An innovative sampling of a small group of climate scientists’ perspectives suggests their views may be more commonly shared among their science colleagues than they had thought.

More than two decades after the Intergovernmental Panel on Climate Change (IPCC) began publishing the latest scientific consensus on the globe’s changing climate, widespread doubts persist in the U.S. over whether there really is widespread agreement among scientists. It’s the primary argument of those who deny basic scientific foundations of warming.

But new and innovative survey results suggest the consensus among scientists might actually be stronger than the scientists themselves had thought.

The battles to define and debunk scientific consensus over climate change science have been fought for years. In 2004, University of California San Diego science historian Naomi Oreskes wrote about a broad consensus she found after studying 928 scientific papers published between 1993 and 2003.

Meanwhile, the blow-up over climate researchers’ hacked e-mails in 2009 fueled speculation among skeptics that “consensus” actually is the closely guarded creation of a small cabal of scientists determined to silence opposing views, accusations now widely dismissed as unsubstantiated. That perspective has been largely debunked, but the beat goes on.

On the heels of a January 26 skeptics letter (“No Need to Panic About Global Warming“) in The Wall Street Journal, there have been several follow-up commentaries. They include a vigorous rebuttal on March 22 in the New York Review of Books by Yale University economist William D. Nordhaus; a follow-up response in the April 26 edition of same journal by climate change skeptics Roger W. Cohen, William Happer and Richard Lindzen; and a second response by Nordhaus.

Now, from Carnegie Mellon University and the University of Minnesota Institute on the Environment, comes a fresh study on the question of scientific consensus. Its findings offer something new: scientists appear actually to underestimate the extent to which they, as a group, agree on key questions related to climate change science.

In sum, the newly released poll results identified surprisingly common points of agreement among climate scientists; and yet for each point, those scientists underestimated the amount of agreement among their colleagues. The results:

• Human activity has been the primary cause of increases in global average air and ocean temperatures in the last 250 years. (About 90 percent of respondents agreed with this characterization, but those respondents estimated that less than 80 percent of their scientist colleagues held that view.)
• If governmental policies do not change, the CO₂ concentration in the atmosphere will exceed 550 parts per million between 2050 and 2059. (More than 30 percent agreed, but those respondents estimated that just over 20 percent of their peers held that view.)
• If and when atmospheric CO₂ concentrations reach 550 ppm, the increase in global average surface temperature relative to the year 2000 will be 2-3 degrees Celsius, or 3.2-4.8 F. (More than 40 percent agreed, but those respondents estimated that less than 30 percent held that view.)
• If governmental policies do not change, in the year 2050, the increase in global average surface temperature relative to the year 2000 will be 1.5-2 degrees Celsius, or 2.4-3.2 F). (More than 35 percent agreed, but those respondents estimated that just over 30 percent held that view.)
• The likelihood that global average sea level will rise more during this century than the highest level given in the 2007 assessment of the IPCC (0.59 meters, 23.2 inches) is more than 90 percent. (More than 30 percent agreed, but those respondents estimated that less than 20 percent held that view.)
• Since 1851, the U.S. has experienced an average of six major hurricane landfalls (> 111 mph) per decade. The total number of major hurricane landfalls in the U.S. from 2011-2020 will be seven to eight. (Nearly 60 percent agreed, but those respondents estimated that just over 30 percent held that view.)
• The total number of major hurricane landfalls in the U.S. from 2041 to 2050 will be seven to eight. (About 35 percent agreed, but those respondents estimated that less than 30 percent held that view.)
• Given increasing levels of human activity, the concentration of CO₂ in the atmosphere can be kept below 550 ppm with current technology — but only with changes in government policy. (Nearly 70 percent agreed, but those respondents estimated that just over 50 percent held that view.)

Vision Prize:  What Do You Think Your Colleagues Think?

