Similar question to what I posed the other person; do you really think you’ve serviced an equal number of low income and high income customers to back up your conclusions?
A bit difficult to make that claim when I imagine you didn’t know each customers background, noted it down, and accounted for how proportinal your compared groups were.