When reading studies like this, it's very helpful to look at the tables of regression coefficients rather than just reading text. Unfortunately, it can often be

When reading studies like this, it's very helpful to look at the tables of regression coefficients rather than just reading text. Unfortunately, it can often be a career limiting move for study text to accurately reflect the contents of the data tables/regression coefficients. The first source you cited (the rand study) shows that being African American is a strong predictor of engaging in violence, even after accounting for other factors. See table 2. The second study doesn't address the question. The third study (see table 2) shows that, among other factors, percentage of a city that is black is a strong predictor of violent crime even after accounting for other factors. One productive line of thought which leads from "something inherent to blackness causes crime" is "since this problem is intractible with current levels of biotechnology/social engineering/etc, we should stop wasting resources trying to solve it and focus on other things."