The climate is changing, and we are the primary cause. These are simple facts that are supported by a vast body of evidence and agreed upon by virtually allcr experts. Nevertheless, many people continue to think that the science isn’t “settled” and there is widespread disagreement among experts. Unfortunately, these myths have been propagated and supported by very active misinformation campaigns, so I want to take a few minutes to explain why they are incorrect. First I will explain what we mean when we say that a topic is “settled” or that there is a “consensus,” then I will demonstrate that such a consensus exists for the topic of anthropogenic climate change.
First, I need to explain what I mean by “settled science,” because there are many people who argue adamantly that science is never “settled” because it is always possible that some future discovery will overturn the current thinking. That is technically true, but it can be misleading and requires clarification.
It is true that science, by its very nature, does not provide “proof.” Rather, science shows us what is most likely true given the current evidence. So to that extent, it is true that science is never 100% “settled,” because it is always technically possible there is something we have missed. However, there is a huge difference between a technical possibility and practical doubt. For example, it is technically possible that we are wrong about smoking causing cancer. It is technically possible that all of the countless studies on smoking and cancer are wrong and smoking is actually safe or even beneficial. Further, you can even find a handful of doctors that argue that we are wrong about smoking causing cancer. Does that mean that the science isn’t “settled” or that there is serious debate on the topic? Of course not! The topic has been so well studied so many times by so many people that the odds that we are wrong are insanely low. They are so low that for all intents and purposes, we can treat them as if they are zero. The notion that smoking causes cancer is “settled” in the sense that it is supported by such a massive and consistent body of evidence that it is extraordinarily unlikely that it is wrong, and we must act as if it is correct until such time as compelling evidence arises to the contrary.
This is true for a very large number of scientific topics. There are many things that have been so thoroughly studied that they are as close to “settled” as science can possibly come, and it does not make sense to talk about them as if there is any practical doubt. It would be absurd, for example, for a politician to say, “science is never settled, therefore we can’t really be sure that smoking causes cancer.” The link between smoking and cancer is “settled” in the sense that it is supported by such a vast body of evidence that it is extraordinarily unlikely that we are wrong about it. As I will demonstrate, the same is true for climate change.
Note: please read to the end before arguing that there used to be consensus that smoking was safe. That is a myth and I deal with it below.
What is a scientific consensus?
There are really two different levels at which we can talk about a consensus, and this can become confusing because most people are bad at specifying the level at which they are talking (I have been guilty of this myself). At one level, there is a consensus of experts. In other words, this exists when the vast majority of experts agree on something. This is what most people think of when they think of a scientific consensus, but it is not actually the best level to look at.
You see, when we say something like, “this is a fact” or “the science is settled,” we aren’t basing that on a consensus of experts, but rather a consensus of evidence (i.e., a large body of studies that all agree with and support each other). The consensus of experts is a secondary by-product of the consistent body of evidence. This is really the level we need to look at when asking questions like, “is there any serious debate on topic X.” Science is not a democracy. It is about evidence, not authority. So simply finding some people with advanced degrees who disagree with X does not mean that there is serious scientific debate about the topic. Rather, if there is serious debate, it will be reflected in the peer-reviewed literature, because people will be publishing papers presenting evidence that X is not correct. So that is really the level we should focus on when we talk about a scientific consensus: the evidence, not the experts.
Having said that, there is value in having a consensus of experts when it comes to the general public. No one can be an expert on everything, so even though we would ideally always look for a consensus of evidence, there are many topics on which a given individual simply is not equipped to do that (this is true for everyone, myself included). So, when encountering those topics, it makes sense to look for a consensus of experts, because, on average, an expert will know more about the topic of their expertise than a layperson will, and a consensus of experts usually reflects a consensus of evidence.
I do want to pause here for a second to emphasize the “on average” bit of my last statement. On pretty much any topic, you can cherry-pick an expert who holds an extreme position. You can find doctors who think HIV doesn’t cause AIDs, immunologist who think vaccines are dangerous, etc. That does not mean that there is not a consensus of experts on those topics. You won’t find 100% agreement among experts on just about any topic. So that is not the standard by which we assess a consensus of experts. Further, the fact that you found an expert who agrees with you absolutely does not mean that your position is legitimate or scientifically valid. It is always possible to cherry-pick experts who agree with you, and it is imperative that you avoid falling into that trap. If you go to 100 doctors and all but one of them says you have cancer, you shouldn’t trust the one who disagrees and proclaim that doctors just aren’t sure about your diagnosis. It is intuitively obvious that you should listen to the 99 who said you have cancer.
