Search This Blog

Tuesday, October 11, 2011

Much ado about nothing ...

Pop quiz!  You are the editor of a major scientific journal whose policy is " disseminate concisely-written, high-impact research reports on major scientific advances", with a rapid turn around between receipt of manuscripts and publication so that such results can be communicated rapidly.  In pursuit of the later goal " be a fast-track and high-impact journal ...", your journal has "... a policy of rejecting papers for which major revisions are required to meet the [journal's] criteria of impact, innovation, and timeliness."  You receive an article which one (of two) reviewers considers to be, with minor revision,  a worthwhile low impact paper, but which he deems would require major revision to be a high impact paper.  What do you do?

Roger Pielke Jr thinks that if it is his paper, at least, and if the journal is Geophysical Research Letters (GRL) you should publish!  I think that claim strains credulity.

Pielke, of course, frames the narrative differently.  He suggests both reviews "...find the paper publishable...", and then asks why the editor chose to reject a paper found publishable by both reviewers.  In evidence he quotes the second reviewer as writing (Pielke's emphasis):
"The work seems essentially sound and useful to the community but lacks in-depth analysis and illustration. It does confront the issue of continued misrepresentation by some of the impact of “climate change” on presently experienced insurance and other losses from tropical cyclones. For that reason it is perhaps (just) publishable but claims of a new homogeneous database (based on JTWC outside of the US) are grossly over-stated as there is much work needed before that can be genuinely claimed. This is especially so in regard to intensity, which the authors treat fairly simplistically in any case. I would like to see that aspect down-played and perhaps the title adjusted to read “Towards a homogeneous database …” or some such."
 Transparently, Pielke's emphasis it to try and spin a case.  Without that emphasis, phrases such as "...lacks in-depth analysis and illustration", and " ... are grossly over-stated as there is much more work needed ...", and " ...which the authors treat fairly simplistically..." stand out in a review which is anything but a ringing endorsement.  Even the purported recommendation of publication, "... it is perhaps (just) publishable ...",  is not a recommendation of publication, but a refusal to recommend against it.

Pielke disputes that, and adduces as evidence the numerical rating given by the second reviewer, 3A.  That requires some explanation.  The A refers to presentation, and indicates that no issue of presentation prevents publication.  It is the 3, which rates the science which is interesting.

The GRL defines their category 1 papers as follows:

"The manuscript meets one or more of the following criteria:
  • High impact innovative results with broad geophysical implications at the forefront of one or several AGU disciplines.
  • Results with immediate impact on the research of others and requiring rapid publication.
  • Instrument or methods manuscript introducing an innovative technique that makes new science advances possible, with immediate applications to AGU disciplines."
These criteria should be compared with the criteria the AGL sets for its publication:

"With this goal, the Editorial Board evaluates manuscripts submitted to GRL according to the following criteria:
  • High impact innovative results with broad geophysical implications at the forefront of one or several AGU disciplines.
  • Results with immediate impact on the research of others and requiring rapid publication.
  • Instrument or methods manuscript introducing an innovative technique that makes new science advance possible, with immediate applications to AGU disciplines."
The criteria are, of course, identical.  That makes it very clear that the AGL seeks to publish only those papers that meet its category 1.

With that in mind, we note that category 2 papers are "...potentially Category 1 [but that] clarification that can be achieved within the GRL space constraints is needed".  In contrast, category 3 papers are "...unlikely to become a Category 1 paper for GRL."

Pielke spins this differently.  He notes that in category 3 papers, the "Science is sound and paper is publishable in the refereed literature ...".  However, the "refereed literature" includes many other journals, including 19 published by the American Geophysicists Union alone.  Interpreting an endorsement for publication in the literature in general as endorsement of the paper meeting the GRL's standards of "impact, innovation and timeliness"  is a stretch.  In this case, it is a stretch that requires us to interpret category 3 papers as being of a higher standard than category 2 papers, which is counter-intuitive to say the least.

Pielke's case is not helped by his ignoring perhaps the most damning comment by the second reviewer, that "The work ... lacks in-depth analysis and illustration."   When the AGU explained the new policies at GRL, they stated that:

"[M]anuscripts are routinely declined if the reviews point to a need for  additional analyses, simulations, or other significant changes to support purported high-impact results or implications.  However, for those submissions that show promise of reaching GRL’s criteria, authors are encouraged to resubmit following necessary revisions."
(My emphasis)

Clearly the second reviewer thought that more analysis was needed for Pielke's paper to be considered "high impact".  His decision to rate the paper as category 3 is, therefore, not without warrant.

Finally, it should be noted that the editor, Noah Diffenbaugh, did in fact invite resubmission of the paper.  We have seen above that category 3 papers are not considered suitable for publication in GRL, even with revision.  That Diffenbaugh invited resubmission shows that in his final judgement, the paper was a category 2 paper.  In fact, he split the difference between the two reviewers.

At the end of his correspondence with the chief editor of GRL (Eric Calais),  Pielke concludes that "...lost all confidence in GRL as a scientific journal."   Well, that is his right.  Part of the process of peer review is the judgement of scientists regarding the quality of the journals they read.  If they do not think the journal adequate, they will not consider papers published in it to be peer reviewed.  If enough of their scientific peers think likewise, the journal is destined for a quick trip to oblivion.  But in this case, I do not think many of Pielke's colleagues will take his view of things.  Rather, they will wonder why he is trying to blow up a fairly normal incident in the peer review process into a case of his martyrdom.

1 comment:

  1. Thanks Tom for alerting me to this post. You have some basic facts wrong -- I dd not claim that my paper should be published, but rather, in the event that an editor judges a paper to require "major revisions" he has an obligation to explicitly state what such revisions would entail, especially if neither reviewer asks for "major revisions." That the editors were unable or unwilling to provide such guidance is that basis of my complaint. Thanks for your interest.