When I was kid, a popular soap brand advertised itself with a slogan similar to the title of this post. I’ve just done an internet search on the subject, and learned that the brand was Ivory, the claimed percentage was 99.44%, and there are multiple stories on where the number comes from. In most versions, the figure comes from a chemical analysis of the product’s ingredients and the .66% remainder is not necessarily an “impurity” in the negative sense (rat droppings, for example), but simply material that does not, strictly speaking, count as soap–fragrance, perhaps.
Anyway, it was an iconic, and, I’m sure, very successful ad campaign, but whenever I saw it I was always distracted by the thought of that .66%. Did I really want to wash with a bar that contained .66% impurities? Why are they boasting about a product that isn’t really completely pure?
Whatever. I have no actual opinion on the brand. It was just an ad campaign.
The point is that the scientific consensus is commonly presented as being even less pure than Ivory Soap: 97% of climate scientists reportedly agree that global warming is caused by humans. Obviously, I need no persuading that anthropogenic climate change is real, but the figure bothers me. It leaves itself too open to well-placed questions–above all, scientists don’t vote on the truth by majority rule, so why is the percent who agree even relevant?
So, let’s take a little time and look at this: where does 97% come from, why isn’t it 100%, and who are those 3% who disagree?
Before we go further, though, I want to point out that, contrary to the posts of certain trolls, climate dissidents do not live in fear of bullying by the establishment majority. I’m sure those people aren’t popular at departmental parties, but it is the scientists who do support the consensus who see their work maligned and ignored, who are personally harassed, and who occasionally receive death threats. Researchers who want the easy way out get out of this field, not into it.
Um, 97% of What?
The figure, 97%, comes from several suitably scientific analyses of science. That is, it’s not a result of a public opinion survey, but rather a series of literature reviews and reviews of reviews conducted by climate scientists themselves. These reviews, conducted over a period of years, present various figures, but most place scientific agreement at somewhere between 90 and 100%. The figures vary because the methods vary–and this is important, because while different questions can yield the same answer, the answers mean different things.
Percent of Papers?
The public statements generally refer to the number of scientists who agree, but in many of these reviews, it wasn’t the scientists who were questioned–it was their papers. The distinction is important for two reasons. Most obviously, the same scientist can write multiple papers. The fact that most papers agree with a certain proposition tells us very little about how many humans agree. It’s simply a separate question.
A more subtle point is that while the number of scientists who agree on something is arguably irrelevant, the number of studies that agree matters very much. Science works on the principle of repeated observation. If I claim that an opossum has gotten into my basement, you might well ask whether anyone else who has gone down there has seen it. We’re not going to vote on the existence of the opossum–it’s either down there or it isn’t, whether the majority agrees or not–but we both know a single observation could be wrong. The light’s bad down there, after all. Maybe I just saw my cat, or some old stuffed animal. But if most people who check my basement also see an opossum (or opossum scat, or opossum hair, or other sign), that is harder to dismiss.
97% of visits to our planet’s basement have involved sightings of the global warming opossum.
Percent of Experts?
Several of the studies did survey individuals, but varied in how they defined the pool of respondents. On one end of the continuum were surveys limited to those who regularly publish peer-reviewed papers on climate. On the other were surveys open to people who work in any science at any educational level. A field tech in geology is a scientist, but does not necessarily know any more about climate than anyone else. Not surprisingly, the percent of respondents who agree with the consensus is higher if the survey is limited to people with the most relevant expertise.
Agree With What? And What is Agreement?
The wording of the surveys varied a lot. Some asked if the respondent believed at least some climate change is caused by humans, others if most is caused by humans. Clearly, for some people, the answers to these questions could be different.
A related issue is that many climatology papers do not state whether anthropogenic climate change is real. An even larger number do not make such a statement in their abstracts. Some studies have counted these as denying climate change, others have simply excluded them from analysis. Arguably, most of these should be counted as supportive of the reality of climate change, since the reason they don’t address the question is that the authors regarded the answer as accepted and obvious. Chemistry papers don’t take the time to note that water is wet, after all. Physicists don’t bother to express a professional opinion on whether gravity is real.
In case it’s not obvious, all of the above came from the two sites I linked to above. Both are worth a read.
So, Who Are the 3%?
I originally set out to write a piece about that three percent–who are they, why do they believe what they do, and are they genuine examples of free disagreement, or are they paid shills?
Unfortunately, I haven’t found an answer, yet. It may be that no one knows the answer–if quantifying the consensus was complicated, qualifying the dissent must be more so. How do you define your sample without including at least some people who aren’t legitimate scientists, and without excluding at least some people who are legitimate but maybe don’t act that way anymore because they disagree with the current consensus and go rogue?
How do you get someone to respond to your survey if doing so could expose them as a paid shill?
I suspect that at least some dissenters are exactly that–paid shills. Others are likely artifacts of the analysis. That is, dissenters who don’t really exist, but only seem to because of how a survey was worded or how papers were coded for analysis. Some may well be both real and genuine. But how many fall into each group and who they are and why they believe thus may be one of those questions science has trouble answering.