August 7, 2020
The Geography of Campus Closures and How to Avoid Them
Mapping At-Risk Metro Areas and Identifying Interventions to Sustain America’s Colleges and Universities
By Shea O’Neill, Vice-President at U3 Advisors and Elliot Felix, Founder of brightspot strategy
By 2019, years of enrollment decline and debt accumulation led Marygrove College to close its doors. Before the Detroit-based institution could join the ranks of over 140 four-year colleges and universities forced down the same path in the last 5 years,1 something remarkable occurred. The Kresge Foundation stepped in to help Marygrove restructure its debt and subsumed the college’s landholdings on to their balance sheet. This intervention afforded Marygrove an opportunity to structure a deal with Oakland University to accommodate former students, and it created an opportunity for Kresge and their partners to repurpose the campus into an early childhood education center and a K-12 community school.2
Why did Kresge take such an unprecedented step for such a relatively small and locally-focused institution? To understand, it is important to place Marygrove in context. While Marygrove is not an HBCU, it served a similar mission for its community. Historically, over 89% of the college’s students were people of color and ~75% were Pell Grant recipients (a traditional barometer of low-income students). Marygrove also occupied 52 acres in one of the most underserved neighborhoods in all of Detroit. Vacancy of Marygrove’s facilities would have dealt another blow to a community that had already suffered so much. These factors were critical for The Kresge Foundation, who has long-standing commitments to the city of Detroit.
Marygrove’s experience reveals what we miss when we discuss college closures solely within the terms of “supply and demand” or “liquidations and acquisitions.” College closures can create profound ripples that reverberate well beyond their doors, impacting local communities, and potentially derailing opportunities for economic advancement.
In recognition of the importance of this subject, U3 Advisors and brightspot have undertaken a long-term research project to track risk factors for 2,3003 four-year private not-for-profit and public institutions4 and to synthesize best practices from colleges working to mitigate these risks. From each institution, we used longitudinal data from IPEDs to grade their competitive advantage and their financial health. For competitive advantage we evaluated short and long-term enrollment changes, investment in online programming, reliance on international students, admission and enrollment rates, and geographies of recruitment. For financial health we evaluated overall solvency, strength of assets (primary reserve ratio, endowment coverage), debt reliance (viability ratio, leverage), tuition dependency, and, for public institutions, short and long-term changes in appropriations.
We found that as many as 560 schools (~25%) were at serious risk for potential closure in the next few years, even before COVID-19 introduced new uncertainty into the equation.
These schools had declining enrollment for multiple years, faced intense geographic competition, and fell below industry-recommended metrics in at least four of the seven financial ratios referenced above. This cohort will likely require transformative strategies to remain operational; but even schools with emerging risk factors will also need to take proactive steps to mitigate certain trends before they worsen.
In this whitepaper, we will describe some key insights that have emerged from our analysis, share publicly-available resources that we have created, and document key strategies that college and universities can pursue.
Why America’s Colleges Are in Trouble
Though Kresge’s response was unique, Marygrove’s story is not. Our research indicates that 55% of four-year public and not-for-profit colleges declined in enrollment over the last five years: over 35% declined at a rate greater than 1% annually. These enrollment declines have happened at a time when institutions, especially private, are increasingly dependent upon student tuition and services as the financial burden of education shifted from the state to the student in the Great Recession: state funding has declined 21% per FTE since 2008 and federal funding has declined 11% per FTE. As of 2019, over 40% of private not-for-profit schools depend on tuition and student-driven revenues for over 75% of their revenues, leaving these colleges very sensitive to even slight shifts in the market.
Now, only 60% of private institutions have sufficient assets to cover their operating expenses in the event of an emergency. The COVID-19 Pandemic has inserted an additional layer of uncertainty, likely expediting potential closures that were on the horizon. In a recent interview that validates this research, Robert Zemsky, co-author of The College Stress Test with Susan Shaman and Susan Campbell Baldridge, estimated that 10 to 20% of institutions face severe risk of closure. By other estimates, 25% of institutions will close in the next 10 to 15 years and 35% of institutions have at least three risk factors for closure.
But COVID-19 did not cause this crisis in higher education; it simply amplified several longstanding and inter-related risk factors.
