Durbin-Watson Statistics Table has three types of critical values for significance at 1%, 2.5% and 5% level. So how to choose which one to use when evaluating Durbin-Watson statistics (e.g. d=1.12)?
Thank you!
Usually 5% significance level is used. However, this test is bit different from others. You need to get both lower limit and the upper limit. Please use the attach guideline to see how it is performed.
Kind regards and good luck with your research.
Thushara
D-W test is interpreted based on its value
if the value is 2- no autocorrelation
less than 2, positive autocorrelation
more than 2, negative autocorrelation
there is no specific probability distribution for DW test.Hence, 'p' values are not generated for it.
Srikanth Potharla is partly wrong. Let us say that the probability distribution was considered as too complex for being described. Note that DW is approximately 2 - 2 r_{1}, where r_{1} is lag 1 autocorrelation. So, in absence of regressors, and using the N(0,1/n) asymptotic distribution for r_{1} in the case of a white noise process, it is possible to obtain rough p-values. Or course they are not valid with regressors and for too small samples. I suppose this has been investigated using Monte Carlo simulations.
The level of significance of 5% is generally accepted. There is no scientific criterion that this is the best. The most important Durbin-Watson test is the number of observations and the number of explanatory variables.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics