Friday, August 10, 2007

NSF shows output of US scientific publications flat for more than a decade

Jeffrey Mervis of Science discusses a National Science Foundation [NSF] study (nsf07320) showing that the overall number of publications by US scientists has not grown in over a decade.

There is an interesting quote from a second NSF study (nsf07204): "data from interviews and meetings are not very useful for considering some possible explanations" for the stagnant publication output.

Mervis presents theories:

1. Aging US scientific workforce that is growing less productive
2. Emphasis on quality, not quantity
3. Increased demands from non-US governments for greater production from their scientists, which production has pushed out some mediocre work from US scientists
4. The steep learning curve associated with collaborative research, an increasingly popular mode of operation.

IPBiz finds all four reasons pretty silly. One can think back to Jan-Hendrik Schon to address points 1 and 2. He wasn't old and his emphasis was on quantity, not quality. The same can be said for the workers copying Schon's themes. Of #3, there is still a pro-US bias among many top-of-the-line journals, so even mediocre work probably can be published. [on a related point, note that the recent "discovery" within the US of Hwang Woo-Suk's parthenogenesis work had already been reported in South Korea in May 2006, but no one seemed to remember. Also, recall the detailed history of Cha's Fertility & Sterility paper.] Of #4, particle physicists solved the collaborative research problem years ago, without any difficulty. Both Jan-Hendrik Schon and Hwang Woo Suk also demonstrated skills in collaborative research.

Curiously, although the "stagnation" problem is presented in terms of the absolute number of publications, the graphs in Science present percentages (relative numbers of publications). Thus, in the time frame 1992 to 2003, academic publishing is "up" 0.8% but non-academic is "down" 0.1%, with private-for-profit down the most, 1.4%.

A separate graph shows output from the EU-15 has exceeded US output of publications since 1998. Of interest to New Jersey, the New Jersey Institute of Technology showed one of the largest growths between 1992 and 2001, 71%.

The Science article did not mention patents, or patent quality.

[see 317 Science 582]

Ron Katznelson emailed the following:

"There are other possible explanations to the flat U.S. scientific journal publications statistics. First, there is recent evidence that patent publications constitute an increasing source of first disclosure of new basic scientific information, gradually fulfilling the traditional role of the scientific journal literature. See my paper on patenting trends at (footnote 60, citing work showing that patents' share as a first information source increased over two decades since 1980 at a rate of approximately 8%-10% per decade, particularly in the chemical and life sciences). In that sense, one should look at U.S. patent disclosures (my Figure 1) that have been doubling every 14 years in the last 25 years. That is certainly no 'flat' statistics.

Second possible reason is that internet publishing without any counterpart journal publication becomes a growing venue for scientists to publish their scientific work. Any analysis that attempts to quantify the national scientific output while ignoring internet scientific publications is simply behind the times and should not be taken as a definitive indicator.


Post a Comment

<< Home