From a comment on Skeptical Science shortly after Legates letter was published in an education journal.
I ran some calculations on the detailed results as released on this site. They show that to obtain a 0.3% "consensus rating", Legates et al had to only count papers rated 1, and then also exclude any papers categorized as "impacts" and "mitigation".
The first step not only excludes every paper that endorses the consensus without explicitly quantifying the contribution of humans, or only implicitly endorses the consensus - it actually counts them and neutral (rating 4) papers as disendorsing the consensus. That follows because they are not rejecting the 32.6% of all abstracts rated as endorsing the consensus in Cook et al, but the 97.1% "among abstracts with AGW position". So, either it is a deliberate strawman by quantifying something they know to belong to a different category (% among all abstracts) or they are tacitly asserting that all abstracts have a position on AGW, and that overwhelmingly that position is a refusal to endorse AGW. Curiously they are willing to assert this without any sign that they themselves have rated the abstracts. They are insisting that their a priori rating is better than Cook et al's empirical rating.
Excluding "impacts" and "mitigation" papers is even more dubious. First, it confuses "endorses" with "is evidence of". A paper about marigolds could "endorse" AGW by simply noting that they think AGW is true. That is not evidence of AGW, and nobody pretends otherwise. It merely indicates the opinion of the authors about AGW (ie, they think it is true). And, of course, Cook et al is not trying to measure the level of evidence, but the distribution of opinions. In fact, it is one of the main arguments of the pseudo-skeptics that a consensus is not evidence, but here they ignore that distinction and pretend that Cook et al by trying to measure consensus is actually trying to measure evidence, the only basis on which excluding "mitigation" papers would be relevant.
It is worse than that, however, for a large portion of "impacts" papers are about the climatological impacts of increasing CO2 levels. They make findings about such things as the likely temperature increase from a doubling of CO2, or from historical and projected CO2 emissions. These are exactly the sort of papers that do provide evidence about whether or not anthropogenic emissions have caused >50% of recent temperature increases. Yet Legates et al want to exclude them as irrelevant (while counting them among "abstracts with [an] AGW position".
The contortion of reasoning involved in their claim is, as you can see, beyond belief.
--Tom Curtis
