Heartland Institute email distorts American Meteorological Society study, admits it’s all about “spin”

Joseph Bast of The Heartland Institute

Joseph Bast of The Heartland Institute

Abstract: The Heartland Institute sent an email that inaccurately reported the results of a study into the scientific consensus about the nature of global warming. The American Meteorological Society objected to the deceptive nature of the email, and so Heartland’s President Joseph Bast defended the email. Instead of accurately reporting the study’s results, both the email and Bast chose instead to distort the study’s findings, quote mine, and ignore inconvenient results in the service of an admitted desire to fool the public into disbelieving that climate change is real, human caused, and likely to be harmful.

On November 26, the Heartland Institute sent a direct marketing email that distorted the results of a study investigating the level and strength of scientific consensus about industrial climate disruption among members of the American Meteorological Society (AMS). In addition to the spam-like tracking features embedded in the email, it also prominently featured the seal of the American Meteorological Society (AMS) and was only identified as coming from Heartland in the footer. Following a public complaint by Keith L. Seitter, the Executive Director of the AMS, Heartland President Joseph Bast published a defense of the email in which Bast claimed that everything in the email was true, that Heartland had done nothing wrong, and more or less told Seitter to quit complaining.

Given Heartland’s long history of deception, dishonesty, and hypocrisy with respect to industrial climate disruption, S&R compared the claims made in the email and by Bast in his defense with the actual study (“Meteorologists’s views about global warming: A survey of American Meteorological Society professional members,” hereafter Stenhouse et al 2013). S&R found that the email and Bast’s blog both fail to accurately describe the results of Stenhouse et al 2013 in multiple ways. Both distort the study’s finding on the scientific consensus among AMS members, both caricature the study’s findings on how political ideology is related to thinking that global warming is happening, the email excises a critical part of a quote and Bast defends the quote mining, and both fail to mention that Stenhouse et al 2013 replicates another study into the scientific consensus.

In addition, Bast’s defense of the email provided a peek behind the curtain and some new insight into Heartland’s long term goal with the email, with a Forbes blog published on November 20 by Heartland Senior Fellow James M. Taylor, and with Heartland’s other industrial climate disruption-denying actions over the past few years – to fool the public into disbelieving that there is an overwhelming scientific consensus that climate is changing, that human activity is the dominant cause of those changes, and that the changes will be disruptive to both global ecosystems and human society.

Heartland sends deceptive marketing email

Screen capture of Heartland email using AMS logo.

Screen capture of Heartland email using AMS logo.

S&R acquired a copy of the email that Heartland sent on November 26. The email is from “2013AMSsurvey@gmail.com” rather than from a Heartland Institute email address. According to Bast’s response, this is “a common practice in email marketing” used to “maximize the open rate.” In other words, it’s common for the sender to use spammer tactics in order to trick recipients into opening the emails they might otherwise leave unread.

The email also used the AMS seal, making the email appear as if it could be from the AMS or that it could be an “AMS approved email” rather than an email from the Heartland Institute. In fact, if the recipient doesn’t read the email’s footer, he or she wouldn’t know that the email had been sent from the Heartland Institute instead of the AMS. It’s this aspect of the email that most disturbed Seitter. Seitter wrote

A disturbing aspect of this e-mail is that it seems some effort was placed in making it appear to have been sent by AMS. It was sent from an e-mail account with AMS in the name (though not from the “ametsoc.org” domain) and featured the AMS logo prominently (used without permission from AMS). Only in the fine print at the bottom was it clear that this apparently came from the Heartland Institute.

Bast’s response to Sietter? That “There was no attempt to deceive recipients about who sent the message,” that using logos in email and on websites without permission is “common practice,” and that the AMS should stop complaining. However, Heartland has a history of impersonating others in order to gain access to information, so it’s possible that Bast is not being entirely truthful.

Bast also wrote that the Heartland email “quotes from the report and provides a link to the document on the AMS Web site.” When S&R looked at the email’s HTML code we discovered that link Bast mentions does not go directly to the AMS website (ametsoc.org). Instead, it goes through a third-party relay site, namely “icptrack.com.” Icptrack.com doesn’t have a website – typing it into a browser window will get a “Forbidden – You don’t have permission to access / on this server.” error. But a “whois” query reveals that icptrack.com is a domain owned by iConnect, the firm that Heartland paid to provide direct email marketing for this email.

This is a form of click tracking, and it permits iConnect (and through them the Heartland Institute) to track the IP addresses of every click on that email. This data enables Heartland to track where the individual clicking on the link works, lives, and how engaged the clicker is with the email (the more engaged the audience, the more clicks). And since every link is unique to the email, Heartland can track which emails were forwarded.

