Jager said:
An interestng byproduct of internet communications (beyond the ease and swiftness with which people oftentimes get pissed at each other) is the sharp amplification of the perception of problems. This affects pretty much any product - check out an internet forum for whatever your favorite thing is and you'll almost certainly find it. It's natural for us to then want to draw conclusions from what seems surely to be a trend, but from a statistical basis that's almost always wrong. Usually grossly so.
Me thinks there's probably a good subject for a dissertation there...
There's actually a whole field there already, sometimes called "psychoeconomics." You can read a good introduction in some of the later chapters of Peter Bernstein's lively and entertaining book
Against the Gods - The Remarkable Story of Risk.
Among other things, he describes how two Israeli psychologists did a series of experiments to develop what they call Prospect Theory, a model of how predictable biases in perception lead people to evaluate probabilities incorrectly.
One that's particularly applicable in Internet fora is that people consistently tend to overestimate the probability of rare but dramatic events, while underestimating the probability of common but mundane events.
An example described in Bernstein's book was based on a survey of medical students asked to estimate what proportion of deaths were likely to be from natural vs. non-natural causes. Not only did the students overestimate the likelihood of non-natural deaths, but the more detailed the categories they were given (not just natural vs. non-natural, for example, but non-natural via car accidents, household accidents, violent crime, etc.) the
more more they over-estimated the proportion of non-natural deaths.
A speculative photography-related example of this: People reading four or five highly-detailed, dramatic reports of a specific type of defect in the new Hypo-Digithetica D77 camera might estimate that a high percentage of D77s were going to fail because of this defect. Meanwhile, a long-term study of
actual D77 failures might conclude that it was many times more likely for the camera to fail as a result of being dropped, or smashed, or licked by aardvarks, or whatever. But because these types of incidents were not reported in dramatic detail, people would be likely to underestimate their risks, while overestimating the risks of the dramatic failures that actually were less frequent.