Testing
There are lots of tests.
Here we focus on testing coefficient restrictions in linear models.
t Tests,
Size and Power
Two-tail Tests
One-tail Tests
F tests
General linear restrictions
Chow tests
How to rearrange to get
H0: beta = 0.
Warning
Testing is only valid given the conclusions, which in turn are only
valid if the assumptions are satisfied by the data at hand.
When is a t test equivalent to an F test?
When can you do an F test, but not a t test?
Why do they call it
Student's
t distribution? (Hint: it has something to do with beer.)
Pitfalls
R2: the real story
