Heartland’s president distorts polls, surveys, and studies in support of the Unabomber billboard

Posted on May 16, 2012


Part four of a series.

On May 3, 2012, the president of The Heartland Institute Joseph Bast wrote an essay originally titled “Our Billboards” to accompany the Chicago billboard that inaccurately suggested actual climate realists (those who accept the overwhelming scientific evidence supporting human-driven climate disruption) were the same as the terrorist Ted Kaczynski. The essay, since moved from the website of the Heartland-organized seventh International Climate Change Conference to the main Heartland website and renamed, contained multiple dishonest claims and examples of both Heartland’s and Bast’s hypocrisy. It also contained a great many examples of distortion and deception, both large and small. Three significant examples of this will be addressed in this article, namely the claim that global warming “believers” are a “radical fringe,” that two published climate disruption consensus studies are supposedly meaningless, and that claims of a general scientific consensus on climate disruption are all wrong.

Additional examples will be addressed in future articles in this series.

125 million Americans aren’t a “radical fringe”

In the “Billboards” essay, Bast claims that “60 percent of the general public (in the U.S.) do not believe man-made global warming is a problem.” On the face of it, this statement is perhaps a little odd, but there’s nothing in the essay to say it’s not correct. Which is, in fact, one of the problems with this claim – there is not a link to an opinion poll that provides independent support for this claim, and so it’s impossible to verify.

However, even if the claim is correct, there’s still 40% of Americans who do believe “man-made global warming is a problem.” The US Census Bureau estimates that the population of the US is about 312 million people today. 40% of 312 million is 125 million people. It’s not realistic that so many people could be described as “more than a little nutty” or as a “radical fringe of society,” as “Billboards” does.

S&R tried to identify where Heartland got the data for this claim. It’s not from the latest Gallup poll, which didn’t ask about whether or not global warming was a “problem” at all. Gallup asked “Do you think global warming will pose a serious threat to you or your way of life in your lifetime?” and found that 61% of respondents answered “no.” Gallup also asked “How much do you personally worry about global warming” and found that 55% of Americans worry “a great deal” or a “fair amount.”

The data is not from Pew Research’s latest poll either, which did ask “How serious a problem is global warming?” Pew found that 65% of Americans felt it was either “very serious” or “somewhat serious.” This is exactly opposite of the claim made in “Billboards.”

The data isn’t from Rasmussen either, which found that “64% of likely voters say global warming is at least a somewhat serious problem, including 30% who say it’s Very Serious.” This is again the opposite conclusion from what “Billboards” claims.

Another possible source is the results of the 2011 poll about Global Warming’s Six Americas by the Yale Project on Climate Change Communication and George Mason University’s Center for Climate Change Communication. The poll divided respondents into six basic categories with respect to their views on climate disruption. Those categories are, in order from most to least concerned about climate disruption: Alarmed, Concerned, Cautious, Disengaged, Doubtful, and Dismissive. In May 2011, 39% of Americans were in the Alarmed or Concerned categories, while 61% were in the other four categories. Allowing for rounding, this is a reasonable source for Bast’s claim.

But the Six Americas poll also didn’t ask whether or not respondents felt global warming was “a problem.” The closest the poll gets to that question is one of two questions: “How important is the issue of global warming to you personally?” or “How worried are you about global warming?” In answer to the first question, 8% of respondents said “extremely important,” 15% said “very important,” and 38% said “somewhat important,” for a total of 61% who felt that global warming was important. In answer to the second question, 9% said “very worried” and 43% responded “somewhat worried,” for a total of 52% who were worried about global warming. In either case, what Heartland claims in “Billboards” is wrong, and it’s deceptive to make a ostensibly data-based claim without providing the references for it.

Bast misrepresents two published studies into the scientific consensus on climate disruption

“Billboards” asserts that one poll supposedly “excluded all but 79 (not a typo!) of the thousands of people who responded to it in order to arrive at the 98 percent figure.” This is true, and being skeptical of the result given the small sample size is reasonable. However, it’s important to remember that polling experts regularly estimate the opinions of 312 million Americans with samples of about 1000 people. The Six Americas study polled only 981 American adults and yet had a margin of error of only 3% (at a 95% confidence level). So, contrary to Bast’s assertion, a small sample size doesn’t automatically disqualify the poll’s results.

As was the case with the source for Bast’s public opinion poll claims, however, “Billboards” doesn’t directly refer to the poll. However, this time the essay does link to another of Bast’s essays, “The Myth of the 98%”, which does identify the poll in question: Examining the Scientific Consensus on Climate Change, by Doran & Zimmerman, 2010 (hereafter D&Z2010). Bast’s “98%” essay alleges that D&Z2010‘s “tiny sample size makes it meaningless,” an allegation that is not automatically correct.

