29 October 2009

Will William Connolley do the Right Thing?

In an earlier post I had speculated -- naively and prematurely -- that one small point of dispute between Steve McIntyre and Michael Mann in the wars over the Hockey Stick might be resolved empirically to the satisfaction of all. After all, if I, a mere political scientist can understand the debate and also (and more importantly) see that others far more qualified than I understand the debate and that they have decided to resolve it in favor on McIntyre, then I thought that -- aha -- here we have an issue that can be unambiguously resolved.

As Lee Corso might say, no so fast my friend. William Connolley, formerly of the British Antarctic Survey and the Real Climate blog, steps in and uses my assertion of a resolution as an opportunity to smear me (blogging on climate is a contact sport, fair enough), but more importantly to pledge allegiance to Michael Mann. What is clear from the subsequent discussion spanning several blogs, however, is that Connolley doesn't understand the substantive debate at all, but he does know where his tribal allegiances lie. This is unfortunate because it reinforces the perception that the "Hockey Team" (as they apparently self-named themselves) will never concede a point, never admit to fallibility, and never break ranks. I have been wondering if Connolley will "do the right thing" which in this case would be to familiarize himself with the facts of the matter and render an independent judgment on McIntyre's claims. Connolley shows some signs of moving in his thinking. How far will he go? Where it ends up is the uncomfortable position of admitting that on this issue, McIntyre is right and Mann is wrong. He's even been given a gracious face-saving way out by Jean S, a Finnish statistician who concurs with McIntyre.

McIntyre helpfully explains the issue for Connolley in a new post.

Why does this small dispute matter? Lets be clear, it is not about the overall validity of various hockey stick reconstructions (which involve many, many other issues in dispute, which is why resolving them one-by-one is important). The issue matters because it speaks to trust and credibility, which are two important factors that non-experts use to evaluate the claims of experts. If members of the Hockey Team cannot or will not admit error when such an error is demonstrably shown to be the case (in a manner that even political scientists can understand), for better or worse, for many observers it will say a lot about their credibility on other matters. An important part of science is being about to admit when mistakes were made, to correct them, and move on to further work with an improved understanding. It is not about fighting for one's allies position regardless of what the facts say, there is another word for that -- politics.

Will William Connolley do the right thing? I expect so.


  1. Do you write about this problem in your book? I mean scientists taking a partisan approach. Using science as a stick in what is a political debate. What can we do as citizens?. Even cultivated people cannot follow the scientific details. To whom believe then?. I agree we can build our opinion based on the openess and honesty of the participants but then the criteria is purely social and political, that is not science. I do feel quite lost.

  2. -1-rjb

    I do.

    I also treat the issue more generally in my book The Honest Broker (see it linked to the left).

  3. As long as scientists are Homo sapiens (or is it Homo tribeans?), this is what science will always was and will be- unless there are sufficient countervailing forces to move toward the ideal.
    There is some hypocrisy in claiming that science knowledge claims deserve special consideration due to their objectivity and then not behaving that way.
    Perhaps we're really "Homo hypocritus."

  4. When scientific reputations and careers are at stake reinforcing failure makes perfect sense, though not perfect science.

  5. Judging by Connolley's comments, he is being deliberately obtuse. I can't believe it is a case of not understanding.

    He is refusing to budge from his postion on the basis that the orientation of the data is simply a linear transformation and hence perfectly acceptable.

    But I am sure he knows that when you bring the model back into the real world for interpretation, such a transformation implies that up is down in the physical world.

    Sort of like taking a photgraph with your camera upsidedown and then showing the resultant image as proof that the sky is green and the field is blue.

  6. The Team cannot back down now. They've gone too far out on the limb to come in again. Acknowledging the details of their errors would be simple enough, but the personal attacks on McIntyre can't be dealt with so easily. They've staked their integrity as human beings on McIntyre being a total fraud - you can't just say "Oops!" to something like that.

    It's a cliche, but none the less true - it's not the crime, it's the cover-up. The Team is All In, as they say in poker. They can only brazen it out, they can't back down. To do so now would be utter humiliation.

  7. #5 and #6: I don't think Connolley did this on purpose.

    He has a very strong predisposition towards believing that Mann is right and McIntyre is wrong and it tripped him up in embarrassing fashion.

    The reason why Connolley is getting targeted like this, is because based on his past record (being an intellectually honest individual), it is genuinely expected that he will eventually acknowledge McIntyre's point.

    Nobody would expect this of Mann or Romm.

