Luca and Zervas' study, which is currently being reviewed by a journal for publication, draws a few large conclusions about who cheats and why. They found that restaurants that have a "weak" reputation (few and/or poor reviews) are more likely to boost their presence with fake positive reviews. Those that faced more regional competition might write positive reviews for themselves and leave negative reviews for competitors. Luca and Zervas also found that chain restaurants, which do not benefit greatly from Yelp, are less likely to commit review fraud.

Luca says he was surprised to find that ethics had nothing to do with it.

"This type of decision was influenced by the amount of competition and incentives," he says. "This wasn't just a story of a few rotten apples that were going to leave a bad review no matter what."

The number of posts Yelp has to filter has increased dramatically in its nine-year existence. Luca says its part of an "arms race" between the review sites and scammers.

"People are finding increasing sophisticated ways to get around the algorithms," he says. And, as Zervas told Boston University, as more consumers make decisions with on crowd-sourced reviews, businesses have a higher incentive to cheat.

But those developing the algorithms are also creating new tools that might make fake reviews less relevant in the first place. Luca's next project is working with health inspectors to tap into a restaurant's reviews and predict which ones might have hygiene violations.

If a restaurant's Yelp page is branded with a "pest infestation" alert, it might be hard to distract with the fake review: "Great nachos, fast service."

Copyright 2013 NPR.

fake reviewsfake reviews

Hunger Games: What's Behind Yelp's Fake Reviews?