S&E Publications in Peer-Reviewed Journals

Output indicators in the form of articles appearing in the research literature are discussed in Chapter 5 because academic researchers account for most of those articles. This section focuses on trends first in the number and share of S&E articles produced by authors affiliated with industry, then in their collaboration patterns with other U.S. sectors and internationally.[34]

Number of Articles

Trends in the number of S&E articles written by industrial researchers that appear in peer-reviewed journals, while not a direct indicator of innovation, are a rough indicator of outputs from research being carried out in industrial settings. This section examines the total number of articles authored by industry researchers as an indicator of overall industrial research activity, and the number of articles by these researchers published in basic research journals as an indicator of the volume of basic research carried out in industrial laboratories.[35]

Articles With an Industrial Author

The number of scientific articles with at least one author in U.S. private industry fluctuated between about 13,000 and 16,000 per year between 1988 and 2005, peaking at slightly more than 16,000 in 1991, then falling to its lowest level just below 13,000 in 2004. During this same period, however, the total number of U.S. S&E articles increased from 169,000 to 215,000 (appendix table 6-27Excel.). Consequently, industry’s overall share of U.S. article output declined from just below 9% to about 6% (figure 6-26figure.).

Six broad fields accounted for about 90% of the S&E literature by U.S. industry authors from 1988 to 2005: biological sciences, medical sciences, engineering, chemistry, physics, and the geosciences. With one exception, the number of industry articles peaked in 1995 or earlier for all of these fields. The exception is medical sciences, for which articles increased throughout the period, peaking in 2005. In four of these broad fields, industry’s share of all U.S. articles in the field declined between 1988 and 2005, from 26% to 14% in engineering, 18% to 8% in physics, 17% to 11% in chemistry, and 7% to 5% in the geosciences. Industry’s share of articles in the biological sciences remained stable throughout the period (between 6% and 8%), whereas its share of articles in the medical sciences increased (from 3% to 5%) (figure 6-26figure.).

Articles in Basic Research Journals

Between 1988 and 1995, the total number of basic research articles having authors in U.S. private industry fluctuated between 3,400 and 4,200 per year (appendix table 6-28Excel.). However, after peaking in 1995, the number declined by 30% through 2005. In contrast, the total number of basic research articles by authors from all sectors grew between 1995 and 2005. As a result, industry’s share of this output declined, from slightly more than 6% to 4% (figure 6-27figure.).

Five broad fields accounted for about 95% of the basic research literature by U.S. industry authors during the entire 18-year period: biological sciences, chemistry, physics, the geosciences, and the medical sciences. The trend in the number of basic research articles by U.S. industry researchers in the biological, medical, and geosciences, as a percentage of basic research articles in those fields, generally mirrored the trend for all fields, with gradual declines in share of about 1 percentage point.

Article output by U.S.-industry authors in physics and chemistry showed notably different patterns. In physics, the total number of these articles decreased sharply from nearly 1,000 in 1988 to about 300 in 2005. As a result, industry’s share of basic research articles in physics dropped by more than 7 percentage points (figure 6-27figure.). 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.

The pattern in chemistry has been different. U.S.-industry authors’ share of basic research articles in chemistry fluctuated between 9% and 13% over the period. Researchers at large pharmaceutical companies continued or increased their already strong publishing traditions in chemistry basic research journals despite consolidation within the industry. The pharmaceutical industry’s far greater reliance on patents and exclusivity for intellectual property protection relative to other industries may have played a role in its continued strong publishing record. Beyond pharmaceuticals, some of the same companies that saw declines in physics basic research articles also declined in chemistry.

Changing Emphasis on Basic Research

Industrial publications tended to shift away from basic research between 1988 and 2005. After peaking at 26% in 1995, the percentage of articles with an industrial author published in basic research journals declined to 22% by 2005 (figure 6-28figure.). [36]This declining emphasis on basic research in industry publications has been especially strong in the biological sciences (from around 50% in the early 1990s to 39% in 2005), in physics (from 31% in 1988 to 20% in 2005), and in the medical sciences (from 10% in the early- to mid-1990s to 5% in 2005). Again, however, the pattern in chemistry has been quite different. The basic research share of industrially authored articles in chemistry increased from around 30% during the late 1980s to 46% in 2005.

Industry Collaboration in Publications

Both in the United States and worldwide, a major increase in collaboration across sectors and countries on S&E publications has been evident during the past decade. (For a more complete discussion of collaboration patterns, see "Coauthorship and Collaboration" and "Trends in Output and Collaboration Among U.S. Sectors" in chapter 5.)

Articles by Institutional Author Type

Articles with one or more authors in private industry can be broken down into five unique types:

  • Single company-single author[37]
  • Single company-multiple authors
  • Multiple companies, with authors from more than one U.S. company
  • Multiple sectors, with U.S. authors from more than one sector[38]
  • International, with at least one foreign author.

Between 1988 and 2005, single company-single author articles declined by almost 60% (to about 2,000) and single company-multiple author articles declined by almost 40% (also to about 2,000) (appendix table 6-29Excel.). Multiple-company articles increased by 20% during this period. In contrast, multiple-sector articles and international articles increased by about 70% and 300%, respectively (about 5,000 in both cases). The net result of these trends were drops from 19% to 6% in the proportion of single company-single author articles and from 30% to 14% for single company-multiple author articles. During the period, international articles increased from 9% to 26% and multiple-sector articles increased from 36% to 47% (figure 6-29figure.).

Industry Collaboration Across U.S. Sectors

Coauthorship data indicate that U.S. industry collaborates more frequently with the academic sector than with other U.S. sectors.[39] Since 1988, more than 60% of the articles that industry authors have coauthored with someone outside their company have had an academic coauthor (appendix table 6-30Excel.). This is not unexpected, because the vast majority of S&E articles with a U.S. author include an author from academia.

Although the number of industry articles not limited to a single company increased substantially between 1988 and 2005, collaboration patterns between industry and other sectors changed very little during that period (figure 6-30figure.). The only sector in which a large change in collaboration has occurred is the private nonprofit sector. The proportion of industry articles coauthored with the private nonprofit sector steadily increased from 9% to 15% from 1988 to 2005.


[34] These articles are identified by at least one author having a private, for-profit institutional address.

[35] In this section, article counts were reported on a fractional-count basis. In the following section's discussion of collaboration trends, articles are reported on a whole-count basis. See the sidebar "Bibliometric Data and Terminology" in chapter 5 for a description of these methods of counting articles and how they are generally used.

[36] In contrast to the decline in emphasis on basic research in industry publications, about one-third of U.S. publications overall were published in basic research journals from 1988 to 2005.

[37] All addresses for a company and its subsidiaries are unified into a single code for the parent company.

[38] Other U.S. sectors in which researchers produced articles are academia, the federal government, state and local governments, federally funded R&D centers, and the private nonprofit sector.

[39] The base for the percentages discussed in this section is the number of industry articles with one or more industry authors minus the number of single company articles.

Right-click on image to save.