Structural Demographic Shifts
Dr. Nathan Grawe of Carleton College in Minnesota has identified that by 2025 the declining birthrate from the Great Recession is expected to translate into 15% fewer 18-24-year-olds in the US. These shifts are not geographically equal. Some states, such as Texas, are projected to buck the trend. In others, such as Michigan, the projected decline in 18-24-year-olds is an amplification of a trend that began years earlier.
Curiously, each state’s demographic projections seem inversely proportional to the number of schools competing in those states. Our research suggests that ~90% of public and private not-for-profit institutions enroll a majority of their undergraduates from their respective state or immediate neighbor. In some states, this creates serious conflicts between supply and demand, best exemplified in Michigan. As of 2019 there were 52 four-year public and not-for-profit institutions in Michigan. Excluding the University of Michigan – Ann Arbor, these schools recruit 95% of their undergraduate populations from Michigan, despite the fact that Michigan’s college-aged population has decreased by 5%5 since 2012.
Zero Sum Competition
Competitive differentiation is important for a school’s success. However, when numerous institutions pursue the same tactics – and the same students – without long-term consideration, it can create a zero-sum game. A good example of this is the rise in tuition discounting. As determined by NACUBO, in 2019 for the first time tuition discount rates for first-year students at private colleges exceed 50%. But this does nothing to solve for a shrinking pool of applicants. This “race to bottom” by attempting to recruit from a diminishing pool of students by discounting tuition also has at least two unintended consequences: increased reliance on international students, who pay full-price but are subject to more uncertainties, and increased “amenitization” of campuses, such as building luxury dorms to attract high-income students also paying full-price.
Unsustainable Cost Structures
Underpinning all demographic and geographic trends are the unsustainable cost structures that support these institutions. To remain financially solvent, colleges must achieve a delicate balance between tuition and other student-focused revenues, assets, and debt (public institutions will also have state/federal dollars, though this pool has shrunk in some states over the last 10-20 years). In the last decade this “balance” has destabilized for many institutions. Our research suggests that ~38% of schools have operated at a deficit at least once in the last three years and that 18% have operated at a deficit for three consecutive years. Since peak enrollment of 2010-2011, institutional debt has grown by 57%, while revenues have only grown by 37%, and new assets by only 32%.
Some institutions may be able to control and mitigate some of these risk factors, especially if the risk factors are newly emergent. Other institutions may need to close. The problem is that the decisions that will push these colleges in one direction or another will likely be made in isolation, in panic, or at the whim of people with their own biases. But the outcomes of these decisions affect not only students, faculty, and staff, but neighborhoods, communities and cities. Consider the following three trends that emerged from our research. Of the 560 schools that we identified as having the most significant combination of risk factors:
- 25% anchor some of the most economically distressed and disadvantaged communities in the United States, as documented by Economic Innovation Group.
- 38% are clustered within only six states (NY, OH, PA, IL, MA, and MI).
- 16% of all students of color and 20% of all Pell Grant recipients currently attend these schools. These percentages increase to 30% and 40%, respectively, if you isolate the six states mentioned above.
When one school is on the verge of closing, it requires individual action; when many schools are on the verge of closing within the same city or state it suggests a need for policy.
What if colleges and universities at risk could take corrective action now to put their institutions on a path to solvency and success for the future?
The path to corrective action begins with data and it ends with transformation. University leadership and trustees must take an unbiased and unflattering look at their finances and enrollment projections and then make the hard calls about how best to proceed. Policy makers and foundations must evaluate the collective state of higher education in their respective markets and propose policies and initiatives that address underlying risk factors and/or buy institutions the time to make thoughtful decisions.
To support this process, we have created a publicly available Tableau Dashboard that synthesizes our research. This dashboard aggregates data on the 2,300 institutions that we have studied at the metropolitan (MSA) and state levels. Users can identify how many institutions are at risk in a given geography and quantify the potential impacts that college closures could have on those markets.
We have also outlined nine potential strategies that institutions can apply – individually or together – to ensure their sustainability before they get to the crisis point where they are considering mergers, large layoffs, or liquidation. These strategies pull directly from our experience working with small colleges and from other practitioners and academics who are studying these issues.
Nine Strategies For At-Risk Colleges and Universities
On the top line, institutions can increase enrollment by becoming relevant to more student segments; monetize their real-estate through sales, subleasing, and rentals; engage their alumni to build support; and offer online programs to expand their access and resilience.
