Is Nick Stokes Trying to Cover up his Mistake?

There’s been a recent chain of blog posts discussing old issues regarding Michael Mann’s infamous hockey stick debate which I played some role in triggering (examples here, here and here). I’ve commented on some of them, but I’m not going to discuss any of their topics here. For this post, the arguments in those post are irrelevant.

What I want to do is see if we can establish one particularly vociferous participant, Nick Stokes, has dishonestly covered up an error he made in a figure he posted after I pointed it out.

The figure in question is the first figure of this post, displayed below:

10-1-Stokes-Altered

When I first saw this image, I was confused by what appeared to be a discrepancy between it and another image displayed in the very same post. After spending a little time trying to figure things out, I posted a comment at Climate Audit saying:

Steve, I have a question. Nick Stokes has a (in my opinion ridiculous) post up which uses an emulation of your MM05 Figure 1 (his emulation was posted . The version he posts has some difference though. Notably, his emulation begins at ~.175 whereas your begins at slightly over .2, and his ends at a higher point than yours. The effect is his emulation shows the most recent part of the reconstructed series is unprecendented whereas yours shows the first 30 years surpass it.

Do you know what causes this discrepancy? My assumption right now is it’s a matter of smoothing, but I’m not sure what smoothing would produce which results.

If you read that comment, it is clear I am discussing something different than the image currently displayed in Nick Stokes’s post. The current figure starts at slightly over .2, not at the ~.175 I claimed. This causes the first 30 or so years of his graph to be higher than the most recent portions of it. It’s difficult to see how I could have simply imagined the numbers I gave. This is especially true since after posting the comment above, I began writing a comment directed at Stokes (which I hadn’t posted as I wanted to do some more testing first):

Nick Stokes, the graph you say you “think this… become a very inconvenient graph” is not the graph you claim it is. You claim to have shown the plot as an appendix in your previous post, and you repeat the same plot below this graph, but but that plot does not match the graph you lead with in this post.

In both the previous figure and your current lead, “MBH Decentered” begins at about -.1. In your current lead, Centered PCA begins at about .175, but the previous figure showed it beginning at a bit above .2. This change has a noticeable effect on the modern/past differential. In the previous figure, the first ~30 years of the reconstruction are higher than any other. In your current version, the most recent years are higher than any past temperatures.

My word is enough to accuse someone of dishonesty, but it is enough to make people look a little closer. I didn’t save a copy of the image/page I saw, and I haven’t found such in any web archives. As such, we do not have direct proof. We do, however, have compelling evidence in the form of this graph:

10-1-EE1

Which Stokes claims is his emulation of a figure created by Steve McIntyre and Ross McKitrick:

10-1-EE0

But as you can see, the bottom panels in the two figures are different. The discrepancy I claim to have seen in Stokes’s latest post is exactly the one seen in these two figures. This is true despite the fact both are shown in the same post, one after the other. Moreover, that post has another highly relevant figure:

10-1-EE2

As you can see, that figure is very similar to the one I claim Stokes has surreptitiously altered. There’s an additional line and the color scheme is different, but otherwise, they’re practically the same figure. The “MBH Decentered” line and the “Centered PCA” lines are almost identical. The only difference I can find in them is the old figure of his begins at ~.175, making it lower than the modern portion of the graph. That’s the exact difference I claim existed in the figure shown in Stokes’s recent post.

So what should we believe? Should we believe I completely imagined having seen two different versions of that figure? Should we believe that figure was always what it is now even though it contradicts previous versions of the same figure he’s posted in the past? Or should we believe Nick Stokes read my comment highlighting a discrepancy in his results, fixed the error then updated the figure without telling anyone in order to cover up his mistake?

I’m going with the latter. He commented on the same page I pointed out the discrepancy in his results. Heck, he commented in the same discussion fork. If my comment was completely delusional, I think he would have said something about it.

By the way, the title of the post I claim this deception happened in is, “What Steve McIntyre won’t show you – now.” It apparently included an image Nick Stokes won’t show you… now.

Advertisements

11 comments

  1. Brandon,
    Someone commented in an email that the lines were indistinct. I upped the line width of the smooths from 2 to 3. That was the only change. I think in your comment you said that the plot corresponds to the numbers in the file that you have.

  2. Brandon,
    You posted this earlier:
    “Posted Sep 30, 2014 at 2:50 PM | Permalink | Reply
    Hrm. I’ve plotted the data series from this file, and I get results about the same as those in Stokes’s figure. It’s mostly similar to what you showed in MM05EE, but there is definitely something different about the endpoints. I’ll try looking through your code to see what smoothing you used in the display.”

    Still true?

  3. Nick Stokes, I’ll repeat what I said at Climate Audit regarding your portrayal of what happened then discuss your new idea separately:

    Brandon,
    Someone commented in an email that the lines were indistinct. I upped the line width of the smooths from 2 to 3. That was the only change.

    To which I said:

    Nick Stokes, you’re free to claim that. I believe the evidence strongly suggests otherwise. I’ve shown the discrepancy exists in past versions of your same “emulation,” including a figure which is a nearly direct predecessor to the one in question. The only way I can see your claim being believable is if you’ve posted a note somewhere informing people you’ve updated your emulation to correct for some problem.

    I don’t even know how I could be wrong about this. I only looked for figures of your “emulation” after I saw a discrepancy. If the discrepancy was purely a figment of my imagination, how did I happen to find the exact same one in the previous versions of your figures? And what, I just happened to correctly call that the figure had been changed, but somehow was completely wrong about what the change was?

