While it's interesting to see which models performed better I wish they did a control. Gave the question to x number of people and had them provide an answer.

bad (or even made up) statistics have been floating around meat space since forever.