Page 1 of 1
probability level for Durbin watson statistics
Posted: Sat Oct 24, 2009 8:47 am
by kelaynak
Hello,
For eviews 7; Is it possible to add probability level (p-value) on DW value in all regression analysis result?
Thanks
Re: probability level for Durbin watson statistics
Posted: Wed Oct 28, 2009 12:05 pm
by tchaithonov
I wonder what you want more from DW = (Sigma(e(t) - e(t-1))^2)/(Sigma(e(t)^2) ..... is there a p-value associating with it? I can't remember there is such a thing ...
Re: probability level for Durbin watson statistics
Posted: Thu Oct 29, 2009 7:29 am
by kelaynak
Yes there is p value for DW. I saw afew algorithm in fortran.
Matlab also calculates it, with "dwtest (R,X, method)" commands
Method:
exact p for n<400
approximate p for n>400