Complimentary approaches can make this thread's point clear. So here's an exploration of county-level results🧵 Firstly: We can explain most of the variance in homicides by just using county race shares, and not even getting particularly detailed with them.