In fact, we can test that claim. D&Z2010 polled 3146 scientists, of which about 157 were self-identified climate scientists. Of those, 79 were identified as having published more than 50% of their recent peer-reviewed papers on the subject of climate change.” It was these super-experts that 97.4% of respondents concluded that

human activity is a significant contributing factor in changing mean global temperatures.

Bast claims in his “98%” essay that there are “tens of thousands or even hundreds of thousands of scientists with real expertise in basic sciences related to climate.” This is certainly true, but no-one with expertise in the “basic sciences related to climate” would qualify as a super-expert according to D&Z2010‘s criteria. But let’s assume that Bast is right, and there are actually 100,000 climate super-experts in the world (there aren’t – D&Z2010 identified a population of 10,257 earth scientists in US federal, state, and academic research). For a standard 5% margin of error and a 95% confidence level (ie a 1 in 20 chance that the actual results will be outside the +/- 5% margin of error), D&Z2010 would have needed a sample size of just 39 super-experts to accurately determine what 97.4% of the hypothetical 100,000 super-experts thought. D&Z2010‘s actual sample size was more than double that.

A second way to look at this is to determine what margin of error 79 scientists actually provides. Again, using a 95% confidence level and a response rate of 97.4%, the actual margin of error is only 3.5%. So for a hypothetical population of 100,000 super-experts in climate disruption, there’s a 95% chance that the actual consensus on the question “Do you think human activity is a significant contributing factor in changing mean global temperatures?” is between 93.9 and 100%.

A third way to look at this is to determine what the confidence level is (how likely it is that the “real” result is within the margin of error of the response). Using the standard 5% margin of error, a hypothetical population of 100,000 climate disruption super-experts, and a response rate of 97.4%, the confidence level is 99.45%. That means that there’s only 1 chance in 182 that the real result is outside the range of 92.4 to 100%. Feel free to check all three cases yourself at this online sample size calculator.

However we look at it, D&Z2010 had a large enough sample size to accurately determine that there is an overwhelming consensus on whether human activity is having a significant contributing factor. And this runs contrary to what Bast wrote in the “Billboards” essay and his supporting “98%” essay.

Beyond D&Z2010, however, Bast also misrepresented another published study, namely “Expert credibility in climate change” by Anderegg, Prall, Harold, and Schneider, 2011 (hereafter Anderegg et al 2011). While Anderegg et al 2011 is also not directly referenced in “Billboards,” it is explicitly mentioned in the “98%” essay. Bast’s main allegation in “Billboards” is that the methodology of Anderegg et al 2011 doesn’t correctly account for potential biases in the publication record of climate scientists, pointing to the “98%” essay for proof that “there are many reasons why realists appear to be published less often than alarmists.”

First, Bast claims in the “98%” essay that “publication bias” leads peer-reviewed journals to publish papers that “find something” as opposed to papers that don’t. Unfortunately, he misapplies this bias. He argues that “finding” a causative link between man-made emissions of carbon dioxide and rising global temperatures would be more likely to be published. The problem is that the causative link is so well established at this point that “finding” the link was wrong would be more likely to be published. Yet Anderegg et al 2011 found that scientists who are “unconvinced by the evidence” were published much less often, in opposition to the expected bias.

Second, Bast writes in “98%” that “resume padding” could lead to biasing the results because the “large number of co-authors” on climate papers allegedly “makes objective peer review difficult or impossible.” Put another way, Bast is alleging that the practice of having lots of co-authors makes every climate paper a case of “pal review,” where a paper is given a pass by friendly reviewers. The problem with this claim is that Anderegg et al 2011 controlled specifically for this effect, as they describe in the second-to-last paragraph of their methodology description:

citation analysis research suggests that the potential of [pal review] patterns to influence results is likely to decline as sample size of researchers, possible cliques, and papers analyzed for citations considered increases (22, 25–28). By selecting an expansive sample of 1,372 researchers and focusing our analysis only on the researchers’ four most-cited papers, we have designed our study to minimize the potential influence of [pal review] patterns. [emphasis added]

Simply put, “pal review” bias is a problem with small sample sizes and when using a complete record – by using only the four most highly cited papers from any author, the record was reduced dramatically so padded and pal reviewed papers would be severely limited.

Third, Bast writes in “98%” that age and academic status could be a source of bias too, since most critics of climate disruption already have tenure. It’s true that “publish or perish” is an issue, but the same approach used to counter resume padding and pal review would address this claim as well.

