The Intersection

i-391627a85ca69c7f8b64b6b20d3304d5-Catarina.jpg

As Roger Pielke, Jr., has already noted, word has it that the new IPCC report will say that hurricanes have measurably intensified due to global warming. Roger warns that, if true, this will cause huge controversy. I would go even further and say that if true–and that’s still a huge caveat at this point–this will be the most controversial part of the report.

However, it’s important to pay attention to the alleged details. According to the same media scoop (by the AP’s Seth Borenstein), the IPCC’s language will merely say that it’s “more likely than not” that changes to hurricanes have occurred. That sounds to me like a fairly low level of certainty to ascribe to the conclusion–barely above 50 %. It’s far less than the certainty level expected to be ascribed to the basic conclusion that humans are causing warming itself (90 % or higher).

Compare “more likely than not” to the recent World Meteorological Organization consensus on hurricanes and global warming (PDF), which stated:

Though there is evidence both for and against the existence of a detectable anthropogenic signal in
the tropical cyclone climate record to date, no firm conclusion can be made on this point.

How much difference, in terms of percentage or level of certainty, is there between “more likely than not” and “evidence both for and against”? I leave that up to scientists to argue about, but I suspect that in spite of the vastly different ways in which the two statements will be read (and rightly so), scientifically they may not be quite as distant from one another.

In fact, let’s go further. The last IPCC Summary for Policymakers (PDF) contained the following footnote:

…the following words have been used where appropriate to indicate judgmental estimates of confidence: virtually certain (greater than 99% chance that a result is true); very likely (90−99% chance); likely (66−90% chance); medium likelihood (33−66% chance); unlikely (10−33% chance); very unlikely (1−10% chance); exceptionally unlikely (less than 1% chance).

By these criteria, wouldn’t “more likely than not” and “there is evidence both for and against” both fall within the range of “medium likelihood”? Surely there is a difference between “likely” and “more likely than not,” and surely the latter is a weaker statement….right?

To me, what all this underscores is the continuing problems that scientists have when it comes to translating their “uncertain knowledge” for a public that doesn’t (and won’t) understand the nature of such knowledge.

Comments

  1. #1 Fred Bortz
    February 1, 2007

    My hypothesis: Nonscientists fail to grasp that scientists speak in terms of uncertainty rather than certainty.

    “More likely than not” is the linguistic equivalent of error bars.

    Yet, more likely than not, the phrase fails to convey what we mean.

  2. #2 John Fleck
    February 1, 2007

    Chris -

    If one assumes that Seth’s characterization is accurate (and I even hesitate to post at all, given that it breaks my own rule about waiting until we’ve actually seen the damn thing), what should the journalistic response be? I’ve argued endlessly that the IPCC-style consensus documents need to provide the foundation for our journalism. What if we have two that say two different things?

    I guess, in a sense, your final paragraph implies an answer?

  3. #3 Thom
    February 1, 2007

    Gotta’ love Pielke: “Third, the IPCC’s lead man on hurricanes and climate change is a fervent partisan in the debate itself.”

    Why is it that everybody but the man who, as the AP reported, the Republicans requested to debate politicization of science is either a “partisan” or an “advocate?”

  4. #4 Chris Mooney
    February 1, 2007

    Fred,
    Exactly.

    John,
    I don’t know if I have an answer, except that scientists need to be more conscious about how these statements are going to be interpreted.

    I posted a shortened version of this to Pielke’s blog:
    “both the WMO and (according to this report) the IPCC are saying there’s lots of uncertainty on this issue. And yet the two statements (again, if the latest report is correct) will now be pitted against each other. How much difference, expressed as a percent chance that the conclusion is true, is there between “more likely than not” (the IPCC, allegedly) and “evidence both for and against” (the WMO)?”

    So far no scientist has answered this question….

  5. #5 Dark Tent
    February 1, 2007

    J Fleck said: “I even hesitate to post at all, given that it breaks my own rule about waiting until we’ve actually seen the damn thing)”

    So, why not wait until the actual report comes out before even throwing out questions?

    By asking the question, you are implying that there may indeed be two conflicting consensus statements, which you can not conclude at all until you have seen the not-yet-released document.

    The “guess what the scientists will say” game only confuses the public.

  6. #6 Chris Mooney
    February 1, 2007

    Folks,
    If it’s in the media, I say it’s fair game to comment on. Once the reporters report their scoops, there is nothing wrong with us analyzing them, with the appropriate caveats (i.e., this report may be wrong).

  7. #7 John Fleck
    February 1, 2007

    Dark -

    Feel free to stay out of the discussion, then.

    I think this raises a fascinating question about journalistic methodology that is fascinating, regardless of the specifics of the final language. And Chris has the standing to comment usefully on the issue.

  8. #8 Dark Tent
    February 1, 2007

    John:

    I did not trealize that I had even entered the “discussion”.

    I merely commented on the fact that you mentioned your rule (which I think is a very good one) and then proceeded to ignore it.

