# The null hypothesis

Are the following statments true: 1. Rejecting the null in Breuch-Pagan test means you have conditional hetero problem 2. Rejecting the the null in Durbin Watson means you have serial correlation 3. Rejecting the null in Dickey Fuller means you do not have unit root problem

1. False 2. False? (these decision rules are complicated) 3. True
1. True?? 2. so i’m pretty sure 2 is false since DW=2(1-r) and if ur rejecting DW=0 you are rejecting r=1…so NO CORRELATION! 3. i think its false…same double negative relationship as above i tihnk.

no sorry 3. true. ahhh

They are all true… If you reject the null hypo for DW it means that your calculated stat ~ 2(1-r) falls outside du and 4-du, therefore you have a serial correlation problem. the null hypothesis doesnt always mean you are comparing to 0…

crap. wow the only 2 i got right was the one i was unsure aobut…the only one i was sure about i got wrong. hmm what does that tell you…

longoncfa - i was under the impression that you fail to reject the DW as it is inconclusive in two states adjacent to du - 4-du because test considered inconclusive. you reject if there is either positive or negative serial correlation.

sorry i am tired and misread you post