I feel like it’s usually a poor understanding of epistemology wrt the conditions of data being gathered (e.g. not understanding how overpolicing warps arrest rates, the difference between conviction and actual guilt, etc.). Could you explain your thinking?