This is what the blog has to say about the 97% consensus:
Surely the most suspicious “97 percent” study was conducted in 2013 by Australian scientist John Cook — author of the 2011 book Climate Change Denial: Heads in the Sand and creator of the blog Skeptical Science (subtitle: “Getting skeptical about global warming skepticism.”). In an analysis of 12,000 abstracts, he found “a 97% consensus among papers taking a position on the cause of global warming in the peer-reviewed literature that humans are responsible.” “Among papers taking a position” is a significant qualifier: Only 34 percent of the papers Cook examined expressed any opinion about anthropogenic climate change at all. Since 33 percent appeared to endorse anthropogenic climate change, he divided 33 by 34 and — voilà — 97 percent! When David Legates, a University of Delaware professor who formerly headed the university’s Center for Climatic Research, recreated Cook’s study, he found that “only 41 papers — 0.3 percent of all 11,944 abstracts or 1.0 percent of the 4,014 expressing an opinion, and not 97.1 percent,” endorsed what Cook claimed. Several scientists whose papers were included in Cook’s initial sample also protested that they had been misinterpreted. “Significant questions about anthropogenic influences on climate remain,” Legates concluded.
0.3% is quite a bit different than 97.1 percent, so how did Legates get such a different number when he "recreated" the study?
It turns out that Legates et al didn't recreate the study at all. They simply decided that Cook's rubric was wrong, and picked a different set of papers from Cook's dataset.
Cook et al 2013 divided abstracts into seven types of endorsement:
(1) Explicit endorsement with quantification
(2) Explicit endorsement without quantification
(3) Implicit endorsement
(4a) No position
(4b) Uncertain
(5) Implicit rejection
(6) Explicit rejection without quantification
(7) Explicit rejection with quantification
https://iopscience.iop.org/article/10.1088/1748-9326/8/2/024024
These are the results:
| Rated by volunteers | | Self-rated by authors | |
Position | % of all abstracts | % among abstracts with AGW position (%) | % of all authors | % among authors with AGW position (%) |
Endorse AGW (1, 2, 3) | 32.6% (3896) | 97.1 | 34.8% (10 188) | 98.4 |
No AGW position (4a) | 66.4% (7930) | — | 64.6% (18 930) | — |
Reject AGW (5, 6, 7) | 0.7% (78) | 1.9 | 0.4% (124) | 1.2 |
Uncertain on AGW (4b) | 0.3% (40) | 1.0 | 0.2% (44) | 0.4 |
Legates et al criticism is threefold:
1. Abstracts with no position should be treated as if they are undecided on AGW or disagree with AGW. ("It was only by arbitrarily excluding those 7930 abstracts that expressed no opinion (but retaining forty abstracts expressing uncertainty) that Cook et al. (2013) were able to conclude that 97.1 % endorsed ‘consensus’.") This reduces the percentage from 97.2% to 32.6%.
2. Explicit endorsements of AGW without quantification, and implicit endorsements of AGW, don't count as endorsements of AGW. ("...consensus hypotheses must be expressed quantitatively.") This reduces the percentage from 32.6% to 0.5%.
3. Of 64 abstracts marked as an "explicit endorsement with quantification", they say that only 41 actually qualify. ("However, the authors’ data file shows that they had marked only 64
abstracts (0.5 % of the entire sample) as endorsing the standard definition of consensus. Inspection shows that 23 of these 64 do not, in fact, endorse that definition. Only 41 papers
(0.3 % of the sample) do so.") This reduces the percentage from 0.5% to 0.3%.
Basically, Legates et al are just relying on a ridiculous interpretation of Cook's analysis in order to disregard the overwhelming majority of abstracts that endorse AGW.
[MENTION=341]angelo[/MENTION], do you agree with Legates that Cook should have reported 0.3% instead of 97.2%?