The specialized poll of scientists earlier this year is a project involving research director Peter Kriss, a doctoral candidate in the Department of Social and Decision Sciences at Carnegie Mellon in Pittsburgh. The poll concluded on March 31.

Dubbed the “Vision Prize,” the method of questioning is designed to poll experts on a range of scientific topics relevant to decisionmakers and financial markets. It asks scientists about their own views on a range of subjects, and asks them to also predict the views of their colleagues on those same subjects.

The poll was publicized on the listserves of various climate research centers, on the Real Climate blog, and by word of mouth — and 172 scientists participated, said Kriss. More than three-quarters of the participants identified themselves as working in academia, with the remainder working in industry, government and non-governmental organizations. Nearly half of all participants in the poll were academicians working in the earth sciences.

Questions for the poll were developed after consulting the IPCC Summary for Policymakers and various climate change blogs, Kriss said. An exchange between Kriss and Real Climate contributor Eric Steig on how to frame one of the questions led to the re-wording of the question on sea level rise, for example (see post #10 on January 23).

It is evident that some participants on the Real Climate blog had problems with various aspects of the poll — mostly with wording of the questions. Those can be found in the comment section of the January 22 announcement of the poll.

The Vision Prize poll used a new polling method designed to elicit carefully considered answers, particularly in regard to the views of others. MIT psychologist Drazen Prelec in 2004 described the polling method in the journal Science.

The method, as described in Science, assigns high scores not to the most common answers but to the answers that are more common than collectively predicted by the group. The scoring adjustment is designed to remove all bias favoring consensus.

A 2004 story on the polling method in New Scientist, titled “Mathematical ‘truth serum’ promotes honesty,” explained that the method encourages truthful responses by asking people questions in pairs.

The first question, suggests the New Scientist story, might be something like, “Will you vote in the next presidential election?” or “Have you had more than 20 sexual partners in the last year?” Then, the second question asks the respondent to estimate how many other respondents would answer similarly.

“It is this perception of what other people’s answer might be which gives hints as to whether the person is telling the truth — especially when their answer is the unusual or unpopular option,” according to the New Scientist story.

“Prelec says if people truly hold a particular opinion, they tend to give higher estimates that other people share it,” the story continued. “So if someone did have more than 20 recent sexual partners — but lied about it — that person would probably assume a higher rate of such behavior in general than someone who had not had so many partners.”

Scientists Predisposed to Assume Disagreements?

While the poll revealed consensus among scientists on key questions related to climate science, the question remains: Why do they believe — incorrectly — that consensus on those questions is lacking?

“It’s not just that they’re inaccurate,” Kriss said of the respondents’ assumption that there is not broader agreement on climate science. “They’re systematically underestimating the extent to which they agree. That’s the part that I found most surprising.”

Research going back decades shows that people generally overestimate the extent to which people agree with them. So, what’s going on with climate scientists?

The poll didn’t address that question, Kriss said. But he speculated that scientists may simply expect their views to be challenged. After all, science advances through vigorous questioning and debate. Perhaps that ingrained view shapes scientists’ assumptions about how much their colleagues agree — or more to the point, don’t agree — with them.

Kriss said he wanted to examine the views of scientists on climate change not only to test the polling method developed by Prelec, but also to inform a major policy topic.

“Our biggest concern is that people may underestimate the amount of agreement among climate experts … and we think that may hinder effective decisionmaking — that if people perceive disagreement, that would be a reason to not act,” Kriss said. “If that perception is inaccurate, it would be nice to correct it.”

As an academic studying behavioral science, Kriss added that the climate issue is an interesting case study of how groups form collective opinions and act or don’t act based on those views. Research shows that it’s a challenge for people to work together toward better outcomes, and Kriss said he hopes to help organizations improve decisionmaking.

“In some sense, it’s the mother of all collective action problems,” Kriss said of the climate challenge. “There’s uncertainty, there’s a long time scale, and there are all sorts of factors that make resolving it difficult. This seems like the biggest real world problem where I thought I could make some contribution.”