Saturday, May 23, 2009

Anyone read this yet?

I was in Barnes & Noble yesterday evening and saw a copy of Plastic Fantastic in their science section. This is Eugenie Reich's telling of the Schön saga. Anyone out there had a chance to read this yet? Steve? Don? I'll have to pick up a copy at some point.


Don Monroe said...

Yes, I read it. I think it's quite good. I think I'll just jump right in, assuming people know the history.

Having read the book, I feel like I understand better what Hendrik thought he was doing--although still not perfectly. The book convincingly documents the idea that Hendrik's great results were crafted precisely to match people's expectations, and that's a big part of why he got as far as he did. There's a great quote from Peter Littlewood on this point: "I didn't believe it could have been fraud because I didn't believe one person could make all that up. Then I realized, we all made it up."

It's worth noting that she never got a full accounting from Hendrik himself, although not for lack of trying. But it would be hard to know what to believe from him in any case.

A couple of things in this book that I've not seen elsewhere are some brief discussions of postdocs and graduate students whose careers were seriously screwed up (very important), and a little more discussion (but not enough) of what the journals did. The description of the events in April and May of 2002 that led to the copying accusations were also very detailed and very interesting to me.

Another thing that probably hasn't been published before was the discussions I had with Bell Labs management in fall of 2001. Of course, I knew all about that part already, so there weren't many surprises. Ultimately, I'm afraid I backed down prematurely--if I'd persevered it might have ended half a year earlier. Interestingly, these "SAMFET" results were not very well constructed to match expectations (they violated many known principles) which led to a lot of pushback, both from inside and outside of Bell Labs.

Overall, I think Eugenie was quite fair to the people concerned, although I'd guess some of the people who made bigger mistakes than I did feel differently. I wish that more of the managers had consented to interviews. I'm sure they felt it was a no-win situation, and perhaps on a personal level that's true. But the community would have been better served by getting more insight into their struggles with how to deal with Hendrik's amazing string of results. It can't have been easy, and others in the future will face similar challenges.

As it is, Eugenie comes down hardest on Cherry, for her explicit later claim that fraud had never occurred to anyone until the copying accusations came out. Unfortunately, we know that's simply not true, even if the claims had been suspended. Since Cherry probably had little knowledge of the events at the time, that may be a bit harsh, but I agree that the claim was revisionist history. Other managers really had more responsibility for how events actually occurred, though.

The book skips the whole investigating committee part of the story, which is fine with me. As a member of the committee, I'd like to think that it's because we did such a good job describing both our conclusions and our process that it would have been redundant. but who knows? Eugenie implies that the formation of the committee and the open dissemination of our conclusions was shrewdly designed to give the best outcome for Bell Labs. In the end, that's probably true, but at the time it wasn't so clear, at least to me, and I felt that the Bell Labs managers were sincerely trying to do the right thing for science.

It's hard for me to know how interesting this book will be for an ordinary person. For those of us that followed it at the time, or who know some of the players, it's a real page-turner.


Don Monroe said...

By the way, Steve posted a more mixed review and promises more to come.

Doug Natelson said...

Thanks, Don. I skimmed a few parts, and what I read was well-written and interesting. I happened to read a particular section that really came down hard on one of the managers in question. I also know of another manager who reportedly said things like, "Well, even if the SAMFET stuff is flawed, he's a genius." So, given that, I'm curious about Steve's forthcoming defense of management on this. Clearly I need to buy a copy and read the whole thing.

Anonymous said...

Just finished it.

The book was a quick, engaging read that did help shed some light on the whole affair and what Schön might have thought he was doing.

That said - and I may be a snob - I found it only passably written; perhaps B work. Also, the descriptions of the science were frequently sorely lacking, using analogies that didn't quite work. I also thought in several instances, the author made unsupported judgments regarding the integrity or motives of some of the Bell managers and then overstated her case.

Overall, a decent read, but not a particularly distinguished piece of writing.

One question I'd love an answer to: how many devices or measurements did Schön actually make? Did he fabricate a device, measure it, and then fabricate results or was he in some cases simply writing the paper?

Steve said...

Hi Don,

I was hoping you would ring in on this matter.

I really need to read the book with a fine-toothed comb to make my case carefully that she is unfair to management. Perhaps I am predisposed to this viewpoint as I found myself defending management one way or the other as the events unfolded. But take the case of Cherry that you brought up, as this was one of the cases where I really wanted to object to Reich’s treatment.

