I don't think he's arguing the disparity of the rate, but rather the implied generalization that former-convict = African-American. Look at the statistics posted above, that is not true. I agree, it needs redesign.