Richard Tol apparently has a new comment about Cook et als. paper Quantifying the consensus on anthropogenic global warming in the scientific literature. The basic motivation is that
Cook et al. (2013) seem to be an outlier in the consensus literature: their results, for a large sample, are in line with other results for small samples but contradict other large samples.
This is despite Richard himself saying
The consensus is of course in the high nineties. No one ever said it was not. We don’t need Cook’s survey to tell us that.
Tol basically looks at a large number of other surveys and tries to argue that Cook et al. is some kind of outlier. Shall we provide some much needed context? For example, Tol quotes a consensus of 75% from Oreskes (2004). What does it actually say? It says
Of all the papers, 75% fell into the first three categories, either explicitly or implicitly accepting the consensus view; 25% dealt with methods or paleoclimate, taking no position on current anthropogenic climate change. Remarkably, none of the papers disagreed with the consensus position.
How is that a 75% consensus?
Tol uses data from Stenhouse et al. (2013), a survey of members of the American Meteorological Society (AMS), which says explicitly
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
So, a survey that was partly motivated by a desire understand why some AMS members do not recognise the widespread scientific agreement, is somehow evidence that there isn’t widespread scientific agreement.
Tol quotes a maximum consensus in Verheggen et al. (2014) of 79%, despite the abstract explicitly stating
90% of respondents with more than 10 climate-related peer-reviewed publications (about half of all respondents), explicitly agreed with anthropogenic greenhouse gases (GHGs) being the dominant driver of recent global warming.
Why the difference? Tol has included all surveyed, even those who responded with “I don’t know”. Well, if I’m trying to survey relevant experts and they don’t know the answer, then they’re probably not relevant experts.
Tol also includes that Verheggen et al. has a subsample with a consensus of 7%. Who were they? They were those “unconvinced of anthropogenic climate change”. Wow, what a surprise. A vast majority of those unconvinced of anthropogenic climate change, are unconvinced of anthropogenic climate change.
What Tol also fails to highlight is that Cook et al. was a survey of the literature itself. Most other large survey were surveys of people, not surveys of the scientific literature directly. Anwyay, that’s probably enough context for now. If you want some more context, you could read this.
...and Then There's Physics 
What’s this obsession with Cook et al—as if I didn’t know? Could it be the fact that they run a very effective website rebutting climate denial memes? Could it be the fact that the 97% consensus is such a powerful argument in countering attempts to swamp climate science in a sea of doubt?
john,
I think that is almost explicitly stated in Tol’s comment
The problem is that the result in Cook et al. is pretty close (as Tol himself acknowledges) to the actual level of consensus. Essentially it is “true”. Of course, it’s going to impact public perception of and the political debate on climate policy, but so will Tol’s attempts to discredit it. The problem, though, is that Cook et al. is a fair representation of the truth, and Tol’s attempts are not. If we’re going to have things influencing public perception and the political debate, I’d much rather they were the former than the latter. Tol appears to disagree.
You bastard! You don’t even credit me for spotting the Oreskes issue. 😛
I really don’t get it why he’s attacking Cook 2013 in this way. It’s nonsensical and Verheggen already pointed out to him on Twitter that what Tol did is not a valid interpretation of their data.
Collin,
Sorry, although it was an obvious thing to check 🙂
That it is nonsensical is no great surprise, as is the fact that what Tol is doing is not a valid interpretation of the data.
“Cook et al. (2013) state that 12,465 abstracts were downloaded from the Web of Science, yet their supporting data show that there were 12,876 abstracts. A later query returned 13,458, only 27 of which were added after Cook ran his query (TOL). The paper is silent on these discrepancies.”
Oh look, the “12,876 vs 12,465”-zombie argument shows up again!
http://blog.hotwhopper.com/2015/03/the-evolution-of-97-conspiracy-theory.html
The original discussions that lead to Sou’s post, and where several people tried to explain some very, very basic stuff to him, here:
http://blog.hotwhopper.com/2015/03/deconstructing-97-self-destructed.html
The man has no shame.
Marco,
Thanks, I was going to point that out too, but decided I’d wasted enough time on this post as it was.
Right, who’s up for a survey of papers mentioning gravity but not explicitly endorsing the inverse square law?
I might actually add what Sou has in her post that Marco links to above
Here is what Tol says in his comment
No, their data does not. It shows that the IDs go up to 12,876 with some gaps. The reason for that is explained above.
When we invent a time machine, we can criticise papers for not reporting something that happened after the data was collected. Until such time, papers should not be expected to comment on things that happen in the future.
Andrew,
That would be quite a nice way to illustrate the idea of a consensus, something Richard is clearly still very confused about.
“I really don’t get it why he’s attacking Cook 2013 in this way”
Simple: because he has so openly attached his reputation to the validity of his criticism. He himself said he could not afford being wrong about Cook et al, because then his reputation would be in tatters. Of course his reputation *is* now in tatters, but he will never admit as much. And therefore he doubles down.
To me the worst part is that ERL will apparently publish this ‘comment’. To call it shoddy quality would be extremely generous. Claiming the Oreskes consensus is 75% is just using an entirely different definition of ‘consensus’ (including abstracts that don’t take a position on AGW) than we used in Cook13 and thus they don’t belong on the same chart and shouldn’t be directly compared.
Tol’s ‘comment’ can basically be boiled down to: ‘by defining “consensus” in a bunch of different ways, most of them totally nonsensical, I can come up with a bunch of different percentages.’ I love that he includes the 7% ‘consensus’ among the sub-sample in Verheggen et al. that were “unconvinced of AGW”. The only surprise there was that the number wasn’t 0%.
I don’t know what the folks are ERL were thinking here. Maybe they got sick of Tol constantly harassing them and finally caved in.
Yes, I can imagine the meeting they had. “If we let him publish this, maybe he’ll stop whining”.
Dana,
Could always point out that Tol has just illustrated that the consensus amongst those “unconvinced of AGW” is only 93%, smaller than the overall consensus with respect to AGW (97%) 🙂
It Brings to mind the drunk who keeps falling over but still keeps wanting to fight, never knowing when to admit defeat …
PLOT SPOILER : He finally admits defeat and puts his hand out in reconciliation (no, not Richard Tol, the Lithuanian drunk in the video!)
The paper raises six issues:
Cook’s estimate is an outlier.
No published test for systematic tests between raters.
No published procedure for guaranteeing rater independence.
No published procedure for guaranteeing no additional information.
Sample size.
No published procedure for guaranteeing that early results did not affect later data collection.
Issues 2-6 are really one: Cook did not publish a survey protocol. These concerns are easily taken away by publishing it now.
The comparison to earlier estimates was news to me, and if this stands up I will have to eat my earlier words.
The exclusion of “don’t knows” is a tricky issue. For instance, Verheggen approached people he thought had expertise. Some of them said they did not know the answer to his question. You may argue that, therefore, Verheggen sampling strategy was wrong; and all his data are rubbish. But if instead you argue that the sample is valid, you have to live with the whole sample.
It would be nice if the authors of the studies Tol cites, at least some of them, write a rebuttal letter to Tol’s comment. That is the sort of “academic” put down that will shut him up. He knows that blog posts decrying his crap do his career no harm.
MIke Pollard: It would be nice if all of them did. Individually.
This incessant addiction to quantifying consensus is pathetic (NB: High School is over). Considering the 3% non-consensus are the Sky Dragon nutters, the 97% is meaningless. What matters is the consensus on filling data gaps, evaluating risk and designing mitigation that can be politically implemented. Like Mosher says, the debate on high-level attribution is over. You consensus mongers keep on fighting while your opponent is out cold. That makes you nutter brothers with the deniers, locked in a dance of gotcha discourse with the passion of a heroin junkie.
… and you are supposed to be the smart guys.
Howard:
” Like Mosher says, the debate on high-level attribution is over.”
No, it’s not. Here in the US, the leadng Republican candidates for President all deny that there is an overwhelming scientific consensus on climate change, and papers evaluating the consensus are claimed to have been “debunked”.
So, politically, it is still very important. Sad, but true.
@Howard :
“Earnest falsehoods left unchallenged risk being accepted as fact.” – Monty Montgomery
Please, do not misinterpret this to mean I think Richard Tol’s falsehoods are honest anymore. That ship has sailed.
> Considering the 3% non-consensus are the Sky Dragon nutters, the 97% is meaningless.
I would not consider Senior as a Sky Dragon nutter.
***
> [T]he debate on high-level attribution is over.
Yet half of the American population has yet to know this.
***
> That makes you nutter brothers with the deniers […]
The Goldilocks framing will never get old. Nor too hot. Nor too cold. Just lukewarm.
***
> What matters is […]
Look, a very important squirrel!
Pingback: Non piove, governo ladro - Ocasapiens - Blog - Repubblica.it
“The Goldilocks framing will never get old. Nor too hot. Nor too cold. Just lukewarm.”
Damn. That has to be the Climateball equivalent of a goal via bicycle kick.
==> “Yet half of the American population has yet to know this.”
Why don’t they know it? How will that change?
From handbag fights in the climate-o-sphere?
No, Joshua – from efforts like this:
http://theconsensusproject.com/
Dan Kahan will take care of everything else after he straightens out the Floridians’ ocean views.
I’m dubious, Willard.
IMO, relatively few people will have their opinions changed by consensus-messaging; the evidence in support of such efforts affecting change is slim, IMO. (1) People are inclined to filter such information so as to reinforce existing beliefs and, (2) People are inclined to identify the messengers in such a way as to reinforce existing beliefs.
IMO, this is maybe 97% about sameosameo/identity politics.
ATTP, yep it is trivial to notice that one. Though it was something I instantly noticed. You just beat me to publicly stating it as I actually have to work.
*runs for his life*
How many places on earth are experiencing unpleasant damning weather? And the nutters are all standing around like Gumby’s saying there’s nothing wrong while is plainly obvious that there is something wrong.
This not only makes them look bad, its makes them look incredibly incredibly stupid.
Dana1981: “Tol’s ‘comment’ can basically be boiled down to: ‘by defining “consensus” in a bunch of different ways, most of them totally nonsensical, I can come up with a bunch of different percentages.’”
Richard Tol focuses on other shortcomings of Cook et al but for me the main one is this: by defining ‘consensus’ in a bunch of different ways, some of them contradictory or overlapping, it came up with a bunch of different percentages that were repackaged to give the expected result.
I’d say that earlier consensus studies played similar games but I wouldn’t have the energy to defend a statement like that.
So I’ll just say that the whole field is skunked, which is self-evidently true. It’s activism, not science. Just step away, people.
Howard, if the world was acting in some linear fashion (must do (1) before (2) before (3)) I might agree with you.
1) get 95% agreement on science amongst politicians and economists, not just scientists, only then
2) develop strategies for mitigation and adaptation, only then
3) implement solutions
But we aren’t acting in a linear fashion.
Society acted on smoking because there was a sufficiency of evidence to act, and unlike with AGW, without even an understanding for the underlying mechanisms [which we do for AGW thanks to Tyndall et al]).
So we have instead parallel activities in train on man-made global warming:
i) still doing the science because we need to refine the science to inform policy, and there are areas of uncertainty (e.g. how fast will Greenland melt) e.g. MetOffice recent report
ii) people are developing strategies and not only COP21, but in many national, state and other areas, including States in the USA.
iii) people are implementing solutions – ok, not fast enough – and some great minds and entrepreneurs at work in a myriad of initiatives (Germany, Musk, etc.)
So agreed, we should not devote excessive time on convincing people like R.T. of the 97%, but I don’t believe that is the case. Some effort is required, because the contrarian strategy is like the smoking industry … hoping to win by repetition of falsehoods and procrastination w.r.t. policy.
Unfortunately, while we have powerful groups and people trying to put off action on man-made global warming, and to some extent succeeding, they cannot be ignored. That does not mean they are winning, because the world is not stuck at (1) – the world is pushing ahead on many fronts (i), (ii) and (iii).
Just not fast enough. So no time to relax, I am afraid.
@Vinny Burgoo
Perhaps you and Richard Tol should get together a group of like-minded people and do your own study to show Cook et al where they went wrong instead of sniping from the peanut gallery. It would have taken far less time and energy than has been expended thus far, and if you are correct, far more damning.
Vinny Burgoo
Misinformation. Just like Tol only different.
“[B]y defining ‘consensus’ in a bunch of different ways, some of them contradictory or overlapping, it came up with a bunch of different percentages that were repackaged to give the expected result.
I’d say that earlier consensus studies played similar games but I wouldn’t have the energy to defend a statement like that.”
I don’t see why not, as no calories were burned to support your first statement.
Personally I like the reasoning behind James Lawrence Powell’s recent approach to ‘quantifying the consensus’ in which he looks for the papers appearing in the peer-reviewed literature that have explicitly *rejected* AGW. In recent years they can literally be counted on the fingers of both hands.
http://www.jamespowell.org
As the run up to Paris ramps up expect more of this type of Gish Gallop to promote doubt, after all that is their product to confuse the masses.
> [R]elatively few people will have their opinions changed by consensus-messaging[…]
That’s an interesting hypothesis, Joshua. Another hypothesis is that the bandwagon effect has influence on many people, whether it’s to choose a toothpaste, a President, or to conform to social norms. I doubt we can snob the advertizing and the political industries the same way we usually snob Al’s weightful efforts. Politics is the Entertainment division of the militaro-industrial complex (pace FrankZ) for a reason.
Logic check
“Why the difference? Tol has included all surveyed, even those who responded with “I don’t know”. Well, if I’m trying to survey relevant experts and they don’t know the answer, then they’re probably not relevant experts.”
do you see what this implies?. It implies that all experts in a field have an answer to every question in the field.They are know it alls
“90% of respondents with more than 10 climate-related peer-reviewed publications (about half of all respondents), explicitly agreed with anthropogenic greenhouse gases (GHGs) being the dominant driver of recent global warming.”
It is possible to be an expert in a field and answer “I dont know” to questions about that field.
For example, you could be an expert in attribution and not know if c02 was the dominant driver.
You might for example believe it caused 60% of the warming but only be 75% confident in that answer and respond “I dont know”. I dont know is an ambiguous response.
Ask any expert in say football who the best team is this year and he might respond I dont know.
Sometime experts know when to keep their mouths shut. Ask him to place a bet and you’ll get an answer.
If you ask asked Peter Thorne to explain the reason why the various land series show differences in Diurnal range… he might say I dont know. That’s a concrete real example.
I certainly doubt that consensus messaging will persuade people who already have a strong opinion on the subject, and/or an inclination to resist the message that AGW is a threat on political grounds, to change their minds. Let’s face it, if someone has paid enough attention to the issue to have a strong opinion on it then they most likely already know full well there is a consensus – they are just looking for reasons to dismiss it.
But for the rest of the public who haven’t paid close attention but have maybe read the odd “sceptical” piece in the media or have seen a Lawson or Ridley on TV and found them superficially persuasive, I don’t see why it shouldn’t be effective.
Isn’t the natural reaction of most people without expertise on scientific issues to defer to the judgement of those who do have it? I mean I know you always get a few “bah, what do the ‘experts’ know” blowhards but thankfully they are a minority.
Richard,
Rubbish. You included the consensus of those who are unconvinced about AGW, which is just stupid.
The rest of your comment is just you JAQing. Also if you publish this with the following statement
then you are publishing a statement that you do not know to be true. I would regard that as having publishing something dishonest. All you know is that there are missing identifying numbers, not missing abstracts.
Anyway, actually trying to engage you in a discussion about this would only be worthwhile if I regarded you as honest enough to actually admit any issues. I do not and hence I won’t bother engaging you. Others can make up my own mind.
Steven,
No, it doesn’t. Read harder!
Vinny,
No, they did not.
No, it is not self-evidently true.
(At the risk of overstaying my welcome, have any “experts in attribution” offered strong support for the conclusion that our emissions are responsible for 60% [or less] of warming since the 1970s?)
Further to my previous comment, a lot of ordinary people don’t have the skills, the time or the inclination to research the subject in depth themselves. They just want to know “is this something I should be worried about”. If “there is a consensus amongst expert opinion that this is a real thing and there is cause for concern” is not a valid answer to that then what is?
And to be honest, as someone without a scientific background, and much as I’ve made a real effort to understand the science to the best of my ability and can argue to an extent some of the scientific points, I’m ultimately pretty reliant on the “consensus” myself. Particularly when it comes to likely impacts of AGW.
andrew adams, a consensus can exist without [redacted] studies seeking to quantify it.
Vinny,
Indeed, but that a study has quantified it doesn’t suddenly show that there isn’t a consensus.
willard –
To know the answer to that, we’d need to evaluate the bandwagoning effect in a polarized context, where any bandwagoning message takes place in a partisan context.
We need real world evidence, not some abstracted principle that makes sense, or comparisons to advertising in non-polarized contexts. The only real world evidence that we have is a measurement of effect, so it doesn’t tell us anything about causality and it doesn’t tease out countervailing influences. Public opinion hasn’t shifted much over time. Why not? Why are there so many people who underestimate the consensus, in particular, libz who underestimate the consensus? Is it because they’ve never heard any consensus-messaging? Hmmm. I’m dubious. Is it because in the real world, there are counter-balancing consensus-messaging campaigns under way? Possibly. I certainly wouldn’t rule that out. But then that brings into question whether further consensus-messaging is going to have some real world impact that is different than past consensus-messaging.
Andrew –
==> “But for the rest of the public who haven’t paid close attention but have maybe read the odd “sceptical” piece in the media or have seen a Lawson or Ridley on TV and found them superficially persuasive, I don’t see why it shouldn’t be effective.”
See above. Why would you assume that the person who finds Ridley superficially persuasive would then determine that he isn’t superficially persuasive simply because they see someone else delivering a counter message. Why do you assume that she hasn’t already seen counter messaging?
==> “Isn’t the natural reaction of most people without expertise on scientific issues to defer to the judgement of those who do have it? I mean I know you always get a few “bah, what do the ‘experts’ know” blowhards but thankfully they are a minority.”
Yes. In non-polarized contexts, which this isn’t. Even people who aren’t strongly identified recognize that it is a polarized, and identity-baggaged issue. Most people at least lean (at least by virtue of identifying with a group of leaners), and they recognize evidence in line with their learning. Many of those who don’t lean probably don’t have much of an opinion one way or the other, and aren’t likely to be persuaded by messaging that they can clearly see is aligned within a polarized constellation.
Some argue that the failure of existing consensus messaging proves that it has no effect. That seems like a really bad argument to me, unless you can identify the differential effect of messaging in opposing directions. However, unless we can see real world evidence that consensus messaging has your hypothesized effect, I for one remain unconvinced. Of course, I recognize that I’m one of the few who isn’t convinced. Partisans on each side are quite convinced that their are contradictory effects, respectively. I always find that a bit amusing (not saying you’re in that category).
Personally, I think that most hypotheses about the causality of public opinion are too simplistic. The causal mechanisms are complex. IMO, consensus messaging isn’t likely to have much differential effect within such a complex dynamic. I don’t think that’s likely to have the kind of blowback that some theorize (piss of “skeptics” and make them skeptical) – because those folks are already polarized; but nor do I think that it’s going to sway many non-aligned. We’re talking about a tiny % of people affected, IMO, in comparison to the more explanatory influences in play.
The dynamics of opinion formation relative to climate change fit into a pattern that exists with other issues as well. Those dynamics relate to the polarized nature of the discussion about climate change – and related patterns in human psychology and cognition. Those dynamics also relate to the nature of the problem; risk assessment of long horizon risk of low probability, high damage function phenomena. Moving opinion requires addressing the following issues – no one seemed impressed the last time I linked this, or even commented, but I’m not a quitter: 🙂 .
> To know the answer to that, we’d need to evaluate the bandwagoning effect in a polarized context, where any bandwagoning message takes place in a partisan context.
Actually we would if we already have evidence that relatively few people will have their opinions changed by consensus-messaging in the same context. Before having to parse what “partisan context” means, here’s some evidence related to the techno wars:
https://mitpress.mit.edu/books/bandwagon-effects-high-technology-industries
The techno wars can get as ugly as ClimateBall.
Some evidence of the bandwagon effect in the music industry:
Click to access moe-and-earl-bandwagon-effects.pdf
Ask your nephews and nieces about the polarizing effects of music preferences.
Well whether or not it will convince someone who was not already convinced, it has convinced me that the science is sound and that there isn’t a lot of disagreement on the key fundamentals. If I thought there was a lot of disagreement or if there were more scientists who believed AGW would be mild with little negative consequences. I think I would have a reason to change and or doubt my current position.
And, to add, I would include the IPCC reports as part of that consensus messaging.
@ Joshua
Why do we need these demonstrations of consensus as in Cook, et al.? Is it not because of ceaseless attempts by the disinformers to portray science as filled with conflict over the issue?
The Oregon Petition, which might have started this, still appears as a gambit in climateball games across the Internet and is only one example of the still-effective fables about significant, fundamental dissent in climate science. As long as these myths have traction in the public mind, it remains important to counter them with the facts.
Joseph –
If you’re here, you’re an outlier.
Extrapolating from one’s own experiences, particularly as an outlier, to hypothesize about the dynamics that affect public opinion is something that I see frequently at “skeptical” websites, and I think it isn’t very skeptical.
The most common argument I’m talking about is one that you’ve seen, no doubt, dozens of times. It goes something like: “I was a believer in mainstream climate science until I started looking more deeply and I found flaws in the science. Further, climate scientists have made many predictions that turned out not to be true, and I realized that I was being sold a bill of goods. My experiences explain why there are so many “skeptics.”
Now I think that often, those arguments are pretty much pure bullshit – because the “skeptic” speaking never had a coming to Jesus moment as they described. But there’s no way to know whether any particular “skeptic” is bullshitting when they present that story. But there is a way to tell that they aren’t being particularly skeptical – because they are extrapolating from an outlier and because they are describing trends over time by projecting in ways that are consistent with their ideological orientation and w/o any actual evidence.
It seems to me like it would be possible to measure how many people who underestimate the consensus have never heard consensus-messaging. My guess is that in the U.S., it’s got to be pretty small. Certainly, I think that there are many who underestimate the consensus and who have heard consensus messaging. Why does that happen? And I doubt that all of them are self-identified as hardcore “skeptics.” Why would someone who isn’t a self-identified hardcore “skeptic” underestimate the consensus even if they have heard consensus messaging? If you’re going to show me that consensus messaging will work, seems to me that you’d have to show that (1) you can deliver the messaging in the real world in a way that will change that person’s opinion or, (2) that there aren’t many people who aren’t self-identified hardcore “skeptics” who underestimate the consensus despite having heard consensus messaging.
Adam R.
==> “Why do we need these demonstrations of consensus as in Cook, et al.? Is it not because of ceaseless attempts by the disinformers to portray science as filled with conflict over the issue?”
I’m not sure that they’re needed. Certainly, disinformers paint a picture filled with more disagreement among experts than what exists. I would think that there probably is a small marginal effect of countering the influence of the disinformation through consensus messaging, but again, I think it is small
==> “As long as these myths have traction in the public mind, it remains important to counter them with the facts.”
This argument makes a very large assumption that others will see the ‘facts” as you see them. In reality, people filter facts so as to conform with their identity orientation. Who is likely to change their opinion about the veracity of the Oregon Petition because Obama says that 97% of climate scientists agree about climate change?
> Now I think that often, those arguments are pretty much pure bullshit – because the “skeptic” speaking never had a coming to Jesus moment as they described.
These ClimateBall ploys are far from being BS. The first objective of a rhetor is to bridge a gap between himself and the audience, and personalization fits the bill alright. As the Chemist candidly avows:
http://achemistinlangley.blogspot.com/2015/06/on-rcp85-and-business-as-usual-scenario.html
Since just about any data can fit any story, what the Chemist says at the end applies trivially.
***
To dismiss what the contrarians are naturally doing would be silly, more so considering that those who wish to speak for establishment would profit from doing the same. One might go as far as to say that the only way to connect with an adversary is to remind of your mutual humanity. Cue to George Marshall, since it’s been a while:
The main problem (H/T Howard) is to find a common ground while standing your own. This still is an open problem.
Willard –
I wouldn’t tell any individual that I think they’re full of shit – because I don’t know. For all I know with any given individual, it could be true. But where they are full of it, for sure, is when they extrapolate from their own experience to project onto the public more generally – as a way to score points in the climate wars.
And it would be a waste of time to try to find common ground with a climate warrior – because . their objective is a scorched earth, zero sum gain. In general, I don’t view “skeptics” as my adversary, but I view them as being beyond my influence. There are some exceptions, of course.
Willard… I do dismiss the Pseudo Skeptics, they aren’t worth the effort.
I do work hard at figuring out whether folks are genuinely skeptical. In my own life, after a few embarrassing family dinners… various folks have come around. A few haven’t, but they also believe in UFO’s, Jewish Overlords, and sundry conservative values… so its not like the conversation is a loss. 🙂
Meanwhile there’s always someone like Richard T(r)ol stirring things up….
Oilman –
Where did that come from?
> [W]here they are full of it, for sure, is when they extrapolate from their own experience to project onto the public more generally – as a way to score points in the climate wars.
Yet that’s how it works:
Just imagine Sir Rud’s comments at Judy’s without any kind of medal shining.
This is not rational enquiry. This is ClimateBall.
By taking deniers seriously, their prestige is increased and their sense of community forged against you watermelon warmistas. All Hail ClimateBall
@ Joshua. “Who is likely to change their opinion about the veracity of the Oregon Petition because Obama says that 97% of climate scientists agree about climate change?”
The fact is that the 97% figure has become widely recognized and repeated. It is an effective counter-meme to the “scientists disagree” idea fostered by denier propagandists, and has the additional virtue of being verifiable.
No Joshua for my experience to be not reflective of at least some others would imply that people don’t rely on scientific consensus to shape their views at all. For example, no doctor would recommend a procedure and then tell you very few studies found that the procedure would work and some found complications. And if they did would you still want the procedure done?
And another question I have for you, if you shouldn’t use consensus messaging when discussing the science, how would you defend statements like CO2 will likely causes warming in excess of 2C this century if no serious mitigation efforts are taken, or there are risks from climate change impacts in the future? I guess you could refer them to the literature and tell them to read it all or you would have to depend on some consensus opinion like that in the IPCC reports, right.
> [T]heir prestige is increased and their sense of community forged against you watermelon warmistas.
The wartermelon warmistas make them do it, “it” being what they do best:
http://georgelakoff.com/2014/11/29/george-lakoff-in-politics-progressives-need-to-frame-their-values/
One does not respond to contrarians to win their hearts or change their minds. One takes contrarians’ concerns seriously to promote one’s own values. Openness implies we remain forever thankful for their concerns and anyone else’s, including Howard’s.
Joshua: xkcd… the source of all truth and knowledge.
https://xkcd.com/386/
Willard: Yeah, the teabagging religious right hates the foot-washing pope. I find it very hard to take that kind of contrariness seriously.
@Wotts
The graph has, in fact, many results and the conclusion does not depend on this one subsample of Verheggen’s. There is a table in the paper too, so that you can readily draw your own preferred graph. And the papers used are all in the public domain so that you can check my numbers.
That said, consensus in the high nineties is only found in Cook’s whole sample and in others’ subsamples. If you compare Cook’s results to others’ USING ALL THEIR DATA, Cook emerges as an outlier. I had always thought that ignoring data is a bad strategy, so I prefer this comparison.
Richard,
But it does put it into the right context (i.e., particularly stupid).
Using all of Cook’s data returns a result of around 35%. Why have you decided it’s okay to drop some data in Cook et al., but not in the other studies? Furthermore, you’ve publicly stated that the consensus is very obviously in the high nineties. That Cook et al. might be an outlier therefore seems rather irrelevant. It’s for this reason that I regard you as a mis-informer. It’s very hard to see how else to describe someone who is committing so much effort to discrediting something with a result you don’t even dispute.
Illustrating your overly simplistic idea of how data analysis should work. This is the kind of thing I expect to hear from an under-informed, self-professed, blog-expert, not a Professor of Economics.
You also haven’t commented on the fact that you may be about to publish something that you do not know to be true.
@wotts
Cook et al cast the net so wide that one may argue that the don’t knows should be ignored.
I only publish things I do not know to be true. I never publish things I know to be untrue, though.
Richard,
Just because you’re clueless, does not mean everyone is clueless.
Huh? I’ll explain again. If you state in your paper that there are ~400 missing abstracts, rather than ~400 missing abstract identifiers, then you are publishing a statement that may well not be true and that the author of the original paper has already explained as not being true. That, to me, is fundamentally dishonest and probably unethical. That you would do such a thing is, however, no great surprise.
You also haven’t explained why you chose not to use the full sample from Cook et al., despite your argument that you were using full samples. To be clear, I don’t expect you to answer this as that would be completely out of character. I’m simply highlighting that what you claim to have done, is not what you’ve done.
Vinny,
andrew adams, a consensus can exist without [redacted] studies seeking to quantify it.
Yes, of course – my point was about the value of communicating the fact that a consensus exists in general, not about any particular paper or survey, although I can see how trying to put numbers on it can make it easier to communicate. Personally I would use the IPCC as my reference point rather than surveys of the literature, but that’s just me.
Joshua,
Why would you assume that the person who finds Ridley superficially persuasive would then determine that he isn’t superficially persuasive simply because they see someone else delivering a counter message. Why do you assume that she hasn’t already seen counter messaging?
If you’re just saying “I’m right about the scientific arguments and he’s wrong about them” to someone who has no obvious reason to accept your authority on the subject over theirs then that is certainly not going to be effective. But what consensus messaging does is delegate authority to those who actually have it, you’re saying “I think this guy’s wrong but ultimately you shouldn’t listen to either of us you should listen to the people who actually know about this stuff.”
It’s also a way to encourage people to cut through the highly polarized nature of the argument and get to the actual facts of the matter. I do think that (at least in the UK) there is a section of the population which doesn’t have any natural inclination to believe one side or the other because of their own political leanings and are open minded on the subject. A lot of them probably aren’t even aware of the extent of the political arguments on the subject. I don’t know how many there are but I suspect there are a fair few of them, and these are the people for whom consensus messaging can be effective. For other people you would need different messaging. You always have to tailor your message to the specific audience you are addressing.
@wotts
In the latest comment, I note the discrepancy in sample size, and invite Cook and coauthors to explain. I do not know what happened, and the various second- and third-hand explanations I’ve seen strike me as implausible.
Richard,
No, you make a statement about missing abstracts, when all you have is some missing numbers.
And when someone makes you supreme ruler of the scientific process, maybe then you could have some kind of right to ask questions and expect an answer. Until such time, you can – I guess – ask questions, but most would simply regard you as JAQing off.
Of course you don’t know what happened. The explanations are, however, entirely plausible. Here’s a comment I wrote in 2013 in which I say
The earlier Cook et al. search returned 12465. Since the database updates regularly, a discrepancy of 0.65% seems reasonable. Certainly no evidence for 411 missing abstracts. That you would continue to question the response that has been given – without any evidence that it was a reasonable response – is, in my opinion, highly unethical.
In fact, I’ve just done the search again. Search term was “global warming” or “global climate change”. Database was restricted to Web of Science Core Collection and under More Setting I selected Science Citation Index Expanded only. It returns 12606 articles.
first-hand, Wotts, first-hand
Richard,
Tell you what. If you want to email John Cook directly, let me know and I’ll try to get hold of him in advance to warn him and to suggest that he doesn’t simply tell you to “piss off”. That’s what I’d do, but maybe John is more polite than I am. There’s always a chance that you’re genuinely interested in an answer, rather than simply trying to mis-inform, but I’ve rather lost any sense that that is the case. Also, you haven’t commented on my point above that my WoK search returns 12606. How did you get 13458?
@wotts
I just ran the same query as they did for the same period to find 13,458; WoS tells me 27 papers have been added later. Note that Cook did not exclude the SSCI, explaining the discrepancy between your results and his.
Richard,
The only way I can get close to 13458 is to use both the Science Citation Index Expanded and the Social Sciences Citation Index. Are you sure you did the appropriate search?
Richard,
Are you sure? When I did my search in 2013 I matched their results by selecting the Science Index only. Cook et al. only had 288 Not Climate Related (one reason for this being Social Sciences), which seems rather small if they were including the Social Science Index. So it seems to me that you may once again be making a statement that you do not know to be true. You do realise that someone not saying explicitly that they did, or did not, do something, does not suddenly mean that you’re justified in claiming that they did, or did not, do this thing?
@Wotts
A later search should find a superset.
Richard,
Huh? A later search returns more than the earlier search. If that’s what you mean by a superset, then it does. Of course, what you’ve failed to address is how – without a time machine – Cook et al. were even expected to comment on a later search returning more than their earlier search. You do realise that we have yet to invent a time machine, don’t you?
@Wotts
WoS is a proper database. It tells you when things were added. So you can reproduce a past query.
Richard,
Well, I can’t work out how to do that. Maybe you can tell me how many would be returned for a search done in March 2012 and in May 2012, excluding the Social Science Index. I got 12547 in April 2013 and 12606 in September 2015.
“first-hand, Wotts, first-hand”
Did you, at any time, specifically ask John Cook, in person, to help you understand an apparent discrepancy you thought you found?
You know, the way a good scientist would go about when he thinks there is something he does not fully understand: ask the corresponding author.
@ Richard,
Care to comment on assigning a 75% consensus to Oreskes 2004. Do you think it is at least potentially misleading? Why leave out the sentence ATTP highlighted – “Remarkably, none of the papers disagreed with the consensus position.”?
[Mod : I know I’ve been rather rude to Richard, but I think I’ll refrain from posting this as me being rude on my blog is probably slightly different to allowing others to do so. To be fair, this comment isn’t bad by climate blog standards, but I’ll err on the side of caution 🙂 ]
Isn’t it obvious the kind of game Tol is playing here? He doesn’t care about the results, but only wants the term “flawed” attached to Cook.
It’s a binary world denierville is.
Same as the Hockey Stick was “flawed”, same as Hansen’s 88 model was “flawed” etc. Yes, in the real world they all were “flawed” in one sense or another. But that’s a long way from total dismissal…
Is Tol(bert) the Donald Trump of global warming?
@-Is Tol(bert) the Donald Trump of global warming?
No, he may share some trichological problems but Tol is not THAT stupid surely…
As tempting as it may seem, let’s not let this degenerate into a Tol bashing thread. I’d really rather not let this site degenerate too far towards the norm on many other sites.
Pigs and wrestling AT.
Who is enjoying this exchange?
Who is benefiting from it?
Sorry, but that kind of response leaves me sort of frustrated as well. In a thread named “more nonsense” from Tol, mind…
So what else should it denigrate towards?
In fact I believe the comparison with Trump is perfectly apt these days. There is no rational appropriate “uncle Bob” response to Tol, same as there’s no rational appropriate response to Trump.
SDK,
I don’t claim to moderate in an entirely consistent fashion. I also sometimes expect better from some than from others. I also think that if I mention someone in a post they should be able to come here and defend themselves, even if I think their behaviour leaves a lot to be desired. My comment was made without prejudice and was simply a request. My behaviour on this thread has not been without fault, but then it’s my blog and Richard can always complain if he really wants to.
vtg,
Yup, I get your point.
@Richard Tol : could you perhaps provide your exact WOS search parameters so that others here with access could recreate it and compare results? I am assuming from the discussion here that it is not in the paper proper.
Just an observation, but Richard Tol grabbing all the papers he can find to indiscriminately support a position seems to be one of his chief modus operandi.
He currently holds his own, old, meager, work on climate costs in the same light he holds the latest work by the best and brightest out there. (Its like he stepped off the elevator on the second floor, and started yelling up the stairs.) I wonder if Richard Tol could comment on the quality of work done Aristotle or Galileo? Why not?
He’s doing this again with is nonsense on the consensus paper. A real man. A real science. Man of true worth, would do some work and attempt to earn some respect. That man is NOT Richard Tol, and I strongly suspect it never will be.
My offer of $1000 of my own cash to Richard Tol actually find his 300 supposed papers still stands. Heck… I’ll up it to $10000. (I think he’s just plain too lazy to do the work and collect, and as the years go by we’ll all watch and know the truth about him.)
Someone on Twitter has reminded me of Richard Tol’s attempt to argue that the data point from his own study was not an outlier.
Re-slicing and dicing a database after the results are in can give you any answer you want by trimming the question to yield the answer.
Richard bandied about the 12,465 / 12,876 numbers months ago, and in a discussion at HotWhopper he was informed just what those numbers represented: Cook did two separate downloads from WoS, and eliminated some 411 duplicates. Of the 12,465 unique abstracts remaining, the authors removed papers not peer-reviewed (186), not climate related (288), or without an abstract (27), resulting in the 11,944 abstracts studied – all of which was stated in Cook et al 2013.
Tol took part in the discussion on HW, and thus cannot claim ignorance – yet he continues to push and now, apparently, _publish_ fallacious claims such as these. That’s both dishonest and unethical.
The rest of his ‘comment’ (I seriously see no reason whatsoever for such a bunch of insinuating BS to published) is nonsense he’s pushed before, just repeats of debunked claims of ‘nefarious intent’ and ‘something must be wrong’.
On the issue of abstracts and papers that didn’t express opinions, I would quote the federal appeals court ruling on EPA regulation of greenhouse gases:
“This is how science works. EPA is not required to re-prove the existence of the atom every time it approaches a scientific question.”
Shame on you, Tol.
Eli: “Re-slicing and dicing a database after the results are in can give you any answer you want by trimming the question to yield the answer.”
Economics has a math problem: http://www.bloombergview.com/articles/2015-09-01/economics-has-a-math-problem
“Economists use math as a tool of rhetoric instead of a tool to understand the world.” A lack of consensus on key issues is evidence that they tolerate “academic politics, where persistent disagreement is encouraged as a way to create distinctive sub-group identities.”
@pbjamm
The parameters are in Cook’s paper. There is some ambiguity which databases were included, but that disappears once you realise that a query in 2015 returns a superset of a query in 2012.
Hi Richard Tol,
Ackerman.
Gremlins.
#freethetol300
Enjoy the attention.
Oh, and a quote from your good self.
You miss that one, Very Tall:
http://t.co/wXd0FjekBE
KR,
I agree, appalling. The idea that Richard can pontificate about research integrity while knowingly publishing claims in his paper that are almost certainly not true is remarkable; and not in a good way. I wish that I could say that I was surprised, but I’d be lying if I did.
This is like a replay of the HW thread that KR links to above. Since others were able to replicate Cook’s search 6 moths ago it would appear that your search does indeed differ in some way
pbjamm,
Essentially it is clear that if you use Web of Science but with the Science Citation Index only, you return a search that looks compatible with Cook et al. (2013). If you use the Science Citation Index and the Social Citation Index you return about 1000 more abstracts. Now Cook et al. did not explicitly state that they used the Science Citation Index only. However, in their paper it’s clear that they define Social Science abstracts as Not Climate Related. Therefore, why would use the Social Science Index if they didn’t want those abstracts to be returned. It only took me a couple of minutes to work this out. However, if your goal is to misinform, then clearly working this out is not in your best interests and it is much better to just continue asking questions so as to imply that Cook et al. were not completely honest. The irony, of course, is that Richard is potentially about to publish a paper with a number of claims that are very obviously not honest. So, we have Tol making potentially dishonests claims in order to imply that others are not completely honest. Bizarre doesn’t even begin to describe this.
Tol writes: “WoS is a proper database. It tells you when things were added. So you can reproduce a past query.”
The Web of Science may indeed have a metadata field available to the employees that create and maintain the academic database. But it is not accessible to end-users, quite apart from the very significant amount of work that would be required (even if a Date Added field was a possible search criteria) to settle a very minor question to which Tol already knows the answer.
Magna,
Thanks, I did waste a bit of time today trying to work out if it was possible to actually duplicate a search done at an earlier time. It became fairly clear, fairly quickly, that you can’t (or, at least, I couldn’t work out how, which seems consistent with what you’ve said).
Willard, indeed.
There’s also L’affaire Irlandais
Our emphasis.
http://www.independent.ie/opinion/analysis/daniel-mcconnell-a-fearless-whistleblower-or-a-disgruntled-crank-26809004.html
More contributions welcome.
That was interesting, vtg. For what it’s worth, I’m starting to notice certain similarities between Richard Tol and Luboš Motl.
vtg,
I wasn’t aware of that one. Eventually it will have to become obvious who the common thread is in all these controversies, won’t it?
“Eventually it will have to become obvious who the common thread is in all these controversies, won’t it?”
Yep. The ‘outlier’ is Richard Tol, advisor to the GWPF along with a host of climate science deniers. Shame he couldn’t find something constructive, or for that matter factual, to spend his time on.
AT,
it’s just possible there are more. Anyway, rather than give Richard the enjoyment of engaging with his latest attention seeking misdirection, I suggest just posting links to his previous ones.
Gelman is particularly good, wouldn’t you agree Richard?
Seems that Richard has decided the answer to that one is yes!
If anyone is bored, I suggest mining this thread. Comedy gold. But also a sad reflection of a wasted intellect.
http://andrewgelman.com/2014/05/27/whole-fleet-gremlins-looking-carefully-richard-tols-twice-corrected-paper-economic-effects-climate-change/
Actually, ATTP, perhaps some ought to bring Andrew Gelman’s attention to this latest misuse of statistics. Hmmm…
vtg,
I get the impression that he’s no longer interested.
Can’t say I blame him
I keep forgetting that I’m no longer really interested either 🙂
That assumes no journals are dropped. Hint: Some are
I meet your ESRI, Very Tall:
http://www.thejournal.ie/esri-paper-unemployed-dole-richard-tol-484632-Jun2012/
Because sometimes EconometricsBall is a work-in-progress.
Willard –
A thought did occur to me around this. Imagine you took papers in physical geology from 1955 to 1985. For each year, you measured the percentage of papers that directly affirmed plate tectonics.
What I’d expect is that the percentage would start off near zero, reach a peak in the late 1960s to early 1970s – when the main details were being thrashed out – and then fall off as the theory simply became part of the background. And pretty much the same in any field in which the main details of a fundamental theory were being thrashed out. Which would allow for ‘consensus dating’.
Richard Tol claims that Cook’s results are an outlier, but Tol is misrepresenting other research in order to arrive at that conclusion.
E.g. regarding Oreskes: You can’t just divide the number of affirmative statements by all papers in the sample, if many papers didn’t actually stake out any position on the question at hand. The latter should logically be excluded, unless you want to argue that of all biology papers, only 0.5% take an affirmative position on evolution, hence there is low consensus on evolution. (https://twitter.com/BVerheggen/status/643806716342222848, as already quoted by Willard just above)
Or even better, let’s ask some Heartland staffers, some skeptical retired economists, Pat Michaels and Roy Spencer what they think of climate change and pretend that the results is somehow a representative indication of the scientific consensus on climate change. That’s in effect what Tol does when he presents our subsample of “unconvinced” (in Verheggen et al).
It’s hard to believe that he’s serious in these claims. I’m still thinking that perhaps he’s trying to pull a practical joke or something.
Bart,
I thought the same, for a very long time. If it is simply a practical joke, it’s been running for a long time, and it’s really no longer funny.
Well, if there’s one thing that there’s no consensus on (and never will be), it’s what constitutes a good sense of humour. And that’s probably a good thing.
True, although comedians who don’t make people laugh do tend to find themselves short of work eventually.
Bart,
It’s not the first time Tol has done this, and I very much doubt it will be the last
Gelman:
http://andrewgelman.com/2014/05/27/whole-fleet-gremlins-looking-carefully-richard-tols-twice-corrected-paper-economic-effects-climate-change/#comment-167769
@Bart
As already noted by Wotts, there is a discrepancy in the way I treat Oreskes (include don’t knows) and Cook (exclude don’t knows). Of course, if I include Cook’s don’t knows then he is still an outlier, but then falling far below the other studies.
Cook and Oreskes strongly contradict each other.
Oreskes, N = 1000, agree = 75%, don’t know = 25%, disagree = 0.0%
Cook, N = 12000, agree = 33%, don’t know = 75%, disagree = 0.1%
Anderegg start the game of subsampling. In their whole sample, the consensus is 66%. They show that they have subsamples with a much higher consensus. By implication, there are also subsamples with a much lower consensus.
While it is valid the mock your 7% as “those who disagree tend to disagree”, it is equally valid to mock your 79% as “those who agree tend to agree”.
Cook’s 75% is, of course, 67%.
Richard,
My goodness you talk an awful lot of crap. I really do think that you’ve now passed into the realms of too ridiculous to actually bother criticising. I shall endeavour to do so in future.
Richard,
you really have picked up Andrew Gelman’s challenge to spend two more decades doing substandard work, just because you can
Please, do continue!
@RichardTol I suspect this has already pointed out to you, but “don’t know” is not a valid characterisation of papers that do not take a position on the particular question that is the subject of the paper. Many papers on climate change are not concerned with whether the cause is natural or anthropogenic (i.e. they are not attribution studies) and hence there is no reason to expect them to express a stance on the issue in the abstract. For this reason, they should not be included in the computation of the consensus. This ought to be obvious.
Of course, what Richard fails to seem to realise is that the real reason that Cook et al. is an outlier is because of the sample size, and not because of the result it gets (just look at the figure in the comment to which I link in the post). So, following Richard’s argument to it’s logical conclusion, one should never do a study with a sample size that is much larger than any other study, because your result will – by definition – be an outlier.
vtg,
The real problem is that the rest of us can’t spend the next two decades pointing this out…oh, hold on, maybe that’s a feature, not a bug.
Of course this is a complete mis-characterisation
It’s not don’t know it’s don’t take a position which should be obvious from the quote from Oreskes that I included in the post. It should also be obvious that don’t take a position is not the same as don’t know, well obvious to virtually everyone.
I suspect that Richard Tol is just putting forth a ‘deliverable’ — paid or not. He can’t very well chum around with his GWPF cohorts without having something to show them.
Richard may not be the sharpest knife in the drawer, but few are that dull or obstinate not to face the facts after years of being beaten by a dead-horse. Yes, being beaten by a dead horse – not beating one 🙂
Someone above mentioned that Richard is the outlier; no, he’s just noise.
@wotts, dm
Replace “don’t know” by “no position” and see whether the numbers change.
Richard,
Okay, let’s see:
Oreskes, N = 1000, agree = 75%, no position = 25%, disagree = 0.0%, therefore 100% of those that took a position agree with the consensus.
