Alliance for Research on Regional Colleges
Scholars, researchers, and practitioners have always faced challenges distinguishing which colleges and universities are regional comprehensive institutions. An adaptability to regional contexts, focus on teaching, and historical service to marginalized populations are RCIs’ greatest strengths but collectively defy classification through quantitative data. The question this paper addresses, then, is how do we identify a group of institutions that is largely defined by mission, rather than by a specific set of data points? More to the point, how do we identify RCIs as a sector among U.S. institutions of higher education? This paper relies on institutional comparison groups, in which RCIs name institutions they view as comparable peers, to provide insight into which colleges and universities RCIs see as similar to themselves.
Scholars, researchers, and practitioners have always faced challenges distinguishing which colleges and universities are regional comprehensive institutions (RCIs). In some ways, it is remarkably easy to spot an RCI. Just like one could pick out a liberal arts college for its small, leafy, residential campus or a research university for its prominent athletics teams or ubiquitous brand, one could identify an RCI through several common traits. An RCI may have a direction in its name (e.g., University of Northern Iowa), or an observer may simply recognize it as their own local, public college—an institution deeply embedded in its community, employing a large proportion of its region’s population.
It is more challenging, however, to identify RCIs as an entire group. While one could find a liberal arts college based on the proportion of liberal arts baccalaureates it awards or a research university by the number of doctorates it grants, there is no single data point—or set of data points—that mark regionals as a sector. There is a simple explanation for this: Regional comprehensive colleges and universities are the Swiss Army Knives of higher education. Just as Swiss Army Knives contain different permutations of tools, each RCI comes with its own set of components that has been collected over time to fit local needs. Schneider and Deane said as much in a 2015 book when they attempted to identify RCIs, saying they “considered degree offerings, patterns of student enrollment, and proximity to population centers, but quickly discovered that these institutions operate on a complex continuum” (p. 6). Their adaptability to regional contexts, focus on teaching, and historical service to marginalized populations are RCIs’ greatest strengths but collectively defy classification through quantitative data.
This paper proposes a new strategy for identifying regional comprehensive institutions (RCIs). This strategy relies on self-identified comparison groups that U.S. postsecondary institutions have submitted to the Integrated Postsecondary Education Data System. Comparison groups are an important institutional decision and demonstrably signal where an institution turns when looking for new ideas. Indeed, they reflect an institution’s sense what it is and what it aspires to be, as well as its understanding of which institutions are similar. They are, in essence, social identity claims about belonging that give organizational outsiders a way to understand what a given institution does. Faced with the questions, “who are you, and what do you do,” a university president can respond, “we are like the institutions in our comparison group, and we do what they do.” In this sense, comparison groups provide an anchor for an institution, relating it to others.
Taken in aggregate, comparison groups can reveal which institutions others see as fitting together. See the example below in Figure 1. In this illustration, Institution A and Institution B both select a comparison group. Institution A selects three other institutions: {University 1, University 2, and University 3}. Institution B additionally selects three institutions: {University 1, University 2, and University 4}. This network of peer nominations is depicted in Column A of Figure 1. Looking at these connections from a different perspective, we can imagine an “associational network” in which institutions are linked because they are nominated together. Figure 1, Column B shows such a network. In this instance, universities share links if nominated in the same group. Universities 1 and 2 have a thicker connecting line because both Institutions A and B nominated them together, where Universities 3 and 4 do not share a connection because they were not chosen in the same comparison group.
Now, consider this process of co-nominating institutions totaled across U.S. higher education. Given many more comparison groups, we could, at one extreme, observe Universities 1 and 2 are frequently co-nominated, where Universities 3 and 4 never are. This pattern would suggest that others see Universities 1 and 2 as similar. In an example from 2019 IPEDS data, the University of North Carolina – Greensboro and Ball State University were co-nominated 10 times (mean = 1.99, sd = 1.58), though UNC Greensboro did not nominate Ball State in its own group. This finding illustrates that, whether or not UNC – Greensboro sees Ball State as a peer, other institutions routinely pair the two together. We thus have a window into which institutions are commonly associated that we could not otherwise observe. Through social network analysis techniques described in the subsequent sections, we can additionally observe which groups of institutions are commonly nominated together.
To find institutional groups, the paper employed a common technique to detect communities in social networks called modularity maximization. This method uses algorithms to identify clusters of institutions that are densely connected (i.e., groups that have many connections between them), with sparse connections between clusters. The goal is to maximize the modularity statistic (ranging between -1 and +1), which summarizes how many more ties exist within a group than would be expected at random.
The analysis yielded eight total clusters in the full associational network. For illustrative purposes, the paper focuses on two key groups of the eight: RCIs and research universities. Table 1 provides a summary of the institutional profiles comprising each group. The RCI group, relative to research universities, is larger with 497 institutions compared to 171 (see the full list of institutions). Identifying RCIs using this strategy includes more institutions than some other methods or well-known classification schemas, such as the Carnegie Master’s group, which in 2018 included 256 public master’s colleges and universities. This is likely because RCI peer groups routinely include a diverse pool of institutions, so the cluster likely includes baccalaureate colleges, doctorate-granting universities, and institutions that award a mix of baccalaureate and associates’ degrees.
A comparison between the two groups yields predictable findings. The RCI community contains a much higher proportion of public institutions. Howard University, an HBCU, several private regional colleges, and two special-focus colleges (culinary arts and health) comprise the nine private institutions. The private institutions in the research university category are well-known, highly selective universities. The RCI group additionally includes HBCUs and, on average, reported lower tuition sticker-prices ($9,161 versus $28,021). With respect to degrees awarded, RCIs demonstrated greater focus on undergraduate education and in 2019, on average, awarded 33 baccalaureate degrees for every doctorate awarded. Research universities, on average, awarded 7 baccalaureate degrees for every doctoral degree. It is also notable that the average RCI awarded 123 associates’ degrees in 2019, illustrating programmatic breadth. Figure 2 depicts the network of RCIs and identifies several example institutions.
RCIs and research universities also have dramatically different enrollment profiles (see Table 2), demonstrating RCIs’ historical commitments to educational access for groups who have been excluded and marginalized. For example, students receiving Pell grants comprised 47.74% of the average RCI’s undergraduate student body in 2019. RCIs as a group, then, enrolled approximately 1.86 million students receiving Pell grants. In comparison, research universities taken together enrolled approximately 863,000 students receiving Pell grants (22.49% of enrollment). We can observe a similar pattern when considering enrollment of underrepresented student groups. Black students, for example, accounted for 15.18% of the undergraduate student body at the average RCI, compared with only 6.05% at the average research university. Examining each group as a whole, then, RCIs enrolled approximately 590,000 Black students, while research universities enrolled approximately 232,000. Finally, the average RCI enrolled a greater proportion of in-state students in 2019 than did research universities, though this statistic likely reflects, at least in part, RCIs’ lower appeal to out-of-state students. Regardless of the cause, however, it is clear that RCIs are critical institutions that educate a significant proportion of their home states’ populace.
