After testing the individual b_(k)'s, we next want to test the entire model with all of the parameters.

Hypothesis Tests

   = 0    -->all regression coefficients are zero, try a different model

   at least one b_(k) differs from 0    -->at least one regression coefficient is not zero,

proceed with that model

Given a Model:

           where k=1,...,p-1

For Testing the Model:

               same as simple linear

   if F > F(crit) : Reject H₀

   if F <= F(crit) : Fail to Reject H₀

        With a fixed alpha in the upper tail.

Now let's try an example.

Regression Tutorial Menu    Dictionary

STATS @ MTSU