    That’s pretty implausible. Show us your updates which informed your readers you changed your “emulation” so they should focus on the new results, not the old ones. Then maybe it will seem a bit more believable.

    I think in your comment you said that the plot corresponds to the numbers in the file that you have.

    I never said the figure you currently display matches up with anything because I was discussing a different figure. Even if you think that different figure was a product of my imagination, you cannot possibly claim me believing it matches a data file’s contents in any way supports the figure you show now.

  4. Nick Stokes, I don’t intend to answer random questions you ask that divert the discussion from the central issue, that all evidence points to you having switched out graphs.

    If you show where you’ve previously explained the difference in your “emulations” so we can see it didn’t happen like I claim, or you acknowledge I’m right about what you did, then I’ll discuss the details of how the figure should appear.

    But if we can’t agree on what figure you’ve displayed, I’m not going to try to reach an agreement on what figure ought to be displayed.

  5. Brandon,
    I’ve checked, and I think you are right. The version you first saw was one that I made a couple of days ago and linked to in a comment at CA. I still have a copy and I have posted it here. It’s the version that agreed with your file.

    As you have noted, I’m basically using the graphics code from that earlier post which compared with and without Gaspe 1400-1450. I think the earlier version was with Gaspe, which agrees with your file. The current version is without, which agrees with MM2005, as you can see. I think the earlier version is actually the right one, since it shows the difference solely due to decentering. But the current one agrees with MM2005, who showed the combined efect. The difference has no relevance to the point I am making about recent hockeysticks.

    And yes, I don’t know how I switched. I’ll check. As I complained in that earlier post, it’s rather indirectly determined in the code.

    I think you could try to be less excitable, though.

  6. Nick Stokes:

    I’ve checked, and I think you are right. The version you first saw was one that I made a couple of days ago and linked to in a comment at CA. I still have a copy and I have posted it here. It’s the version that agreed with your file.

    I actually figured out where my confusion came from once I went back and looked at the data files I had loaded. There were two files showing the data for MM05EE Figure 1. One file had a column to show the effect of removing NOAMER (but not Gaspe) even though it wasn’t shown in the paper. The other file did not have that column.

    There being a different numbers of columns in the two files caused me to plot the wrong series. I thought I was plotting data from one file when I was really plotting data from the other.

    I think you could try to be less excitable, though.

    You insisted I was wrong when had you even looked at the figure you originally plotted, you would have known I was right. I’m willing to wager if I hadn’t written this post, you wouldn’t have admitted your mistake. You would have just insisted:

    Someone commented in an email that the lines were indistinct. I upped the line width of the smooths from 2 to 3. That was the only change.

    This wasn’t a matter of me being excitable. It was a matter of me being tired of you making bold statements which are obviously untrue because you refuse to even consider views you don’t like and taking the opportunity to hammer you for it.

    Plus, the timing is still suspiciously coincidental. You kept the same mistake for over six months. I point it out, and within hours, you fix it. It strains credulity.

  7. Oh, I forgot to point out the thing which made me decide to turn this into a post. I was really on the fence until I reread the post and saw:

    I think this has become a very inconvenient graph.

    There was just something too silly about calling something “What Steve McIntyre won’t show you – now” and suggesting it is “a very inconvenient graph” all while hiding an erroneous graph. The silliness compounded with your reaction, and it may grow even further depending on what (if any) updates you make to your posts because of this.

    Your post was stupid through and through, but stupidity is generally boring. Absurdity like seen in this case is much more interesting.

  8. Nick Stokes has added an update which misrepresents things:

    Update – this graph changed slightly, as Brandon noted. It comes from an earlier post comparing what happens in 1400-1450 with and without Gaspe cedars. The first posted was with, which corresponded to MBH, and so is more appropriate, in my view. The second agrees, as you can see, with M&M. Since the difference doesn’t affect my point, I’ll let that stand. The change was not intended – I was widening the smoothed curves for visibility. The first version is here.

    The earlier post does not just compare what happens “with and without Gaspe cedars.” It does give some focus to that, but it also shows another issue:

    The bottom panel shows the effect of changing the offest mean to centered. I’ll say more about that in a future post.
    Now I’ll plot the same data but superimposed….
    I’ve made the panel 3 plot faint, since it isn’t the current subject.

    Stokes claims to have emulated the figure by Steve McIntyre and Ross McKitrick, but his third panel is clearly different than theirs. He knows this as he’s acknowledged the point in reference to his recent post. He offers no explanation as to why, despite knowing his earlier post has errors in it, he’s not adding any sort of note or disclaimer to the post. Instead, he pretends the old post simply didn’t cover the issue where he messed up.

    Additionally, Stokes says:

    The first posted was with, which corresponded to MBH, and so is more appropriate, in my view. The second agrees, as you can see, with M&M.

    Which is difficult to interpret as there’s apparently a word missing after “was with.” While we may not be able to tell just what Stokes was trying to say, we can be fairly certain it’s not true as there was nothing “which corresponded to MBH” in his first version. All he did was show the effect of fixing the PCA calculation and not extending the Gaspe series separately, choosing not to show the combined effect.

  9. Not the only misleading graph he has had up at his site. Zeke was doing adjustments at Lucia’s and put up a graph of USHCN station numbers. Nick linked to his site where the graph only went up to year 2000. Pointed this out and it changed to near current ? 2012. Then a few weeks later and in subsequent posts he went back to using the 2000 misleading graph again. What gives with this guy’s honesty meter?

  10. Angech, I find it fascinating Nick Stokes didn’t go back and add a note to his old post where he first made the mistake. That means a month from now, people are just as likely to see the post with only the mistake as they are the post with the note about it.

    I really don’t get it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s