See this post for the difference between deferring to experts and appealing to authority.
See this post for more information on why you should avoid cherry-picking experts, including discussions of some of the handful of climatologists who deny anthropogenic climate change.
The consensus on climate change
With the semantics now out of the way, let’s look at climate change. The most famous (or infamous depending on your point of view) study looking at the consensus on climate change is the Cook et al. 2013 study that produced the 97% statistic that we have no doubt all heard. This paper has been widely criticized by the good people of the internet, mostly for invalid reasons. Nevertheless, there are some issues that are worth talking about (I have talked about it at length here, so I’ll be brief).
In short, this study did not actually look at a consensus of experts, but rather looked at a consensus of evidence (studies), from which a consensus of experts (authors) was inferred. It took 11,944 papers (by 29,083 authors) and scored them based on whether or not their abstracts stated agreement, uncertainty, or disagreement with the notion that humans are causing climate change. They found that of the papers that made an explicit statement about humans causing climate change (agree, uncertain, or disagree), 97% agreed that we are causing climate change.
That seems like pretty straight forward and compelling evidence of a strong consensus of evidence, so why the hoopla over this study? It mostly comes from the large number of papers that did not explicitly express a view one way or the other. 62.5% did not express an opinion, so the 97% agreement figure comes only from the subset of papers that did express an opinion, and that fact has drawn criticism from both sides.
On one hand, climate change deniers often erroneously claim that Cook et al. “threw out” nearly two thirds of the studies, and actually only 36.9% of the papers agreed that we are causing climate change. This is a faulty argument because it tries to treat papers that did not express a view as if they expressed uncertainty or disagreement, which is clearly false. If they did not express a view, then we cannot draw any conclusions about whether or not the authors agree with anthropogenic climate change, therefore they must be removed before calculating percentages. If we translate this into a more standard survey, if you sent a survey to 10,000 people asking if they thought smoking causes cancer, and 1% said it does not, 49% said it does, and 50% didn’t respond, you would not conclude that only 49% of people think smoking causes cancer, because that would assume that the 50% who didn’t respond are all either uncertain or think that smoking does not causes cancer, which is clearly absurd. It’s the same thing with the Cook et al. study. You can’t conflate “did not express an opinion” with “does not have an opinion” or “disagrees with the consensus,” yet that is exactly what this criticism does.
Further, a lack of explicit statement is precisely what we’d expect on a topic that has reached a consensus. Most studies on vaccines, for example don’t include a statement of “vaccines are safe” in their abstract, because that is so well-established that there is no need to explicitly state it (abstracts have tight word limits). Nevertheless, by the criteria used in this study, we would relegate those studies to the “did not express a view” category, even though the authors almost certainly think that vaccines are safe. Indeed, Cook et al. actually found that the number of explicit statements decreased over time, which is exactly what we expect from a growing consensus.
This does, however, lead to the other criticism. Namely, that Cook et al. actually severely underestimated the consensus because most of the papers in the “did not express a view” category probably were by people who do actually accept anthropogenic climate change. This is actually a fair criticism (though it is often stated too forcefully and unnecessarily denigrates the work of Cook et al.). An awful lot of those studies were on the impacts of climate change, and it is quite a stretch to think that most of those authors disagree with fundamental scientific facts about climate change. Indeed, many of those papers were by authors who are known to agree with anthropogenic climate change, and some of their other papers were categorized into the “accepts the consensus” category. So, it does seem extremely likely that Cook et al. underestimated the consensus.
To solve this problem, James L Powell took a different approach. He argued (I think correctly) that if there is actually disagreement on a topic, that disagreement should be prominent in the literature. It should be easy to find papers that explicitly reject anthropogenic climate change, whereas if there is a consensus, most papers simply won’t make an explicit statement one way or the other (just as they don’t for most “settled” topics). Therefore, if we want to look for a consensus, we should simply count the number of papers that explicitly reject anthropogenic climate change. As an example, he cited 500 recent studies on plate tectonics, none of which either explicitly endorsed or rejected the theory. Based on the Cook et al. criteria, this would erroneously lead to the conclusion of no consensus, whereas based on the criteria of explicit rejection, we would correctly conclude that there is a strong consensus.
