Header 1

Our future, our universe, and other weighty topics

Monday, July 3, 2017

Exaggerations Abound When People Talk About a Scientific Consensus

One of the most powerful argumentative techniques in favor of some truth claim is to assert that there is some consensus of opinion among scientists that the claim is correct. But often such assertions are unwarranted. Quite a few of the times that people claim that there is a scientific consensus on something, there is no actual majority of scientists who assert such a thing. Below are six reasons for thinking that quite a few claims of scientific consensus are exaggerated, and are not matched by an actual majority opinion of scientists on the matter in question.

Factor #1: Claims of Consensus Are Often Made Before a Consensus Is Reached

Let's imagine that there's some theory that is starting to get traction in the scientific community. Imagine you are some advocate for the theory, trying to get even more people to accept it. What is your easiest route to such a goal? It is to claim that there is a scientific consensus in support of this theory you support. Many people will meekly fall into line and accept your theory, as soon as they hear a claim that most scientists have accepted the theory. The temptation to claim “most scientists believe this” is so great that people often make such claims even when no such consensus has been reached.

Factor #2: Most People Who Claim a Scientific Consensus Offer No Evidence

The great majority of statements claiming a scientific consensus on something offer no evidence to back up for such a claim. For every time that someone claims that most scientists agree on something, and tries to back up such a claim by referring to some opinion poll or study of scientific opinion, there must be ten or twenty times that someone claims that most scientists agree on something without offering any evidence to back up the claim.

Factor #3: There Typically Exist No Formal Processes for Identifying the Opinions of Scientists on Theories

Given the fact that people are often claiming that most scientists think such-and-such a thing, it is rather surprising to consider that there typically exists no systematic process for having scientists state their opinions on whether particular theories are true. In the world of science, there is nothing equivalent to the voting booth. For example, scientists are not sent annual questionnaires in which they are asked to rate the likely truth of different theories on a scale of 1 to 10, with 10 being certainty about a particular theory.

So when people claim that most scientists believe this or most scientists believe that, and try to back up their claims with some evidence, they may refer to opinion polls or a survey of the scientific literature. These are very imperfect measures of opinion, for reasons discussed below.

Factor #4: People Often Self-Censor Private Opinions Conflicting With Perceived Norms

In November of 2016 there was a startling result in the American presidential election. Donald Trump won a victory in the electoral college, despite losing the popular vote by millions. This was despite both late election polls showing him losing by a substantial margin, and also Election Day exit polls showing him losing in some of the key states he won. A reasonable idea to explain this is the idea of self-censorship. This is the idea that when people hold opinions that differ from perceived norms, they often never publicly state such opinions, and will only express them in something like a secret ballot. It may be that a significant percentage of voters planned to vote for Trump, but told pollsters otherwise, as they regarded their support of Trump as something that conflicted with perceived norms.

We have no idea how much self-censorship plays a role in scientific opinion. Many a scientist may disagree with theories that are supposedly supported by a majority of scientists. Such scientists may engage in self-censorship, figuring that it is not a good career move to speak in opposition to some theory that many other scientists are supporting. This makes it harder to determine just what the majority of scientific opinion really is. 

An example of self-censorship

Factor #5: It Is Very Hard to Unravel the Level of Support for a Theory Based on Scientific Papers

Since scientists have no formal process for voting on the truth of theories, some people have attempted to use studies of scientific papers to draw conclusions about a scientific consensus on some topic. Such attempts can be problematic.

An example of an analysis of scientific papers that offers limited insight is this study, which has been widely although inaccurately summarized as reporting a 97% consensus about anthropogenic global warming. It is probably correct that a majority of scientists do believe that mankind is the main cause of global warming, although the study does not back up the claim of 97% consensus. For one thing, the study was based only on abstracts, those short summaries that appear at the top of a scientific paper. Secondly the study actually reported that 66% of the abstracts reported no opinion about man-made global warming. The 97% figure was from a second phase that sent a questionnaire to those who had already stated an opinion in their abstract about whether humans cause global warming. Of those people, only 14% responded; and of those 14%, 97% supported anthropogenic global warming either explicitly or in a weaker implicit sense. It is not correct to extrapolate from such a fraction of a fraction and make the same 97% claim about the scientific community in general, particularly given the dubious business of getting that 97% by lumping together explicit endorsements of anthropogenic global warming and merely implicit endorsements that may be more nuanced and ambiguous.

Page 15 of this Pew poll of scientists indicates that only 89% of them agreed that earth is warming mostly due to human activity, and that only 77% of them agreed that global warming is a very serious problem. This suggests a consensus about this topic much less than the 97% figure cited (I agree with the 89% on this topic).

It would also be extremely problematic for someone to draw conclusions about a scientific consensus based on an analysis of scientific papers on topics such as cosmic inflation or string theory. Let us consider a physicist who has become familiar with the arcane speculative mathematics of string theory or cosmic inflation theory. Such a physicist learns that he can make a comfortable living grinding out speculative papers offering yet another twist on these theories. But suppose this scientist publishes five papers on such a topic. Does it mean he actually believes the theory is likely to be true? We cannot tell. It could be that the physicist is simply interested in the mathematics, and finds that he can fulfill his yearly quota of scientific papers by writing on the topic. Such a thing does not tell us whether the scientist believes such theories to be true.

