Okay, fair enough, that's a good answer.

As long as you're equally skeptical of all such things, then that's a very respectable outlook. I personally tend to think that when people try to match equivalent conditions (obviously imperfectly) and things get above a few percentage points, the tentative presumption should be that there's a real bias unless you can find unaccounted-for mitigating factors, but it's also reasonable to go the other way. What really matters isn't the slogan-level characterization but the acceptance of the result as tentative and a matter for further investigation.

(I have looked at at least one paper on gender-disparity in sentencing and race-disparity, and my recollection is that the methodological issues of the type you mention cannot be ruled out in either.)

It would only be the inconsistency of strongly endorsing one thing as evidence of systemic racism but the other as not evidence of systemic gender bias that would strike me as problematic, because that inconsistency would literally be bias against some group (i.e. a biased view of who gets presumption of bias vs presumption of extenuating circumstances).