Statistic to measure no.of cancelled members

These percentages indicate how many male vs female members cancelled a subscription we offer.

Male             Female
22%              13%


Looking at the numbers it might seem like more men cancelled our subscription than females. This is not true however; because we have more men subscribed to our service than women.

Is there a statistic I can use that will allow me to definitively come to a conclusion that more men cancelled than women irrespective of the no.of men or women we have in our system. So, basically is there a statistic that ignores the underlying population of men and women? It is currently not being ignored because I am calculating this percentage by dividing the no.of cancelled men by the total no.of men.

• Given there are more men and the male percentage is higher, it is an ineluctable conclusion of arithmetic that more men cancelled than women. You seem to be looking for something beyond this obvious statement--but what? – whuber Mar 19 at 17:32

If $$m>w$$, then
$$0.22m > 0.13w$$