Factor #6: Opinion Polls Of Scientists Can be Misleading or Confusing Because of the Way They Are Phrased

Pros in the political field know that the way questions are worded can have gigantic effects on the results. For example, if a question asking about support for abortion is worded from a pro-choice perspective, it will get some answer suggesting a very high support for allowing abortion. The same question worded from a “protect the unborn child” perspective may show a vastly different level of support for allowing abortion.

The same principle holds true in regard to polls of scientists about scientific theories. For example, a Pew opinion poll asked a question of scientists that seemed designed to produce the highest level of response: a question asked whether they agree that “humans and other living things have evolved over time.” That got a 98% yes response. But “evolved over time” could mean small-scale stuff, what is known as microevolution. A scientist believing in small-scale evolution may answer “Yes” to such a question, even though he doesn't believe in the origin of species from more primitive species, or does not believe that such a thing is mainly caused by natural selection. 

Very absurdly, the Pew poll question gave respondents a choice between asserting that "Humans and other living things have evolved over time" and "Have existed in their present form since the beginning of time."  Such a choice forces anyone believing in a 13-billion-year-old universe to choose the first answer, since there is no option such as "Humans originated for unknown reasons about 100,000 years ago, long after the Big Bang." This is a classic pollster's goof: make it seem like almost everyone believes in choice A by offering a choice between choice A and some choice B that almost no one would accept. 

What if these questions were asked:

Is it true that humans have evolved from ape-like ancestors?
Is it true that humans have evolved from ape-like ancestors mainly because of Darwinian natural selection?

These are the questions Pew should be answering, but it doesn't. On page 28 of this full report, it does ask the respondent to choose between the choices shown below:

A poll of scientists (with dubious aspects discussed above and below)

It is interesting that despite constant indoctrination to the contrary, nearly two-thirds of the public reject the claim that humans have evolved over time due to natural processes such as natural selection. It is also interesting here that about 10% of scientists do not believe that evolution occurs mainly because of natural processes such as natural selection.  The survey was made only of American Association for the Advancement of Science members, a subset of scientists more likely to be "old guard" thinkers conforming to ideological orthodoxy.  A full survey of scientists might have yielded a number greater than 10% doubting the "party line" on this topic.

Here we also have a case where there is a large chance of significant self-censorship, as the prevailing academic culture declares deviation from Darwinian orthodoxy as a taboo. The actual percentage of scientists rejecting the Darwinian explanation may be much higher than the 10% indicated in this survey, and could easily be as high as 15% or 20%. The people who responded to the Pew survey were people who responded after being  mailed a letter with the AAAS masthead, signed by the head of the AAAS.  That must have maximized the peer-pressure "fall in line with the majority" effect. A secret ballot without such a "Big Brother is watching" effect might have produced a very different result. 

But the poll still doesn't tell whether there is any consensus about natural selection as an explanation for evolution. The poll asks about “natural processes such as natural selection,” but does not tell us what percentage of scientists are satisfied with the "prevailing party line" claim that natural selection and random mutations can explain the mountainous amounts of biological complexity we observe. Is that percentage 70%? 60%? Or less than 50%? We don't know. Although we sometimes hear claims that almost all scientists believe the idea that Darwinian natural selection explains the origin of species and biological complexity, we don't have polls backing up such a claim. We don't know whether this supposed overwhelming majority is even a 50% majority.

What about fields such as neuroscience? Is it really true that an overwhelming majority of neuroscientists believe that the mind is purely a product of the brain? We don't know, because there is no institutional scientific process for voting on such a thing.


From the discussion above, two general conclusions may be drawn:
  1. When it is claimed that there is a scientific consensus on something, the consensus is often much weaker than is claimed, with a substantial minority rejecting the majority opinion.
  2. Although some claims of a scientific consensus are warranted, it is often claimed that most scientists agree on some topic, when there is actually no clear evidence that such a majority of opinion exists.
Consider also this confusing fact: you may be inclined to believe something if you hear "there is an overwhelming scientific consensus in favor of this," but you may well suspend judgment about such a thing if you hear that "a substantial minority of scientists disagree with the claim."  But both phrases can be used to describe a situation where 90 percent of scientists accept some doctrine, and 10 percent of scientists disagree with that doctrine.

So what are you going to do, when the waters are so muddied in regard to what scientists think? The answer is simple: decide based on facts, logic and evidence, rather than following an “I'll think like most of them think” strategy. Since the insular tribes of academia are often ideological enclaves very much subject to dubious thought customs, inappropriate hero worship, bandwagon effects, sociological influences and groupthink, it's not a good idea to simply follow an “I'll go with the crowd” principle. “Follow the facts” and "follow the logic" are better principles than “follow the crowd.”