That's what I was going to say. Big conclusions based of tiny sample sizes are ridiculous.