On the bottom line, institutions can increase their space efficiency, rethink how the supply of space meets the demand from people, deliver their academic and administrative services more efficiently, and review and prioritize their academic programs.
Or they can find ways to intentionally partner or merge with other institutions with enough time to transition well.
1. Broaden Appeal (Without Losing Focus) and Redesign Support Services
By 2025 the declining birthrate from the Great Recession is expected to translate into 15% fewer 18-24-year-olds in the US. With enrollment falling, few colleges and universities can continue to recruit the same types of students from the same geographic area. So, to maintain or grow enrollment, institutions need to either appeal to the same types of students in other geographic areas or market to other types of students. Competition for “more of the same” is fierce and will get tougher; for instance the National Association for College Admission Counseling recently revised their code of ethics to permit marketing to students enrolled at other institutions. So, many institutions will need to broaden their appeal to recruit new types of students such as adult learners, part-time learners, first-generation students, transfer students, and returning students. For instance there are 36 million people in the US who left college with no degree; and institutions such as Portland State University that accommodate transfer students as a mainstay rather than afterthought are positioned well to help returning students.
Institutions must resist trying to be all things to all people. Institutions should focus on their most relevant, most distinctive, and strongest academic programs and recruit more diverse students for those programs. In broadening their appeal, colleges and universities must also recognize that they will need to redesign their student services to meet the needs of a more diverse student body. This includes making services like advising, tutoring, and financial support more visible, digital, accessible during nights and weekends, and easier to navigate physically.
2. Use Space More Efficiently
Overall the growth of facilities in higher education is outpacing the growth of enrollment, according to data compiled by Sightlines. Among baccalaureate institutions, facilities have grown by an average of 9% versus 1% for enrollment growth; among master’s level institutions these rates are 17% and 4.5%, respectively. Only among research universities have these rates remained relatively aligned (14.5% facilities growth vs. 15.5% enrollment growth). While some campuses use their space efficiently, often there are opportunities to reduce cost without compromising quality.
Two kinds of spaces that are ready for rethinking are classrooms and offices. Classrooms are generally utilized 30 hours per week, nine months of the year, and, by our rough estimate, there are about 208,000,000 net square feet of classrooms among four-year public and private nonprofit institutions.6
Many institutions can reduce classroom space by better matching courses to room sizes and better utilizing existing spaces. Large classes may not need classrooms at all, as they can be taught online – a change that has accelerated in response to COVID-19. Most institutions’ academic plans for fall 2020 assume online instruction for large classes where social distancing isn’t feasible.
Office space is generally occupied 30-50% of a 9-to-5 workday, and, by our rough estimate, there are about 555,600,000 net square feet of office space among four-year public and private nonprofit institutions.7 Office space can be rethought to account for the reduction in paper-based processes and the increased collaboration implicit in flexible work programs. The experience during COVID-19, while challenging, has already revealed just how many faculty and staff can work remotely. In addition to using on-campus space more efficiently, institutions can also relocate functions to lower cost locations, such as storing infrequently used library books in an offsite warehouse or locating administrative functions with limited student and faculty contact (e.g., purchasing or facilities) to locations off campus.
3. Monetize Real Estate
Colleges that experience a misalignment of facilities and enrollment growth may also turn to monetization as a mechanism to confront these trends simultaneously. Institutions typically monetize real estate in one of two ways: the sale or lease of stabilized assets (i.e., housing operations to a third party) or the sale/ground-lease of land for construction. Depending on the structure of the transactions, monetization can provide short-term cash infusions and/or create long-term revenue sources and long-term reductions in operating costs. Real estate asset monetization is a complicated subject that institutions should always approach with careful consideration of the myriad tradeoffs involved. The first step is to consider several crucial factors:
- Understand the Impact on the Institution’s Mission: An institution’s assets should be a reflection of their mission and priorities. When considering a sale, a college/university should reflect on how it will impact all programs associated with the asset (i.e., if the gains realized from the sale would be offset by the need to reinvest in a program elsewhere).
- Know the Market: Each institution’s monetization strategy should be defined by their geography. Schools located in growing cities with active commercial and residential markets will have less trouble finding buyers. Schools in weaker or emerging markets must consider the lost opportunity costs if they sell too quickly.
