But this didn't surprise Airbnb data scientist Dave Holtz. "The ratings in most reputation systems are overly positive," Holtz said Thursday in San Francisco, speaking at Import.io's Data Summit. This is a problem for Airbnb, because its system relies on reputational data in order to function properly. For example, the company doesn't want to keep directing renters to a home with a mildewy bathroom and a shabby mattress as a result of other users' providing inaccurate reviews or none at all. So how does the overly positive bias creep into the review system? It's all about psychology, Holtz explained during his talk entitled, "If You Don't Have Anything Nice to Say, Please Say Something." First, it's important to understand how the old review system worked. Get the full story at Inc.com