Closed saref closed 1 year ago
Yes, this is how that pvalue was computed (though 1-pvalue would still be a pvalue here as it's a two-sided test).
This repo is not the source for the labs, but will propose a change upstream to the source to correct. Hence, I will close, after having taken note of this. Thanks.
I have fixed a minor error in the text description for re-sampling p-value in ch13 Jupyter file.
P-value is the fraction of the times that the test statistics obtained via re-sampling exceeds our observed test statistic (in absolute value).
It was written the other way around. The code correctly matches with the revised description.