Skip all navigation and go to page content

Chapter 5. Academic Research and Development

Outputs of S&E Research: Articles and Patents

Chapter 2 of this volume discusses the human capital outputs of higher education in S&E. The present chapter focuses on the S&E functions of U.S. colleges and universities, including funding and performance, physical infrastructure, and human capital devoted to R&D. This section examines the intellectual output of academic S&E research using indicators derived from formal research articles and U.S. patent data.

Researchers have traditionally published the results of their work in the world's peer-reviewed S&E journals,[31] and article-level data from these journals are used here to explore total S&E research output by countries and—within the United States—by sectors of the economy.[32] These so-called bibliometric data are also useful for tracking trends in S&E research collaboration using coauthorship measures between and among departments, institutions, sectors, and countries. (See sidebar "Bibliometric Data and Terminology.") Finally, citations in more current research articles to previous research offer insight into the importance and impact of previous research.

The 2008 edition of Indicators (NSB 2008) focused attention throughout these bibliometric indicators on three large geographic units: the United States, the 27 members of the European Union, and a group of 10 fast-growing countries in Asia. This edition adjusts that particular organization of the data to focus instead on five S&E article-producing countries/regions that together account for more than four-fifths of the world total: the United States, the European Union, China, Japan, and eight countries/economies together referred to as the "Asia-8" (India, Indonesia, Malaysia, Philippines, Singapore, South Korea, Taiwan, and Thailand).

S&E researchers publish the results of their work in the peer-reviewed literature, and symbolic payment for their work is a citation to their article when it is used by future researchers (see Merton 1973). This recognition is uniquely valuable inside the scientific community, where it can be critical to career advancement, and does not necessarily reflect the value society might place on particular scientific findings.

In contrast, when researchers file for patent protection for a practical advance over "prior art" and the claim is granted in a successful patent, the patent owner obtains certain rights to the potential value of the advance. This chapter uses the patenting activities of U.S. academic institutions as another type of indicator of the outputs of academic S&E research. (Chapter 6, "Industry, Technology, and the Global Marketplace," discusses patenting by other sectors in "Global Trends in Patenting.") Because citations to the S&E literature in successful patents indicate the use of past research in practical advances, literature/patent linkage data illuminate patterns of the impacts of academic S&E research on potential technological development.

S&E Article Output

Between 1995 and 2007, the total world S&E article output as contained in the journals tracked by the Science Citation Index (SCI) and Social Sciences Citation Index (SSCI), which are analyzed in this chapter, grew at an average annual rate of 2.5% (table 5-14 ). Scientists and engineers in institutions in the member countries of the European Union authored or coauthored 32% of the world total in 2007,[33] followed by the United States with 28%. China, Japan, and the Asia-8 accounted for another 22% of the world total (appendix table 5-25 ).[34]

Growth in S&E article output across these five countries/regions has been uneven. Twelve-year growth rates in the mature economies of the U.S. (0.7%), Japan (1.0%), and the European Union (1.9%) have been lower than in the rapidly growing economies of the Asia-8 (9.0%) and China (16.5%) (appendix table 5-25 ; figure 5-20 ).

Exactly 200 countries or other entities[35] receive credit for publishing S&E articles (appendix table 5-23 ). A small number account for most of the publications.[36] Thus, table 5-14 shows that five countries (the U.S., China, Germany, Japan, and the United Kingdom) accounted for more than 50% of the total world S&E article output in 2007. The 45 countries in table 5-14—less than one-quarter of the countries in the data—produced 98% of the world total of S&E articles. Nevertheless, the data are constantly evolving to reflect changes in the makeup of nations around the world or the sudden appearance of an author from a heretofore non-publishing country.[37]

Trends in Country and Regional Authorship
Steadily increasing investments in S&E education and research infrastructure in many countries, especially in Asia, have led to increased R&D in those countries and laid the groundwork for increased research productivity. As a result, scientists and engineers in those countries are increasingly prominent contributors to international peer-reviewed journals.

Differences in recent rates of growth in article production are striking. In the major Asian countries, average annual growth rates between 1995 and 2007 were highest in China, at 17%. Across the Asia-8 countries, growth rates have been 9% annually for the same period (appendix table 5-25 ), led by Thailand (15%), South Korea (14%), Singapore (11%), and Taiwan (9%) (table 5-14 ). These growth rates mirror those in R&D expenditures and number of researchers. Japan's growth in article output averaged a modest 1% annually between 1995 and 2007. China's rapid growth rate in S&E article output propelled it in 2007 past the United Kingdom, Germany, and Japan to rank as the world's second-largest producer, up from 5th place in 2005 and 14th place in 1995 (appendix table 5-25).[38] During the same period, South Korea went from 22nd to 10th place.

These high rates of growth in S&E article authorship in Asia contrast with much slower rates for the world as a whole (2.5%), for countries with mature S&E infrastructures such as the United States (0.7%), and for the European Union (1.9%) (appendix table 5-25 ). Russia's article output decreased over the period, from 18,600 in 1995 to 14,000 in 2007, as did Ukraine's, from 2,500 to 1,800. Many of the other former republics of the Union of Soviet Socialist Republics (USSR) experienced negative growth on this indicator as well.