    Judith Curry commented on Prometheus:

    “The second lesson is not to provoke trouble and controversy by prematurely reporting the contents of the IPCC assessment (business as usual for a journalist, but I would hope that scientists would hold off on commenting on this until the IPCC actually releases its statement).”

    That it may be “business as usual for journalists” does not make it right, even for them.

    After all, “journalistic” business as usual (not properly checking out sources and merely repeating hearsay) also got us into the huge mess in Iraq.

  9. #9 Roger Pielke, Jr.
    February 1, 2007

    Chris-

    The IPCC characterizes “about as likely as not” to indicate that the certainty is 33-66%. So “more likely than not” is >66% certainty.
    http://www.ipcc.ch/activity/uncertaintyguidancenote.pdf

    The WMO “evidence for and against” might be interpreted as 50-50, but perhaps the IPCC’s language is more precise 33-66%. Clearly, the WMO did not take a side in the discussion, so if accurately reproted, this does represent a difference in view.

    From a journalistic standpoint the IPCC is in a no win situation. If the report by the AP was in error then the IPCC will be seen as backing off. If the AP is correct, then the IPCC has misrepresented the state of scientific opinion.

  10. #10 Gerald Spezio
    February 1, 2007

    Has anybody read and/or genuflected in gratitude to Bayesian physicist Edwin T. Jaynes? Jaynes’s seminal PROBABILITY THEORY: THE LOGIC OF SCIENCE delivers what an increasing number of hard-nosed scientists consider to be THE very best answer to the problem of induction. Simply stated; “a probability is a measure of a state of knowledge and may change as this state is updated.” Just as historian of science, Niaomi Oreskes, reminds us in her recent WAPO article posted in an adjoining blog.

    Lest we forget, David Hume was literary man! There is an answer. It isn’t absolutely certain or infallible, but we can (and we had better) put a probabilistic number on it. Deductive syllogisms are wonderful, too, in their place.

    An earlier online version of Jaynes’s PROBABILITY THEORY is available here http://www-laplace.imag.fr/Jaynes/cc01p.pdf. If you only start with the preface, you will almost surely begin to see that probability theory need not suffer from endless Humean masturbation about induction. As Steven Weinberg thoughtfully suggests; we cannot solve the problem of induction by induction. We can, however, according to both Weinberg and Jaynes get on with our scientific searching just as Laplace and James Bernoulli postulated long ago. Tough Aussie philosopher, David Stove, also took Hume and his philosophical fog to the woodshed for a good thrashing.

  11. #11 Thom
    February 1, 2007

    Oh boy….

    Pielke: “[T]he IPCC is in a no win situation. If the report by the AP was in error then the IPCC will be seen as backing off. If the AP is correct, then the IPCC has misrepresented the state of scientific opinion.”

    I think I finally get it. Whenever an assesment or study agrees with the Pielke view of the world, then it’s fine. Otherwise, it’s “playing politics,” or “taking sides.”

    Well, if the IPCC comes out and Pielke chooses to advocate for the WMO opinion, then I see nothing wrong with that. But for others, it would probably be best to stick with the international assessment.

  12. #12 Chris Mooney
    February 1, 2007

    Roger if that’s true, then “more likely than not” is the same as “likely”…which in turn is the level of confidence that the IPCC ascribed to its 2001 conclusion attributing global warming to human causes. I don’t see how the attribution of hurricane changes to global warming, today, could be expressed with the same level of confidence that was used to describe the attribution of GW to human agency in 2001….so something doesn’t add up.

    Does anyone else get the sense that we had just better wait and see what the report actually says?

  13. #13 Roger Pielke Jr.
    February 1, 2007

    Chris-

    “Likely” = 66-90% (from the same IPCC document), so it is bounded on the up side.

    The report itself won’t be out until May, so the basis for the statement in the SPM (if true) won’t be known for a pretty long time.

  14. #14 Benny
    February 1, 2007

    Agree with Mooney. Just got an email from a friend in Paris and learned that Seth Borenstein is a bit obsessed with the hurricane thing. This report may be premature and things could change. Pielke is sort of losing it when he attempts to imply that a change would be seen as “backing off.”

    Scientists don’t sit around biting their fingers and worrying what happens to appear on the AP news wire. I doubt that any of them have even seen the story.

  15. #15 Chris Mooney
    February 1, 2007

    Folks,
    Andrew Dessler seems to agree with the original post, and doesn’t think the (presumed) IPCC statement and the WMO statement are as far apart as all that…
    http://gristmill.grist.org/story/2007/2/1/111155/3674

    So, for that matter, does Josh Rosenau
    http://scienceblogs.com/tfk/2007/02/ipcc_on_hurricanes.php

    As for me, I’m still scratching my head about what “more likely than not means”…it seems to me it has to be close to “the balance of the evidence” (line from the 1995 IPCC SAR)

  16. #17 Greg
    February 1, 2007

    “a public that doesn’t (and won’t) understand the nature of such knowledge”

    One need only read the exchanges above to obtain some sympathy for why the public *can’t* understand.

