I think the amount and frequency of comments should play a role as well. Even if a user doesn't particularly like a link (as evidenced by a down vote) if the article is generating comments and discussion (vs. a preset benchmark) it shows that it's at least interesting and notable in some way.