The HTML code also revealed a tracking pixel embedded in the email. It’s a 1×1 pixel that has the same color as the background and it also contains a unique identification number. This number permits iConnect (and through them the Heartland Institute) to track every single instance of that email being read in an HTML-enabled browser. This pixel provides similar information as click tracking does, but it also tells Heartland which emails were forwarded and opened and provides a measure of how much the email is being discussed – every time the email is replied to, the tracking pixel is re-downloaded from icptrack.com. All of this information is collected and put into a database that permits Heartland to tie every single discussion of the email performed in an HTML-enabled (as opposed to plain-text) email program.

The Heartland email likely doesn’t technically qualify as spam according to iConnect’s strict anti-spam policy (although no-one at S&R is a lawyer, so we can’t say for certain). But given all the tracking features concealed within the email, it’s fair to say that Heartland intended to gather a great deal of information about every recipient of this email.

Heartland’s deception by the numbers

The first distortion S&R identified in Heartland’s email was the deceptive claim that the January 2012 AMS survey found “52 percent believe global warming is happening and is mostly human-caused, while 48 percent do not.” What makes this claim deceptive is that it is specifically correct, but ignores a couple of important and relevant facts.

Relevant fact #1: The 52% result is likely too low. The survey used a 150-year period and that likely biased the survey’s results. From Stenhouse et al 2013:

Six respondents sent emails to notify us that their answers would have been different if we had asked about the most recent 50-year time frame rather than the 150-year time frame; the time frame used in the question may have also influenced other respondents. Our results therefore may represent a more conservative estimate of the consensus on global warming than would have been obtained had we asked about a 50-year time frame. (emphasis added)

Bast also fails to mention this important caveat in his defense of the Heartland email. S&R demonstrated previously that Taylor knew about this caveat yet chose not to include any mention of it in his blog, and as such we called Taylor’s failure to mention it “dishonest.” S&R can’t say for certain that Bast has read Stenhouse et al 2013 well enough to understand this issue, so his repetition of the email’s deceptive claim may be ignorant rather than dishonest.

Relevant fact #2: Both the email and Bast fail to mention just how few respondents thought that global warming was either not occurring or thought the causes were mostly natural. Only 5% thought that the causes of global warming were mostly natural, and only 4% of respondents thought that global warming wasn’t happening at all.

In his defense of the email, Bast makes several additional numerical mistakes that, taken in combination, create an inaccurate perception of both Stenhouse et al 2013 and the 2012 survey upon which it’s based – “American Meteorological Society Member Survey on Global Warming: Preliminary Findings” (hereafter the Maibach et al survey). First, Bast incorrectly claims that the Maibach et al survey found that “76 percent of those who believe in man-made global warming also believe it is ‘very harmful’ or ‘somewhat harmful’ (emphasis added)” What the Maibach et al survey actually found was that 76% of those who think global warming is happening, regardless of the cause, also think it will be somewhat or very harmful. Question #4 of the Maibach et al survey asks about the harm/benefit of global warming, but it was only asked of respondents who answered “Yes” to whether or not global warming is happening.

Bast also neglects to mention the other responses to the harm/benefit question. Specifically, an additional 12% of respondents thought global warming would generate both harm and benefit roughly equally while only 2.4% thought that the global warming would be somewhat or very beneficial.

Bast’s more significant error, however, is his calculation that “39.5 percent of all AMS members say they believe man-made global warming is dangerous.” Bast asked readers to check his math, so S&R did.

Bast arrives by his 39.5% number by multiplying the following: 52% of AMS members who say global warming is mostly human-caused multiplied by 76% of AMS members who, according to Bast, say “man-made global warming” is very or somewhat harmful. 0.52 x 0.76 = 0.3952 or 39.5%. Pretty straightforward, except that it’s wrong.

First, if Bast is going to do the calculation, he needs to stick with using just numbers from Stenhouse et al 2013 or just numbers from the Maibach et al survey – he can’t mix and match them. Since Stenhouse et al 2013 doesn’t have calculations for the harm question, Bast should stick to using just the Maibach et al values. In this case, the updated calculation would be 59% x 76% = 44.8%. But even that calculation is still wrong.

To do the calculation properly, one must have a basic understanding of statistics, specifically statistical independence. In this case, the way that the survey questions were asked creates statistical dependence – Question #3 (human activity) was only asked if the answer to Question #1 (global warming is happening) was “yes.” Similarly, Question #4 (harm/benefit) was only asked if the answer to Question #1 was “yes.” But because the relationship of the two questions to each other has not been determined, it’s not possible to know exactly what percentage of AMS members believe that man-made global warming will be harmful without having access to the the raw data. S&R does not have the raw data and, so far as we can tell, neither does the Heartland Institute.

