Thursday, June 12, 2014

Informational Cascades and the Bubble of False Causation



Information technology amplifies irrational group behavior
Apr 11, 2013, phys.org

...information technologies can mislead us by magnifying social processes that distort facts and make us act contrary to our own interests.

Informational Cascade
Curiously, an old book entitled Love Letters of Great Men and Women: From the 18th Century to the Present Day, which in 2007 suddenly climbed the Amazon.com bestseller list, provides a good example of group behaviour set in an online context:

"What generated the huge interest in this long forgotten book was a scene in the movie Sex and the City in which the main character Carrie Bradshaw reads a book entitled Love Letters of Great Men – which does not exist. So, when fans of the movie searched for this book, Amazon's search engine suggested Love Letters of Great Men and Women instead, which made a lot of people buy a book they did not want.

...review goes on to talk about group polarization, echo chambers, and radicalization via information selection algorithms in online forums or searches

via
"Infostorms", Pelle G. Hansen, Vincent F. Hendricks, Rasmus K. Rendsvig, Metaphilosophy, Volume 44, Issue 3, pages 301–326, April 2013
DOI: 10.1111/meta.12028

see also:
Margaret Heffernan: Dare to Disagree
ted.com, Aug 2012


Research funding has become prone to bubble formation
Nov 22, 2013, phys.org

...past decade's massive investments in cognitive neuroscience as a potential bubble

"These investments have been preceded by a dramatic rise in fields that attach 'neuro' to some human behaviour or trait with promises that the techniques of neuroscience will explain it – and into game-changing explanations of the human mind," Budtz Pedersen

due to ... traditional scientific incentives such as academic capital and reputation are being replaced by monetary incentives and competition

These structural reforms of research management may amplify social phenomena like "pluralistic ignorance" and "lemming effects" ...

"When researchers choose to ignore their private information and instead mimick the actions of researchers before them, they initialise a so-called lemming effect in which everybody publishes in the same journals and applies for funding for the same type of projects. Such a scientific bubble will eventually bust when the programmes' scientific explanations are put to the test, but the problem is that they may already have drained the research system from resources. And then the system will be faced with an investor confidence crisis."

"Science Bubbles", David Budtz Pedersen, Vincent F. Hendricks. Philosophy & Technology, Nov 2013

No comments:

Post a Comment