Next, set the sample size to something relatively small, like 15 or 20. Is it possible to make the probability of a Type II error effectively zero even for a very small sample? Explain what that means in practical terms.