Rating Systems

In class last week we were asked to bring in one form of rating website, or wikipedia-esque website. The different examples of rating systems brought in were yelp, tripadviser, Hostels, and Netflix just to name a few. These rating systems brought up an interesting conversation centered around who posts reviews on them. It became clear that most people only post a rating review when they are extremely pleased or displeased with some service or place. This leaves the average or mediocre ratings and experiences out. Some students said that they look at the worst and the best ratings to compare what a business might thrive in versus what they lack. Another opinion is that the best rating system is a 1-5 rating or giving it a certain number of stars. This is so that all of the reviews and ratings and be summed up in one overall average that is easy to interpret.

The other issue when it comes to rating systems and websites is finding a real rating. One cannot discount the fact that anyone can post on most of these with very little regulation, if any. An owner of a restaurant could simply go on and post wonderful reviews about themselves, and no one would know. In addition, the opposite scenario could be true. A competitor could easily go on the website and post something negative about their competition. On Amazon, when reviews are done of the books it could quite possibly be the author's friends and collaborators who are the ones posting reviews, which while maybe not intentional could give a bias review. Also frequent reviewers who post long positive reviews with frequency can be sent free early copies of books, or other things giving them incentive to post a positive review. A young artist, Sam Adams, is said to have bought his own music on itunes thousands of times so that his song could be ranked in the top charts. These examples are just a few ways that these systems can be manipulated.

Professor Lazer brought up the way his profession reviews papers. Two people within the profession are carefully chosen to comment and review a paper with end remarks such as revise and edit, or rewrite and revise, etc. The author does not know who the two people are, therefore no one will feel compelled to slant their reviews one way or another to please a collaborator. I think this is one of the few systems that seems like it would be an effective rating and review of someone's work.