cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar

Durbin-Watson

I'm doing a regression with a dependent variable of price and an independant variable of date.

I know I need to check for autocorrelation with the Durbin Watson test but I don't know which number I should be looking at.

The one under "Durbin Watson" or the one under "Autocorrelation"? What is the threshold (greater than / less than) that I should be worried about?

Thanks.

Charlie
1 REPLY 1

Re: Durbin-Watson

In the JMP output, the "Durbin-Watson" value gives the test statistic (d), which is testing whether the residuals have first-order positive autocorrelation. You can get the p-value associated with this test by clicking on the section red triangle menu and selecting the "Significance P-Value" option (note that the computation of this exact probability can be memory and time-intensive if there are many observations). A significant p-value (based on your chosen significance level, commonly 0.05) allows you to reject the null hypothesis that the residuals are independent and conclude that there is first-order positive autocorrelation.

 

The value under AutoCorrelation simply displays the calculated autocorrelation of the residuals.

Rules of thumb regarding the value of d are discussed in the Durbin–Watson statistic Wikipedia article: Since d is approximately equal to 2(1 − r), where r is the sample autocorrelation of the residuals, d = 2 indicates no autocorrelation. The value of d always lies between 0 and 4. If the Durbin–Watson statistic is substantially less than 2, there is evidence of positive serial correlation. As a rough rule of thumb, if Durbin–Watson is less than 1.0, there may be cause for alarm. Small values of d indicate successive error terms are positively correlated. If d > 2, successive error terms are negatively correlated.