First, if you are in high management of a government lab, like Cherry (or Eric) was, there is no question that you are not going to grant an interview. Some journalists are not intending to be impartial, and although Reich made a reasonably honest attempt to be fair, a dishonest journalist will take an interview and cut quotes in a way to make you look like a crook or an idiot. Not knowing in advance what Reich’s intentions are, it is pretty clear that Cherry was not going to grant an interview.

The one line that Reich quotes several times is from the nature materials article (April 2003) coauthored by Saswato:

“In neither case did it occur to his managers that deliberate fraud might have taken place.”

Here is a larger excerpt

“… in two cases, members of technical staff and management discovered anomalies in Schön’s work. In October 2001, statistics in one of the papers seemed too good, and in April 2002, identical curves were noticed in two papers. When confronted, Schön produced explanations on each occasion that seemed reasonable, though it made his work look very sloppy. The management was upset with his sloppiness and lack of accurate record keeping, and demanded that he send in an explanation to the journals. In neither case did it occur to his managers that deliberate fraud might have taken place.”

I think there is a poor choice of wording here… but the fact that he was “confronted” sounds to me pretty clear that the possibility of fraud had been raised at least implicitly. Schön produced an explanation and the possibility was then dismissed. I certainly would have preferred it if the final sentence above read

“Once reasonable explanation had been given for the two anomalies in his work, the proposal that deliberate fraud had taken place was evaluated and concluded to be unlikely. “

These two sentences (the actual one and my proposed replacement) are not very different and Reich tried to hang Cherry for it. The phrase “Once reasonable explanation had been given” is absent from the direct quote, but in the full paragraph it seems almost implied. “Did not occur to his managers” is too strong a statement, because any thinking person will try to consider all possibilities, but it was not taken seriously at the time because the excuses seemed honest.

From the way Cherry is quoted “I stand by the statements of that article” it sounds to me like Reich tried to ask Cherry more about it and Cherry simply said “I stand by the statements, end of interview.” (see above comment about not talking to journalists).

Melissa said...

I read the book and wrote a bit about my reaction. I'm significantly more distanced from the situation than some of the other commenters, coming at it from the perspective of someone who started her graduate career trying to chase some of these results.

Don Monroe said...

I see your point, but I don't buy it.

To me the phrase, "In neither case did it occur..." is clearly intended to suggest that "at no point during the entire discussion did it occur." If it occurred once but was later dismissed, then it still occurred. The context does not explicitly say that fraud was discussed in the beginning (actually, it quickly became the only serious contender discussed for the SAMFET statistics). So the wording is very misleading, if an inferrence that fraud was raised is meant to trump an explicit statement that it was not.

The phrasing clearly suggests that the first hint of fraud was in the spring, and that's just not true.

I'm not second guessing the decision not to be interviewed, but your post seemed to me to suggest that it was a failing on Reich's part.

Steve said...

(Another thing that I am now finding annoying about the book is that it is very badly indexed. I can’t find the pieces I want except by thumbing through page by page, which is very tiresome).

Don: Valid points.

I admit that this one sentence is certainly not correct the way it is written. No doubt Cherry should have chosen words more carefully. But in the context of the rest of the paragraph I still think there is of room to interpret the statement as a badly phrased attempt to state what I describe above.

Why I blame Reich for taking Cherry to task for this? Because Cherry did not have a real chance to clarify this statement (whether it was her decision not to be interviewed or her management’s I do not know). Reich took her decision to “plead the fifth” as evidence of guilt – which it should not be.

At any rate, this was far from the only thing in the book that made me think that Reich was looking for a bad guy (besides Schon), and the Bell management was an easy one to take aim at.

Carl Brannen said...

Okay, I'll buy it.

I was peripherally involved in the famous incident where Joe Weber kept discovering gravity waves that weren't there; I took his class in general relativity at U. Cal., Irvine. So I found the book that resulted from this, "Gravity's Shadow," and found it absolutely fascinating.

Joe Weber was a wonderful guy. I don't think there were many accusations of fraud, he just didn't do his statistics right.

Doug Natelson said...

Alright - I'll try to buy a copy tomorrow and read it so that I can comment intelligently about this. Anonymous, all I know is that in the two years I was a postdoc at Bell, overlapping with Schon's most "productive" period, I never saw him sign out a liquid helium dewar, and I never saw him in the cleanroom doing processing. At the time, the claim was that a lot of the measurements were being done at Constanz. In hindsight, I think that the vast majority of all of his "data" was fictional.