When we apply this rejection criterion to the climate change literature, we find almost no studies that argue against the position that humans are causing climate change. For example, Oreskes (2004) reviewed 928 papers published between 1993 and 2003 and failed to find a single one that rejected anthropogenic climate change. Similarly, Powell has looked at this at several time points, always with the conclusion of a very strong consensus. For example, he examined 13,950 articles published from 1991 to November 2012, and only found 24 that rejected anthropogenic global warming. That’s a 99.83% agreement among studies. He later followed that up by looking at the 2,258 climate change papers published from November 2012 to December 2013. This only revealed 1 paper that rejected anthropogenic climate change (a 99.96% consensus). Admittedly, neither of those were published in peer-reviewed journals (but you are welcome to replicate his results), but a subsequent analysis of papers in 2013 to 2014 was peer-reviewed. In it, he examined 24,210 papers by 69,406 authors, and found a grand total of 5 articles published by 4 scientists that rejected the notion of anthropogenic climate change (Powell 2015). That gives us a consensus of evidence (studies) of 99.98%, and a consensus of experts of 99.99%. To put that another way, for every 1 publishing climatologist who disagrees with anthropogenic climate change, there are 9,999 who agree with it. That’s a pretty extraordinary consensus of experts.
Powell also examined the same papers used in Cook et al. 2013, only 24 of which rejected anthropogenic climate change (99.78% agreement; Powell 2016; many of these were the same papers in his non-peer-reviewed analysis). Finally, part way through last year (2019), he examined all of the studies on climate change that had been published so far that year (11,602), and not a single one rejected anthropogenic global warming (Powell 2019).
Additionally, Lynas et al. (2021) took 88,125 papers published between 2012-2020, randomly subset the dataset to 2,718 papers (a representative sample; see note), and examined those abstracts for stated agreement or disagreement with anthropogenic climate change. Of these, only four rejected anthropogenic climate change. That’s a 99.85% consensus of evidence. Similarly, if we apply the Cook et al. criteria, we get a 99.53% consensus of evidence.
These surveys of the literature are extremely compelling evidence that a consensus has been reached and the topic is “settled.” If there was actually serious debate, if actual evidence existed discrediting anthropogenic global warming, we would see that in the literature. We would see numerous studies publishing evidence against anthropogenic climate change, but we don’t see those studies because that evidence doesn’t exist. All of the available data very clearly shows that we are causing climate change. The scientific consensus on this topic is truly overwhelming. Nevertheless, I am sure many people are preparing to fire off responses, so I want to spend the rest of this post preemptively dealing with them.
Note on Lynas et al. They started with a random selection of 3,000 papers, but some were not actually about climate change (database searches for papers often have false positives), so they were removed. Also, it’s worth mentioning that for their sample, they calculated the 99.999% confidence interval to be 99.212%–99.996%. In other words, were are 99.999% sure that the true consensus of evidence (for all papers) is >99% of climate change studies.
“But what about those petitions/letters where thousands of scientists said we aren’t causing climate change?”
There have been many attempts to discredit the consensus of experts by accruing lists of signatures, but if you examine those lists, their fraudulent behavior becomes apparent. Probably the most famous is the “Oregon Petition” which (depending on the source commenting on it) received signatures from 16,000, 30,000, 31,000 or 32,000 scientists. That sounds impressive, until you do even a modicum of fact checking, at which point you’ll realize that this petition is a fraud.
First, there was virtually no verification process. As a result, there were lots of fake signatures, including celebrities and fictional characters. Further, even for the signatures that were real, the only requirement was a B.Sc. in science, which hardly makes someone a scientist, and certainly doesn’t make them an expert on climate change. A huge portion of people who get undergraduate degrees in science never actually use their degrees. Further, even for those who went on to obtain additional degrees and pursue careers in science-related fields, many were experts in totally unrelated fields. For example, how does an orthopedic surgeon, veterinarian, or mechanical engineer qualify as an expert on climate change? When you cut through all the crap, you are left with only 39 people who actually have relevant degrees and expertise in climatology. That is hardly an impressive number and certainly doesn’t discredit the notion of a consensus of experts (more details here, here, and here).
A more recent attempt was a letter to the UN, supposedly signed by 500 scientists, arguing that there is no climate emergency. Once again, however, when you start looking at the signatures, most weren’t even scientists, let alone climatologists. Further, many of them had conflicts of interest, and the claims made in the letter aren’t supported by actual scientific evidence (more details here and here).
There is also another more fundamental issue with these attempts to discredit the science. Namely, they are only about the consensus of experts, not the consensus of evidence. Science is not a democracy, and even if there were hundreds of climatologists who rejected climate change, that would be irrelevant unless they actually had data to back up their position, which they don’t.
“But what about the list of 500 studies showing that climate change isn’t happening/is natural?”
Science deniers love lists of studies. I have, for example, written extensively about the lists anti-vaccers have assembled. The problem is that these lists are inevitably assembled without an actual understanding of the science, and when you look at the papers, they don’t say what the science-deniers think they say. For example, when I went through anti-vaccers’ lists of 160 studies that supposedly showed that vaccines cause autism, I found that 33 of their studies weren’t about autism, 82 weren’t about vaccines, multiple studies explicitly stated that vaccines don’t cause autism, and only 13 were actual human trials arguing that vaccines caused autism (all of which were riddled with problems). The exact same thing has been true of every list of papers I have ever been shown that supposedly discredits anthropogenic climate change. The lists are inevitably filled with papers that talk about regional trends (not global), talk about past climates (without addressing the current warming), talk about how natural climate forcing work (without discussing the current warming), etc. As with the anti-vaccine lists, nearly all of them are misrepresented, most are irrelevant, and many actually argue the opposite of what science-deniers are claiming.
“But what about the few studies that do actually argue against anthropogenic climate change?”
At this point, you might be thinking, “fine, most climatologists agree, and very few studies disagree, but there are a few studies that disagree, and in science, any position can be overthrown by new evidence, so what about those studies?” This is a fair question given two caveats: first, we always have to examine all studies in the context of the broader literature. Given that the context in this case is literally thousands of studies with numerous lines of evidence showing that we are the cause, the evidence in the dissenting studies had better be pretty good.
This brings me to the second point. We always have to critically examine studies rather than assuming that they are valid, and when we do that, we find that these studies used weak designs, shoddy statistics, and are full of problems (Benestad et al. 2016). So they do not in any way discredit the overwhelming mountain of evidence.
“But scientists are just following the ‘dogma’ of their field”
This well-worn trope argues that lots of scientists actually have evidence against anthropogenic climate change, they just don’t publish it because in science it is supposedly forbidden to go against the “dogma” of your field. This is one of those fundamental misunderstandings of science that just will not die. Science is extremely adversarial. We love to prove each other wrong. Further, every scientist who was ever considered great was great precisely because they discredited the views of their day. No one gets anywhere in science by blindly going with the “dogma” of their fields. If anyone actually had compelling evidence that we weren’t causing climate change, they would publish in a high-ranking journal and collect their Nobel Prize. No one has done that precisely because those data don’t exist.
“But scientists have been wrong before”
This is another trope that I have dealt with many times before, so I’ll be brief. First, there are few (if any) examples where modern science has been wrong about something with the same level of evidence that we have for climate change. The “evidence” that was used for things like the sun orbiting the earth is not even remotely comparable to the evidence for climate change.
Second, past mistakes do not automatically negate the evidence for climate change. If it did, then you could use it any time that you wanted to discredit any scientific study. “You think that smoking causes cancer? Well science has been wrong before, so I don’t have to accept that.” See how stupid that is? You need actual evidence to discredit climate science.
Third, this argument is inherently self-contradictory, because it is only through science that we know that previous scientists have been wrong, but based on this argument, we can’t trust science. Therefore, we have no more reason to trust the evidence that the earth moves around the sun than we do for the discredited evidence that the sun moves around the earth. In other words, if the fact that scientists have been wrong before means that we can’t trust scientific discoveries, then we can’t trust the scientific discoveries that were used to show that scientists had been wrong before. It’s a paradox.
Also read this post before arguing that “most scientific studies are actually wrong”
“But there used to be a consensus that smoking was safe”
This is just a special case of the “science has been wrong before” argument. Further, it’s not even true. Tobacco companies certainly ran a good misinformation campaign (much as fossil fuel companies do today), but actual scientific studies have consistently shown that smoking is dangerous. Indeed, scientists suggested that smoking was dangerous way back in the early 1900’s, and essentially all of the research since then (minus a few industry-driven papers) confirmed their suspicions (you can find an overview of this history in Proctor 2012).
