Introduction
This research represents the first phase of a multi-part project investigating the collaborative dynamics between Carnegie R1 and R2 institutions. Here, we present our initial findings on the rate of co-authorship in scholarly journal articles, establishing a baseline for subsequent, more in-depth analyses. This preliminary study lays the groundwork for future investigations into the qualitative aspects of these collaborations, including the impact of resource disparities, the role of specific disciplinary cultures, and the long-term outcomes of inter-institutional partnerships. Our aim is to develop a comprehensive, longitudinal understanding of how institutional classification shapes the contours of scholarly collaboration within the contemporary research university.
The Carnegie Classification of Institutions of Higher Education, while at face value is simply a taxonomy in which to sort universities, serves as a crucial framework for understanding the diverse landscape of American higher education institutions. These classifications provide a standardized and widely recognized metric for categorizing institutions based on their research activity, degree offerings, and other key characteristics. This framework informs policy decisions, resource allocation, and even public perception of institutional prestige. Understanding the nuances of these classifications is essential for accurately interpreting trends within higher education, including trends in scholarly collaboration.
In this initial analysis, we report collaboration between Ph.D. granting R1 and R2 institutions in the United States by way of journal article coauthorship over a decade: 2014-2023. We mined the Academic Analytics commercial database for all journal articles matched to faculty authors in each annual database within our timeframe, and we tallied the number of articles with co-authors from both R1 and R2 institutions. It is important to note that the analyses presented in this phase of the project predate the most recent restructuring of the Carnegie classifications, and are based on pre-2025 classifications. However, we do note that some institutions changed classifications over our timeframe, and those changes are taken into account (e.g., Ohio University changed from R2 to R1 in 2022; Ohio University thus is classified in our dataset as an R2 institution in years 2014-2021 but is classified as R1 in 2022 and 2023).
Overall Collaboration Rates
The table below shows summary data from the first (2014) and final (2023) year of data in our sample. All metrics show an increase from 2014 to 2023, but vary in the magnitude of that increase. For example, the number of faculty at R2 institutions in our study increased more than 31%, compared to an increase in tenure-line faculty of only ~11% at R1 institutions, suggesting that R2 institutions may have seen much greater growth in faculty head-count over the previous decade. The number of faculty who are authors of at least one journal article also increased at a much higher rate among R2 institutions, suggesting an emphasis on increased research activity among R2 institutions in our timeframe.
The number of journal articles with at least one author from an R1 institution increased 22.2% from 2014 to 2023, while the number of articles with at least one R2 author increased by 32.6%. This is consistent with other studies that document an increase in journal article production over the past decade (e.g., Hanson et al., 2024; Savage & Olejniczak, 2022), and we wrote about this phenomenon and the concomitant increase in the number of scholar journals in our post to this weblog last month. Although growth in the publication of journal articles written by authors at R2 institutions outpaced growth in article production by R1 institutions, growth in the number of articles written collaboratively was greatest of all – the number of articles collaboratively authored by scholars from both R1 and R2 institutions increased by more than 56% between 2014 and 2023, suggesting deepening ties between these institutional categories.
There is nuance in these numbers, however. The percentage of articles authored by faculty from R1 institutions that also have R2 coauthors increased only ~1.25% (from 4.75% in 2014 to 6.01% in 2023). In other words, the vast majority (94-95%) of articles written by R1 scholars do not represent collaborations with R2 peers. Taken together these initial results seem to imply that the increase in R1-R2 co-authorships we observed stems more from an increase in research activity among R2 institutions than an explicit increase in the number of collaborations between institutions of different Carnegie classifications.
Some Comments on Individual Disciplines
Looking more closely into the data concerning R1 authorships with R2 co-authors, a few interesting variances emerge among broad fields of study. Within the Academic Analytics Taxonomy of disciplines, individual authors are classified according to their departmental affiliation. Departments are then aggregated into 11 broad fields such as Business, Engineering, Humanities, Physical and Mathematical Sciences, etc. As mentioned earlier, while the vast majority (94-95%) of articles written by R1 scholars do not represent collaborations with R2 peers, where there are collaborations, the percentages show interesting patterns.
The dearth of collaborations between R1 and R2 researchers across most broad fields is obvious. The Humanities tradition maintains a publishing ideal of the single scholar working alone. Bench sciences and engineering labs are under constant pressure to publish ahead of their competitors. The larger percentage of collaboration in Social and Behavioral Sciences may be a result of the growth of big data and the need for team research and publication in many of those fields. The anomaly is Business. Why were so many R1 authors seeking out R2 co-authors in 2014 and why did that increase in 2023? We leave these questions for the next installment of this blog.
R2 researchers are far more interested in collaborating with R1 researchers than vice versa. The percentages of R2 authors with R1 co-authors provide evidence for that. However, what is interesting is that some disciplines appear to place much more emphasis on collaboration than others. Looking at the Biological and Biomedical Sciences, the average R2 with R1 co-author percentage for the entire broad field is 4.7%. Yet, the single discipline of Epidemiology has a rate of 24.3% in 2014 and 23% in 2023.
Across all 19 Physical and Mathematical Sciences disciplines, the 2014 average R2 with R1 co-author percentage for the broad field is 9.9%. Three disciplines stand out with far greater percentages.
The 2014 Social and Behavioral Sciences average percentage for R2 with R1 co-author is 10.0%. Yet two disciplines display much greater percentages.
The 2014 Business average percentage for R2 with R1 co-author is 14.8%. Two related disciplines show far more interest in collaborations.
These numbers suggest something interesting is taking place. We will take a closer look in the next installment of our blog.
Future Directions
This initial foray into the dataset provides insights into the changing dynamics of R2 institutions, and their increased research activity profile. While we are far from ascribing causality to this result, we recall recent work by (Wapman et al., 2022), who documented the “prestige hierarchy” of US higher education, noting in particular that the faculty hired by R2 institutions are likely to have earned their Ph.D. at an R1 institution. In addition to (or coincident with) strategic plans among R2 institutions to increase their research output, hiring faculty from among the graduates of R1 institutions, who maintain their network of collaborators at their Ph.D.-granting institution, may have played a role in the outsized pace of research growth among R2s.
In future months, we will continue to grow this study. In particular, we plan to focus on the following areas:
Understanding the importance of author order on the articles we include in the study: are R1 or R2 authors the first, last, or corresponding authors on collaborative studies?
Studying temporal trends by discipline and broad field of study.
Some schools moved from R2 to R1 classification in our study period – does rate of collaboration with R1 peers prior to their reclassification have a relationship with their promotion to R1 status?
What do these data say in light of the new Carnegie structure revealed earlier this year?
We are eager to know whether you have other suggestions as we move forward, and we look forward to presenting “Part 2” here on this weblog and eventually as a preprint and scholarly article.
References
Hanson, M. A., Barreiro, P. G., Crosetto, P., & Brockington, D. (2024). The strain on scientific publishing. Quantitative Science Studies, 5(4), 823–843. https://doi.org/10.1162/qss_a_00327
Savage, W. E., & Olejniczak, A. J. (2022). More journal articles and fewer books: Publication practices in the social sciences in the 2010’s. PLOS ONE, 17(2), e0263410. https://doi.org/10.1371/journal.pone.0263410
Wapman, K. H., Zhang, S., Clauset, A., & Larremore, D. B. (2022). Quantifying hierarchy and dynamics in US faculty hiring and retention. Nature. https://doi.org/10.1038/s41586-022-05222-x
