There being more male construction workers doesn't make it a sexist system.
There being more female teachers doesn't make it a sexist system.
There being more asian doctors doesn't make it a racist system.
Let people do what they want to do, you can look at demographics but stop trying to read racist tea leaves with it.
If you also collect demographics of applicants, it may tell a different story.
If your employee demographics look skewed, but they're consistent with applicant demographics, you have an easy out when accused of ism.
Of course, if you collect the information, you might also use it for ism purposes. Or you might just lose or discourage applications from non-favored people. It's hard to show.
(It also doesn't help that the federal categories feel poorly chosen and ill-defined)
Are you sure? There being more male programmers is widely used as evidence that it's a sexist system.
If females don't want to go into programming then don't make them.
There are really good female programmers, but males tend to be more frequent.
We all have different traits that make us good at different things, that's the beautiful thing about humans.
But if there is something to address, address the pain points specifically and directly.
Don't try to guess a "correct" ratio and work backwards with affirmative action (denying one race to boost up another).
For all you know the current ratio is the realistic ratio and where we should be.
In reality, the proportion of women in STEM directly matches the proportion of women in STEM majors, which directly matches proportion of women who say they're interested in STEM.
If it doesn't, you have an issue. The only way to know for sure is to measure.
E: To be clear, assuming equally qualified candidates, you would want expect hiring proportions to match the same A:B ratio as you get from applications.
No, I wouldn't expect this at all as I do not expect the height distribution of basketball player to reflect the height distribution of the general population. As I do not expect the personality trait distribution of pop stars to be the same of programmers, and again I do not expect the latter to be similar to the general population. Personal inclination, innate intelligence, talent, conscientiousness, and of course demographics, parenting, generational wealth all play a role.
Skin color is not an advantage when programming. It is also a characteristic by which it is illegal to hire a programmer.
That can be measured (imperfectly, but well enough), by filtering only for qualified candidates and then comparing the rate at which both A and B are hired and the rate at which they appear in the filtered list. E: This of course requires the filtering to be done _only_ with knowledge of a candidates skills/accomplishments, and association with demographics (including name, location, etc.) to be done only after the sorting.