“But in the 70s there was a scientific consensus on global cooling”
No, there wasn’t. There was certainly media hype about this, but it was never a prominent scientific position. Indeed, there were a grand total of 7 papers on it, compared to 42 during the same time span that argued that we were causing global warming (Peterson et al. 2008).
“But scientists don’t agree about the extent to which we are causing climate change”
This is a very common tactic among science deniers: taking a minor disagreement and conflating it with a major one. There is some disagreement among analyses about exactly how much we are contributing to climate change, but they all agree that the majority of the change is being caused by us. There is no serious disagreement that we are the primary cause. If there was, this would, once again, be easy to find in the literature, but good luck finding many studies that argue that we are only playing a minor role. They are virtually non-existent.
There are tons of other invalid counterarguments that I’m sure I’ll get assaulted with, but I have already addressed most of them in previous posts so please read them before making inane comments. Also, if you want to more information about why simply looking for papers that reject climate change is a good approach for testing a consensus, read Powell’s papers (cited at the end) as they explain things in much greater detail.
- This post covers most common arguments and counter points.
- This one explains the evidence that makes us so certain that the current warming is not natural and is being caused by us
- This one goes over the evidence that climate change is already having serious consequences
- This one debunks the absurd notion that scientists are just in it for money
- This one talks about Cook et al. 2013 in more detail and discusses other attempts to estimate a consensus
In short, there is an overwhelming consensus that we are causing climate change. This consensus exists both among studies and among scientists. Indeed, recent estimates put it at over 99.9% agreement that we are causing the climate to change. Thousands of studies have confirmed that we are the cause, and virtually none argue that we aren’t. Further, the handful of contrarian studies are riddled with problems and are easily debunked. Every shred of evidence confirms that we are causing climate change, and there is no serious debate among experts.
This level of consensus is important, because it means that there is no valid reason for doubting the reality that we are causing climate change. The level of consistent evidence for it is on par with the evidence for things like smoking causing cancer. Both topics have been extremely thoroughly studied, both topics have a huge and remarkably consistent body of evidence (i.e., a consensus of evidence), and for both topics, that body of evidence has resulted in a nearly unanimous consensus among relevant experts. Nevertheless, on both topics, it is possible to cherry-pick studies and experts that disagree with the consensus, but doing so is folly! As I explained, it is always best to look at the evidence itself, but for most people, that’s not possible, in which case it is rational to simply listen to experts, but why would you choose to listen to the 0.01% of experts who disagree with the consensus? You wouldn’t do that on a topic like smoking causing cancer, so why would you do that with climate change? If 9,999 doctors diagnosed you with cancer and told you to immediately start treatment, but 1 doctor told you that you had nothing to worry about, would you blindly follow that one doctor? I highly doubt it, so why would you do that with climate change? Why would you listen to the 1 scientist saying we aren’t causing it rather than the 9,999 who are saying that we are causing it and need to change our actions?
- 25 myths and bad arguments about climate change
- Anti-vaccers, climate change deniers, and anti-GMO activists are all the same
- Dear Americans, stop using China and India as climate change scapegoats
- Extreme weather: The effects of climate change are already here
- Global warming hasn’t paused
- Global warming isn’t natural, and here’s how we know
- The Rules of Logic Part 6: Appealing to Authority vs. Deferring to Experts
- Yes, there is a strong consensus on climate change
- Benestad et al. 2016. Learning from mistakes in climate research. Theoretical and Applied Climatology 126:699–703.
- Cook et al. 2013. Quantifying the consensus on anthropogenic global warming in the scientific literature. Environmental Research Letters 8:024024
- Lynas et al. 2021. Greater than 99% consensus on human caused climate change in the peer-reviewed scientific literature. Environmental Research Letters 16:114005
- Oreskes. 2004. The scientific consensus on climate change. Science 306: 1686.
- Peterson et al. 2008. The myth of the 1970s global cooling scientific consensus. Bulletin of the American Meteorological Society 89:1325–1337.
- Powell 2015. Climate scientists virtually unanimous: Anthropogenic global warming is true. Bulletin of Science, Technology & Society 35:121-124.
- Powell 2016. The consensus on anthropogenic global warming matters. Bulletin of Science, Technology & Society 36:157-163.
- Powell 2019. Scientists reach 100% consensus on anthropogenic global warming. Bulletin of Science, Technology & Society
- Proctor 2012. The history of the discovery of the cigarette-lung cancer link: evidentiary traditions, corporate denial, global toll. Tobacco Control 21: 87-91