Null hypothesis significance testing: If we want to test a hypothesis about a model (and, therefore, the parameter estimates within it) using the framework described in Section 2.9 then we assume that the parameter estimates have a normal distribution.