- Evaluate Short-term Gain vs. Long-term Loss: If an institution is pursuing this strategy to solve a short-term cash problem, has the institution considered all other alternatives? This is particularly important for land-constrained institutions who may be priced out of their market if they try to acquire property in the future. If a college sees no other alternative, then it is imperative to structure operating agreements that exert some control over how an asset is ultimately developed and used (i.e., a school with significant low-income student population might not wish to partner with a developer who is adding only market-rate and above units on a formerly college-owned asset).
4. Engage Alumni to Grow Support
A closure crisis can rally alumni support, such as in the case of Sweet Briar College. However, colleges and universities can identify, analyze, and engage their alumni before a crisis to build their support. First, institutions need to define engagement – the different ways alumni can interact with and support their institutions. The Council for Advancement and Support of Education (CASE) provides useful categories of four types engagement: volunteering, experiential, philanthropic, and communication. Next, they can create an engagement strategy; the University of Waterloo provides a thoughtful example in their case study “The Science behind Alumni Engagement.” Key steps involved in this process include: identify who alumni are; estimate their giving potential (i.e., correlate job title and location with salary range); analyze their engagement across a range of dimensions (e.g., financial contributions, event attendance, social media activity, etc.); and develop a strategy to build relationships accordingly.
As colleges and universities seek to better engage their alumni, two important considerations come to mind: time horizon and types of loyalty. First, colleges and universities need to take the lifetime view that Christine Tempesta, director of information systems and volunteer services at the MIT Alumni Association espouses in this interview. Similarly, institutions need to create a “culture of giving” as Elise Betz, Executive Director of Alumni Relations from the University of Pennsylvania advocates for in this talk. Second, as they engage their alumni, colleges and universities need to think carefully about two kinds of loyalty that they can build: transactional and emotional. Transactional loyalty comes from the sense of what you get (i.e., a discount for alums for a continuing ed course) whereas emotional loyalty is earned through involvement, empowerment and recognition (i.e., your institution is thought of more as a friend or family member). There are a variety of social platforms that can help build both – for instance, Switchboard can help alumni ask for or offer something to other alums as part of a community.
5. Increase Access, Flexibility, and Resilience with Online Programs
Online education is growing at a rate of 6% a year and, according to NCES data, 33% of enrolled students take at least one course online. While there are some programs like Georgia Tech’s $7,000 online masters of computer science that provide a lower cost delivery option with limited real-time student-faculty interaction, the reality is that, done right, online education often costs about the same or more to deliver than on campus instruction, as the WICHE Cooperative for Educational Technologies recently concluded after surveying 200 institutions. Several states also mandate that institutions charge less for online courses. Tuition costs should come down over time with economies of scale and maturity of technology. While fees are often greater for online courses, there are potentially diminished costs of attendance such as housing and transportation.
In the immediate future online learning should be seen not as a way to cut costs but rather as a way to expand accessibility to students from discrete geographic areas, demographics, or both. The geographic pool might not expand as far as you think, as about two-thirds of students taking online courses enroll within 50 miles of home. In addition to potentially reducing costs and broadening access for enrollment, online programs also provide more resilience for institutions by providing the flexibility for students to learn when, where, and how they learn best. While larger public universities have led the way with large investments in the staff and technology for online learning, small colleges like Peirce College in Philadelphia have shown that they too can embrace the flexibility of online learning.
6. Find Creative Ways to Rethink Supply and Demand
Most colleges or universities allocate space assuming students are on-campus two semesters (or three quarters) annually and that each person needs their share of classrooms, labs, offices, study space, athletics and recreation space, event and dining space, and residential space. Taken together, these add up to an average of about 315 net square feet per (NSF) student at private not-for-profit institutions, 240 NSF per student at public four-year institutions, and 65 NSF per student at public two-year institutions, according the the Society of College and University Planner’s (SCUP) Campus Facilities Inventory.
