Critical F value for testing Full vs Reduced models.

cgllll

New Member
Full Model: Y=b0+b1x1+b2x2+b3x3+b4x4

Supposed the sample size is 16. What is the critical F value for testing the full versus the following model: Y=b0+b2x2+b3x3+b4x4?

ondansetron

TS Contributor
What are the pertinent formulae for computing F-statistics? What is the formula for testing individual coefficients with a T-test? Do you need one, both, or neither? Why?

hlsmith

Not a robit
I know I am a hard as s on this but both seem pretty saturated.

j58

Active Member
For the F-statistic, the numerator degrees of freedom is the difference between the number of parameters in the full model and the number of parameters in the reduced model, and the denominator degrees of freedom is n - p_f, where n is the sample size and p_f is the number of parameters (including intercept) in the full model.

But wouldn't it be easier just to look at whether b1 is significant?

Last edited:

ondansetron

TS Contributor
For the F-statistic, the numerator degrees of freedom is the difference between the number of parameters in the full model and the number of parameters in the reduced model, and the denominator degrees of freedom is n - p_f, where n is the sample size and p_f is the number of parameters (including intercept) in the full model.

But wouldn't it be easier just to look at whether b1 is significant?
If you look at my post, I wanted the OP to come to this conclusion after seeing what results from calculating and F stat and a t-stat. I assumed this is homework so I took a harder stance 