    As reported elsewhere, those who are paid to obfuscate have already their lies lined up. Indeed, some have chosen to attack pre-emptively.

    Those who misguidedly believe truth is not served unless it is over-shadowed by a generous dispute have their reports written, awaiting only a few snappy context-free quotes.

    No ethical public could possibly choose among the contradictory claims pressed by so many credible authorities.

  17. #18 Steve Bloom
    February 1, 2007

    The upshot is that “more likely than not” is a term used in the AP story and not in the AR4 (unless a new category has been invented and we haven’t heard about it). Even if Roger were correct in his hair-splitting, the AP story still would be forgotten about in a couple of days.

    Regarding the WMO statement, simply calling something a consensus doesn’t make it one. If it does, then I hereby declare that this comment is a consensus. Seriously, trying to place the WMO statement on the same level as the AR4 is laughable, and imagining that even if the AR4 turns out to consider AGW as “likely” having strengthened hurricanes there is any chance whatsoever of the WMO picking a fight over it is even more laughable (notwithstanding Roger’s efforts to talk himself into a job as referee).

    It’s rather telling that the AP story quoted Kerry Emanuel (a signer of the WMO statement) as apparently not finding a fundamental conflict between the AR4 and the statement. He apparently thinks that while the AR4 language is “strong,” it’s defensible.

    All of that said, the sea level issue is much more important.

  18. #19 marlowe johnson
    February 1, 2007

    I’m as puzzled as you are Chris. It’s the “as not” part that’s got me confused and it’s not referenced in the IPCC guidance document, and I while I disagree with Roger that it means the same thing as > 66% (i.e. “likely”), I do think that it’s unfortunate that they strayed from their own agreed upon terminology.

  19. #20 Chris Mooney
    February 1, 2007

    Folks, from Andy Revkin et al at the Times:

    “Scientists involved in the discussions said today that the U.S. delegation, led by political appointees, was pressing to play down language pointing to a link between intensification of hurricanes and warming caused by human activity.

    “They have tended to highlight uncertainties on certain issues,” a scientist involved in the negotiations said in an e-mail message sent to a reporter today. The scientist sent the message on the condition of anonymity.”

    I guess we’ll see what the doc says when we see it…in just a few hours.

  20. #22 Marlowe Johnson
    February 2, 2007

    Chris,

    FYI, the SPM says that “more likely than not” = >50% — so it looks like they’ve used the commonsense definition afterall…

  21. #23 Gerald Spezio
    February 2, 2007

    Many of us will always worship at Richard Feyman’s successfull science in isolating the “O” rings as the critical variable in the causal chain leading to the fatal Challenger space shuttle explosions. Would anybody nitpick now about ascribing a probability of 99+ % to the “O” ring hypothesis, demonstration, and subsequent extensive testing?

    Feynman made a final comment on the homicidal stupidity of the administrative decision to fly on that cold morning against the opposition of all fourteen Thiokol engineers. Feynman posited; “FOR A SUCCESSFUL TECHNOLOGY, REALITY MUST TAKE PRECEDENCE OVER PUBLIC RELATIONS, FOR NATURE CANNOT BE FOOLED.”

    In the name of science and the quest for truth, REFUSE TO BE MYSTIFIED! Some of the above linquistic micro-analysis and epistemic relativism makes me want to puke.

  22. #24 Chris Mooney
    February 2, 2007

    Folks,
    I’ve had a slow start this morning morning–tried to sleep in because I’ve been sick, and have some conference calls. But it looks like the hurricane part is going to be less controversial than we might have thought. Pielke, Jr., has more.

  23. #25 Gerald Spezio
    February 2, 2007

    Here is a better url for access to E. T. Jaynes, Probability Theory – The Logic of Science; http://www-laplace.imag.fr/ENGLISH/PRESENTATION/index.html

    I beg you to at least try the preamble. The sincerity of the man comes across in the sincerity of the writing. Mr. Bortz, are you conversant in Jaynes’s Bayesian probabilities?

  24. #26 Chris Mooney
    February 2, 2007

    As this post has been widely linked, I want to point readers to my follow up post, which just went up:
    http://scienceblogs.com/intersection/2007/02/the_ipcc_hurricanes_and_global.php

  25. #27 Dark Tent
    February 2, 2007

    “Those who misguidedly believe truth is not served unless it is over-shadowed by a generous dispute have their reports written, awaiting only a few snappy context-free quotes.”

    I love that, Greg.

    News is not news at all, just recycled Mad Libs.

    And, after all the empty, misguided speculation about the IPCC report, who could blame the IPCC scientists if they purposely planted a few bogus “consensus whispers” here and there to send the paparazi off chasing wild turkeys ?(in Turkey, of course)

    Don’t get me wrong. I’m not implying that anyone actually did this, but it certainly would be entertaining to watch.