States invest heavily in academic research with the expectation that these efforts will advance scientific knowledge, support innovative industries, and strengthen local talent pipelines. Comparing research performance across state lines is difficult due to differences in academic landscapes: some may have large medical schools with high-cost labs, while others have research-active public universities in lower-cost fields or are more pedagogically focused.
Article output is one way of measuring academic research. This edition of Useful Stats standardizes science and engineering (S&E) article output for peer-reviewed documents (e.g., articles, reviews, and conference proceedings) published in refereed scientific journals at the state level by two complementary numerators. These numerators showcase different facets of academic research productivity: output per million dollars of academic S&E R&D spending, and per 1,000 science, engineering, and health (SEH) doctorate holders employed in academia. Although similar, each highlights a distinct part of the research ecosystem, with one reflecting the intensity of research spending and the other the publishing activity of academic researchers themselves.
All data used in this article are from the National Center for Science and Engineering Statistics (NCSES), available here. Refer to the data notes section at the bottom of this article for more details on the data and its limitations.
Costs per article appear to be increasing
Academic S&E article output on a per million USD of academic S&E R&D basis has steadily dropped almost every year from 1998 to 2023, from nearly eight articles per million USD to just over three, across all captured jurisdictions except Puerto Rico. Breaking down the standardized metric into its two component parts, academic S&E article output and academic S&E R&D, reveals that both have increased over time. The R&D value, however, has outpaced that of article output, leading to a decrease in the number of articles each dollar nets. This may be due to the nature or area of the research changing to decrease publication emphasis, larger portions of the R&D investment being sponsored by private corporations, and simply the cost of doing research having increased.
Refer to the section below on each component metric for a more detailed breakdown, as well as Figure 1 for a visual representation of the trends.
The states with the largest relative reductions in output per million USD are the EPSCoR states of Delaware (-75%), Maine (-72%), and Montana (-67%).
Notably, 25 states, including many of those with higher R&D expenditures, such as Texas (-58%), Washington (-60%), California (-60%), Pennsylvania (-62%), New York (-63%), and Ohio (-64%), all performed worse than the national value.
Adjusting for inflation, the picture is slightly more positive with a national decrease of -21% with eight EPSCoR jurisdictions—Puerto Rico (107%), Idaho (34%), Nevada (18%), Oklahoma (15%), South Carolina (14%), South Dakota (7%), Hawaii (5%), and Arkansas (1%)—experiencing an increase in output per million of constant USD. One state, Wyoming, remained unchanged, while the remaining 43 jurisdictions decreased.
It is, however, very important to note the typical lag between funding and publication; publication can be a time-consuming process in some cases, which influences the data to a certain degree by masking impact for as long as multiple years. Research itself may also not be completed in the same year as the funding is received. These masked impacts, in the case of the standardized metric, may cause a mismatch between the numerator and denominator—more specifically, the R&D dollars expressed may not be reflective of the articles output for the same year.
Figure 1 below includes a drop-down menu to toggle between the current and constant USD standardized metrics, as well as the component metrics for each. Note that each line chart has a separate y-axis, which is better suited for comparing trends across state lines rather than directly comparing values.
Figure 1: Academic S&E article output per million USD of academic S&E R&D
Academic S&E article output on a per 1,000 SEH doctorate holders in academia basis has, unlike the per-R&D metric, slowly increased over time. This increase is due to article output outpacing the growth of doctorates in academia.
In the 26 years from 1998 to 2023, academic S&E article output per 1,000 SEH doctorate holders in academia increased 13% nationwide. Individual jurisdictions ranged from North Dakota (110%), Mississippi (78%), New Mexico (76%), and Nevada (74%) on the high end to Iowa (-13%), Oregon (-14%), Arizona (-14%), and Puerto Rico (-35%) on the low end.
A total of eight jurisdictions decreased over the 26 years, notably including California, while 43 increased, and the remaining one, New Hampshire, stayed the same.
