Members of Congress love to talk about investing in science and technology so that American researchers can outperform those in other countries. A new study suggests that approach may miss the point -- because a growing percentage of the papers by American scientists are written with authors from other countries.

The study was released last week by Britain's Office of Science and Innovation, which, like Congress, worries about its country's relative competitiveness in science. The study compares the scientific output from nine countries in research published in seven fields: clinical sciences, health, biological sciences, environmental sciences, mathematics, physical sciences and engineering. The authors note all kinds of ways that one could quibble with this method for measuring collaboration, but conclude that this is one meaningful way to approach the question.

Here are the results for two periods compared, which show relatively little change in total output, but considerable change in the rates of collaboration across national borders.

Research Output and Collaboration

The report also notes patterns within the various collaborations. For example, American scientists publish roughly the same number of papers with British and German scientists, but about twice as many with colleagues from those two countries as with those from France. American collaboration with Britain leans heavily toward biomedical research, while projects with Germany tend to focus on the physical sciences and engineering.

American scientists are highly sought as partners by those in many other countries. German scholars have twice as many collaborations with American scientists than with their British counterparts. But an exception to that rule, noted in the report, is France, which in most fields has more collaboration with Britain than with the United States.