(This is a slightly edited version of a post I made to the WTA-Talk email list on July 31, 2009.)
James Hughes posted [on the WTA-Talk email list] the results of one particular survey that reported apparently strong agreement on something or other. James especially highlights the figure of 97% agreement. That does indeed sound very impressive. I do think that such a tight consensus among a group of scientists would be something to give considerable epistemic weight to -- at least in the absence of major objections, say from a neighboring discipline. I’m perfectly willing to be persuaded that a consensus on some clear point exists that I currently disagree with. So far, however, I haven’t been given sufficient reason to do so. Let's look at little more closely at this particular survey and my reasons for doubt.
>Two questions were key: Have mean global temperatures risen compared
>to pre-1800s levels, and has human activity been a significant
>factor in changing mean global temperatures?
>About 90 percent of the scientists agreed with the first question and 82 percent the second.
These numbers are lower than the most impressive one of 97%, but still high.
>The strongest consensus on the causes of global warming came from
>climatologists who are active in climate research, with 97 percent
>agreeing humans play a role.
Agreement was lower in certain groups:
>Petroleum geologists and meteorologists were among the biggest
>doubters, with only 47 percent and 64 percent, respectively,
>believing in human involvement.
Why would agreement be higher among the climatologists than among other scientists, including meteorologists and physicists? One plausible answer is that it's because the climatologists can make better judgments. (Although evidence-based forecasting shows that expert forecasts of future changes cannot be trusted with this kind of problem.) Another plausible answer is that groupthink is at work, as it is in so many areas of human activity. This is hardly an arbitrary suggestion, given all the accusations of "denial" and "planetary traitors" and the strong pressures being exerted against skeptics.
Of course there are other surveys, which produce different results. Climatologists are only one group qualified to answer these questions. But l'll set that aside here.
One question that comes to mind is; How were the people to be questioned selected? What percentage of the total does the 3,100 or so represent? From what I've seen, some 10,200 earth scientists were contacted. Only 3,100 replied. Now, these may be representative, or they may not be. Anyone with an academic background in the social sciences, or statistics knows that samples can and often do misrepresent the whole. Given the thousands of scientists who have signed dissenting opinions, I'm not terribly confident that the percentages of respondents in this survey accurately represent the whole group. It seems, for instance, that earth scientists working in private industry were ignored. Given that government-funded scientists may have an incentive (above and beyond the obviously heavy peer-pressure) to agree, the results may not give an accurate picture of all relevant scientists.
These questions come to mind especially because of the highly politicized nature of this discussion. Also, specifically, because of misrepresentations such as seen with the IPCC report, where a small group of people claim to speak for a much larger group. (Compare the summary of the IPCC report to the actual details of the report...)
Other surveys have yielded different percentages. You can see that just from the Wikipedia article.
But, set aside these concerns.
Much more troubling are the questions and the conclusions so quickly drawn from them. Consider the questions. What exactly were those surveyed being asked?
1. "Have mean global temperatures risen compared to pre-1800s levels?"
1800 was around the time that we began to recover more quickly from the Little Ice Age. So what does this tell us? Not much about today or about human activity. It does show that climate scientists agree that the global temperature changes over time. Who is going to disagree with that?
2. Has human activity been a significant factor in changing mean global temperatures?
So, 82% said yes to this. Is this anything to get excited about? Should it impress those of us who are a bit skeptical about warming catastrophe stories? Suppose you are entirely certain that carbon dioxide released by humans is not the cause of global warming. You would still easily grant that global mean temperatures has risen due to the urban heat island effect.
In addition, the question is very vague, certainly if "significant" is taken in the sense of statistical significance (as it presumably is by these scientists). If those climate scientists believed that only 2% or 5% of observed warming could be attributed to human activity, they would still agree with that statement.
How many would still agree if the question was:
-- Do you agree that warming was almost certainly primarily due to human activity? (Not just "significant".)
-- Is global warming principally or quantifiably due to human activity?
-- Are you certain or almost certain that human activity would cause a degree of future warming that constituted a catastrophe?
-- Do you believe that large cuts in carbon dioxide would be effective or cost-effective?
-- Do you believe that the Kyoto Protocol is a sensible solution?
Claiming consensus -- even if entirely justified -- on such vague questions that few skeptics would disagree with is an easy victory that gets us nowhere with any discussion that matters. Once again, dumbing down the issue to a "consensus" of some vague kind isn't useful.
Aside from the foregoing points, I have to say that given the inaccuracy of climate models (as shown comparing them to the past), being impressed by a supposed (or even real) consensus of climate scientists doesn't look too different from relying on a consensus of astronomers. (I would have equally harsh things to say about economists, when they model whole economies...) Granted, that's overstating it. But not by a whole hell of a lot. Again, see my previous post pointing to an audit of the forecasting methodology of the IPCC report, which is considered the gold standard.
I just can't see climate modeling as having attained the status of a hard science at this stage. Even if there was a rock solid consensus on some point of interest (rather than on statements that I have no problem with at all), I would not feel rationally compelled to assent to it as I would, for instance, in the case of a consensus among particle physicists who tell me not to worry about strangelets as they start up the Large Hadron Collider.