I wonder if we would get better outcomes if the jury were randomly selected. Otherwise based on this link we don’t seem to be getting a jury of our peers, but a jury of selected peers with specific biases not present in the same proportion as the general population.