Determining groups that have been historically disadvantaged is generally not difficult.

The part that is difficult is how you correct for that. There are many things you need to deal with. The most obvious ones being

* Fields having relatively small populations run into the standard low population vs. overall population statistics

* Fields that have historically restricted access to a group can legitimately point to a lack of "qualified" members in the restricted group

Neither has an easy solution. The only thing we know is that a failure to actively work on correcting the effect of historical discrimination extends the impact of that discrimination beyond the point when the discrimination "ended".

Like I said in the original post this particular example is clunky, but I also don't have any better ideas :-/