Perhaps you're right. I read their methodology and it seems fair. They did solicit demographic info from the participants, but didn't mention if before or after word testing. There's still wiggle room there for bias, but probably not.