  8. One would be tempted to say who cares what Connolley thinks? Except that apparently he was Brad DeLongs chosen expert on climate science and in that role had reprimanded Levitt & Dubner for not choosing the "right" climate scientist to talk to. Why of course DeLong should trust people who don't know up from down - that would be pretty much standard behavior for an economist in todays loony upside-down world where we cure a problem that was caused by too much debt by getting into more debt and the same people who caused the problem are given the job of fixing it.

    Like Sharon I wouldn't call this politics though, it's typical academic hubris and it's far from the only area where academics will argue black is white or up is down or will refuse to admit to an error; it's just the most visible one at the moment.

  9. Conolley is indeed being deliberately obtuse. The statement, ""Multivariate regression methods are insensitive to the sign of predictors", is true and entirely beside the point.
    For example, imagine one were trying to relate the consumption of electricity to the growth in population in small town. One would expect that an increase in the population would increase the demand for electricity and vice-versa. When one runs an regression one would likely find that population growth is positively correlated with electricity demand.
    Now imagine that when entering data on population growth, one mistakenly flips the sign so that positive population growth is entered as a negative number and negative population growth is entered as a positive number.
    This will not make a difference in the outcome of the regression because the mistake will be cancelled out when the regression is run. Now the regression will show that there is negative correlation between population growth and electricity consumption. As long as one continues to flip the sign on population growth when one makes a prediction, the results will come out the same.
    This is what I understand Connolley to be saying at his blog when he writes:
    "The sign going in is the point at issue here. If the sign going in is wrong, then the correlation has the same absolute value but the wrong sign, which means the series gets added back in to the reconstruction with the sign wrong twice, ie correctly."
    This has nothing to do with error the SM was identifying and it is a red-herring to suggest that it does.
    Amazingly, he goes on to write:
    "If by "the correlation is aphysical" you mean, of the wrong sign, then you are correct. BUt this is irrelevant"
    In fact, what he calls "irrelevant" is exactly what SM is identifying as the problem.
    Connolley is playing games.

  10. Just an addition to the previous post, Connolley writes, "If by "the correlation is aphysical" you mean, of the wrong sign, then you are correct. BUt this is irrelevant"
    This requires a little more discussion. SM is indeed arguing that the relationship between the proxy and temperature is "aphysical", but he means something more than simply "wrong sign". He is saying that because of distortion during the calibration period the relationship established by the regression is the opposite of the physical relationship that actually existed in the reconstruction period.

  11. William in comment 3:

    "[There are two possibilities here. The charitiable one is that Roger just didn't understand what is going on. The rather less attractive possbility is that he did understand (the third, that I'm wrong, can of course not credibly be entertained). I'm going to go with the first until proved otherwise. I think by now it is rather likely that he does understand that his original post was a mistake but that he will refuse to admit error -W]"


    William in comment 25:

    "Aha! Thank you. This is the first time someone has made a coherent arguement over this (perhaps it has been said before but lost in the noise, if so my apologies for missing it)."


    It will be interesting to see if "his original post was a mistake but that he will refuse to admit error".

  12. From Mann's supporting information:


    What it shows is that, contrary to Mcs insinuations, Mann et al knew about the data quality issues in Tijinder and countered them as best they were able.

    So Steve Mc has been doing nothing but re-picking a nit Mann et al origonally picked over a year ago.

    And we all know Stoat was right on the narrow math issue, right?

    "Potential data quality problems.

    In addition to checking whether or not potential problems specific to tree-ring data have any significant impact on our reconstructions in earlier centuries (see Fig. S7), we also examined whether or not potential problems noted for several records (see Dataset S1 for details) might compromise the reconstructions. These records include the four Tijander et al. (12) series used (see Fig. S9) for which the original authors note that human effects over the past few centuries
    unrelated to climate might impact records (the original paper
    states ‘‘Natural variability in the sediment record was disrupted
    by increased human impact in the catchment area at A.D. 1720.’’
    and later, ‘‘In the case of Lake Korttajarvi it is a demanding task
    to calibrate the physical varve data we have collected against
    meteorological data, because human impacts have distorted the
    natural signal to varying extents’’). These issues are particularly
    significant because there are few proxy records, particularly in
    the temperature-screened dataset (see Fig. S9), available back
    through the 9th century. The Tijander et al. series constitute 4
    of the 15 available Northern Hemisphere records before that

  13. -12-bigcitylib

    Thanks for this comment. Should I ever need evidence to support the claim that there are "deniers" on both sides of this issue I now have this to point to ;-)

    Still waiting for William to do the right thing.

  14. bigcitylib #12 --

    > [The Mann et al SI shows that] contrary to [McIntyre's] insinuations, Mann et al knew about the data quality issues in Tijinder and countered them as best they were able.


