Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 591 Bytes

File metadata and controls

10 lines (6 loc) · 591 Bytes

Reading: Testing Nested Models

Please read this section on using the F-test to test nested models.

In tests of this form:

  • The intercept-only model is the "null" model.
  • The fuller model that you have estimated is the "alternative" model.

The test is whether the F-statistic that this change in parameters creates is larger than the change that we would expect under the null hypothesis that in the population there is no actual relationship between outcomes and those variables. That is, that $\beta_{1} = \dots = \beta_{k} = 0$.