A little geek talk about understands what the null hypothesis means in a couple of common research settings. One of my longer blog posts.
When an audit or financial accounting researcher studies a mandatory disclosure regulation it is important to think about what is the null. The alternative hypothesis is simple, there is a reaction to the disclosure.
Case 1. You find the disclosure of new information to both markets and to managers affects real operations of the firm. In other words the null is rejected. What did you reject?
The null is NOT no disclosure would not affect real operations. If the new information is insightful to managers, whether they disclose it or not, they might use it in their operating decisions. Hence, the null is that the disclosure of new information to both market and managers does not affect operating decisions, which is quite different from disclosure of new information to managers but not to the public means no effects on operating decisions! The problem is that without the disclosure-being made, or researcher inside access to the firm, is that the researcher cannot tell the difference between the two possible nulls! Read this slowly to fully appreciate the difference.
In Case 1, misunderstanding the null could cause you to recommend disclosure must happen or managers would not use the new information. Maybe they would not, but maybe they would still use it in managing! The problem is that the researcher without inside access can not tell unless the disclosure is made! Next day, what if you find no effect?