That's a very important distinction that the "easy-mode" metaphors don't really take into account. A good word for it is "intersectionality" (though please note that many people will define intersectionality as only applying to marginalized groups): it's the various ways all of one's "labels" (gender, class, ethnicity, race, etc) interact, and it's something to keep in mind while having these discussions.
The reason is that there isn't one axis along which we can measure "more" or "less" oppressed or disenfranchised. Men tend to lose custody battles more frequently, for instance.
That said, you can still look at a group of people and say something along the lines of "in such-and-such context, members of the first group tend to be more successful at this goal than members of a second group". Phrasing it this way, I think, tends to deflect the "us-vs-them" or "men-vs-women" mentality that crops up fairly often in these discussions (especially when they take place online, where anonymity and lack of non-textual clues make it more difficult to have a civil conversation).
The point of all this is this: if you are a man, you are more likely to kill yourself or be homeless, and that is a HUGE PROBLEM. If you are a black man, by some accounting, you are more likely to be raped than a white woman (because black men are disproportionately sent to prison), and that is a HUGE PROBLEM. But neither of these problems make the discrimination faced by women in the STEM fields less of a problem, it just means that we have many huge problems on our hands.
When you are a member of group A, and group A tends to be more successful than group B (so, male programmers as group A and female programmers as group B), I think it is your obligation to ask _why_ there is this statistical skew, and if the answer is "systematic discrimination due to power imbalance", work to correct it, irrespective of how your group fares in other such comparisons. That doesn't mean that your group faring poorly in these comparisons is not unfair, because it totally is. It just means that there is a complex web of privileges* and disadvantages that we _all_ need to help untangle.
(*You didn't think I was going to get through five paragraphs about identity politics without whipping out the P-word, did you?)