Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.
|
|
Luis A. Afonso
Posts:
4,292
From:
LIsbon (Portugal)
Registered:
2/16/05
|
|
Jarque-Bera: a disastrous test
Posted:
Aug 24, 2012 12:03 PM
|
|
Jarque-Bera: a disastrous test
I repeatability advised here that the so-called Jarque-Bera test of normality does not fulfill the expected requirements a test must obey, namely it skips completely the p-values skewness and kurtosis knowledge. In fact, it deals with a two parameter?s sum and consequently an abnormal high of one of them can be hidden by a sufficiently low of the other in order that the sum is not rejected. So we *accept* normality when it must be rejected by the commonly asserted standards, the test ignores.
What absolutely amazing is this fatal feature passed undetected (as far as I know) since early 80´s, and, notwithstanding the error, being heavily used by economists till today.
Luis A. Afonso
|
|
|
|
