Researchers, Expenditures, and Publications within the Academic Sector

Although research funds, research personnel, and articles all increased within the academic sector from 1994 to 2011, they did so at markedly different rates (figure 1). After adjustment for inflation, academic S&E R&D expenditures doubled, rising from $24.5 billion to $49.7 billion. Meanwhile, academic research personnel and article counts grew more slowly, increasing by approximately 30% during the same period.[23]

FIGURE 1. Academic R&D publications, researchers, and expenditures: 1988–2011
FIGURE 1. Academic R&D publications, researchers, and expenditures: 1988–2011.

NOTES: Whenever a year is noted, it refers to the publication year. Data for academic research personnel are lagged by one year, and data for R&D expenditures are lagged by two years, to account for the passage of time between funding, conducting, and publishing results of the research. Thus, the data shown in the figure for the estimated researchers in 1994 are actually 1993 estimates, and data shown in the figure for expenditures in 1988 are actually 1986 expenditures. Data on academic researchers prior to 1993 are not presented because data from 1988 to 1993 are not comparable to data from 1993 on. Total publications are computed using whole counts.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and Colleges and Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 1 Source Data: Excel file

When the timeline is extended back to 1988, it becomes clear that the growth in expenditures varied over the full 23-year period. Average annual growth in expenditures was very strong from 1999 to 2006 at 7%, with 5% annual growth prior to that and 2% annual growth from 2007 to 2011.

Although both of the input factors are very good predictors of the article output, expenditures are more tightly coupled with article output than research personnel.

Overall Trends in the Ratios of Expenditures to Articles and Articles to Researcher

Over the period 1988–2011, there was a steady increase in the ratio of spending to S&E articles, regardless of the counting method used (figure 2). After adjustment for inflation, this ratio (using whole counts) increased by over 80% from 1988 to 2011 ($122,000 to $224,000). Growth rates were highest during the late 1990s and early 2000s, coinciding with the near-doubling of NIH's budget. From 2006 to 2011, the spending-to-publications ratio was relatively stable, dipping slightly and then rising slightly.[24]

FIGURE 2. Ratios of academic R&D expenditures to publications: 1988–2011
FIGURE 2. Ratios of academic R&D expenditures to publications: 1988–2011.

NOTE: Constant 2005 dollars are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and Colleges; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 2 Source Data: Excel file

Possible explanations for the increasing trend in the ratio of spending to publications through the mid-2000s are numerous. For one, research leading to publications may have become more complex and costly. For example, interdisciplinary research often requires extensive communication to synthesize different perspectives on a problem into a research project.[25] Or, in response to changing funding priorities, universities may have focused spending on activities and investments that did not necessarily result in publications. Another possible explanation could be that changes in university culture, specifically toward leadership that thinks increasingly in business terms, have resulted in a growing emphasis on converting research into patents.

Changes in administrative responsibilities associated with conducting R&D may also have affected the spending ratio.[26] Survey results from the Federal Demonstration Partnership—a program sponsored by the National Academies to reduce administrative burdens associated with research grants and contracts—indicate that the percentage of time that federally funded principal investigators spend on administrative matters (excluding proposal writing) has more than doubled over the past two decades, rising from 18% in the early 1990s to 42% in 2010. Bienenstock suggests that a variety of factors may have contributed to this increase.[27]

The ratio of articles to academic researchers was relatively stable over the period of analysis. From 1994 to 2011, an average of about one article annually was associated with every researcher. From 1994 to 2002 there was a slight dip in the publications-to-researcher ratio. Output per researcher then grew steadily from 2003 to 2008, only to dip again slightly from 2009 to 2011. The relatively constant pattern over time was evident whether articles counts were computed using whole counts or fractional counts (figure 3).

FIGURE 3. Ratios of academic publications to researchers: 1994–2011

FIGURE 3. Ratios of academic publications to researchers: 1994–2011.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 3 Source Data: Excel file

Trends by University Type

The pattern of an increased academic R&D expenditures-to-publications ratio was replicated at all university types (figure 4). This ratio rose fastest at less research-extensive academic institutions (3.2% average annual increase). Increases were greater at public universities (2.7% average annual increase) than at their private counterparts (1.8% average annual increase). The gap in spending between public and private schools that appeared in the early 1990s has widened since then, with the spending-to-publications ratio being about the same for public and private schools in 1988, about $13,000 higher for public schools in 1994, and $38,000 higher for public schools in 2011. Unlike for public universities, in private schools the ratio of spending to articles produced was relatively stable throughout the 1990s. The lowest average annual increase was at the Research I private universities.

FIGURE 4. Ratios of academic R&D expenditures to publications, by university type: 1988–2011

FIGURE 4. Ratios of academic R&D expenditures to publications, by university type: 1988–2011.

