interpreting the Anderson darling test scipy

Question

Interested in knowing how to Interpret the result of the Anderson darling test in python.

It seems as though the AD stat has to be lower then the critical value at its associated significance level though I am not sure how to properly determine this from the return on the function.

this is the result of the function

AndersonResult(statistic=1.383562257554786,
               critical_values=array([0.574, 0.654, 0.785, 0.916, 1.089]), 
               significance_level=array([15. , 10. ,  5. ,  2.5,  1. ]))

score 8 · Accepted Answer · answered Jan 05 '19 at 07:29

The Anderson Darling tests whether the sample from the population follows a particular distribution. If the test statistic is higher than a critical value, the null hypothesis is rejected at the corresponding significance level (i.e there is evidence to suggest that the population doesn’t follow that particular distribution).

Looking at your output, the null hypothesis is rejected at all significance levels here as the test statistic is higher than all the critical values.

You can check out the documentation on the AD test here https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.anderson.html.

score 1 · Answer 2 · answered Apr 02 '20 at 13:21

In case you want to deepen your understanding of the details of the test you can reach out beyond the SciPy documentation to NIST reference or an in-depth article on Computation of Probability Associated with Anderson–Darling Statistic.

In case you have a mathematical question about the test validity or formulation please search (or post) Cross Validated forum.

interpreting the Anderson darling test scipy

2 Answers2