A Frequent Misuse of Significance Tests
CESifo, Munich, 2001
CESifo Working Paper No. 549
![](https://cesifo.org/DocImg/cesifo_wp549.jpg?c=1689237113)
Economists sometimes interpret the failure of a significance test to disconfirm a hypothesis as evidence that this hypothesis is valid. Six examples of this are cited from recent journals. But this is a misinterpretation of what significance tests show. While in general it is correct that every failure to disconfirm a hypothesis adds to its credibility, the term "disconfirm" is defined differently for this purpose than it is in the context of significance tests.