Institutions can combine some of the strategies mentioned earlier and add others to rethink their “supply” of space and their “demand” of people. In terms of demand, they could decrease demand for space by introducing co-ops so that students are on-campus for reduced periods of time. They could increase online learning, which still needs certain spaces like faculty and staff offices and studios but generally doesn’t require spaces like classrooms, dining, or dorms. They can also move entire cohorts online; for instance, in the fall of 2020 hundreds of institutions like Stanford, MIT, Emory, and Bowdoin will only have some of their students on-campus at a time with the rest learning online. Between 16% and 33% of institutions are committed to this “hybrid” solution8 for the fall of 2020, according to the College Crisis Initiative or the Chronicle of Higher Education, respectively. In terms of the “supply” of space, institutions have a variety of options for what to do with the excess facilities that result from these “demand” strategies. For instance, K-12 schools, senior living communities, community groups, coworking spaces would benefit from co-location with a university. Unused space could be decommissioned to save operating costs, rented as an additional revenue stream, or sold as mentioned earlier. Lastly, institutions can “grow in place” by increasing enrollment without adding space.
7. Deliver Student Services More Efficiently
Public and private non-profit colleges and universities in the US spend approximately $38 billion annually on student services like admissions, financial aid, course registration, academic and career advising, libraries and technology, and health and wellness, according to NCES data (see Tables 334.10 and 334.30). The cost of student services has increased 22% in the last 10 years, and the number of professional staff has increased about 40% since 1990.9 This is generally because institutions have been adding offices and centers to meet expanding and increasingly specialized student needs: for instance, an Office of International Students or a Center for First-Generation Students. But, adding functions like this often results in student services that are harder to navigate, less holistic, and more costly.
Now is the time for institutions to rethink what services they offer and how/where they deliver those services. Institutions can consolidate functions into student service hubs that are more effective for students and more efficient to staff – these one-stop-shops have been shown to increase graduation rates from 3% (in this Rand Study) to 6% (in this Single Stop study). These hubs should not only be physical, but digital as well by consolidating services into redesigned student portals. Many of the transactional aspects of student services such as paying a bill should happen online to make them more efficient for students and staff alike, while also freeing up staff time to do more impactful, consultative work.
8. Review and Prioritize Programs
The financial model at many colleges and universities is not sustainable. In the last 30 years, tuition at public, 4-year colleges has increased 3.1x at public institutions and 2.3x at private, non-profit institutions, and student debt is $1.5 trillion nationally. Institutions must review and prioritize their programs relative to their positioning and their performance, including degree, athletics, and student life programs. For example, Paul Quinn College famously made the tough decision to end their football program, Keystone College discontinued their geology and visual arts programs and Oglethorpe University has relentlessly tried to match the affordability of public flagship universities. Institutions should focus on those programs that are the most unique, relevant for the communities they serve, and impactful in terms of enrollment and reputation. In addition, institutions need to design rigorous and repeatable methodologies for program review.
The decision to stop, start, or continue an academic program must consider more than just costs and must accept that some programs will be profitable (e.g., biomedical engineering) and subsidize others that cannot be profitable but are still mission-critical (e.g., philosophy). To review the revenues and costs of their programs, colleges and universities can use the activity-based costing model that Granof, Platt, and Vaysman developed.10 To apply the ABC model institutions should determine the unit of analysis (e.g., a degree program), allocate costs and revenues are directly associated with a program, and equitably allocate indirect costs and revenues that are shared across programs (i.e., by the number of students per program). Institutions can then understand which programs are running at a profit or loss and make informed decisions to continue or stop them – as well as understand pricing and enrollment needs for potential new programs.
9. Explore Strategic Partnerships or Mergers
Mergers and consolidations may be among the most far-reaching tactics on which there is the least agreement. The Chronicle of Higher Education best encapsulated this divide by releasing complimentary articles within the span of four months: A Merger Won’t Save Your College and A Merger Can Save Your College, But Don’t Wait Too Long. In each piece, the authors analyze data from over 150 recent closures, ~57 of which merged or consolidated. One key finding was the difference in experience for public vs private institutions. Approximately 33 consolidations involved a public institution that was subsumed into its larger system. There is evidence to suggest that the consolidation of public institutions can achieve a balance between cost savings and academic enhancement. Lauren Russell, an associate professor in the Fels Institute of Government at the University of Pennsylvania, documented consolidations within the University System of Georgia and found that retention and graduation rates improved for the consolidated schools, likely due to administrative savings that translated into enhanced advising and academic support services.
Jeffery Selingo, a New York Times bestselling author on trends in higher education, expressed similar insights in his piece on The Networked University.11 While his research focuses more on strategic partnerships, he arrives at the same conclusions regarding cost sharing and efficiency. There is, however, an important distinction in his research: each of the principal case studies that he uses to illustrate efficiencies are partnerships between healthier and more stable institutions that are pro-actively working to tackle one or two discrete problems (i.e., affordability), rather than to sustain overall operations.