Figure 2 below includes a drop-down menu beneath the title to toggle between the standardized metric and its components. Note that each line chart has a separate y-axis, which is better suited for comparing trends across state lines rather than directly comparing values.
Figure 2: Academic S&E article output per 1,000 SEH doctorate holders in academia
Below is a very brief look at each component metric:
Academic S&E article output
Academic S&E article output has increased across every state and territory captured by the NCSES data over the full range of years from 1998 to 2023. The national value increased 66% over these 26 years, with jurisdictions ranging from more than doubling (North Dakota, 108%; Florida, 114%; Nevada, 119%; and South Dakota, 184%), to much more modest increases (Alaska, 29%; New Mexico, 28%; Hawaii, 28%; Iowa, 16%; and Louisiana, 11%).
In terms of increase in raw output, big states with lots of institutions of higher education top the list: California increased by nearly 13,000 S&E articles from 1998 to 2023, followed by Texas (11,500) and New York (10,500).
On the other end of the spectrum, smaller jurisdictions with fewer institutions of higher education, like Alaska (64), Puerto Rico (104), Wyoming (150), and Montana (199) reasonably have smaller increases in terms of count. However, these increases still represent impressive growths within each—the specifically mentioned states had percent increases of approximately 30-50%.
Refer to Figure 1 above for additional details on these trends and more.
Academic S&E R&D
Academic S&E R&D has increased in all jurisdictions over the 26 years from 1998 to 2023, with the national total jumping by nearly 300% or $76 billion. All captured jurisdictions except for Puerto Rico more than doubled, with North Dakota (512%), Delaware (499%), Maine (468%), and Tennessee (408%) topping the nation with more than 400% increases.
Adjusting for inflation, however, shows a more modest increase in the academic S&E R&D values; the U.S. total still more than doubled, at 110%, along with 28 states. Only Puerto Rico’s value decreased (-37%).
Refer to either Figure 1 or 2 above for additional details on these trends and more.
SEH doctorate holders in academia
From 1998 to 2023, SEH doctorate holders in academia have increased 32% nationwide, with nearly 112,000 new SEH doctorates within academia in just 26 years. While much of the increase is from states like California, Texas, New York, Pennsylvania, and Georgia, jurisdictions with the largest relative increases include Arizona (50%), Puerto Rico (48%), Georgia, and Idaho (each 47%).
Only three states’ SEH doctorate holders in academia decreased: Mississippi (-3%), Wyoming (-5%), and New Mexico (-40%).
Refer to Figure 2 above for additional details on these trends and more.
Data limitations
Articles were fractionally allocated among states as determined by the institutional addresses of all named authors to prevent double-counting. The publication total for the national value referenced in the analysis includes Puerto Rico and all articles without a determined location.
As per NCSES:
“Publication counts are based on the number of articles that appear in a set of journals tracked by Elsevier’s Scopus database. The journal set consists of S&E publications (including publications on the natural sciences, the applied sciences, the medical sciences, and the social sciences but excluding the arts and humanities). Only documents published in referred (sic) scientific journals are counted (mostly articles, reviews, and conference proceedings), as these documents were reviewed by peers prior to being accepted for publication.”
Academic R&D is only included for institutions with R&D over $150,000 each year.
Academic institutions include 2-year colleges, 4-year colleges and universities, medical schools, and university-affiliated research centers.
While the SEH doctorates in academia metric data are less complete compared to R&D, with only select years available between 1997 and 2023, SSTI has computed a 1998 value, calculated as the average of 1997 and 1999 values, for a more direct comparison with the other data available.
Refer to the methodologies for all above data here.
This page was prepared by SSTI using Federal funds under award ED22HDQ3070129 from the Economic Development Administration, U.S. Department of Commerce. The statements, findings, conclusions, and recommendations are those of the author(s) and do not necessarily reflect the views of the Economic Development Administration or the U.S. Department of Commerce.