Confused at a higher level

The view from Carleton College's physics department

  • Archives

  • Stats

    var sc_project=3293756; var sc_invisible=0; var sc_partition=21; var sc_security="d61881ba";
    free hit
  • Subscribe

  • Recent Posts

  • Follow me on Twitter

The F word in physics… Fraud

Posted by Melissa on May 24, 2009

When it came out earlier this month, I immediately picked up Eugenie Samuel Reich’s book about the Hendrik Schön scientific fraud case because my graduate work began as part of the Minnesota efforts to reproduce some of Schön’s results (there are a few pages on the Minnesota collaboration in the book). Being touched by the fraud personally, I found this book provided a satisfying look at how it all happened. Although you can never get into Schön’s mind, I began to understand how fraud of such magnitude could unfold, namely Schön asked others what they would expect to see, and then created results that matched those expectations. That he aimed to meet people’s expectations, combined with his quiet and amiable personality, meant that the red flags didn’t go up immediately.

One theme that Reich kept raising was whether the self-correcting nature of science worked in this situation. Personally, I found this to be a distracting framework in which to place the Schön case. I consider the self-correcting nature of science to refer to the back and forth of different researchers, theorists and experimentalists, as they try to arrive at the correct understanding of a particular physical phenomenon. Researchers often begin by putting forth what turn out to be incorrect interpretations. As others consider the work, the interpretations are refined and refuted until consensus is reached about what is the appropriate description. To me, this is the self-correcting nature of science. I don’t think the primary purpose of this process is to catch people who are trying to dupe the system, and to ask whether the self-correcting nature of science worked in the Schön case is not an appropriate question.

I was also interested to learn about Schön’s graduate work, which showed an early tendency towards sloppy practices. The book mentions the graduate student Schön fiddling to produce a line of best fit that matched the scientific literature better than what he would have gotten without fudging, and Reich also finds an early example of data manipulation in published work related to Schön’s dissertation. Is it a slippery slope towards fraud? Do we convey forcefully enough the importance of good record keeping and honest practices, even at the undergraduate level?

I’ll admit that starting my graduate work chasing fraudulent results has indelibly shaped my view of the scientific endeavor. I think I am somewhat more cynical about science as a result. I also tend to be wary of the flashy results that show up in Science or Nature, appreciating the more in-depth technical papers appearing in Phys Rev B that actually require significant discussion of how results were obtained and what they might mean. Nevertheless, I recognize that science is ultimately about trust, and I’ve got to trust that most scientists are dedicated to the advancement of science over personal advancement.

The question of whether there is a way to prevent this type of fraud in the future is worth considering, and I generally agree with Steve at Complex Matters that for the most part the system works, although there is always the chance that someone will come along and take advantage of a trusting system. (Of course, I can say this because the Schön episode only impacted the very start of my graduate career so I didn’t get burned badly. Those whose careers were seriously damaged might feel differently.) I did walk away from Reich’s book wishing it had more discussion of the role of the journals, reviewers comments, and editors decisions in potentially helping or hindering fraudulent work.

Georg at Life on the Lattice wonders if the open science movement would have prevented Schön’s fraud. Of course, more transparency is always better, but despite the trusting nature of scientists, there is still a tendency to play things close to the chest for fear of being scooped and to maintain an advantage for continued priority of discovery. When in the trenches trying to replicate Schön’s work, missing information about experimental techniques regarding the deposition of aluminum oxide was problematic. There were e-mail exchanges in which Schön happily dished out information about what he supposedly did, but ultimately, we decided he used some experimental trick that we didn’t. I’m not sure tricks of the trade will ever be freely shared when researchers are hotly competing with each other to be the first to get results.  But then again, maybe that’s just my cynicism coming through.

I’d be interested in hearing others’ thoughts. Do we simply live and learn with each fraud case? Or can we do more to prevent fraud in the future? Is science too trusting? Or not trusting enough?


3 Responses to “The F word in physics… Fraud”

  1. Don Monroe said

    Hi Melissa, thanks for posting this.

    In biology, there is an expectation of sharing reagents and materials, not just data. This is a condition of acceptance of a manuscript at some journals. Can you imagine a procedure for sharing the gated organics that would be acceptable to researchers?

    By the way, the name on your post links to http:///, which, needless to say, doesn’t get very far.

  2. Melissa said

    That’s an interesting idea. I wasn’t aware of that practice in biology. If such expectations were in place, I think it would have revealed problems with Schön’s work earlier. Deciding with whom one must share devices, particularly if there are few devices and they are difficult to produce, might be a thorny issue. You’d also have to consider if there are different expectations of a researcher working in an industrial setting like Bell Labs, where the company has paid for the research with hopes of future returns on their investment, as compared to University labs where the research is funded by NSF.

  3. Don Monroe said

    thanks for fixing your link. Actually googling “arjendu” led to you in a couple of clicks anyway.

    On one of your other points: did self-correction work in this case? My answer is that it was in the process of working, but not quickly enough. The damage done, in terms of prizes, jobs, and the wasted time of researchers like yourself was really significant.

    I think the question becomes more interesting when you realize how smugly complacent most researchers are about it. There is a really nice treatment of this in the excellent book “Betrayers of Truth,” by New York Times science writers Nicholas Wade and William Broad.

    One point they make is that, although most scientists will point to repetition as the gold standard for establishing scientific validity, in many cases (like this one, which hadn’t happened yet) persistent non-replication is met with various alternative explanations and excuses. In this case it did not ultimately play any direct role in launching the investigation.

    Again, perhaps it would eventually have worked, but not quickly enough. How long might this have continued if it were not for the duplications?

Leave a Reply

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

You are commenting using your 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 )

Connecting to %s

%d bloggers like this: