7.2 Final note about hypothesis testing

When you read journal articles, you’ll note that they rarely discuss the null or alternative hypothesis. They may explain their research questions or their hypotheses (these hypotheses are their alternative hypotheses), but they rarely discuss the null.

This is not necessarily a bad thing. Rather, what may be problematic with it is if researchers apply NHST without critically thinking about what their null hypothesis is or whether they have a one-sided hypothesis, which leads researchers to use defaults when the defaults may not be most appropriate. However, it would probably be a better thing if everyone clearly specified their alternative and null hypotheses if they are doing NHST.

Also, you may have heard some things about p-values not being reliable or desirable. We’ll discuss what p-values are and how some researchers want to move away from them in the next chapter. In the meantime, just know that p-values are often misunderstood and therefore misused, and a lot of the recommendations for moving away from p-values are just replacing one potentially problematic approach with another.