    The SI doesn't show that, in my (lay) opinion. Mann et al's handling of the Tijander varve proxies is the topic under discussion.

    > we all know Stoat was right on the narrow math issue, right?


    On certain narrow issues, he's right. Which one did you mean? Whether his view of the math issues involved in treatment of the Tijander varve proxies by Mann et al is correct, is one of the subtopics being discussed.

  15. Amac,

    "Mann et al's handling of the Tijander varve proxies is the topic under discussion."

    And the state of the discussion is Steve Mcs insinuation that Mann had no clue as to the data quality issues with Tijander. That Mann was basically stupid, put in data upside down and etc. The SI shows this to be a false insinuation. Whether Mann et al were able to correct sufficiently for the problems with the Tijander proxies ( using methods given in the SI) is an entirely different question from whether they were aware of them.

    The only question now is, will Pielke Jr. do the right thing?

  16. And Roger Jr., your last comment veers rather too closely to outright gibberish.

  17. -15-bigcitylib

    aka MJ Murphy from Toronto

    You are welcome to comment here, and to try your best to distract with strawmen and other techniques from the issues being discussed, but do know that your credibility is exactly zero, e.g.,


    I've been waiting for you to do the right thing since that strange outburst of yours not long ago.

  18. You are giving Real Climate, Connolly and The Hockey Team far too much benefit of the doubt. Despite having Ph.D.'s they are not serious scientists but propagandists who are being generously financially rewarded for their activities. They will never admit that their science is wrong because they are not about right and wrong. They are about politics and religion. This should be obvious in 2009 to any rational informed observer.

  19. Roger Jr.,

    Then why don't we do a swap? I'll apologize for my intemperate language, and you apologize for lending credence to Steve MCs latest defamatory silliness.

  20. -17-Roger sez:

    "bigcitylib... know that your credibility is exactly zero"

    I second that motion!

    But, Roger is -- as usual -- far too generous.

    I would rate BigCityLib's credibility well below zero.

  21. bigcitylib #15 --

    Your paraphrase of McIntyre's stance is likely quite wrong.

    What I think: Mann et al were clearly aware of some issues with the Lake Korttajarvi series post-1720, because (as you point out) they discuss them in SI Methods, pg. 2.

    I doubt they were aware of the extent of those problems.

    Despite the SI Methods text, I doubt they realized that human-activity effects during the calibration period (1850-1995) had the effect of causing their CPS algorithm to select the varve-quality-to-temperature correlation that was the reverse of (i.e. upside-down with respect to) the scheme that Tiljander used in her paper.

    If they had recognized this contrarian pattern, I think they would have investigated further, and either cut the Lake Korttajarvi series from the reconstruction, or picked a different calibration period. At the least, they would have mentioned the issue in the Methods.

    I suspect that McIntyre would largely agree with what I've written. But, again, I could be wrong.

  22. To paraphrase Big City Lib says whether Mann was able to correct the problems in the proxies is entirely different from the question whether he was aware of them.

    The difficulty is that the only way to deal with problems with the Tijander proxies is to discard them. If you don't understand this, you don't understand the problem. If Mann were fully aware of the problems and used the Tijander proxies anyway, this would be a more damning allegation against him than the allegation that he was merely careless or incompetant.

  23. Mann et al. were "aware" of the issues. But what's the point? They don't "matter". Throw out the varves, toss in bristlecones. Throw both those out and toss in Arabian Sea G Bulloides. The methodology just needs something stick like to get the shape. So Mann can do a "sensitivity test" just throwing in another garbage stick creating proxy, and the "problematic" proxies "don't matter".

  24. bcl: You remind me of some kids I knew growing up who thought that professional wrestling was a real sport because it went through many of the motions of a real sport. Mann's papers are endlessly fascinating (and appalling) to me because they often acknowledge issues like this in the text, but then completely ignore them in the actual analysis -- this being simply one case in point.

    McIntyre's point is simply that Mann ended up using the Tiljander proxy in the opposite sense from what Tiljander presented. It does not matter that Mann stated that he looked at the issue -- he still ended up using the proxy in the opposite sense from that in the original study (with absolutely no explanation for why Tiljander was wrong).

    Another telling example from the same paper. In the introduction, Mann states that a key motivation for the paper was to respond to concerns of the NAS review panel. This panel said that stripbark bristlecone pine tree rings were too problematic to use, but Mann used them anyway in the paper!

  25. Big City Lib is just wrong. Steve McIntyre's original comment on Oct 2 2008 accuses Mann of not being aware of how his methods used the varve data, not of being unaware of the problems.

