October 24, 2007

Number of U.S.-Based Published Science Articles Hits a Plateau

Print More

The face of scientific publishing has changed dramatically in the past 15 years. While it was formerly common to see only two names on a research paper, now one is likely to see an article listing up to 12 authors.
The need for increased collaboration and the rising cost of doing research have contributed to the plateaued number of science and engineering journal articles from the U.S. since the 1990s. This is a marked contrast from the previous two decades, which saw a rise in publication numbers. This unexpected development led the National Science Foundation’s Division of Science Resource Statistics to produce a quantitative study of the patterns and trends of U.S. journal articles from the past 15 years.
The first study of this three part effort, “Changing U.S. Output of Scientific Articles 1988-2003,” takes a quantitative look at the science and engineering article output of the world’s four major scientific publishing centers: the U.S., the E.U.-15, Japan and East Asia. This section also looks at the breakdown of national publishing counts within different science and engineering fields.
The second article, “The Changing Research and Publication Environment in American Research Universities,” is a qualitative report focusing on interviews with researchers and administrators across a wide spectrum of disciplines at nine major research universities in California, Illinois, Massachusetts and North Carolina.
According to the study, while the number of journal articles published is an indicator of research output, it may be “an admittedly imperfect one.” There are many other ways scientists disseminate research today, including databases and online archives. These methods of storing information are expensive to produce, but are extremely important and useful to researchers.
Prof. David Wilson, molecular biology, described one such example of the use of databases in science today. “The human genome project cost a lot of money, and it probably in the end was worth it, but it probably only produced four or five papers. And yet it cost several hundred million dollars to do that project,” he said.
Another large factor affecting research paper counts may be based policies at research institutions. U.S. researchers interviewed for the study perceive less of a focus on quantitative output measures at their universities than at institutions in other countries. The focus at U.S. universities seems to be based more on quality, not quantity.
“What really counts is the quality, what is the impact of a given paper that you write,” said Prof. Charles Williamson, mechanical and aerospace engineering.
Williamson recalled speaking with a colleague at a university in Australia. “I remember one conversation where he was lamenting the fact that they kept counting numbers of papers and weighing up how effective a person was based on number of papers. He kept saying, ‘If only they would take note of how important each paper was,’” he said.
The current technicality and complexity of most research has led researchers to increase collaboration with those in other areas of expertise, as well as in other countries. The use of the internet has also helped to facilitate these collaborations.
“Because it’s so complicated, you need computer people to lay out your design; you need people who can do the etching of the design; you might need people who are chemists; you need a whole crew. So now your paper has 23 people on it. And plus, you’ve got another guy over from Belgium now. I think the internet’s had a huge impact on that,” said Laurel Southard, director of undergraduate research and outreach. “We used to joke, because all of a sudden you’d see articles with 40 or 50 authors.”
NSF plans to release a third report in the series that will explore the inputs and outputs of research and funding. It hopes to determine which of these factors is most relevant to the overall trend towards the plateau of journal article numbers in the U.S.