Crowds are Not Always The Answer

There is an article in the Science News site commenting on the putative "crowd" behavior of ratings. They state:

The “wisdom of crowds” has become a mantra of the Internet age. Need to choose a new vacuum cleaner? Check out the reviews on Amazon. Is that restaurant any good? See what Yelp has to say. But a new study suggests that such online scores don’t always reveal the best choice. A massive controlled experiment of Web users finds that such ratings are highly susceptible to irrational “herd behavior”—and that the herd can be manipulated.

 As to Amazon and many other sites it is often not the crow that manipulates but the real manipulators manipulate. Shills are often used to praise a topic so it gets high points or to drive out a bad review if it detracts from the story to be pushed. I have argued and demonstrated that again and again. They are active "politically" driven sites.

The piece concludes:

The “wisdom of crowds” has become a mantra of the Internet age. Need to choose a new vacuum cleaner? Check out the reviews on Amazon. Is that restaurant any good? See what Yelp has to say. But a new study suggests that such online scores don’t always reveal the best choice. A massive controlled experiment of Web users finds that such ratings are highly susceptible to irrational “herd behavior”—and that the herd can be manipulated.

 I argue that all too often it is not this random crowd but deliberate groups pushing an agenda. I have seen several late. The first was the pro-Government controlled Internet crew who pushed a book by some contender for a FCC position. The group wrote dozens of positive reviews while voting down any negative reviews. The second is a review of a book about the HeLa cancer cells. This set of reviews slammed any negative comments as well.

The solution is simple. Always go for the most negative review and work upwards. The solution is that there should be no anonymous reviews at all and the identity of the reviewer and their bona fides should be available. Otherwise any character can praise of slam anything. Is there a crowd effect? Possibly, but many times it is a "group" rather than "crowd" effect.

Now for the purpose of this discussion let's define the two terms:

Crowd: A crowd is a collection of independent but manipulable individuals who have a propensity for commenting on some topic.

Group: A group is a collection of individuals who are bound in some manner by a common world view which they have a need to reaffirm or express to others.

Crowds just like to belong whereas Groups desire to influence and self affirm an a priori world view. 

Crowds as contrast to Groups are malleable and tend to reaffirm an a posteriori consensus.

Now to the MIT researchers. How does one differentiate between a Crowd and a Group. Hint, timing and intensity. Groups have a semblance or organization. They tend to jump on an issue and they tend to be more clustered in intensity and position. In contrast the Crowd is more diffusion like, slowly building and following. 

Given these general characteristics the interesting question is to apply some form of pattern recognition to the responses to ascertain whether the responses are Crowd or Group driven.This has been done in the intelligence world and one would suspect that the Amazons of the world would find this useful, Google certainly would.