Countries within the European Union showed different trends in S&E article output between 1995 and 2007. Growth rates below 3% were common, for example, in Austria, Belgium, Denmark, Germany, and the Netherlands. Among the lowest rates of growth on this indicator were the United Kingdom (0.3%), France (0.5%), and Sweden (0.5%). Relatively high growth was experienced by the Czech Republic (5.4%), Greece (7.6%), Ireland (6.1%), Portugal (10.9%), and Spain (5.3%). Although not a member of the European Union, Turkey experienced one of the fastest growth rates in S&E article output in the world: 14.4% annually (from 1,700 articles in 1995 to 8,600 in 2007 (appendix table 5-25 ).

The countries in Central and South America together increased their S&E article output between 1995 and 2007 at an annual rate of 7.8%. Among the Central and South American countries that had more than 1,000 articles in 2007, Brazil had the highest growth rate (10.9%), followed by Mexico (6.7%), Chile (5.8%), and Argentina (4.6%). Brazil is also steadily increasing in rank among the world's S&E article producers: it was 23rd in 1995 and 16th in 2007 (table 5-14 ).

Across North Africa and the Middle East, only Egypt (2.8% annual growth since 1995), Israel (1.2%), and Iran (25.7%) produced substantial numbers of S&E articles in 2007. Iran's growth rate was the fastest of all nations (see sidebar "S&E Publishing Trends in Iran").

Research Portfolios of Top Article-Producing Countries/Regions
The S&E article output of the United States, the European Union, China, Japan, and the Asia-8 together accounted for 82% of the world total in 2007 (appendix table 5-25 ). A field-by-field comparison across these five countries/regions provides a snapshot of their research portfolios, and strong differences are evident. China, Japan, and the Asia-8 emphasize the physical sciences more than the United States and the European Union. China's S&E research articles in chemistry and physics accounted for almost one-half of its total article production in 2007 (table 5-15 ). In Japan, these two fields accounted for 36%, and in the Asia-8, 37%, compared with 17% for the United States and 25% for the European Union. The proportions of all five portfolios in astronomy (≤1.5%) and the geosciences (4.0%–5.5%) were similar.

These portfolios also vary in their emphasis on the life sciences (the biological, medical, agricultural, and other life sciences): the U.S. output in these fields accounted for 57% of its total, compared with 49% for the European Union, 25% for China, 45% for Japan, and 34% for the Asia-8 (table 5-15 ).

A third strong contrast across the five countries/regions is the emphasis on engineering: S&E research publications with authors in Asia are relatively more heavily concentrated in engineering (China at 16%, Japan at 11%, and the Asia-8 at 19%) than those with authors in the United States (7%) or the European Union (8%).

Coauthorship and Collaboration

Coauthored, collaborative articles with authors from different institutions and different countries have continued to increase, indicating increasing knowledge transfer or knowledge sharing among institutions and across national boundaries. [39], [40] The discussion begins with consideration of broad trends in coauthorship for the world as a whole, moves to regional patterns, and ends with an examination of country-level trends, including selected country-to-country coauthorship patterns and indexes of international collaboration. (Indicators of cross-sector coauthorship, which are available only for the United States, are examined below in the section "Trends in Output and Collaboration Among U.S. Sectors." Indicators of collaboration using different data are discussed earlier in this chapter under "Collaborative Research" in the "Doctoral Scientists and Engineers in Academia" section. For a consideration of the limitations of bibliometric techniques in identifying interdisciplinary S&E research, see the sidebar "Can Bibliometric Data Provide Accurate Indicators of Interdisciplinary Research?"

Article Author Names and Institutions
Between 1988 and 2005, the number of S&E articles, notes, and reviews grew by 60%, while the number of institutions and the number of author names on them both grew by more than 100% (NSB 2008, 08-01, figure 5-29). The trend continued in 2008. In all broad fields, the number of author names per article increased (table 5-16 ). The average number of authors per paper was more than five in astronomy, physics, the biological sciences, and the medical sciences. Growth in the average number of coauthors was slowest in the social sciences (from 1.4 authors per paper in 1988 to 1.9 in 2008) and in mathematics (from 1.5 in 1988 to 2.0 in 2008). Unpublished NSF analyses show that in 2008, 90% of all S&E articles had at least two author names.

A closely related indicator, coauthored articles (i.e., articles with authors in different departments or institutions), has also increased steadily. Coauthored articles grew from 40% of the world's total S&E articles in 1988 to 64% in 2008 (figure 5-21 ). This growth has two parts. Coauthored articles that list only domestic institutions in the byline grew from 32% of all articles in 1988 to 42% in 2008. Articles that list institutions from more than one country, that is, internationally coauthored articles (which also may have multiple domestic institutional authors) grew from 8% to 22% over the same period. The remainder of this section focuses on these internationally coauthored articles.

International Coauthorship From a Regional Perspective
From the perspective of large article-producing countries/regions, interregional coauthorship has increased steadily.[41] From 1998 to 2008, interregional coauthorship increased as a percentage of the total article output of the United States (from 20% to 30%), the European Union (from 21% to 29%), Japan (17% to 26%), and the Asia-8 (22% to 27%) (table 5-17 ). Notably, China failed to increase on this indicator: as a percentage of its total S&E article output, China's interregionally coauthored articles declined from 26% in 1998 to 25% in 2008.

As a percentage of the world's interregionally coauthored articles, the shares of articles with a U.S., European Union, or Japanese institutional author declined slightly, giving way to a rise in the share of articles with an institutional author from China (from 6% to 13%) or the Asia-8 (from 9% to 14%). These changes in share of the world's interregional articles are similar to the changes in each region's share of all the world's articles.

The other regions identified in table 5-17 tend to have a less-developed S&E infrastructure, and scientists and engineers in those regions tend more often to coauthor articles with colleagues in the more scientifically advanced regions/countries. For example, in 2008, 41% of all S&E articles with an institutional author from the Near East/North Africa (which includes Israel) had an author from another region, as did 61% of S&E articles with an institutional author from Sub-Saharan Africa (which includes South Africa). The following sections look more closely at coauthorship patterns of specific countries and country pairs.

International Coauthorship Patterns From a Country Perspective
When the region-level data discussed in the previous section are disaggregated to the country level, a richer picture of international S&E article coauthorship emerges. Table 5-18 displays the international coauthorship rates of countries that had institutional authors on at least 5% or more of the world's internationally coauthored S&E articles in 2008 (see also appendix tables 5-39 and 5-40 ).

The sheer volume of U.S. internationally coauthored articles dominates these measures: 30% of U.S. articles in 2008 were internationally coauthored, and U.S.-based researchers were coauthors of 43% of the world's total internationally coauthored articles. The next highest percentages of the world's coauthored articles were held by Germany and the United Kingdom, each at 19% of the world total.

Even higher rates of international coauthorship are evident among the countries of the European Union and in Switzerland. Both Japan's and the Asia-8's international coauthorship rates have increased over the past 10 years.

What accounts for specific coauthorship relationships? Narin and colleagues (1991) concluded that "the direction of international coauthorship is heavily dependent on linguistic and historical factors." Coauthorship data suggest that geography, cultural relations, and the language of particular pairs or sets of countries play a role (Glänzel and Schubert 2005; Schubert and Glänzel 2006), and these preferences have been evolving over time (Glänzel 2001). In more recent years, European Union policies and incentives that foster intra-European Union, cross-border collaboration are also partly responsible for some high rates of coauthorship. The discussion below in the section "International Collaboration in S&E" identifies strong coauthorship relations in specific country pairs across the world, based on the strength of their coauthorship rates.

International Coauthorship With the United States
Table 5-19 lists the 31 countries whose institutions appeared on at least 1% of U.S. internationally coauthored articles in 2008. U.S. authors are most likely to coauthor with colleagues from the United Kingdom (13.9%), Germany (12.7%), Canada (12.0%), and China (10.4%—up from 3.5% in 1998).

Table 5-18 shows that the rate at which U.S. researchers participate in international collaboration is below that of many countries with smaller science establishments. However, because of the size of the S&E establishment in the United States, the share of U.S. internationally coauthored articles that were coauthored with any other country is lower than the share of the other country's internationally coauthored articles that were coauthored with U.S. researchers (table 5-19 ). For example, 3% of U.S. scientists who coauthored internationally in 2008 collaborated with Israeli counterparts; the corresponding figure for Israel, with its much smaller scientific infrastructure, is 52%. Again, 51% of scientists and engineers in Canada who coauthored internationally collaborated with U.S. colleagues, but only 12% of U.S. international coauthorship was with colleagues at Canadian institutions;[42] linguistic, geographic, and other ties combine to facilitate these collaborations.

Notable changes in these patterns of U.S. international coauthorship parallel changes in other indicators discussed in this section. As China's total S&E article output grew rapidly, so did its coauthorship with U.S. authors: the U.S. share of China's internationally coauthored articles increased about 6 percentage points over the past decade, and China's share of U.S. internationally coauthored articles increased almost 7 percentage points (table 5-19 ). U.S. scientists and engineers lost relative share of international coauthorship with some countries/economies, notably India, South Korea, Taiwan, and Japan, as their counterparts in those countries/economies broadened the geographic scope of their collaborative relations.

International Collaboration in S&E
In developing indicators of international collaboration between countries and across regions, researchers have developed statistical techniques that account for unequal sizes in countries' S&E article output and coauthorship patterns (Glänzel and Schubert 2004). One of the simplest is the index of international collaboration (table 5-20 ), which is defined as the ratio of country A's rate of collaboration with country B divided by country B's rate of total international coauthorship (Narin, Stevens, and Whitlow 1991). Indexes above 1 represent rates of coauthorship that are higher than expected, and indexes below 1 indicate rates of coauthorship that are lower than expected. For example, if country B produces 12% of internationally coauthored articles, and 12% of country A's coauthored articles are with country B, the index of international collaboration is 12%/12% = 1.0. The indexes for all pairs of countries that produced more than 1% of all internationally coauthored articles in 2008 are shown in appendix table 5-41 .

Table 5-20 lists the international collaboration index for selected pairs of countries. In North America, the Canada-United States index shows a rate of collaboration that is slightly greater than would be expected based solely on the number of internationally coauthored articles shared by these two countries, and the index has changed little over the past decade. The Mexico-United States index is just about as would be predicted and is also stable.

Collaboration indexes between pairs of countries on opposite sides of the North Atlantic are all low and have changed little over the decade. In Europe, collaboration patterns are mixed, but most have increased over the decade, indicating growing integration across the European Union in terms of S&E article publishing. Among the large publishing countries of Germany, the United Kingdom, and France, collaboration was less than expected but grew in all three countries over the decade.

The Scandinavian countries[43] increased their coauthorship indexes with many countries in Europe (appendix table 5-41 ), and within Scandinavia, the indexes are among the highest in the world and, overall, have been growing stronger (table 5-20 ).

Cross-Pacific collaboration patterns are mixed. Japan-United States collaboration fell below the expected value over the decade, while the China-United States index rose to near 1. U.S. collaboration indexes with South Korea and Taiwan declined but remained higher than expected in both cases. Canadian scientists and engineers were less likely than their U.S. neighbors to have coauthored with colleagues in Asia. Mexico's collaboration with Argentina is almost four times higher than expected, at 3.74 in 2008. In South America, the collaboration index of Argentina-Brazil, at 5.32, is one of the highest in the world.

Collaboration indexes within Asia and across the South Pacific between the large article producers are generally higher than expected but with only minor changes over the past decade. Australia's coauthorships are strongly linked to New Zealand, at nearly four times the expected rate of coauthorship. Two strongly coauthoring pairs are South Korea-Japan and Australia-Singapore. India's collaboration index with South Korea grew from 1.61 to 2.19 over the past decade.

Trends in Output and Collaboration Among U.S. Sectors

In the U.S. innovation system, ties between and among universities, industry, and government have been beneficial for all sides. These ties include the flows of knowledge among these sectors, for which research article outputs and collaboratively produced articles are proxy indicators. S&E articles authored at academic institutions have for decades accounted for more than 70% of all U.S. articles (76% in 2008) (appendix table 5-42 ). This section contrasts U.S. academic authorship with nonacademic authorship, including output trends by sector and the extent of coauthorship, both between U.S. sectors and between U.S. sectors and authors abroad.

Article Output by Sector
Total annual S&E articles by authors in U.S. nonacademic sectors changed little over the past decade, ranging from 50,000 to 55,000 articles[44] per year between 1995 and 2008 (appendix table 5-42 ). The number of articles produced by scientists and engineers in the federal government and in industry was more than 15,000 in 1995 but slowly declined to range between 13,000 and 14,000 each through 2008 (figure 5-22 ). State and local government authorship, dominated by articles in the medical and biological sciences, remained constant across the decade. Between 1995 and 2008, scientists and engineers in the private nonprofit sector increased their output from about 15,000 to about 18,000.

Federally funded research and development centers (FFRDCs) are research institutions sponsored by federal agencies and administered by universities, industry, or other nonprofit institutions. FFRDCs have specialized research agendas closely related to the mission of the sponsoring agency and may house large and unique research instruments not otherwise available in other research venues. Although authors at FFRDCs published articles in all of the broad S&E fields considered in this chapter, articles in physics, chemistry, and engineering together represented 69% of publication by this sector in 2008, reflecting its specialized research programs. Physics articles accounted for 39% of the FFRDC total (9% of the total for all sectors); engineering articles for 15% (7% of the total for all sectors); and chemistry articles for 16% (8% of the total for all sectors (appendix table 5-42 ).

The 16 FFRDCs sponsored by the Department of Energy dominated S&E publishing by this sector. Across all fields of S&E, DOE-sponsored labs accounted for 83% of the total for the sector in 2005 (NSB 2008). Scientists and engineers at DOE-sponsored FFRDCs published 96% of the sector's articles in chemistry, 95% in physics, and 90% in engineering (see "S&E Articles From Federally Funded Research and Development Centers," NSB 2008, p 5-47). Nine other federal agencies, including the Departments of Defense, Energy, Health and Human Services, Homeland Security, Transportation, and Treasury; the National Aeronautics and Space Administration; the Nuclear Regulatory Commission; and National Science Foundation also sponsor another 23 FFRDCs (NSF/SRS 2009a).

In contrast, articles published by authors in the private nonprofit sector are primarily in the medical sciences (55% of the sector's articles in 2008) and biological sciences (25%) (appendix table 5-42 ). Federal government authors show a similar pattern, with 30% in the biological sciences and 27% in the medical sciences.

Trends in Sector Coauthorship
This section considers coauthorship data as an indicator of collaboration at the sectoral level between U.S. institutional authors and between U.S. sectors and foreign institutions.[45] These data show that the growing integration of R&D activities, as measured by coauthorship, is occurring across the full range of R&D-performing institutions internationally as well as domestically.

Between 1998 and 2008, coauthorship within sectors increased for all U.S. sectors.[46] Coauthorship within academia rose from 38% in 1998 to 45% in 2008. FFRDC-FFRDC coauthorship increased 5 percentage points (table 5-21 ).

U.S. cross-sectoral coauthorships show a mixed pattern between 1998 and 2008. Coauthorship between FFRDCs and industry decreased. (Articles authored by industry physicists have been declining gradually across the period. Since a strong emphasis of FFRDC-authored articles is in physics (39%), it may be that fewer and fewer physicists are available in industry for potential coauthorship with physicists in FFRDCs.)[47] The largest gains in all sectors (6.8–9.8 percentage points) were with coauthors in academia, by far the largest sector with the largest pool of potential S&E coauthors. Cross-sector coauthorship with academic authors was higher in 2008 (54%–74%) than intrasector coauthorship within academia (45%), and cross-sector coauthorship with academia was higher in all sectors than any intrasector coauthorship (table 5-21 ).

Except for the decline in coauthorship between FFRDCs and industry, the indicators presented in this section hint at increasing integration between and among the different types of U.S. institutions that publish the results of R&D in the scientific and technical literature. Growth in coauthorship has been particularly strong between U.S. authors in academia and in all other sectors. Because of the predominance of the academic sector in S&E article publishing in the United States, academic scientists and engineers have been on the forefront of the integration of S&E research across institutions, both nationally and internationally.

International collaboration increased rapidly in the United States. International coauthorship rates rose by 7–10 percentage points between 1998 and 2008 (table 5-21 ). Authors at FFRDCs reached the highest rate of collaboration with foreign authors, at 42%, followed by industry and academia at 29% each. Astronomers in most U.S. sectors increased their rates of international coauthorship the most rapidly, and geoscientists, mathematicians, and physicists in most U.S. sectors also increased their collaboration with international colleagues at a higher-than-average pace (NSB 2008, 08-01A, p A5-66).

Trends in Citation of S&E Articles

Citations indicate influence. When scientists and engineers cite the published papers resulting from prior S&E research, they are formally crediting the influence of that research on their own work. Like the indicators of international coauthorship discussed above, cross-national citations are evidence that S&E research is increasingly international in scope. Between 1992 and 2008, international citations grew faster than total citations: 5.8% annually versus 4.6% (figure 5-23 ). By 2008, international citations were two-thirds of all citations.[48]

Two other trends accompanied the steady growth of international citations in the world's S&E literature: changing shares of total citations across countries and changing shares of highly cited S&E literature. These are discussed in the following sections.

Citation Trends in a Global Context
Shares of the world total of citations to S&E research articles have changed concurrently with shares of the world total of these articles. Appendix table 5-43 shows, for example, that between 1994–96 and 2004–06, the U.S. share of world S&E articles declined from 34% to 29% across all fields;[49] the U.S. share declined in every broad field, although the decline varied in size. Table 5-22 shows the parallel trends for the U.S. share of citations and indicates an even larger decline, from 47% to 38%.

China's share of both total world S&E articles and citations increased over the same period. However, in contrast to the global trend of increasing international citations, China's pattern has been different. Unlike the United States and other large article-producing countries, China's share of international citations decreased between 1998 and 2008, from 64% to 51%, suggesting that much of the use of China's expanding S&E article output—as indicated by citations to those articles—is occurring within China.

Trends in Highly Cited S&E Literature
Another indicator of performance of a national or regional S&E system is the share of its articles that are highly cited. High citation rates can indicate that an article has a greater impact on subsequent research than articles with lower citation rates.

Appendix table 5-43 shows citation percentiles for 1998 and 2008 by field for the top five S&E article-producing countries/regions.[50] In that table, a country whose research influence was disproportionate to its output would have higher numbers of articles in higher citation percentiles, whereas a country whose influence was less than its output would suggest would have higher numbers of articles in lower citation percentiles. In other words, a country whose research is highly influential would have higher shares of articles in higher citation percentiles.

This is the case in every field for U.S. articles. In both 1998 and 2008, as displayed in appendix table 5-43 , the U.S. share of articles in the 99th percentile was higher than its share in the 95th percentile, and these were higher than its share in the 90th percentile, and so forth, even while U.S. shares of all articles and all citations were decreasing. In 2008, U.S. articles represented 29% of the world total of 2 million articles in the cited period shown; the U.S. authored 52% of the rare 19,500 articles in the 99th percentile and 25% of the 1.2 million articles below the 50th percentile. This broad pattern was unchanged from the 1998 pattern.

Citations to the European Union's S&E articles displayed a different pattern: it had higher percentages of articles in the lower percentiles across all fields of S&E except in the agricultural sciences (appendix table 5-43 ). Figure 5-24 displays these relationships for all five countries/regions. Only U.S. publications display the preferred relationship of strongly higher proportions of articles in the higher percentiles of article citations; when cited, articles with authors from the European Union, China, Japan, and the Asia-8 are more often found in the lower citation percentiles. These data are summarized in appendix table 5-44 . As the U.S. share of all articles produced declined between 1998 and 2008, its share of articles in the 99th percentile (i.e., the top 1%) of cited articles also declined, particularly in some fields. Shares in the top percentile increased for the European Union, China, Japan, and the Asia-8.

When citation rates are normalized by the share of world articles during the citation period to produce an index of highly cited articles, the influence of U.S. articles is seen to have changed little over the past 10 years. Between 1998 and 2008, the U.S. index of highly cited articles barely changed (from 1.83 to 1.78) (figure 5-25 ; appendix table 5-44 ) and remained well above the expected index value of 1. During the same period, the European Union increased its index from 0.73 to 0.89, and China, Japan, and the Asia-8 increased their index values but remained below their expected values. In other words, the United States had 78% more articles than expected in the 99th percentile of cited articles in 2008, and the European Union had 11% fewer than expected. China had 58% fewer articles in the 99th percentile than expected in 2008, and Japan 42% fewer.

The United States experienced notable gains on the index of highly cited articles in engineering and computer sciences (although with relatively low counts in the latter) and a decline in chemistry (appendix table 5-44 ). The European Union reached its expected value in chemistry, physics, and the agricultural sciences. China achieved an index value of 1 in engineering and mathematics. Japan did not achieve its expected value in any broad field.

Academic Patents, Licenses, Royalties, and Startups

Other indicators of academic R&D outputs reflect universities' efforts to capitalize on their intellectual property in the form of patents and associated activities. The majority of U.S. universities did not become actively involved in the management of their own intellectual property until late in the 20th century, although some were granted patents much earlier.[51] The Bayh-Dole Act of 1980 gave colleges and universities a common legal framework for claiming ownership of income streams from patented discoveries that resulted from their federally funded research. To facilitate the conversion of new knowledge produced in their laboratories to patent-protected public knowledge that can be potentially licensed by others or form the basis for a startup firm, more and more research institutions established technology management/transfer offices (AUTM 2009).

Efforts to encourage links between university-based research and commercial exploitation of the results of that research have been widely studied by researchers. Mowery (2002) notes the strong growth in funding by NIH and the predominance of biomedically related patenting by universities in the 1990s. Branstetter and Ogura (2005) identify a "bio-nexus" in patent-to-paper citations, and Owen-Smith and Powell (2003) explore the effects of an academic medical center as part of the "scientific capacity" of a research university. In a qualitative study of two research universities that would appear to have similar capacities, Owen-Smith and Powell (2001) examine the very different rates of invention disclosure of the two campuses. Stephan and colleagues (2007) found strong differences in patenting activity among university scientists by field of science; a strong relationship between publication activity and patenting by individual researchers; and patenting by university researchers in only a small proportion of the potential population.

The following sections discuss overall trends in university patenting and related indicators through 2007–08.

University Patenting Trends
U.S. Patent and Trademark Office (USPTO) data show that annual patent grants to universities and colleges ranged from 2,950 to 3,700 between 1998 and 2008 (appendix table 5-45 ). In 2008, just over 3,000 patents were awarded to colleges and universities in the United States.[52] (Data in the next section on invention disclosures and applications suggest that patent grants to academic institutions may increase in the coming years.)

The top 200 R&D-performing institutions, with 96% of the total patents granted to U.S. universities during the 1998–2008 period, dominate among universities and university systems receiving patent protection.[53] College and university patents as a percentage of U.S. nongovernmental patents fell from 5.2% in 1998 to 4.3% in 2008. Among the top R&D-performing institutions that received patents between 1998 and 2008, 19 accounted for more than 50% of all patents granted to these institutions (although these included a few multicampus systems, including the Universities of California and North Carolina).

Between 1998 and 2008, three technology areas dominated U.S. university patenting: chemicals (19%), biotechnology (15%), and pharmaceuticals (14%) (appendix table 5-46 ). In numbers of patents, all three of these technology areas have declined from previous highs (figure 5-26 ). The next three highest technology areas over the period were semiconductors and electronics (6%), measurement and control equipment (5%), and computers and peripherals (5%), each accounting for about 200 patents in 2008 (appendix table 5-46).

Patent-Related Activities and Income
Data from the Association of University Technology Managers (AUTM) indicate continuing growth in a number of patent-related activities. Invention disclosures filed with university technology management offices describe prospective inventions and are submitted before a patent application is filed. These grew from 13,700 in 2003 to 17,700 in 2007 (notwithstanding a small decline in institutions responding to the AUTM survey over the same period) (appendix table 5-47 ). Likewise, new U.S. patent applications filed by AUTM respondents also increased, from 7,200 in 2003 to 10,900 in 2007. The AUTM survey respondents reported 348 startup companies formed in 2003 and 510 in 2007. The AUTM 2007 survey also found 3,148 cumulative, operational startup firms associated with U.S. university patenting and licensing activities (AUTM 2009).

