I was just reading through the Adobe docs regarding contribution
analysis and I noticed an "anomaly" (irony intended) in the formula used
to calculate Carmer's V stat.In the above formula, degrees of freedom
has not been accounted for. Given that (C+A) represents the entire
dataset, subtracting degrees of freedom from rows and columns is
required to account for the constraints that arose in using this data.
I'm assuming that this what the below formula is achieving, however,
it's unclear. Thanks...