Just how weak is the case of the anti-science disinformers? In his written testimony for the recent House hearing on climate science, leading science denier Patrick Michaels of the pro-pollution Cato Institute, devoted two pages to the most unscientific ‘evidence’ possible — an online poll.
Michaels, who recently said Big Oil funds some 40% of his work, based a key part of his testimony on the ‘results’ of an online poll by Scientific American that was gamed by the deniers themselves, as SciAm has documented.
Please put two head vises on — from ear to ear and from chin to pate — before reading this excerpt.
Remember, Michaels is trying to make the most persuasive and credible case he possibly can to members of Congress:
Testimony Objective #3: The definition of science as a public good induces certain biases that substantially devalue efforts to synthesize science, such as those undertaken by the IPCC and the U.S. Climate Change Science Program (CCSP).
Visitors to the website of Scientific American have been invited to participate in an ongoing survey on global warming. This survey finds””despite the general environmentalist bent of its readership””that only a tiny minority (16%) agree that the IPCC is “•an effective group of government representatives, scientists, and other experts”–. 84% agree, however, that it is “•a corrupt organization, prone to groupthink, with a political agenda”– (Figure 9). The concordance between the IPCC and the bizarre one-sidedness of the CCSP Synthesis would compel the respondents to say the same about it, if asked.
I’ll spare you the figure, one of two (!) from the poll that Michaels relies on as evidence in support of his “testimony objectives” (see pages 11 and 13 here).
I asked SciAm editor-in-chief, Mariette DiChristina, to comment on what Michaels did. She replied:
The testimony of Patrick J. Michaels demonstrates the climate-skeptic strategy of systematically selecting portions of a number of actual scientific studies, not just this poll, to make an argument that is counter to the prevailing body of scientific evidence about climate change. The use of an Internet poll that was clearly not scientifically conducted””and made no claims to be””is in keeping with that deceptive practice. The portrayal of graphs and other scientific-looking images are known to improve receptivity to arguments by listeners/viewers. I personally deplore such misrepresentations of science and was dismayed to see Scientific American’s good name put to that purpose.More broadly, if this country invested more uniformly in quality science education starting in the youngest grades, its citizens would be better equipped to grapple with complex topics in the face of such obfuscation.
And this isn’t even an isolated instance of the pro-pollution crowd grasping at this particulars straw in an effort to deceive the public. The Wall Street Journal did the same on their op-ed page.
Philip Yam, Online Managing Editor for SciAm, debunked that in a post earlier this week titled, “Do 80 percent of Scientific American subscribers deny global warming? Hardly.”
Readers of the Wall Street Journal may have been surprised by an editorial that appeared Tuesday. We editors at Scientific American certainly were.
“Republican politicians are apparently lower in climate skepticism than readers of Scientific American, which recently discovered to its horror that some 80 percent of its subscribers, mostly American scientists, reject man-made global warming catastrophe fears.”
First, fewer than 10 percent of our subscribers are scientists. Second, the 80 percent climate denial number is not to be believed.
For that 80 percent figure, I’m guessing Gilder relied on a poll that we created for an October 2010 article on Judith Curry. Question number 3 in particular asked visitors, “What is causing climate change?” The poll results show that 77.8 percent responded “natural processes”; only 26.4 percent picked “greenhouse gases from human activity.”
Ignore for the moment that this poll was not scientific (nor was it meant to be) and that it was open to all who have access to the Internet, not just to our subscribers, as Gilder implied.
Rather, the big problem was that the poll was skewed by visitors who clicked over from the well-known climate denier site, Watts Up With That? Run by Anthony Watts, the site created a web page urging users to take the poll.
It sure worked. Our traffic statistics from October 25, when the poll went live, to November 1 (the latest for which we have data on referrals) indicate that 30.5 percent of page views (about 4,000) of the poll came from Watts Up. The next highest referrer at 16 percent was a Canadian blog site smalldeadanimals.com; it consists of an eclectic mix of posts and comments, and if I had to guess, I would say its users leaned toward the climate denier side based on a few comments I saw. Meanwhile, on the other side of the climate debate, Joe Romm’s Climate Progress drove just 2.9 percent and was the third highest referrer.