Most royalties from licensing agreements accrue to relatively few patents and the universities that hold them, and many of the AUTM respondent offices report negative income. (Thursby and colleagues [2001] note that the objectives of university technology management offices include more than royalty income.) At the same time, large one-time payments to a university can affect the overall trend in university licensing income. In 2007, the 161 institutions that responded to the AUTM survey reported a total of $1.9 billion in net royalties from their patent holdings (appendix table 5-47 ).

Between 2003 and 2007, the inventory of revenue-generating licenses and options across all AUTM respondent institutions increased from 9,000 to 12,500 (appendix table 5-47 ). New licenses and options executed grew over the period from about 3,900 in 2003 to 4,400 in 2007.

Patent-to-Literature Citations

Citations to the S&E literature on the cover pages of issued patents are one indicator of the contribution of research to the development of practical innovation.[54] This indicator of science linkage to practical advance increased sharply in the late 1980's and early 1990's (Narin, Hamilton, and Olivastro 1997), due at least in part to developments in U.S. policy, industry growth and maturation, and court interpretation. At the same time, patenting activity by academic institutions was increasing rapidly, as were patent citations to S&E literature produced across all sectors (NSB 2008, pp. 5-49 to 5-54).

Between 1998 and 2008, growth on this indicator was much slower. Of utility patents awarded to both U.S. and foreign assignees, the number citing S&E articles (11% of total utility patents awarded in 2008) grew 1.4% annually over the 10-year period, compared with 0.7% annually for all utility patents (appendix table 5-48 ). Much of the growth in S&E citing patents was in patents awarded to non-U.S. assignees: these grew 3.1% annually.

Five broad S&E fields (the biological sciences, the medical sciences, chemistry, physics, and engineering) accounted for 97% of the total citations in these patents (appendix table 5-49 and figure 5-27 ). Citations to the biological sciences have decreased from their high of 58,000 in 1998 and 1999 but have more recently stabilized at around 50,000 per year. Citations to the medical sciences have increased since 2005 to about 26,000 in 2008.

The data discussed in the previous three paragraphs were heavily influenced by U.S. patents awarded to foreign assignees and references in those patents to non-U.S. S&E articles. Considering only citations to U.S. articles, overall growth in citations has been flat over the past 10 years (appendix table 5-48 ). Change in citations to articles authored in both the private nonprofit and government sectors has been negative over the period. Growth in citations to academic papers (0.9% annually) and to FFRDC papers (4.6% annually) shows that citations to papers in these two sectors have been replacing declining citations to articles in other sectors. Citations to academic articles account for most of this replacement, despite the slower rate of growth in these citations. Of total citations to U.S. articles in 2008, 64% were to academic articles, compared with 2% to FFRDC articles.

Figure 5-28 summarizes the increasing role of citations to U.S. academic articles in the science linkage to U.S. patents. Across all fields, academic articles made up 58% of total citations to U.S. articles in 1998 and 64% in 2008. Of the five broad fields of S&E that accounted for 97% of all patent citations to U.S. academic articles, increased shares of academic citations were notable in engineering (from 46% to 59%) and physics (from 43% to 65%). These increasing shares of patent citations to U.S. academic S&E articles are parallel to the increasing shares of academic S&E articles as compared with other sectors, as discussed above in the section "Trends in Output and Collaboration Among U.S. Sectors."


