TOST(sample1, sample2, d1, d2, alpha)

Two One-Sided Test (TOST) for equivalence

As per Roger et al, 1993, this function tests if two samples come from distributions with different means, against the alternative hypothesis that the distribution means are the same. I.e.,
H0: the two distributions have different means (the difference between the means falls outside of an equivalence interval [d1, d2])
HA: the two distributions have equivalent means (the difference between the means falls within the equivalence interval [d1, d2])
This function performs two one-sided t-tests on the input samples. The outputs of this function include two p values (one per t-test) and a 90% confidence interval. The null hypothesis should be rejected if max([p1, p2]) > alpha, or if the confidence interval falls outside of the equivalence interval.