Cook, N = 12000, agree = 33%, no position = 75%, disagree = 0.1%, therefore 97% of those that took a position agree with the consensus.
See, that wasn’t even all that difficult. You should try it yourself.
@wotts
Pearson and stuff.
You cannot claim that two data sets contain the same information if a statistic agrees. Sufficient statistics must agree. The consensus rate is not a sufficient statistic.
More on Ackermann:
http://frankackerman.com/tol-controversy/
Beyond the bounds of acceptable debate is the title.
***
Do you still have what Richard wrote to your U, AT?
AT,
you seem to be engaging on the substance of claims made by a “troublemaking crank” doing “substandard work”
Perhaps rather than attempt to help such an individual, offering Andrew Gelman’s advice that he should “should take a deep breath and think more seriously about what you’re doing.” might waste less of your time and also help Richard more. As he already knows what he did to arrive at his bogus graph.
Pigs and all. Just saying.
Richard, the “no position” papers don’t tell you anything about the consensus, and so shouldn’t be included in the analysis. This isn’t rocket science or brain surgery, just common sense.
Richard wrote : The consensus rate is not a sufficient statistic.
It also needs to be calculated correctly. The consensus rate is the proportion of those papers that take a position that support the mainstream view on the particular question.
“You cannot claim that two data sets contain the same information if a statistic agrees. ”
I don’t see anyone claiming they contain the same information, just that they agree on the consensus rate (when correctly defined), so that is a rather obvious red-herring.
Richard,
I made no such claim. I simply reported the results in those two studies in a manner that is consistent with the actual data in those two studies.
What? Oreskes is a study of 1000 papers. 100% of the paper in that study that took a position agreed with the consensus. Cook et al. is primarily a study of 11944 abstracts. 97% of the abstracts that took a position agreed with the consensus. Statements about statistics and whether or not they must, or must not, agree does not justify you mis-representing the results in these studies.
What you’ve said is also nonsense. There is no requirement in the scientific method that if you do use multiple methods/datasets to determine something, that there has to be some kind of statistical equivalence in the data that these different methods/studies use. You’re using complicated words to try and mis-inform. It is genuinely despicable. If I hadn’t mentioned you in this post I would simply start deleting any further comments, since there are plenty of other sites that welcome your sort of mis-information. Since I have mentioned you, I feel that it is wrong to delete your comments on this post. That, however, doesn’t mean that I can’t point out where you’re being dishonest and where you’re simply using nonsense to try and mis-inform.
What do you mean by consensus rate? All that is being reported in these studies is some fraction/percentage of people/papers/abstracts that agree with a consensus position. Unless rate has another meaning about which I’m unaware, these are not rates, they’re simply percentages.
I should add, that I don’t expect you to provide some kind of reasonable answer to my question. I’m only asking so that others can recognise that you’re mostly talking nonsense.
@wotts, dikram
Q: Do the Cook data agree with the Oreskes data?
A (Tol): No: The distributions are completely different.
A (W,D): Yes: There is a transformation of the data that is the same.
Now what would Pearson say?
Spot the outlier:
Source: http://www.washingtonpost.com/blogs/monkey-cage/wp/2015/05/23/the-gremlins-did-it-iffy-curve-fit-drives-strong-policy-conclusions/
The only dot clearly above 0 comes from Richard’s own analysis, of course.
Is Richard’s own outlier a sufficient statistic to drive his strong policy conclusions?
Willard, remind me, was that outlier robust? Or did the gremlins affect even that number?
Richard,
Given that you are now mis-representing me, as well as all the other consensus studies, I will take vtg’s advice from now on. I did not say the above. I pointed out that 100% of Oreskes data that took a position, agreed with the consensus, and that 97% of Cook’s data that took a position agreed with the consensus.
I suspect that Pearson would point out that you can’t do a chi-squared test on data that is known to not be drawn from the same population.
vtg –
And it’s still an amateur analysis, a pro would have put the cutoff at +2.4C and done a linear regression.
@wotts, dikraN
“Q: Do the Cook data agree with the Oreskes data?”
A (D) In that both show that the consensus rate, when correctly calculated is at least in the high nineties, yes.
“A (Tol): No: The distributions are completely different.”
Only an incompetent statistician would expect them to be the same (according to the sufficient stats) as they are not samples from the same population.
A (W,D): Yes: There is a transformation of the data that is the same.”
I have already explained that you are calculating the consensus rate incorrectly and to call it “a transformation of the data” is pure sophistry.
“Now what would Pearson say?”
“Stop digging, the hole is deep enough already”?
Andrew Dodds/VTG when working with a small dataset, it is often instructive to look at the model fits resulting from leave-one-out cross-validation. If a feature of the model (e.g. curvature) depends on the presence of a single datapoint, it is usually an indication that the feature is spurious (and possibly that the datapoint is an outlier and needs checking/investigating).
dikran,
Grant McDermott did that here. Doesn’t make much difference with the original data in Tol (2009) but does – I think – when you add the new data (or, maybe, its influence becomes very diminished).
ATTP, interesting, however if you left each datapoint out in turn, I suspect it might have shown how dependent the model fit is to the two patterns at 3.0 as well (the cluster of points at 2.5 is likely to act as a fulcum about which the curve will flex). I suspect heteroskedasticity (non-constant noise variance is likely to be an issue as well). But at least he didn’t try something really risky, such as a non-parametric model…
Ze question, dear Very Tall, is not what would Pearson say of Richard’s gremlins. It’s rather, what does “division by zero” mean:
Click to access Nordhaus_comment_on_Tol.pdf
Dikran et al,
yes, I think everyone agrees the quadratic fit is, well, not entirely robust.
But I also recall that the one positive value, from Tol, also was Gremlin afflicted. But I forget, so many errors from Tol it’s hard to keep up. Willard will enlighten us I expect.
VTG I think it is one of those rare occasions where fitting a model actually detracts from just looking at the data, simply because of the issues in the dataset that depart from the usual model-fitting assumptions, e.g. heteroscedasticity, discretisation of the independent variable, paucity of data outside 2.5. Makes a bit more sense for the version with more datapoints, which apparently didn’t have a model fit!
I think John Tukey might have been a better statistician than Pearson to ask.
A boilerplate, Very Tall:
Inspiration: http://cstpr.colorado.edu/prometheus/archives/climate_change/000996william_nordhaus_on_.html
@ Andrew Dodds,
> consensus dating
Interesting quote from Roger Cohen, head of theoretical sciences at Exxon Corporate Research Laboratories. Not saying directly comparable to the consensus studies but…
“Over the past several years a clear scientific consensus has emerged,” Cohen wrote in September 1982, reporting on Exxon’s own analysis of climate models. It was that a doubling of the carbon dioxide blanket in the atmosphere would produce average global warming of 3 degrees Celsius, plus or minus 1.5 degrees C
“There is unanimous agreement in the scientific community that a temperature increase of this magnitude would bring about significant changes in the earth’s climate,” he wrote, “including rainfall distribution and alterations in the biosphere.”
http://insideclimatenews.org/news/15092015/Exxons-own-research-confirmed-fossil-fuels-role-in-global-warming
ht Eli
Incidentally, had Richard read Cook et al. with sufficient assidulousness, he would know that Cook et al. explicitly compare their sample with that used by Oreskes and explain the difference in the proportions:
“Our sample encompasses, those surveyed by Oreskes (2004) and Schulte (2008) and we can therefore directly compare the results. Oreskes (2004) analyzed 928 papers from 1993 to
2003. Over the same period, we found 932 papers matching the search phrase ‘global climate change’ (papers continue to be added to the ISI database). From that subset we eliminated
38 papers that were not peer-reviewed, climate-related or had no abstract. Of the remaining 894, none rejected the consensus, consistent with Oreskes’ result. Oreskes determined that 75% of papers endorsed the consensus, based on the assumption that mitigation and impact papers implicitly endorse the consensus. By comparison, we found that 28% of the 894 abstracts endorsed AGW while 72% expressed no position. Among the 71 papers that received self-ratings from authors, 69% endorse AGW, comparable to Oreskes’ estimate of 75% endorsements.”
