how to calculaute p-value for durbin watson statistics?

For questions regarding programming in the EViews programming language.

Moderators: EViews Gareth, EViews Jason, EViews Moderator, EViews Matt

ecofine
Posts: 28
Joined: Fri Sep 26, 2014 2:24 am

how to calculaute p-value for durbin watson statistics?

Postby ecofine » Mon Apr 23, 2018 12:26 am

Could anybody know how to calculate p-value for durbin watson statistics?
I 've been always wondering why eviews does not produce p-value for durbin watson statistic.

HIS

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: how to calculaute p-value for durbin watson statistics?

Postby startz » Mon Apr 23, 2018 6:26 am

The exact distribution of the Durbin-Watson is hard and can only be found by simulation. That's why tables give upper and lower bounds. I don't think any program gives p-values. (In any event, for a reasonably large sample you are better off with the Breusch-Godfrey.)

ecofine
Posts: 28
Joined: Fri Sep 26, 2014 2:24 am

Re: how to calculaute p-value for durbin watson statistics?

Postby ecofine » Mon Apr 23, 2018 6:59 pm

Thank you for your reply.
Well, stata provides a command to get a p-value of DW statistic.
The table for Dw is given and i guess eviews can also provide p-value.

HIS

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: how to calculaute p-value for durbin watson statistics?

Postby startz » Mon Apr 23, 2018 7:41 pm

ecofine wrote:Thank you for your reply.
Well, stata provides a command to get a p-value of DW statistic.
The table for Dw is given and i guess eviews can also provide p-value.

HIS

I stand corrected. I didn't know Stata could do that. Is it Stat's dwe command that you use? If it is, let me note that the Stata documentation states, in part,

We at Stata are not fans of the original Durbin-Watson test because the
test's p value is known to be heavily dependent on the
normality-of-the-residuals assumption. A far better test is Durbin's
alternative test, available in Stata by tying -estat durbina- after
estimation by -regress-.

As others on the list have noted, they have requested p values and we have
not responded. Other than (1) doing homework or (2) reproducing the results
of others, we know of no good reason why anyone should want the p value that
assumes normality of the Durbin-Watson statistic.

ecofine
Posts: 28
Joined: Fri Sep 26, 2014 2:24 am

Re: how to calculaute p-value for durbin watson statistics?

Postby ecofine » Tue Apr 24, 2018 11:25 pm

Thank you very much for your kind information.
I will look up to stata information and come back here again if I would have furthermore questions.

Have a nice day!

HIS


Return to “Programming”

Who is online

Users browsing this forum: No registered users and 14 guests