Suppose there are two sub-populations, 1 and 2, and two sexes in each population, m and f. Then the aggregate sex ratio, (m1 + m2)/(f1 +f2), is not equal to the average sex ratio .5*(m1/f1 + m2/f2). That’s obvious if the two sub-populations have very different sizes and very different sex ratios. But even if ratios are weighed by population size, ratios don’t aggregate linearly.

In considering a distribution of sex ratios across jurisdictions, the aggregate sex ratio is the sex ratio in the total population, ignoring jurisdictional distinction. The median sex ratio is the sex ratio with nearly as many jurisdictions having a higher ratio as have a lower ratio. The average sex ratio across jurisdictions is less meaningfully interpretable. It characterizes the sex-ratio distribution in a more generic statistical sense.