How should the PDS Constant Variance Test results be interpreted?
What is the inference of the results are less than 1%?
Per development: If the pValue is less than 1%, then the variance of the error term is not constant. Now, let me provide a bit more extensive explanation about this inference: t-Test is a hypothesis test to test for Type I Error (Significance Test), meaning the hypothesis is true, but is erroneously rejected. To be more specific, in this constant variance t-test, we set the following hypothesis: H0 : Errors have constant variance (NULL hypothesis) H1: Errors have non-constant variance Let's set a risk of 1% (0.01), as mentioned in this email. We basically say that we are willing to take 1% risk that the hypothesis is true, but is erroneously rejected. If pValue is greater than 0.01, it means the "actual" Type I Error is more than the risk we set, and it means there is insufficient evdience to show non-constant variance (accepting NULL hypothesis). On the other hand, if pValue is less than 0.01, it means the "actual" Type I Error is smallerthan the risk we set, and the NULL hypothesis is rejected with a 99% level of confidence that the error term has a non-constant variance (here we are talking about 1% of chance that the error term in fact has a constant variance). Whether to accept or to reject is determined based on the risk that you are willing to take. |
||
|