Reproducibility probability estimation for testing statistical hypotheses

The applications of statistical tests often generate problems, because of their randomness. Here, stemming from the variability of statistical tests, the reproducibility probability of a statistically significant result () is estimated first. Then, statistical tests are defined on the basis of estimation. Lower bounds for , and then pointwise estimates are used. It is emphasized that the threshold for the pointwise estimator to define statistical tests turns out to be 1/2. Hence, taking the perspective of estimation, one can compute statistical tests and have suitable interpretations of their results.