The recent attention to a forthcoming paper ‘replicating’ the Oreskes procedure of 2004, [ostensibly contradicting it, but there are some problems...], and the timely return of Bray & von Storch, shows that there is still interest in the question of whether there is a scientific ‘consensus’ on climate change.
A couple of days ago, I suggested that it was useful to know what scientists think as a group because it adds to public confidence in the results and their implications. I am now less sure of this argument; speaking with rational non-scientists over the weekend has led me to believe that this is not a fair assumption to make: the problem of trust between the public and science may be much deeper than this.
If this is the case, then a ‘unanimous’ scientific voice is of lesser importance in the public debate. It may sway opinion indirectly, or serve to move the frame of debate somewhat, but it cannot change the underlying uncertainty which besets many people when confronted with unpleasant or uncomfortable scientific discoveries or ideas.
That being said, there is still value to assessing the degree to which climate scientists’ views on climate change represent a ‘consensus’.
The first observation is that, in one form at least, consensus already exists. Inasmuch as each of the several reports by major bodies over the past decade represent the product on a consensus decision-making process of some kind, these products are, ipso facto, the consensus of the group, panel, committee, or the participants.
It is known, though, that there are some [number not specified] climate scientists who have not been happy with the consensus as it stands. Some of these individual have published their objections or misgivings. What is also evident is that there are a number of groups or individuals who, for whatever motive, have a desire to discredit one or more [preferably all] of these examples. One of the methods for doing this is to provide evidence that, in spite of appearances, the product was not a true consensus, but some form of compromise, which did not resolve the differences between all the participants prior to being published.
Another important concern is that products such as the AR4 may be sufficiently compromised in terms of the consensus [or the way in which the consensus is reached] , as to render any claim that it represents a fair summary of current opinion as debatable.
Aligned to these worries are the informal observations which point out an ‘unease’ or ‘concern’ about the procedures or products of large reports, which are one part of a more difficult challenge to the authority of the reports, that to some extent it may be corrupted by feelings of force majeure, insecurity, hijacking or similar pressures. These may be subtle, imaginary, or real but insidious, if they exist at all. But if some of the participants feel as if they are under this kind of pressure, real or not doesn’t matter; the effect is to render the consensus less certain.
Given this, it then becomes meaningful to ask whether or not the reports, or the more general ‘feeling’ about climate change (or any issue, for that matter), are genuinely representative of a sufficiently strong majority agreement, such that they can be called a consensus at all.
This matters to policy makers in particular. If the reply to their questions about what, how much and when, relative to climate change , are framed in terms such as ; ‘well, a good bunch of us are happy to say this, but a few of us aren’t so sure…’, or ‘most scientists agree that…’ or, at the extremes; ‘the science is settled’ or ‘there is no agreement about any of it’, then it becomes much less likely that any concerted agreement amongst nations (which would have to be both a consensus on action and a commitment to act, plus a trust that each will keep its promise) can be reached in time for action to be valuable or meaningful.
The committees, panels, etc., are not well placed to test the hypothesis that their consensus is compromised by the sort of uncertainties illustrated above. If the ‘published consensus’ is to be accepted as the authority on the issues, then doubts about this should be looked at and, where possible, tested.
In a ideal situation, every climate scientist would be able to express, privately, and without fear of reprisal, their exact and unmodified opinion about the science of climate change, and about the apparent consensus. This information could then be collected, analysed and presented. Again, in an ideal case, opinions should be sought on specific matters; key statements by the panel (such as William suggests) or specific key findings, as Eli offers.
If such a test showed that the report represents a ‘true consensus’, then all well and good. If, though, it showed that at least some uncertainty exists which is not expressed in the report, then this represents a limit to the extent to which the authors can claim a high level of agreement.
It would be necessary, in doing this, to establish what would be acceptable as as sufficient number of dissenters (assuming that unanimity is almost never going to be a result) to render the product no longer a ‘true consensus’. This can probably be most fairly agreed upon by using statistical analysis methods, which can assign a order of standard deviations; this way, the test of consensus would be comparable to a test of other findings, and should be acceptable to any scientist.
Being of a known opinion myself, I know what I would like such a test to find; that a sufficient number of scientists are in sufficient agreement with certain statements, such as to state that a consensus exists about these statements. If it turns out, though, that this criterion is not met, this, too is important. It is particularly important to organisations such as the IPCC, as it can inform future practice and lead to ever greater accuracy and transparency in both the process and the product. In either case, something important and useful is learned.
This is unlikely to get much interest, as I seem to have written an essay rather than posted a blog entry, but sometimes the old man gets carried away…