A tutorial on how to conduct and interpret F tests in Stata. First, we manually calculate F statistics and critical values, then use the built-in test command.
Jesus christ dude. I cannot describe how lost I was with my econometric homework. You explained literally everything perfectly and clearly, and I had no trouble following along with my own models. Do you have a venmo so I can tip you a 6-pack?
@sebastianWaiEcon thanks very much for this great video. It is extremely helpful. I have looked for a source for the F statistic and P-value calculations in order to reference but I cannot find one that matches this. Have you got a reference please? Secondly, how would this calculation differ if you have two models with exactly the same parameters?
Apologies for missing this. I based this video off Introductory Econometrics by Jeffrey Wooldridge. For an F test, you are always comparing what happens when you remove variables.
thank you so much but you did not give us the cut off F-value. You jus mentioned tht 6.25 means that something in our regression is explaining the dep var (salary). My question is what is the cut off value to check if the F-value is too low to explain the dep var
There is no one "cutoff" to be used in all situations. Critical values for the F distribution depend on the degrees of freedom, so will be unique to each test. Further, the actual significance threshold is always going to be subjective. Typically, people look at 90, 95, and 99% significance, but the choice is yours.
Hi, SebastianWaiEcon. I would like to ask you about F statistic. If i ran a regression and the F(... , ...) for my unrestricted model is very high (216.86) and my Prob > F = 0.0000, I've never seen such high numbers from all the youtube tutorials I've watched. Does this indicate something is wrong with my model or is it more of a common occurences than I thought? Thank you in advance.
@@sebastianwaieconOh I see. Indeed, my sample consisted of +-2800 observation, maybe that is what caused the fstat went high. Thank you so much for your reply, SebastianWaiEcon!!!
This means the p value is so small that it doesn't even show up with 4 decimal places. The p value is not zero -- that is impossible -- it is just a very, very small number.
![[TF2] The Russian Bear Trap](http://i.ytimg.com/vi/YFNSwz2_rOA/mqdefault.jpg)
Good sir, I cannot BEGIN to tell you how many lives you've saved with this video.
This internet gentleman is a better teacher than my tenured econometrics professor
You saved my life!! I am about to have a test on stat regrading regression and your videos are so well explained!
still helpful 3 years later!!!!!
Alas good sir from Durham University you have saved my life
Jesus christ dude. I cannot describe how lost I was with my econometric homework. You explained literally everything perfectly and clearly, and I had no trouble following along with my own models.
Do you have a venmo so I can tip you a 6-pack?
I'm just happy you found this helpful!
You are my new favorite human being.
That's a great way to explain the joint significance of explanatory variables.
Extremely helpful Mr. Sebastian.
Thank you with all my heart
Thank you very much for the nice explanation, keep it up.
You are my hero
Thanks a lot!!! quite practical and useful
This is so helpful! Thanks!!
Thank you very much
it is very clear, thanks
Thank you!
Life saving!
thanks so much!!!
@sebastianWaiEcon thanks very much for this great video. It is extremely helpful. I have looked for a source for the F statistic and P-value calculations in order to reference but I cannot find one that matches this. Have you got a reference please? Secondly, how would this calculation differ if you have two models with exactly the same parameters?
Apologies for missing this. I based this video off Introductory Econometrics by Jeffrey Wooldridge. For an F test, you are always comparing what happens when you remove variables.
thank you so much but you did not give us the cut off F-value. You jus mentioned tht 6.25 means that something in our regression is explaining the dep var (salary). My question is what is the cut off value to check if the F-value is too low to explain the dep var
There is no one "cutoff" to be used in all situations. Critical values for the F distribution depend on the degrees of freedom, so will be unique to each test. Further, the actual significance threshold is always going to be subjective. Typically, people look at 90, 95, and 99% significance, but the choice is yours.
Hi, SebastianWaiEcon. I would like to ask you about F statistic.
If i ran a regression and the F(... , ...) for my unrestricted model is very high (216.86) and my Prob > F = 0.0000, I've never seen such high numbers from all the youtube tutorials I've watched. Does this indicate something is wrong with my model or is it more of a common occurences than I thought? Thank you in advance.
I can't think of a specific reason why that would indicate anything wrong. F and t stats can get quite high if you have a very large sample.
@@sebastianwaieconOh I see. Indeed, my sample consisted of +-2800 observation, maybe that is what caused the fstat went high. Thank you so much for your reply, SebastianWaiEcon!!!
Hi how do I categorise Sic codes by industry in stata?
Sorry but I dind't understand how to choose the variables to remove to do the restricted model. Can you explain me that? please
This depends on your null hypothesis. You remove the variables for which you want to test for joint significance.
Thank you！！
Whenever i enter invFtail into stata, it is saying that it’s an “unrecognised command” please help!
You need to add "display" in front (can abbreviate to "di"). See the video at about 5:30 to see an example.
di invFtail(3,170,.01) try with no spaces between values.
Cool !!!
Is it possible to use the test command for interaction terms, in order to check whether a particular set of interactions has explanatory power or not?
Yes, just list them normally like you would any other variables.
@@sebastianwaiecon I tried but it shows an error "invalid matrix stripe". However, the command testparm works!
Need to balance the comments out. Not to impressed with this even though I learned a lot:)
What's the f test command in stata?
See 7:00 in this video for the test command.
Is this the guy from Khan Academy?
I teach at Indiana University.
How do you interpret „Prob> F = 0.0000“?
This means the p value is so small that it doesn't even show up with 4 decimal places. The p value is not zero -- that is impossible -- it is just a very, very small number.