What little data is available paints a rougher picture for struggling financial institutions who may seek partnerships as a panacea to salvage their operations. The majority of recent mergers between private institutions involved small, niche, and specialized schools (arts, faith-based, etc.) that merged into larger institutions. The co-authors of Strategic Mergers in Higher Education suggest that this may be because university presidents equate mergers with failure or an inability to manage their respective college, and thus do not evaluate these options until it is too late.
Marlboro College may represent one of the few counterpoints. While their consolidation plan with Emerson College has received strong pushback from certain stakeholders, it resulted from an honest look at years of decline. The terms they negotiated are made with an eye toward preserving key values for which the college stood: students and faculty will continue their experience at Emerson; the land will soon belong to Democracy Builders, who will use the campus to launch a bold plan to become one of the country’s first colleges explicitly designed to serve low income and first-generation students.
Institutions that are financially healthy, those with stable enrollment, or those with government support have the flexibility to be more proactive with the partnerships they seek. More desperate institutions may find that willing partners are either unavailable or present terms that make it very clear who is the “buyer” and the “seller.” This is one reason why we feel that the participation of foundations and policymakers could be so valuable in helping to define and forge new partnerships for private institutions. In a way these entities could mimic those elements that work in the public system: acting as an independent third-party, able to ask tough questions and draw unbiased conclusions, with the overarching goal to preserve what makes schools special even if the colleges cannot exist in the way that they once did.
How to Get Started Ensuring Solvency and Success
So, how can you get started if your college or university is at risk because of its enrollment, tuition dependence, tuition discounting, limited endowment, or all of the above? We recommend a four step process that you can implement, learn from, and repeat:
- Look internally to audit the areas noted above such as enrollment, operations, endowment to understand the current state and recent trends.
- Look externally at how you compare to peer and aspirant institutions to get a sense of your positioning as well as where your space, services, and staffing might be out of whack.
- Synthesize and prioritize these internal and external views to identify the problems and plot them relative to effort and impact; for instance low effort high impact items might be some initial quick wins. This is where hard choices will have to be made.
- Implement a mix of initiatives from different areas, thinking about them as a kind of balanced portfolio of initiatives. This will help you manage the change process so that things can be sequenced and learned from instead of too much happening at once.
To learn more about this research and how your institution can rethink your offerings, operations, and organization to optimize your position please contact Elliot Felix (email@example.com) and Shea O’Neill (firstname.lastname@example.org).
Endnote 1: Chronicle of Education: College Closures, 2014-2018; Education Drive: A Look at Trends in College Consolidation Since 2016
Endnote 2: Kresge Foundation: New Cradle-to-Career Educational Partnership to Serve More than 1,000 Detroit Children on Marygrove Campus
Endnote 3: From this point forward, all references to “our research” incorporate data that we pulled from Integrated Postsecondary Education Data System (IPEDS) from the National Center for Education Services, unless otherwise stipulated.
Endnote 4: Our analysis focuses only on Title IV, degree-granting four-year public and private not-for-profit institutions. It does not focus on for profit schools. From this point on we will refer to this cohort as “colleges/universities,” “schools,” or “institutions,” unless otherwise noted.
Endnote 5: U.S. Census Bureau. American Community Survey, 1 Year Survey Data, 2012 to 2018
Endnote 6: This calculation is based on SCUP Campus Facilities Index of 43.1, 50.5, and 10.6 NASF Office/Student x NCES enrollment data of 7.1M, 3.5M, 6.4M public 4yr, private 4yr, and 2yr colleges, respectively
Endnote 7: This calculation is based on SCUP Campus Facilities Index of 43.1, 50.5, and 10.6 NASF Office/Student x NCES enrollment data of 7.1M, 3.5M, 6.4M public 4yr, private 4yr, and 2yr colleges, respectively
Endnote 8: It should be noted that there is some ambiguity as to what hybrid means; some institutions are classifying themselves as hybrid since their instructional model is a mix of online and in-classroom classes with all students back on campus where as other institutions take hybrid to mean only some students are on campus (most commonly international students, low-income students, and entering and/or graduating students) and the rest are online.
Endnote 11: The Networked University