  26. Will William Connolley do the right thing? I just wrote this at Climate Skeptic:
    Michael Mann is the perfect case study for the adage — “a little bit of knowledge is a dangerous thing.”

    Leaving him unsupervised with statistics is more dangerous than an unsupervised child with a loaded gun — more people get hurt.

    Doing the right thing here would be to take statistics out of Mann's hands to stop the damage (to science, to public policy, to the poor, to the world). Somebody needs to put a stop to it. I'm not sure Connolley will or even that he can.

  27. In my experience with him on wikipedia he has never admited a mistake. So it would be the smart thing to do, but he won't. His ego gets in the way.

  28. Postscript.

    A few of the recent emails hacked from the University of East Anglia’s Climate Research Unit--and now acknowledged as genuine--are relevant to the topic of this post.

    I submitted the following to the comment thread of Delayed Oscillator's "Talk about Tiljander" post of November 3, 2009 at 4:41 AM. It entered the moderation queue in postion to be Comment #9 on November 21, 2009 at 1:33 PM EST. I also placed it in the queue at Climate Audit's Connolley Endorses Upside-Down Mann.

    Delayed Oscillator -- Your enthusiasm for Stoat's handling of the Tiljander proxies might have some baiss, after all.

    Recently-added Comment #39 to "Tiljander, again" (see link in the first comment here [link]) directs to an alleged series of emails on the Upside-Down X-Ray Density Varve Proxy.

    Bolding added.

    - - - - - - - - - -

    On Thu, Sep 3, 2009 at 4:45 PM, Nick McKay wrote:

    Hi all,

    I haven't checked the original reference [Tiljander et al., 2003] for it's interpretation, but I checked the code and we did use it in the orientation that he [McIntyre] stated. He's also right that flipping doesn't affect any of the conclusions. Actually, flipping it makes it fit in better with the 1900-year trend.

    I've attached a plot of the original, and another with Korttajarvi flipped.


    - - - - - - - - - -

    ... continues...

  29. ... continuing from above...

    On Sep 4, 2009, at 5:24 PM, Nick McKay wrote:

    The Korttajarvi record was oriented in the reconstruction in the way that McIntyre said.

    I took a look at the original reference - the temperature proxy we looked at is x-ray density, which the author interprets to be inversely related to temperature. We had higher values as warmer in the reconstruction, so it looks to me like we got it wrong, unless we decided to reinterpret the record which I don't remember.
    Darrell, does this sound right to you?

    This dataset is truncated at 1800, so it doesn't enter the calibration, nor does it affect the recent warming trend.

    The attached plot (same as before) shows the effect of re-orienting the record on the reconstruction. It doesn't change any of our major or minor interpretations of course.


    - - - - - - - - - -

    From: Darrell Kaufman
    To: Nick McKay, Caspar Ammann, David Schneider, Jonathan Overpeck, Bette L. Otto-Bliesner, Raymond Bradley, Miller Giff, Bo Vinther, Keith Briffa
    Subject: Arctic2k update?
    Date: Sat, 5 Sep 2009 08:44:19 -0700


    I received my first hate mail this AM, which helped me to realize that I shouldn't be wasting time reading the blogs.

    Regarding the "upside down man", as Nick's plot shows, when flipped, the Korttajarvi series has little impact on the overall reconstructions. Also, the series was not included in the calibration. Nonetheless, it's unfortunate that I flipped the Korttajarvi data. We used the density data as the temperature proxy, as recommended to me by Antii Ojala (co-author of the original work). It's weakly inversely related to organic matter content. I should have used the inverse of density as the temperature proxy. I probably got confused by the fact that the 20th century shows very high density values and I inadvertently equated that directly with temperature.

    This is new territory for me, but not acknowledging an error might come back to bite us. I suggest that we nip it in the bud and write a brief update showing the corrected composite (Nick's graph) and post it to RealClimate. Do you all agree?

    There's other criticisms that have come up by McIntyre's group:

    [Snipped Paragraphs #1 through #5 on matters unrelated to the Lake Korttajarvi varve proxy.]

    Please let me -- better yet, the entire group -- know whether you think we should post a revision on RealScience, and whether we should include a reply to other criticism (1 through 5 above). I'm also thinking that I should write to Ojala and Tiljander directly to apologize for inadvertently reversing their data.

    Other thoughts or advise?


  30. My error in suggesting at 11:59am, supra, that any of the hacked emails in particular have been "acknowledged as genuine."

    RealClimate.org quotes the University of East Anglia as stating, "Because of the volume of this information we cannot currently confirm that all of this material is genuine."

    Thus, caveat lector.