The first section of this article discusses reviews. Thought it might generate an interesting discussion. As quoted:
"An often neglected point that author James Surowiecki made in his popular 2004 book The Wisdom of Crowds is that a group is “far more likely to come up with a good decision if the people in the group of are independent of each other.” In other words, the crowd has a better chance of being wise when they do not have access to “the same old data everyone is already familiar with.” On Yelp, TripAdvisor, Amazon, and other hives of user-generated ratings and reviews, people are not acting alone. The review that they write will be seen by many other people, and before writing their review, they were probably exposed to the opinions of many other people.
Reading about strangers’ experiences is not necessarily a bad way to predict our own experience. As University of Virginia psychologist Timothy D. Wilson and colleagues note, the process of “surrogation” (learning vicariously through others) can lead to “more accurate forecasts about one’s own enjoyment than receiving a description of that experience.” This is in part because we might discount our own biases in thinking about how much we might enjoy something like a hotel, and in part because we might tend to think our own opinion is more unique than it is.
But others’ opinions can create biases of their own. Massachusetts Institute of Technology management professor Sinan Aral and colleagues have found in experiments, through the mechanism of “social influence bias,” that the presence of a positive review, for example, can inflate the number of subsequent positive reviews. An analysis of Amazon reviews, by contrast, found that later book reviews tended to diverge from earlier book reviews; one problem is that readers’ expectations have been influenced by previous reviews—people began to review other reviewers. And reviewers, of course, tend to be those who are most motivated—those who had the best, and worst, experiences. This helps explain the famous “J-shaped distribution” of Internet reviews: Mostly positive, with a sharp negative tail, and a dip in the middle. Selection bias begins before reviewing, of course; “purchasing bias” implies the people who are more likely to like something were those who purchased it in the first place.
There are other caveats. A property may be popular on TripAdvisor because it is cheap, or because it has a lot of reviews—not necessarily because it is the “best” place to stay. As a study in The Journal of Consumer Research found, comparing reviews of products on Amazon across various categories, the things that got the best reviews rarely converged with the products deemed best in Consumer Reports testing.
Lastly, because of ordering effects, we may simply never find the things that would be our favorites. On a recent trip to a coastal Mexican town, a person pointed me to a low-key fish shack on the beach. It was easily the trip’s most memorable meal. But when I looked for it on review sites, it was buried dozens of places away from the “top” restaurants."
So finally, the author arrives at the same conclusion as many TER veterans:
"Glance at the overall rating, and number of reviews, but don’t wade too deep into the thicket of reviews—you will become quickly confused by the conflict between varying peoples’ expectations and experiences. Ignore one or five star reviews, and focus on the action in the middle, where people are more authentically grappling with how they felt. Look at user-submitted photos more than reviews, so you can draw your own conclusions."
-- Modified on 9/8/2017 4:33:18 PM