Anonymous said...

I haven't read the book, but I remember seeing Batlogg give a talk fairly early on before too many questions were being asked. Even then the results were pretty stunning. I always felt that Schon's co-authors got off pretty lightly. I am a theorist, but if my junior collaborator (in the same lab) were producing amazing results, I would want to see the experiment in action. That seems like an important quality control step that should have happened before multiple papers were sent to Nature and Science and doesn't seem to be emphasized so much in discussions of the affair. Does this get covered in the book at all?

Doug Natelson said...

Ok - I read it. I'll write something more in-depth in the next day or two, after my blood pressure relaxes to its normal level. After seven years it still makes me simultaneously angry and disappointed.

Anonymous said...

Re: Batlogg

Bertram Batlogg is very smart guy, but he had been working with junior collaborators for many years in a mode when he would mostly come to the lab to get new data before flying off for a different conference. And so I find it completely plausible that he never saw a lab notebook.

But my suspicion is that he knew something was not quite correct from early on. Although what he knew and when he knew it is not clear.

I remember asking him at a conference in 2000 about access to samples, because I had a number of ideas to do such and such measurement on them. He replied that they were going to do them themselves. This was impossible, but I replied, "well what about XXX measurement?" And he again replied that they had their own plans... It wasn't the answer itself, but the mannerism of how he replied that struck me as odd. He was very guarded and it was clear that they weren't letting anyone else near the samples. He also said that it was impossible at any rate, because they always pushed the devices to failure ultimately. I asked "All the devices?!", and he replied "100% of them". "Why not just test a few well below breakdown?", I asked and he said that it was just their protocol and in this fashion they could learn the most from them. I commented something to the effect of, "Yeah, well sure, but you could learn something also by leaving some of them functioning after testing." And he replied that that was not the way they were doing it. The whole thing just seemed to be cover up even then, but perhaps he was just parroting what Schon had told him.

So I think there is real smoke there. However, despite the ETH position, in the end I am not sure if Bertram has gotten off lightly or not. He is basically shunned and seems like a shell of a man every time I have seen him in recent years. I think he's suffered tremendously... But whether it is too much or not enough, I can't judge.

Anonymous said...

It is too bad about the shunning of Batlogg. Hendrik Schoen was a pathological liar and it would have been a difficult situation for any PI. We often attack people when we see blood (lack of confidence), but overly confident people, we seem to trust. Why is self-confidence seen as a trait of an outstanding scientist?
Scientists are human. We are not god. I hope Batlogg recovers and does not remain a shell of a man. If the physics/science community does not forgive him, shame on us.

Anonymous said...

ReRe: Batlogg

If I was Batlogg (thinking that the results were real), I would prefer to keep the samples and try to publish many Science and Nature papers. Lieber did somewhat the same with nanowires. It was very difficult to get samples from him.

Doug Natelson said...

Anon. @3:22 - I completely agree. When I asked Batlogg and Schon about getting some substrates coated with their incredible gate dielectric, I got two different responses. Schon told me that the machine was down, but gave me sputtering parameters to try (which appear, in hindsight, to have been pulled out of a hat). Bertram was much cagier, basically saying that the dielectric was great stuff and that they wanted to work on it themselves some more. I think this latter response was the typical reaction of someone with past experience in the ultracompetitive world of early high-Tc, not anything more sinister.

Don Monroe said...

Here is the section of Science magazine's Conditions of Acceptance dealing with sharing of materials:

Materials sharing
After publication, all reasonable requests for materials must be fulfilled. A charge for time and materials involved in the transfer may be made. Science must be informed of any restrictions on sharing of materials [Materials Transfer Agreements or patents, for example] applying to materials used in the reported research. Any such restrictions should be indicated in the cover letter at the time of submission, and each individual author will be asked to reaffirm this on the Conditions of Acceptance forms that he or she executes at the time the final version of the manuscript is submitted. The nature of the restrictions should be noted in the paper. Unreasonable restrictions may preclude publication.

I know that biologists abide by these terms. Why not physicists?

random anon said...

Very interesting to read the perspectives of people with first hand knowledge of the Schon saga.

I once happened to meet someone who had worked at Bell Labs and he claimed the following remarkable thing: Prolific as he was, Schon was *not* the most prolific researcher at Bell Labs at the time. There were others there who were publishing at an even greater rate!
Is this correct? If it is, I guess it would (partly) explain why Schon's publication rate didn't attract more attention and scrutiny.