RI = Research I university.

NOTE: Whole article counts and constant 2005 dollars are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and College, special tabulations; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 4 Source Data: Excel file

In general, private universities and colleges have only a slightly higher publications-to-researcher ratio than their public counterparts, but among the most research-extensive academic institutions, the difference is somewhat greater (figure 5). This could reflect greater focus at private institutions on publishing, as well as the broader missions of public institutions. The nation's Research I institutions were credited with over twice as many publications as less research-extensive institutions. Trends in publication rates are relatively flat for most university types but show an increase from 2004 to 2009. The greatest fluctuation over the period was for the nation's Research I private institutions, which dropped from 2.4 articles in 1998 to 2.1 in 2003 before rising back up to 2.4 again by 2009.

FIGURE 5. Ratios of academic publications to researchers, by university type: 1994–2011
FIGURE 5. Ratios of academic publications to researchers, by university type: 1994–2011.

RI = Research I university.

NOTE: Whole article counts are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 5 Source Data: Excel file

This increase in publication rates during the 2000s throughout the academic sector likely reflects the near-doubling of the National Institutes of Health's (NIH's) research budget between 1998 and 2003; during the middle years of the 2000s, academic researchers in the life sciences would have completed and begun to report the results of research paid for by grants resulting from NIH's increased research budget.[30] With the life sciences constituting well over 50% of total academic R&D expenditures, any changes affecting NIH's research budget have a corresponding effect on total academic R&D.

Trends by Broad Field

Publication practices vary across the various S&E fields. For example, social sciences researchers tend to publish fewer journal articles and more books than do researchers in health sciences. Also, in the data under analysis there are ten times as many life science articles than social sciences articles but only twice as many researchers. Because of these field differences, aggregate comparisons across all fields of S&E of the publications-to-researcher ratio are affected when universities specialize in certain fields. Similarly, field differences can also complicate comparisons across fields in the spending-to-publications ratio. Fields requiring more sophisticated research equipment require greater expenditures. Thus it is important to look at trends within individual fields. Aggregate academic sector trends over time were not replicated at the level of each broad S&E field (figure 6).[31]

FIGURE 6. Ratios of academic R&D expenditures to publications, by major field: 1988–2011
FIGURE 6. Ratios of academic R&D expenditures to publications, by major field: 1988–2011.

NOTE: Whole article counts and constant 2005 dollars are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and Colleges, special tabulations; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 6 Source Data: Excel file

Trends in life sciences publishing, making up the majority of academic publishing, mirrored trends for the academic sector as a whole throughout most of the time period. The percentage increases in the spending-to-publications ratio from 1988 to 2011 were highest in life sciences (120%), followed by social sciences (just under 70%) then engineering (just under 40%). The spending ratio in physical sciences remained relatively flat (less than 15% increase).[32]

Interesting differences exist by field in the relationships between academic researchers and publication counts over time. The ratio of publications to researchers rose in the physical sciences and somewhat in engineering, remained fairly stable in the social sciences, and decreased slightly in the life sciences (figure 7).[33]

FIGURE 7. Ratios of academic publications to researchers, by major field: 1994–2011
FIGURE 7. Ratios of academic publications to researchers, by major field: 1994–2011.

NOTE: Whole article counts are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 7 Source Data: Excel file

Trends by University Type and Field

As already discussed, public universities and colleges had a higher spending-to-publications ratio than their private counterparts (figure 4). This is true for each of the broad fields except for engineering, where private institutions had a higher spending ratio but less growth over the past 15 years (figure 8)

FIGURE 8. Ratios of academic R&D expenditures to publications, by major field and institutional control: 1988–2011

FIGURE 8. Ratios of academic R&D expenditures to publications, by major field and institutional control: 1988–2011.

Whole article counts and constant 2005 dollars are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and Colleges, special tabulations; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 8 Source Data: Excel file

Also, with the exception of engineering, public schools showed similar spending-to-publication ratios in each field during the first few years of this analysis. The spending-to-publications ratio in the physical sciences varied somewhat between public and private institutions. It increased very slightly at public universities and colleges while dropping somewhat at their private counterparts. From the late 1990s through 2005, this ratio for social sciences publications grew rapidly at public schools while showing essentially no change at private schools.

In every broad field, the spending-to-publications ratio was generally higher for Research I institutions than for less research-extensive institutions (figure 9). The trend in each field and institution type was toward an increasing spending ratio. The one exception was physical sciences in Research I institutions, where the spending to publications ratio held steady throughout the period.

FIGURE 9. Ratios of academic R&D expenditures to publications, by major field and university type: 1988–2011
FIGURE 9. Ratios of academic R&D expenditures to publications, by major field and university type: 1988–2011.