So it seems the difference in the proportion of “no position” papers is due to the more cautious criteria used by Cook (i.e. not assuming that “mitigation and impact papers implicitly endorse the consensus”).
Of course unless Oreskes had a time machine, Cook et al. also sampled from a much larger population where the proportion of papers stating an position on the question might reasonably expect to change over time. It would not be at all surprising then for the “sufficient statistics” to show they were different distributions even if the counts were compiled to the same criteria (which they weren’t).
Although this exchange has been interesting and entertaining, this equation seems self-evident:
(Tol’s time expended)/(Σ Tol responders’ time expended) << 1
Is it mere chance that you can’t spell ‘troll’ without ‘t’ ‘o’ and ‘l’?
@dm
As I said, Cook and Oreskes disagree.
Richard,
I guess if you keep repeating that over and over again, you might eventually even convince yourself that it’s true
If you have every dealt with Donald J. Trump, understanding Richard Tol is simple. Richard’s way of dealing with anyone who questions Richard is full throated Kzinti roar and attack. For the most part this succeeds because few want the joy of dealing with an axe murderer, and Richard does an excellent impression. This discourages criticism.
However, over a long enough time, the act wears thin, and a few start to speak up, but this does not deter Richard who simply does not pause, relying on sensible people eventually losing interest, As long as he retains a foothold in his field, this also works. Someday though Richard will find his Richard.
Now as to his economics, it is clear that a) it is formally correct and b) deals with another planet where logical consistency is optional
“As I said, Cook and Oreskes disagree.”
No, they don’t. They use different classification criteria, but draw the similar conclusions. The quote from Cook et al. demonstrates they agree on the key point, and your one-liner non-sequitur bluster just tells everyone you know that perfectly well “The consensus is of course in the high nineties. No one ever said it was not. We don’t need Cook’s survey to tell us that. ” Both Cook and Oreskes agree with you on that one.
Stop digging.
@-As I said, Cook and Oreskes disagree.
No they don’t.
with slightly different samples they find slightly different results for the consensus.
Both results reveal that, as you have acknowledged, the scientific agreement over the attribution of global warming is over 90%.
It is incorrect and either misleading from ignorance, (ESL? mathematical formalism??) or maliciously distorting, to use “disagree” when they hardly differ.
What Magma said.
Richard remains the only person enjoying this exploration of what Andrew Gelman termed rightly termed his “substandard work”
vtg,
I’m certainly not enjoying it, but it is surreal. How does someone who talks as much obvious nonsense s Richard, get an undergraduate degree, let alone a PhD and then a Professorship at a perfectly respectable university? Amazing really.
@-“How does someone who talks as much obvious nonsense s Richard, get an undergraduate degree, let alone a PhD and then a Professorship at a perfectly respectable university? Amazing really.”
By saying what some audiences want to hear?
In politics ‘Obvious Nonsense’ is usually a feature not a bug. It has a purpose, beyond making the bystander think the purveyor is an idiot.
ATTP,
Richard is very clever and capable, and makes clear choices when and how to apply his intellect.
Magma, there’s a crossword clue just waiting to be set free.
Odd troll (3), perhaps.
Yawn…. Are we there yet?
vtg,
Yes, I suspect that you’re quite right. That just seems worse, though.
verytallguy: “Richard is very clever and capable, and makes clear choices when and how to apply his intellect.”
Many people would prefer him to do some real work for a change instead of being an obnoxious science lookyloo (look it up) who claims to be an expert instead of actually being an expert.
Why doesn’t he do work for a living? Why can’t he find his 300 papers? Is this a distraction from his clear and obvious failure to locate his missing papers? Make no bones about this… if someone claimed something as ignorantly wrong as he has in a company, he’d be fired, or eating crow by now.
Any rational reasonable human would be fired for acting like he does, and as badly as he does. He’s a shame and a disgrace.
hey guys, don’t underdo it — I was appointed full professor at 29, and am ranked 124th (out of 44903) in the world
Richard,
Yes, I know. I’m as amazed as you are.
@-(out of 44903) in the world.
forty four thousand, nine hundred and three what?
economeretricians?
Just ask this Scientician!
Richard Tol: Mistakes happen…
> I was appointed full professor at 29
Before getting his PhD to boot, and notwithstanding the Nobel prize:
http://www.ivm.vu.nl/en/people/researchers/environmental-economics/tol/index.asp
Well played!
***
Oh, and Richard: what would be Pearson’s statistically robust answer to 2000 – 1969?
Many thanks!
Richard, truly, you are a thing of great beauty.
Some of us remember what happened to the Kzinti when they met humans.
Or possibly that was a result of them meeting Person’s Puppeteers.
VTG –
And a work of art.
A Van Dyck perhaps, Joshua?
The Tolgraph is worse than you can think. Each of the studies uses a different time period to set T=0. Tol has not corrected for this. The graph is meaningless drivel
Clever fellow this Tol. A Ph.D. in economics (1997) from the Vrije Universiteit Amsterdam (on his University of Sussex page) and a PhD Economics, VU University, Amsterdam, The Netherlands in 2000 (on his Institute for Environmental Studies at VU University page). Does he have two PhDs from the same institution or can’t he get the dates right?
Well, we could look at it this way:
If time is infinite, there will be other versions of Tol — the current one being kind enough to allow for the possibility of those future versions to be competent, humble and intellectually honest.
So maybe he’s just being good to future Tols!
> I was appointed full professor at 29, and am ranked 124th (out of 44903) […]
You’re being too modest, Richard:
https://ideas.repec.org/top/top.person.all.html#pto90
You need to accept that you’re an author, first and foremost. Witness the FAQ:
https://ideas.repec.org/t/ranking.html
Do you think Pearson would say this is about professoring?
@Mike
One PhD only. 1997. I did not write the other entry.
Gremlins, obviously.
[canned laughter]
… and still no 300 papers. Yup. $10,000 waiting for RIchard Tol if he ever does some real work and finds them.
I gotta laugh.
““You make no sense” is also a common ClimateBall move. I call it the Chewbaccattack. It’s a more aggressive move than the Chewbacca defense:”
too funny.
ATTP (way up thread): “Yes, I can imagine the meeting [ERL] had. “If we let him publish [his comment], maybe he’ll stop whining”.”
Right. That would be like a woman saying of the annoying guy at work, “Maybe if I sleep with him he’ll stop hitting on me.”
Pingback: Official Socialist Webzine - 2015 SkS Weekly Digest #38
Pingback: Scientists Respond To Tol's Misrepresentation Of Their Consensus Research - Real Skeptic
“there is a discrepancy in the way I treat Oreskes (include don’t knows) and Cook (exclude don’t knows). ”
Says it all. You cannot be trusted.
Bart Verheggen already pointed out some of the mistakes Tol makes here, but I contacted other cited authors and they also weren’t happy with how Tol used their results/data: http://www.realskeptic.com/2015/09/21/scientists-respond-to-tols-misrepresentation-of-their-consensus-research/
I think Tol really made a mistake with what he did with the comment that he submitted to ERL. It’s not often you see cited authors respond in the manner they did.
Collin,
Good post. I suspect that Tol thinks that everyone else is wrong.
Thanks, took a bit of time to get all the responses.
The responses from Tol so far have indeed indicated that he thinks that everyone else is wrong, including the authors of the studies he cited (see the exchange with Bart in this very comment section). Though I haven’t seen Tol say anything anymore about this, seen him move on to other subjects like the VW scandal.
Pingback: Consensus on consensus | …and Then There's Physics
Pingback: Ignoring adaptation? | …and Then There's Physics
Pingback: Non piove, governo ladro – L'archivio di Oca Sapiens