As an aside, I suppose it is a good sign that CP readers didn’t waste as much time as WUWT readers trying to skew what was transparently an unscientific poll. At the time, I thought a result skewed by the science deniers would be bad, but who could have guessed that the antiscience crowd would embrace the unscientific so wholeheartedly?
So we were horrified alright — -by the co-opting of the poll by Watts Up users, who probably voted along the denier plank. In fact, having just two sites drive nearly half the traffic to the poll assuredly means that the numbers do not reflect the attitudes of Scientific American readers.
I’m not sure what the poll numbers ultimately mean. (The poll also showed that 68 percent think science should be kept out of the political process — when did we officially go back to medieval thinking?) Given how the poll has become meaningless and skewed, I have taken it offline.
We certainly took our lumps from all sides about this online poll, and we learned from the criticisms and will aim to do better next time.
And George, if you must know, in another poll of 21,000 readers we conducted earlier this year, 40 percent of respondents said that over the past year they became “more certain that humans are changing climate”; 46 percent said their views were “unchanged” and only 14 percent were “more doubtful that human activity is affecting the climate.”
Philip Yam, Managing Editor, Online
Former SciAm editor John Rennie had a prescient critique:
And for SciAm to do an online poll about site visitors’ views on a contentious subject like global warming? Sheer folly. Nothing good could come of it. The likelihood that SciAm’s name would be associated with gamed results that nobody really believed but that would be trotted out embarrassingly hereafter would border on a dead certainty.
In my October post, I urged the editors to stop using online polls. In my email interview of DiChristina, I asked her, “In the light of this testimony, and the widespread criticism of Scientific American for using such unscientific polls, including by the former editor in chief, will you state for the record that SciAm will discontinue the practice of using such polls?” She replied:
As I have said publicly [http://www.scientificamerican.com/blog/post.cfm?id=misreading-climate-change-on-scient-2010-10-28], polls are a way to interact with readers, and they are used for that purpose by all consumer media. Scientific American has not claimed that such polls are “scientific” or anything more than a poll of self-selected respondents. In fact, we said they were not scientific. After this experience, we at Scientific American have learned a lot in a hurry about how such polls can be manipulated by groups who have an agenda; a climate-skeptic site encouraged voting and accounted for 30 percent of the total traffic to that poll, as Scientific American reported yesterday [http://www.scientificamerican.com/blog/observations/]. We have removed the poll from the site.
It may interest you to know that another nonscientific poll [http://www.scientificamerican.com/article.cfm?id=in-science-we-trust-poll] conducted earlier this year, with 21,000 respondents, showed very different results: 40% of respondents said they became “more certain that humans were changing climate” in the past year, 46% said their views were “unchanged” and a mere 14% were “more doubtful that human activity is affecting the climate.”
In any case, I don’t think we will make the same mistakes twice with an editorial poll about such a highly charged and important topic. But I cannot say Scientific American editorial will never run any poll ever again. That would be like saying you will never write opinions again in your blog after you’ve found that one of your posts (such as the one that speculated that Scientific American has “jumped the shark” on its acceptance of anthropogenically induced climate change) was, in hindsight, not correct. The nature of blogs is to include opinions, and the nature of magazines and consumer media is to include some level of reader engagement and ongoing conversation with the audience, using a variety of means.
I am honored to have my ‘not correctness’ compared to a mistake by perhaps the most distinguished of the science-popularizing magazines (and I am the third to acknowledge that I need a new phrase for “jump the shark”).
I am delighted that the magazine that has led the way in informing the public of the dangers of human-caused global warming (and the myriad cost-effective solutions) will continue in that vein. One of the things I have been most impressed about SciAm over the years was that it was an early and consistent champion of energy efficiency.
But I retain the right to say I “told you so” to any science-based publication or body that uses an unscientific, easily gameable online poll. As I wrote back in March (!): “Memo to media, science museums, homo ‘sapiens’: Enough with the online polls!” Nothing good can come from them!