Information cascades – present or not?
The idea of “crowd [un]intelligence” arising from information cascades has often been cited as a primary reason against the reality of Global Warming. Either as an elaborate conspiracy among a select minority of scientists, world leaders and policy makers, or simply a consequence of poor academic practices; skeptics often argue that the “evidence” for Global warming has perhaps only been questionably observed by the few initial and preliminary studies, and cascades down the “information pyramid”. With the majority of people being “non-experts” and are thus, incapable of having an independent personal signal that has a high conditional probability of being correct given that they believe it is real, these allegations have been treated by some to be a very real possibility. Furthermore, there is an intrinsic bias even among the experts towards establishing “forced climate change” because doing so would lead to greater funding and cited research, providing further monetary and practical incentives to follow the cascading signals even among the individuals capable of generating reliable independent personal signals.
Similarly, information cascades have also been blamed for a huge variety of myths and believes that are simply factually or logically false. These range from all forms of astrology and fortune telling, to obscure believes that are prevalent in tight social communities, where external independent signals are scarce or deemed implicitly to have low conditional probability of being reliable. The question thus arises, how does one differentiate the above two scenarios? What are the tell-tale signs of an information cascade at play, and when exactly, is a popular opinion not just the cumulation of an irrational crowd behavior, but a reflection of the reality?
I believe the crux of the problem lies in observing the behavior of truncations in the information cascade. When individuals choose to act counter to the previous signal, he must be acting upon a strong personal signal. Though the reason for this strong personal signal may be unbeknownst to the subsequent persons in the cascade, the knowledge that the previous person dissented from the original cascade, is a strong opposing signal, creating a conflict between the now two plausible scenarios. As such the subsequent behavior becomes unclear, and a period of erratic randomization might follow suit. If the alternative view turns out to be well substantiated, the conditional probability that the next person in line will follow given that the person immediately after the dissenter also dissented becomes increasingly large, and an information cascade of the reverse notion would thus ensues. Hence, features and behavior at the truncations, such as reversals in the information cascades, or randomizations are strong indicators that could be studied.
Sources:
http://www.nationalreview.com/planet-gore/14278/cascade-unreason/edward-john-craig#
http://www.timeshighereducation.co.uk/story.asp?storycode=409454