Heartland caricatures the study’s findings about ideology

The second distortion identified by S&R in the Heartland email is dishonest, not merely deceptive. The email incorrectly claims that Stenhouse et al 2013 “found that scientists with professed liberal political views were far more likely to believe global warming is human-caused than others (emphasis added).” What Stenhouse et al 2013 does say is that political ideology is correlated with respondents’ views on global warming. But the correlation is simply with the statement that “global warming is happening,” not with “global warming is happening and it’s human caused” as Heartland’s email claims. All Stenhouse et al 2013 found was that liberals are more likely to acknowledge that global warming is occurring, while conservatives are more likely to deny it.

Not only did Bast repeat the dishonest claim in his defense of the email, he tried to take it one step further. Unfortunately for Bast, that one extra step turned a false claim into an example of equivocation (a logical fallacy where two different words are treated as if they mean the same thing). What Bast wrote is that, since Stenhouse et al 2013 “identifies political ideology as the strongest or second strongest factor in ,em>determining a scientist’s position” on global warming, that “one has to suspect” that “all or just nearly all of the AMS members who believe man-made global warming is dangerous self-identify as being liberals” Bast’s used the word “determining” to allege that liberal politics is “the cause of or reason for” accepting human-caused global warming. But what Stenhouse et al 2013 actually found is that liberal politics was a “predictor” of acceptance of human-caused global warming, not a “cause.” Essentially, Stenhouse et al 2013 said that liberal politics was correlated with acceptance global warming, it’s human causes, and it being harmful. But Stenhouse et al 2013 specifically pointed out that their data could not determine which factor was the cause and which was the effect. It’s theoretically possible that acceptance of human-caused global warming could lead to a more liberal political ideology rather than the other way around (or that it goes both ways), as Bast would have his readers believe.

Furthermore, Stenhouse et al 2013 points out that perception of consensus, perception of conflict, political ideology, and climate expertise – the four factors being analyzed – all combine to explain only 37% of the variation in whether or not global warming is happening and only 29% of the variation in whether or not global warming will be harmful. And their methodology didn’t permit them to estimate how much variation in human-caused vs. natural could be explained.

Simply put, the Heartland email and Bast distorted Stenhouse et al 2013‘s ideology findings in multiple ways.

Heartland’s dishonest quote mining

The most flagrant distortion in the Heartland email is an extended quote of recommendations from the abstract of Stenhouse et al 2013. The following is the complete, unaltered list of recommendations:

We suggest that AMS should: attempt to convey the widespread scientific agreement about climate change; acknowledge and explore the uncomfortable fact that political ideology influences the climate change views of meteorology professionals; refute the idea that those who do hold non-majority views just need to be “educated” about climate change; continue to deal with the conflict among members of the meteorology community. (emphasis added)

The section that is emphasized above was not included in the Heartland email. Given that the purpose of the email was to point out that meteorologists supposedly don’t agree with the scientific consensus behind industrial climate disruption, having a recommendation that mentioned “widespread scientific agreement” about it would have diluted the message. In this case Heartland chose to be dishonest rather than accurately report the Stenhouse et al 2013 recommendations.

Replicating the findings of Doran & Zimmerman 2009

from Doran & Zimmerman 2010

from Doran & Zimmerman 2010

Finally, no-one associated with Heartland has acknowledged what is perhaps the most important finding of Stenhouse et al 2013 – the replication of a study by Peter T. Doran and Maggie Kendall Zimmerman titled “Examining the Scientific Consensus on Climate Change” (hereafter Doran & Zimmerman 2009). Doran & Zimmerman 2009 found that scientists with the greatest climate expertise were the most likely to acknowledge that global warming was happening and that human activity was a significant driver of global warming. Specifically, 90% of all respondents acknowledged that global warming was happening, but 96.2% of climate “super-experts” did, and 82% of all respondents thought human activity was a significant driver of global warming while 97.4% of climate “super-experts” did.

Stenhouse et al 2013 found that 93% of AMS members who could be identified as climate “super-experts” acknowledged that global warming was happening and that human activity was at least somewhat responsible (88% felt that human activity was at least 50% responsible). This compared to 73% of all AMS member respondents who acknowledged that global warming was happening and that human activity was at least somewhat responsible (62% felt that human activity was at least 50% responsible).

In general the percentages for AMS members in Stenhouse et al 2013 are somewhat lower than for the database used in Doran & Zimmerman 2009, but the trend is identical – the more knowledge about climate a particular expert has, the more likely he or she is to accept that global warming is both happening and driven by human activity.

This replication plays a key part in a response to Taylor’s Forbes blog by Stenhouse and his co-authors that was recently published on Climate Science Watch. In it, Stenhouse et al accuse Taylor of ignoring or distorting their findings “for ideological ends” and of creating a “caricature of the results.”

