Friday, April 18, 2008

Negative results

James Correia sent this interesting article on publishing negative results. This articles makes the case that you shouldn't need to publish negative results. I have thought that we didn't make enough effort in atmospheric sciences to publish negative results. We always publish the big storms, but null cases that were forecast to be big storms are rarely ever published. Here are two counterexamples, plus a third study showing that a number of factors do not affect storm longevity.

Richter, H., and L.F. Bosart, 2002: The Suppression of Deep Moist Convection near the Southern Great Plains Dryline. Mon. Wea. Rev., 130, 1665–1691.

Doswell, C.A., D.V. Baker, and C.A. Liles, 2002: Recognition of Negative Mesoscale Factors for Severe-Weather Potential: A Case Study. Wea. Forecasting, 17, 937–954.

MacKeen, P.L., H.E. Brooks, and K.L. Elmore, 1999: Radar Reflectivity–Derived Thunderstorm Parameters Applied to Storm Longevity Forecasting. Wea. Forecasting, 14, 289–295.

What are your thoughts?

Monday, April 14, 2008

What happens when reviewers have different opinions?

Recently, I have handled several manuscripts as editor where reviewers offer widely different opinions. For example, Reviewer A might like the paper, recommending only minor revisions. Reviewer B might request major revisions, and Reviewer C might recommend rejection. What is going on? How does an editor make sense of this?

Several things could be going on.

1) The paper is controversial with diverging viewpoints. The number of papers that are truly like this are less common that one would think.

2) One or more of the reviewers is biased for or against the paper, giving a review that would be atypical. In these cases, an editor can generally see through the problem person and make a reasonable judgment.

3) The reviewers do not take time with the manuscript or may be inexperienced, and so don't give a fully adequate assessment of the paper (usually minor revisions, on a paper that requires much more work, as noted by other reviewers).

4) The author has failed to adequately state the purpose and utility of the work. Without doing so, reviewer opinions become diverse, if not plain confused. Some may want one improvement, whereas another reviewer wants the opposite. In these cases, the authors need to more clearly present their work and defend it. Then, reviewers can take a stand for or against the work.

In all cases, good editors can sort through the different opinions and make a reasonable decision.

Scare tactics by publishing companies

As I have been describing in this blog, publishing companies are afraid of open-access journals and other open-access bills being considered by Congress. One of their arguments is that open access will harm the peer review process.

Such undue government intervention in scholarly publishing poses inherent risks and problems, including:

  • Threats to the economic viability of journals and the independent system of peer review
  • The potential for introducing selective bias into the scientific record

"Open access" does not equal "no peer review." That is one of the distortions that the publishing companies use to scare Congress into killing reasonable bills that allow taxpayer-funded research to be accessed for free.

In fact, the highest ranked journal in the atmospheric sciences in terms of impact factor is Atmospheric Chemistry and Physics, published online only by the European Geosciences Union. The journal has a unique peer review process. You submit the paper, and after an initial assessment of quality by an editor, gets posted online for all to read and comment upon nonanonymously. At the same time, anonymous peer reviewers assess the manuscript. After a specified time, the author then must respond to all comments, anonymous and nonanonymous. The manuscripts at ACP eventually get published or rejected, as we do. Page charges are smaller than the AMS (covering web hosting, etc.), rejection rates are only 16%, and there is no cost to access the article. Therefore, the author pays all charges. The readers pay nothing to read the article. This is why open access is so popular among scientists. How many articles would you be able to read online if you didn't have to pay? I know I would read a lot more.

The AMS sits somewhere in between big publishers and open access. The articles are restricted to subscribers for five years, after that, they are free to everyone. That, in my opinion, isn't a bad option. Some journals will allow their otherwise restricted articles to be open access. For an extra $1000 or so in page charges, authors can make their published article free to everyone, not just journal subscribers.

Although I am not a total fan of e-publishing for some contexts, open access is one thing I cheer loudly for. As a fan of science, so should you.