Just one more example of the effects of reporting bias in the scientific literature – and another warning to be wary, even about “consensus.”
The journal, Nature, now reviews its own blogs on a web page titled the “From the blogosphere,”a subheading of the “Author” web page., on the homepage of the journal’s website. The “From the blogosphere” heading was on this morning’s “headlines” that were chosen by my Google search page. Unfortunately, the blogs are behind a paywall.
One of the blogs, “Peer-to-peer: for peer-reviewers and about peer review” has a discussion about the controversy over a meta-analysis published in Science magazine by the science historian, Naomi Oreskes, based on this opinion piece – that is available for free – at the UK Guardian, by Jonathan Wolfe. I believe that the point of Mr. Wolfe’s commentary is that non-experts should “shut” our mouths, because we flat don’t know enough.
Neither Mr. Wolfe nor the Nature blogger make any mention that the report by Oreskes was severely flawed and inaccurate. However, as the one comment at “Peer to Peer” reports, “This is very odd. The main critic of Oreskes’ work was Benny Peiser, who is not a blogger or a think-tanker, but a member of faculty at John Moores University and a fellow of the Royal Astronomical Society.”
I hope Mr. Wolfe or the Nature bloggers will follow up on the Oreskes article and the controversy surrounding it, because there appears to be a secondary theme: bias can lead to error, even in peer reviewed, scientific journals.
Oreske’s article, originally published in Science in December, 2004. This review of the scientific literature is often quoted to support the position that there is no disagreement among scientists about whether the earth is warming due to the increase of “greenhouse” gases, and that those greenhouse gases are due to the influence we humans have on our environment. However, the problem appears to be a flaw in both the professor’s methodology and her reporting. She mis-reported her search terms, and those terms – the use of the three words, “global climate change,” rather than “climate change” – make a huge difference.
Here is the “Erratum” published in January, 2005 by Science:
Essays: “The scientific consensus on climate change” by N. Oreskes (3 Dec.2004, p. 1686). The final sentence of the fifth paragraph should read “That hypothesis was tested by analyzing 928 abstracts, published in refereed scientific journals between 1993 and 2003, and listed in the ISI database with the keywords ‘global climate change’ (9).” The keywords used were “global climate change,” not “climate change.”
The choice of search terms seems to make a huge difference. From a web page entitled, “The Letter Science Magazine refused to publish,” by professor of anthropology, Benny Peiser, we learn,
On December 3rd, only days before the start of the 10th Conference of Parties of the United Nations Framework Convention on Climate Change (COP-10), Science Magazine published the results of a study by Naomi Oreskes (1): For the first time, empirical evidence was presented that appeared to show an unanimous, scientific consensus on the anthropogenic causes of recent global warming.
Oreskes claims to have analysed 928 abstracts she found listed on the ISI database using the keywords “climate change”. However, a search on the ISI database using the keywords “climate change” for the years 1993 – 2003 reveals that almost 12,000 papers were published during the decade in question (2). What happened to the countless research papers that show that global temperatures were similar or even higher during the Holocene Climate Optimum and the Medieval Warm Period when atmospheric CO2 levels were much lower than today; that solar variability is a key driver of recent climate change, and that climate modeling is highly uncertain?
These objections were put to Oreskes by science writer David Appell. On 15 December 2004, she admitted that there was indeed a serious mistake in her Science essay. According to Oreskes, her study was not based on the keywords “climate change,” but on “global climate change” (3).
Her use of three keywords instead of two reduced the list of peer reviewed publications by one order of magnitude (on the UK’s ISI databank the keyword search “global climate change” comes up with 1247 documents). Since the results looked questionable, I decided to replicate the Oreskes study.
According to Oreskes, 75% of the 928 abstracts she analysed (i.e. 695) fell into these first three categories, “either explicitly or implicitly accepting the consensus view”. This claim is incorrect on two counts: My analysis shows that only 424 abstracts (or less than a third of the full data set) fall into these three categories.
It also shows that many abstracts on “evaluation of impact” and “mitigation” do not discuss which drivers are key to global climate change, instead often focusing exclusively on the possible effects of elevated CO2 levels on plant growth and vegetation. Many do not include any implicit endorsement of the ‘consensus view’ but simply use certain assumptions as a basis for often hypothetical impact assessments or mitigation strategies.
Quite a number of papers emphasise that natural factors play a major if not the key role in recent climate change (4). My analysis also shows that there are almost three times as many abstracts that are sceptical of the notion of anthropogenic climate change than those that explicitly endorse it (5, 6, 7).
I guess our lesson should be to be skeptical of “consensus,” just as we are becoming skeptical of “peer reviewed” journals that rush to print on “hot” stories about cloning and stem cells.