Ever wonder how there came to be a consensus on anthropogenic global warming in spite of the paucity of evidence? Here's one theory. Follow the link above to read the original article that inspired this version.
The peer-review system for accepting papers to be published and for approving requests for research grants constitutes a feedback mechanism that is intended to drive research, and ultimately scientific consensus, toward conclusions that are supported by objective data. This peer-review system provides a mechanism for filtering out subjectivity that naturally occurs by virtue of the human nature of researchers. It also tends to amplify previous results. This filtering and amplification is accomplished by the clustering of approvals and grants around previously-obtained results, which creates a peak that "automagically" converges on objective reality.
When the data are of sufficient quality and quantity, the output of the system is correct. A lack of data on a particular topic usually results in a corresponding lack of output or conclusion for that topic. However, the system appears to be susceptible to the creation of erroneous output through the injection and amplification of extraneous signals when the data is sparse and/or contains no clear signal. Example: a favorite pet theory of a prominent researcher might "get legs" because the reputation of the researcher constitutes an extraneous signal into the system. This susceptibility constitutes a hazard -- when a political motive exists, the system can be exploited by the deliberate injection of an extraneous signal into the system on a subject for which the data is sparse.
In order for an injected signal to "take", it must be helped along at first. This is accomplished by subtle means that might not be intentional, and which, to succeed, are not detected. One such method might be the improper selection of arcane or complex statistical methods that coax an erroneous result from sparse or noisy data. Regardless of the method, once even a tiny peak appears in the output signal -- in publications and research -- it will, in the absence of strong data to refute it, be amplified over time and eventually become an erroneous consensus.
The greater the inertia any consensus acquires, whether erroneous or correct, the more plentiful and clear the data must be to overcome it. Inertia comprises many components, including the number of publications, the number of researchers in the field (spending), and may even include public perception. At some point the inertia becomes so great that the only refutation necessary for contrary positions is a simple, "There's a consensus."
Case in point: AGW, or the hypothesis that humans are harming the earth's climate to an appreciable extent. The data for establishing or refuting AGW is extremely sparse, both spatially and temporally. The political motive is the accretion of political power and wealth. Thus, the system is susceptible to manipulation through the injection of an extraneous signal, which signal is something on the order of "humans are bad for planet earth".
This political signal was initially amplified by the fact that it aligns with local maxima, figuratively speaking, which are that the physics of the absorption and re-radiation of escaping terrestrial heat by atmospheric CO2 supports the hypothesis, and that human fossil fuel consumption adds CO2 to the atmosphere. The earth's climate system is far more complex than that, but once the hypothesis has been nursed through the feedback control system of peer review a few times, it takes on a life of its own and we eventually get hockey sticks, tipping points, and "There's a consensus".
It is my observation that there are not enough data to support the conclusion that humans are destroying the earth's climate. The consensus that this is so is an erroneous one, and it should be recognized as such.
Scientists need to be cognizant of the metascientific process and how it can be manipulated in order to achieve political ends, and refuse to play the game. However, it is not clear there is any turning back on the subject of AGW without some breakthrough of sufficient magnitude to wipe the slate clean. Any researcher who will not play the game according to the AGW scenario will simply not be allowed to play.
In other words, we have passed the metascientific tipping point on AGW.
