and I told you that I found a big \(F\)!, say an \(F\) of 6!. And, that the \(F\) of 6 had a \(p\)-value of .001.
