[Gretl-users] Re: Cointegration
Am 25.08.2020 um 02:21 schrieb Mike @me.com:
Sven,
I may have answered my own question. I found this: "A pvalue of less than or
equal to 0.05 is often used to indicate whether there is evidence against
the null hypothesis".
Would you agree with that in my case?
I agree that the statement is correct, in particular the "often used".
Whether you should be using that threshold value is an entirely
different question which I cannot answer for you. People are using the
"customary" significance levels --without reflecting why-- far too often
(IMHO).
But as you are far beyond that value at below one per mille (thousand)
you should be OK indeed. (Provided of course you don't have any glaring
misspecification in your model setup.)
cheers
sven