{original post unavailable}

You don't skew the group ahead of time; you get a random sample,
some who like it, are neutral, or don't like it. That's the public
opinion measurement biz.

jms