Finally, Bast alleges in “98%” that editorial bias is likely, and he uses the published CRU emails as his proof. The problem is that Bast’s allegation of manipulation of peer review was thoroughly debunked by the Independent Climate Change Email Review (as reported by S&R here)as well as many, many others. In fact, this was identified as a dishonest allegation by S&R in the prior article in this series a few days ago.

For a more detailed response to criticisms of Anderegg et al 2011 by one of co-authors, Stephen Schneider, please check out this link.

Criticisms of the general consensus among scientists also go astray

Beyond misrepresenting two studies of the consensus among climate scientists about the human-driven nature of climate disruption, “Billboards” also misrepresents the more general consensus among scientists as being wrong. Specifically, Bast alleges that

[s]ources used to document this claim invariably fail to do so, while more reliable surveys and examinations of the literature reveal that most scientists do not believe in the key scientific claims upon which global warming alarmism rests.

In support of this allegation, Bast refers to another of his essays titled “You Call This Consensus?”

First, D&Z2010 is mentioned in the “Consensus” essay (point #13), but as with the claims above, Bast misrepresents its conclusions. Bast repeats many of the same sample size arguments he made in “98%,” but the same criticisms of those arguments apply here. D&Z2010 found that the overall consensus among earth scientists was 82% (1.3% margin of error at 95% confidence level, or greater than 99.9999% confidence at 5% margin of error).

Second, Bast mentions a 2010 survey of meteorologists that was published in the Bulletin of the American Meteorological Society which supposedly found that “one in four AMS broadcast meteorologists agrees” that climate disruption is mainly the result of human activity (point #3). S&R reported on this very paper back in 2010, and found that Heartland was distorting the paper’s conclusions. Specifically, S&R found that the survey was not statistically applicable beyond the 121 meteorologists who responded. In addition, S&R found that Heartland was blurring the lines between weather and climate, and that Heartland was ignoring the purpose of the survey – to identify who needed to be educated on what climate science actually says.

Third, in the “Consensus” essay, Bast refers to a Scientific American web survey and says “only 26 percent of Scientific American readers” believed human emissions of carbon dioxide were responsible for global warming (point #7). According to Scientific American, however, the poll was a) not scientific, b) open to anyone on the Internet and not just subscribers, and c) was blog-bombed by Wattsupwiththat readers who overwhelmed the poll. Online polls are automatically self-selected, and so are by definition not applicable beyond the group of people who actually responded. Bast incorrectly identified the web survey “less scientific” when it’s not scientific at all.

Finally, Bast uses the the Petition Project of the Oregon Institute of Science and Medicine to claim that there is no consensus, saying that the Project was “often-mocked but never refuted” (point #5). S&R has analyzed the OISM Petition Project results and found that the Project’s 31,000 supposed scientists represented 0.3% of everyone who met the OISM’s own criteria (anyone with a Bachelor of Science degree in certain fields, many of which have little or nothing to do with climate science, such as metallurgy and medicine). Furthermore, when S&R compared just the scientific disciplines that might provide authentic climate expertise, S&R found the following:

  • The OISM counted only 152 atmospheric scientists, or 2.1% of the American Geophysical Union’s (AGU) members who are atmospheric scientists.
  • The OISM counted only 22 hydrologists, or 0.4% of the AGU hydrology group’s members.
  • The OISM counted only 83 ocean scientists, or 1.2% of the AGU ocean sciences group’s members.
  • The OISM counted only 341 meteorologists, or 2.4% of the AMS’ membership.

No matter how you look at it, this data refutes the OISM’s claims.

When Bast writes of the Project that “there is no comparable survey attesting to a widespread embrace of the alarmist position,” he’s correct. But there’s a reason for that – no-one who seriously researches the consensus among scientists in general and among climate scientists in particular would ever consider allowing such an unreasonably broad swath of Bachelor’s of Science degrees to claim to be a practicing scientist. And no serious consensus researcher would equate the climate expertise of a publishing climate scientist with that of a veterinarian or a metallurgist.

In “Billboards,” Bast misidentifies 125 million or more Americans as a “radical fringe,” distorts the actual results of two published studies into the scientific consensus on climate disruption, and fails to prove his case that “sources used to document [claims of scientific consensus] invariably fail to do so in at least four of his 13 specific examples. Bast has been distorting and misrepresenting studies and polls like those mentioned above since at least 2010, even in the face of multiple corrections from people who know climate science better than he does. This means that there are exactly two possible reasons for his ongoing distortions – a profound ignorance that disqualifies him from making the very claims about climate disruption that he does, or intentional deception.

And Bast’s own writings suggest that he’s not at all ignorant of the particulars of climate disruption, only that he doesn’t like them.

Image Credits:
The Heartland Institute
Yale/George Mason University