Table 1 from Stenhouse et al 2013 (click to enlarge)

Table 1 from Stenhouse et al 2013 (click to enlarge)

In Bast’s defense of the Heartland email, he specifically points out that the results being distorted in the Heartland email are from Table 1 of Stenhouse et al 2013. But Bast fails to mention that Table 1 was specifically created to illustrate how Stenhouse et al 2013 replicated Doran & Zimmerman 2009. Given that Bast wrote an essay in 2012 titled “The Myth of the 98%” where he claims to disprove Doran & Zimmerman 2009, it’s not surprising that Bast wouldn’t want to admit that Doran & Zimmerman 2009 had been replicated. S&R reviewed Bast’s essay over the course of the Heartland Unabomber billboard debacle and proved that Bast’s criticisms were based on an erroneous understanding of the statistics involved.

A peek behind Heartland’s curtain

The day after Heartland sent out their email, AMS Executive Director Keith L. Seitter wrote a post for the AMS blog in which he recommends that people read the Stenhouse et al 2013 study for themselves rather than relying on Heartland’s deceptive email. Seitter also wrote that he found it “disturbing” that Heartland had apparently tried to make the email “appear to have been sent by AMS.” As mentioned above, Bast in his defense of the email essentially told Seitter that Heartland hadn’t meant to imitate an AMS email and to quit whining about it regardless.

Seitter also wrote in his blog post that the Heartland email did not accurately report the results of Stenhouse et al 2013, suggesting that Heartland was trying to “spin” the results. Bast essentially agreed with Seitter, admitting that the purpose of Heartland’s email was to report Stenhouse et al 2013‘s results far differently than was reported in by the study itself. Bast also admitted that the email’s purpose was to attack the “belief” that there is an overwhelming scientific consensus on global warming.

Since 2004 at least five other independent studies have all looked specifically at whether or not there is a scientific consensus about industrial climate disruption – “The Scientific Consensus on Climate Change” by Naomi Oreskes, “Expert credibility in climate change” by Anderegg et al, the aforementioned Doran & Zimmerman 2009, “The Temporal Structure of Scientific Consensus Formation” by Shwed and Bearman, and “Quantifying the consensus on anthropogenic global warming in the scientific literature” by Cook et al (which also replicated the Oreskes results). While each has faced some amount of criticism, the fact remains that multiple different studies, using different approaches, and performed by independent researchers have all come up with the same result – that scientists in general and climate experts in particular overwhelmingly agree that the Earth is heating up, that human emissions of greenhouse gases are the dominant driver of the heating, and that the changes will be disruptive to global ecosystems and human society. And that’s doesn’t even include the Fifth Assessment Report of the Intergovernmental Panel on Climate Change which found that the warming since 1950 is “unequivocal” and that there is a > 95% probability that the dominant cause for that warming is greenhouse gas emissions.

It’s not often that someone in Bast’s position publicly admits that his organization is trying to convince people to believe something that is essentially untrue. But his admission provides an explanation for why the Heartland email was designed the way it was, for what Bast wrote in his defense of that email, for what Taylor wrote in his Forbes blog, and for the Heartland Institute’s entire approach to climate science.

Given that Heartland’s goal is to fool the public into believing that global warming isn’t happening, using a spoofed sender address on the email makes sense – the more people who open their email, the better. And if people are tricked by the presence of the AMS logo into believing that the AMS either sent or approved of the email, so much the better. And for those people who aren’t fooled by Heartland’s near-spammer tactics, Heartland still has an opportunity to build a database on both their allies and their opponents.

Heartland’s goal of fooling the public also explains the distortions. Of course Heartland wouldn’t mention that Stenhouse et al 2013 identified a possible conservative bias to their results. Of course Heartland would ignore the fact that only 2.4% of AMS members thought global warming would be beneficial, that only 5% of members thought that global warming was mostly natural, and that only 4% thought it wasn’t happening at all. Of course Heartland would caricature the correlation between liberal political ideology and acknowledgment that global warming is happening. Including any of those points would only dilute Heartland’s message and make it harder to fool people.

Heartland’s goal also explains why Bast didn’t mention that Stenhouse et al 2013 replicated Doran & Zimmerman 2009‘s results. Mentioning that replication would have meant admitting that a previous consensus study was right and that would run counter to to the goal.

Just like including the first recommendation from the AMS study – “attempt to convey the widespread scientific agreement about climate change” – would have run counter to Heartland’s goal of fooling people into disbelieving the very real consensus on global warming.

Bast’s admission essentially demonstrates that he and his associates at the Heartland Institute aren’t really interested in facts. They’re not even interested in the truth insofar as they know what is is so they can better distort it. The Heartland Institute’s global warming initiatives exist for purely political reasons – to convince people that industrial climate disruption isn’t real when all the available evidence says it is.

Bast wrote “This is all about ‘spin’ and not, as Seitter says later in his comment, ‘transparency and scientific integrity.'” Indeed.

5 replies »