How it works...

We run this simulation multiple times, compute the t-tests/Welch tests, and accumulate them into a rejected variable. As we explained previously, we should expect to get 5% of cases rejected (this works as an accumulator when using a critical 0.05 p-value.