**4. Summary**

The point in the ASA statement is not that *p*-values give the wrong answer; the point is that *p*-values usually commit what (Raiffa (1968), attributing the idea to John Tukey) called "errors of the third kind: solving the wrong problem." Not always, of course. For example, in a capital punishment case, we might well be interested only in controlling for Type I error against a null of not guilty, as distinct from deciding whether the accused is innocent or guilty. But in most cases, we do care about what the data tells us with regard to the probability of the studied hypothesis. As a practical matter, the *p*-value cannot be expected to be a good guide for this probability.

**Funding:** This research received no external funding.

**Acknowledgments:** Elements of the arguments made here are drawn from Startz (2014). Helpful suggestions from the academic editor and referees are gratefully acknowledged.

**Conflicts of Interest:** The author declares no conflict of interest.