[31] Publication traditions in broad S&E fields differ somewhat. For example, computer scientists often publish their findings in conference proceedings, and social scientists often write books as well as publish in journals. Proceedings and books are poorly covered in the data currently used in this chapter.
[32] The U.S. sector identification in this chapter is quite precise; to date, sector identification has not been possible for other countries.
[33] European Union data include all member states as of 2007 (see appendix table 5-23 for a list of member countries).
[34] The Asia-8 includes India, Indonesia, Malaysia, Philippines, Singapore, South Korea, Thailand, and Taiwan.
[35] For example, Vatican City is not strictly a country; the Union of Soviet Socialist Republics (USSR) and Hong Kong are contained in the data in earlier years, but the USSR no longer exists and Hong Kong data are now reported as part of China. See appendix table 5-23 for a list of the countries represented in the data.
[36] Distributions of data in which a small percentage of cases accounts for a significant amount of the total value across all cases belong to a group of statistical distributions collectively referred to as power law distributions (Adamic 2000). Other phenomena with such distributions include, e.g., earthquakes (among a large number of earthquakes only a few have great power) and Internet traffic (visits to a relatively small number of sites account for a very large proportion of visits to all sites).
[37] For example, Montenegro appeared in the data in 2006 for the first time as an independent country; the tiny Pacific island nation of Niue appeared in 2007 for the first time because a coauthor from that country appeared in the data.
[38] See also NSB 2008, table 5-21, for detail on field level ranks and changes in rank since 1995.
[39] Coauthorship data are a broad, though limited indicator of collaboration among scientists. Previous editions of Indicators discussed possible underlying drivers for increased collaboration, including scientific advantages of knowledge- and instrument-sharing, decreasing costs of travel and communication, national policies, and so forth (NSB 2006). Katz and Martin (1997), Bordons and Gómez (2000), and Laudel (2002) analyze limitations of coauthorship as an indicator of research collaboration. Other researchers have continued using these data (Adams et al. 2005; Gómez, Fernández, and Sebastián 1999; Lundberg et al. 2006; Wuchty, Jones, and Uzzi 2007; Zitt, Bassecoulard, and Okubo 2000).
[40] The reader is reminded that the data on which these indicators are based give the nationality of the institutional addresses listed on the article. Authors themselves are not associated with a particular institution and may be of any nationality. Therefore the discussion in this section is based on the nationality of institutions, not authors, and makes no distinction between nationality of institutions and nationality of authors.
[41] The coauthorship data discussed in this paragraph are restricted to coauthorship across the regions/countries identified in table 5-17; that is, collaboration between or among countries of the European Union, e.g., is ignored. Intraregional coauthorship is discussed in the following sections.
[42] Readers are reminded that the number of coauthored articles between any pair of countries is the same; each country is counted once per article in these data. However, countries other than the pairs discussed here may also appear on the article.
[43] Finland is included here as one of the Scandinavian countries; Iceland is not.
[44] Article counts in this section are based on the year in which the article appeared in the database, not on the year of publication, and therefore are not the same counts as in the earlier discussion of total world article output.
[45] Identification of the sector of the non-U.S. institution is not possible with the current data set.
[46] Readers are reminded that coauthors from different departments in an institution are coded as different institutions.
[47] Referring to the declining share of industry's basic research articles in physics, the National Science Board (NSB) noted, "Most of this decline is accounted for by widespread restructuring of a few large corporations during this period, including closure, downsizing, or reorientation of large central research laboratories. Increased globalization, intensified competition, and commercial priorities may have contributed to the decline in publishing by companies and their researchers" (NSB 2008, p 6-36).
[48] This chapter uses the convention of a 3-year citation window with a 2-year lag, e.g., 2008 citation rates are from references in articles contained in the tape for 2008 to articles contained in the 2004, 2005, and 2006 tapes of the Thomson Reuters Science Citation Index and Social Sciences Citation Index databases. Analysis of the citation data shows that, in general, the 2-year citing lag captures the 3 peak citation years for most fields, with the following exceptions: in astronomy and physics, the peak citation years are generally captured with a 1-year lag, and in computer sciences, psychology, and the social sciences with a 3-year lag.
[49] The reader is reminded that articles in this section are counted by the year they entered the database, not by year of publication. Therefore article counts, and percentages based on them, are different from the data presented earlier in this section.
[50] Percentiles are specified percentages below which falls the remainder of the articles, e.g., the 99th percentile identifies the number of citations 99% of the articles failed to receive. For example, across all fields of science, 99% of articles from 2004 to 2006 failed to receive at least 22 citations in 2008. Matching numbers of citations with a citation percentile is not precise because all articles with a specified number of citations must be counted the same. Therefore, the citation percentiles discussed in this section and used in appendix tables 5-43 and 5-44 have all been counted conservatively, and the identified percentile is in every case higher than specified, i.e., the 99th percentile is always greater than 99%, the 95th percentile is always greater than 95%, and so forth. Actual citations/percentiles per field vary widely because counts were cut off to remain in the identified percentile. For example, using this method of counting, the 75th percentile for engineering contained articles with three to four citations in 2004 through 2006, whereas the 75th percentile for the biological sciences contained articles with five to eight citations.
[51] For an overview of these developments in the 20th century, see Mowery (2002).
[52] It is unclear whether the recent downturn in patents granted to universities/colleges is a result of changes in USPTO processing. For example, in its Performance and Accountability Report Fiscal Year 2008, USPTO reported an increase in average processing time ("patent average total pendency") from 29.1 months in 2005 to 32.2 months in 2008 (USPTO 2008).
[53] The institutions listed in appendix table 5-45 are slightly different from those listed in past volumes, and data for individual institutions may be different. In appendix table 5-46, an institution is credited with a patent even if it is not the first assignee, and therefore, some patents may be double counted. Several university systems are counted as one institution, and medical schools may be counted with their home institution. Universities also vary in how they assign patents, e.g., to boards of regents, individual campuses, or entities with or without affiliation with the university.
[54] Patent-based data must be interpreted with caution. Year-to-year changes in the data may reflect changes in USPTO processing times (so-called patent pendency rates). Likewise, industries and companies have different tactics and strategies for pursuing patents, and these may also change over time.

Patent citations to S&E research discussed in this section are limited to the citations found on the cover pages of successful patent applications. These citations are entered by the patent examiner, and may or may not reflect citations given by the applicant in the body of the application. Patent cover pages also contain references to scientific and technical materials not contained in the article data used in this chapter, e.g., other patents, conference proceedings, industry standards, etc. Analyses of the data referred to in this section found that nonjournal references on patent cover pages accounted for 19% of total references in 2008. The journals/articles in the SCI/SSCI database used in this chapter—a set of relatively high-impact journals—accounted for 83% of the journal references, or 67% of the total science references, on the patent covers.

Science and Engineering Indicators 2010   Arlington, VA (NSB 10-01) | January 2010