Do we need to only know DW = 2(1-r) or also the other cheesy residual term formulae … has someone come across the formula being tested ??
i’m sticking with the one you put in…
I think you had better. They might have someting in there of the DW is equal to 1.65. What is the implied serial correlation of the error terms in the multiple regression. This would be ranked a moderate questions i would think, but easy to solve with a simple formula. r = .175 or they might say based on the data present what is the DW test equal to: a) 4.5 b) -4.5 c) 1.65 d) -1.65 here the only option that makes sense is C and you needed no data!
Just remember the DW line looks like this 0----2----4. Close to the (0) positive serial correlations…closer to 4…negative serial correlation. Close to 2, no correlation. In between the 0-2 and 2-4 are your dl/dh that they will most likely give you. If it’s between these dl and dh it’s INCONCLUSIVE. In order from left to right of this line just remember …positive, inconclusive, zero, inconclusive, and negative correlation. That’s how i remember the details, I dont know if I make any sense to you, but hope it helps.
Dummy use n-1 dummy variables… WE ALL KNOW use of n variables … will cause collinearity (atleast I didn’t know uptil know)
heeralm Wrote: ------------------------------------------------------- > Dummy > > use n-1 dummy variables… WE ALL KNOW > > use of n variables … will cause collinearity > (atleast I didn’t know uptil know) yeps, see this link - hope it will help you to nail down this concept. http://www.analystforum.com/phorums/read.php?12,730201,730313#msg-730313