RI = Research I universities.

NOTES: Whole article counts and constant 2005 dollars are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Research and Development Expenditures at Universities and Colleges, special tabulations; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 9 Source Data: Excel file

As previously stated, the publications-to-researcher ratio was highest in the physical and life sciences; however this ratio was rising in the physical sciences and falling in the life sciences (figure 7). This falling ratio in the life sciences is attributable mainly to private and Research I universities and colleges, both of which dropped by 0.3 publications per researcher over the time period, while rates for public and less research-extensive institutions changed very little from the beginning to the end of the period (figures 10 and 11).

FIGURE 10. Ratios of academic publications to researchers, by major field and institutional control: 1994–2011
FIGURE 10. Ratios of academic publications to researchers, by major field and institutional control: 1994–2011.

NOTE: Whole article counts are used.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 10 Source Data: Excel file

FIGURE 11. Ratios of academic publications to researchers, by major field and university type: 1994–2011

FIGURE 11. Ratios of academic publications to researchers, by major field and university type: 1994–2011.

RI = Research I universities.

NOTES: Whole article counts are used. Fields refer to the fields of first doctorate in science, engineering, or health.

SOURCES: National Science Foundation, National Center for Science and Engineering Statistics, Survey of Doctorate Recipients; Patent Board analysis of Thomson Reuters Science Citation Index and Social Sciences Citation Index publications data.

Figure 11 Source Data: Excel file

Although the publications-to-researcher ratio in physical sciences has surpassed that in the life sciences for public and Research I institutions, this crossover has yet to occur for private institutions and only recently occurred for less research-extensive schools. Further research is needed to explore the dynamics affecting life sciences and physical sciences R&D in public and private universities. For the two remaining broad fields, engineering and social sciences, institution type has little impact on publication rates—rates and trends are nearly identical for public and private universities and colleges.

Notes

[23] Estimates of personnel growth rates were somewhat higher from 1994 to 2011 (just over 40%) for the smaller set of academic researchers whose primary responsibility was to conduct research.

[24] Higher rates and somewhat greater growth in recent years for fractional counts can be attributed to the trend toward increasing collaboration; recall that fractional counting credits only a portion of each coauthored article to each participating institution.

[25] See Newell WH, 2007. Decision-making in interdisciplinary studies. In Morcol G, editor, Handbook of Decision Making. Boca Raton, FL: CRC Press.

[26] See National Science Board (NSB). 2014. Reducing Investigators' Administrative Workload for Federally Funded Research. NSB 14-18. Arlington, VA: National Science Foundation. Available at http://nsf.gov/pubs/2014/nsb1418/nsb1418.pdf.

[27] Bienenstock A. 2009. Administrative burdens stifle faculty and erode university resources. APS News. http://www.aps.org/publications/apsnews/200907/backpage.cfm/.

[28] NIH's share of total federal obligations for academic R&D grew by 10 percentage points in the 20-year period spanning 1989–2009, rising from 54% in 1989 to 64% in 2009.

[29] Bienenstock A. 2009. Administrative burdens stifle faculty and erode university resources. APS News. http://www.aps.org/publications/apsnews/200907/backpage.cfm/.

[30] There was a 10% increase in life sciences publications from 2000 to 2005, likely reflecting the increase in NIH's budget. For the 5-year period prior to 2000 (1994 to 1999), life sciences publications increased by 2%.

[31] See Rosenbloom JL, Ginther DK, Juhl T, and Heppert J. The Effects of Research and Development Funding on Scientific Productivity: Academic Chemistry, 1990–2009. Working Paper 20595. Cambridge, MA: National Bureau of Economic Research. Available at http://www.nber.org/papers/w20595. This paper analyzes trends in the relationship between spending in chemistry (including spending in chemical engineering) and publications and finds that spending per publication decreased over time. The paper also reports less spending per publication at the nation's Research I universities than at the other institutions it covers. These findings differ from those in our analysis. Two important differences in methodology are (1) National Bureau of Economic Research (NBER) limited their analysis to a subset of about 150 universities and colleges that received the most federal R&D funding over the period 1990–2009, and (2) NBER's publications data set added many journals over time. In addition, NBER did not attempt to control for growing academic R&D collaboration over time.

[32] The expenditures-to-publications ratio for social sciences increased until around 2005, when this spending dropped sharply until 2008 and then leveled off. The data sources used in this analysis do not illuminate this atypical pattern involving the social sciences. One explanation may be that Thomson Scientific added an unusually large number of social sciences journals to their Social Sciences Citation Index coverage in 2008. In ensuing years, Thomson dropped many of the newly added journals.

[33] Trends were generally similar using fractional counts.