Well, this is one reason to consider normalizing up/down votes by views. Is there any data on the number of views each comment received (number of times each wa

Well, this is one reason to consider normalizing up/down votes by views. Is there any data on the number of views each comment received (number of times each was scrolled into view, for example)? You'd also have to factor in the likelihood of each person up/down voting (if it's viewed by many people who weren't going to vote anyway, then no votes are nothing out of the ordinary).