Wednesday, November 08, 2006

Another Publication Scandal?

OK, so this is not quite exactly on the same disaster level as the infamous Schon debacle from a few year ago. However, it is still making the headlines, especially in this week's Nature.

The paper involves a landmark experimental work from published in Nature in 1993 titled "Atomic-resolution chemical analysis using a scanning transmission electron microscope" (N.D. Browning, et al. Nature 366, 143 (1993). According to Nature, while one referee endosed it, the other referee was more hesitant, and expressed the opinion that the manuscript had "disquieting questions" that remained unanswered. Nature's editor decided to publish the paper at that time.

In this week's issue of Nature (Nature 444, 123-124 (9 November 2006)), a correction has been made to this paper more than 10 years after it was published. It is the result of a series of rather unusual event. Let me see if I can outline it from the way I understood it for those of you who do not have access to Nature.

1. 2 of the authors of the Nature paper submitted a preprint to ArXiv in 1995 (Varela, M. et al. (2005)). In the first version of the paper, there were data what appeared to be identical to the one they published earlier for a conference, but using different technique than the one used for the ArXiv paper. There was also an issue of a data being duplicated to make a "mirror image".

2. Others started noticing this issue by Spring of 2006 when a version of the paper was submitted to Nature Physics.

3. Within days of the authors being alerted to the possible concerns of the paper, the authors submitted another version in which references were given for all but one set of data. The duplicated mirror image data were also removed or cropped.

4. This led to Oak Ridge appointing an investigative committee. The panel found that while there were errors in judgement, no deliberate falsification or fabrication of data occured.

5. Here's where I have to make my own inference. Due to the investigation, the authors realized mistakes in the earlier Nature paper. In particular, they did the identical mistake in which they included experimental data from an earlier conference paper using different analysis/background subtraction. At that time, the authors had provided Nature's editor with a firm reassurance over the referees uneasiness.

6. A correction is now published in the current issue of Nature indicating the data set that were differently handled.

While the overall result of the paper right now did not hinder the progress of the field, many physicists that were interviewed loudly wonder why the paper, with some of the main data that are now questionable, is not withdrawn. Nature has offered an explanation over why it did not see fit to ask for a retraction (you have to read the issue, it's too long to type).

At best, this whole debacle has caused many to distrust the work and results coming out of this group.


No comments: