Simulating the Academy: Toward Understanding Colleges and Universities as Dynamic Systems

Citation: What Higher Education is Doing Right, W.F. Massy and J.W. Meyerson, eds., Princeton University 107–120 1997 120

Keywords: Higher Education, Dynamic Systems

Areas of Research: The Scientific Enterprise

See the game that resulted from this research – Virtual-U

Colleges and universities are complex and arcane enterprises. They create and archive fundamental and pragmatic knowledge. They educate our young in preparation for adult life and society’s various endeavors. They interpret and critique culture and influence our world views. We expect these institutions to be all things to all people: generator of inventions for industry, spur for regional economic development, surrogate home for the young, guarantor of good jobs and high incomes, professional developer of those in mid-career, entertainer on Saturday afternoons, equalizer of social opportunity, and political refuge. As important as colleges and universities are to us, however, they are not well understood at a systems level even by those who live and work inside them.

This lack of understanding mattered less when the academy was held in high esteem and resources flowed to it at rates sufficient to maintain internal stability. But times have changed. The gleam of the ivory tower has dulled. A growing number of critics now believe that while educational services are central to America’s successful future, existing colleges and universities are failing to adequately manage their affairs, adapt to changing student needs, and exploit technological possibilities. Internal strife, from heightened competition for scarce resources among a heterogeneous mix of campus constituencies, makes governance increasingly difficult. Tools that can provide leaders both inside and outside the academy with a greater shared insight and und erstanding of our institutions of higher learning as dynamic systems are needed. This paper proposes one such tool.

The Need

Overemphasizing higher education’s importance in America is hard. It is a huge and influential enterprise. Roughly half of all young people enter a higher education institution. About 15 million students currently enroll. Faculty number about 900,000. In 1995, spending totaled close to $180 billion.

As a result of mounting difficulties in raising revenues, rising expectations for the role of universities in social and economic development, technologies that extend the ways in which education can be delivered, and shifts in student demographics and graduate labor markets, many academic leaders are seeking to move beyond incremental change and embark on more fundamental restructuring. Debate has in fact spilled beyond the borders of academe into the contents of best-selling books, lawsuits, and other areas on such issues as:

• the fundamental roles of the university (mission, vision)
• the relative emphasis on research, quality of teaching
• the need, quality, and character of basic research
• the appropriateness of applied research on campus
• faculty responsibility and behavior
• (excessive) management and (lack of) leadership
• costs, especially tuition and overhead rates
• the rationale for and length of time to the doctoral degree
• employment terms for the academic workforce
• curriculum content and knowledge structure
• elitism, social stratification, and diversity
• academic standards (e.g., admission policies, attrition rates, grade inflation, sports)
• the rise of foreign student population and links to foreign firms

The desirable response from higher education seems clear enough. From community colleges to research universities, they should raise productivity, modernize administrative and support services, and improve accountability while preserving autonomy. Upon identifying priorities, they should recast incentives and allocate resources accordingly. But how to go about achieving this response is less clear. A major obstacle is that key stakeholders appear not to recognize or accept facts about how colleges and universities work. They do not view the institution as a system or internalize the linkages between cause and effect. The resulting gaps in knowledge and credibility form major barriers to experimentation and reform. For example, seemingly logical proposals to close marginal departments and redeploy their faculty are vigorously fought. Conversely, incentive programs for early retirement are readily accepted even though they may generate unintended and undesired consequences (e.g., those who leave may be among the institution’s more productive faculty since they are more likely to have compelling career alternatives).

A large share of the problem owes to the fact that universities are complex, both to understand and to manage. Considering their functions and interactions with government, industry, and society in general, we can hardly be surprised. Universities are systems with many independent parts and interactive processes. Outcomes frequently depend on powerful but obscure second-order effects. One example is how the expansion of graduate education created a market for low-cost provision of labor for research and teaching, which now strongly influences the size of graduate programs and admissions, including the admissions of foreign students.

Some of the complexity in managing the system stems from lack of agreement or clarity among the various stakeholders about purposes, measures of performance, and productivity. Furthermore, the professional workforce and relatively flat organizational structure limit the exercise of direct management control, leaving institutional leaders to reconcile conflicting objectives as each stakeholder presses his or her own agenda. Stakeholders often attend only to their own values and needs, not stopping to see their institution in broad perspective. The decision-making process becomes volatile when emotionally charged issues such as tenure, academic freedom, and diversity are perceived to be at stake. Choices ultimately made may not be congruent with the institution’s long-run interest.

Most academic leaders come to their jobs lacking deep experience about economic and management matters. Intelligence and motivation can offset their inexperience to some extent, but organizational complexity limits the offset. The difficulty of achieving a comprehensive view also applies to trustees, faculty, staff, students, and others involved in university decision-making. Trustees are typically grounded in a business or professional field but often lack recent first-hand exposure to the higher education environment. Faculty rarely view their institution holistically, and the same is true for students, alumni, and other stakeholders.

Higher education needs innovative devices that help institutional leaders focus their thoughts, and then communicate with stakeholders without appearing manipulative or quickly raising defense mechanisms. Traditional devices such as conferences, commissions, and editorials help, but people rarely internalize complex scenarios by passively receiving information. A program to understand the college/university as a (complex) system, synthesized in a leadership strategy simulation game, can provide people at several levels with an opportunity to deepen their understanding of how colleges and universities work, motivating and engaging them without imposing the difficulties and risks that come with real life.

People concerned with higher education need to understand the decision-making process of the major actors–administration, faculty, students, and other internal and external stakeholders–and how these processes interact. Modeling the behavior of various subsystems within the university, their interactions, and the influence of external forces upon them can contribute to such an understanding. The parts of the system to be analyzed and modeled in detail will depend on what are considered the most important issues. Since controllability of the entire system is of paramount concern, priority would attach to building a model of the behavior of the entire complex system–yielding an intriguing, though necessarily rough, view of the whole.

A Leadership Strategy Game: “SimU”

The development of a college/university simulator, or “SimU,” would draw on three streams of activity: (a) management education games now widely used in management education (e.g., MIT and Carnegie Mellon distribute corporate management games popular in business schools); (b) special-purpose simulations developed to meet educational objectives in enterprises of various kinds (e.g., military battlefield simulations, nuclear power plant operations simulations); and (c) more purposely entertaining simulation games developed for a broader commercial market. Games such as SimEarth and SimCity (both developed by Maxis Software, the latter with sales of 2.5 million units) have proved to enlighten as well as amuse. Although formal evaluation of their effectiveness is hard to obtain, their enthusiastic adoption in school and university courses suggests their educational value. More explicitly serious simulations in this genre such as SimHealth (developed by Thinking Tools, Inc., to explore the reorganization of the US health care system implied by the Clinton health care reform proposals) have also had reasonable commercial success (sales of tens of thousands of units).

The games use continuous computation and constantly changing color graphics as well as sound to sustain user interest. Individual players “play against the model.” SimCity, for example, confronts a single player with zoning, infrastructure, transport, security, and fiscal issues played out over a sweep of time sufficient for long-term effects to become apparent. Efforts at developing multiuser online simulation games (e.g., the Internet-based “President ‘96” and “Reinventing America,” developed by Crossover Technologies under grants from the Markle Foundation to simulate the political and policymaking processes) are also now attracting considerable interest.

The authors believe that existing research and data are sufficient to build a simulation, both educational and entertaining, that will allow users to grapple with issues such as:

1. strategic positioning of the institution
2. academic performance and faculty morale
3. administrative and support service performance and staff morale
4. incentives and rewards
5. goals and perceptions of students, parents, donors, research sponsors, community members, employers, and government
6. comparative performance with respect to similar institutions
7. tuition, financial aid, and overhead rates
8. financial performance, including capital assets and liabilities (e.g., endowment, physical assets, and deferred liabilities)

The target market for such a simulation would be, broadly, anyone with an interest in how colleges and universities work as systems and, more specifically:

• higher education administrators
• faculty, especially those in leadership roles (e.g., department chairs)
• trustees
• education analysts, writers, and policymakers
• students of higher education, and in general
• alumni and interested public

How SimU Might Work

One of the biggest challenges in building the simulation will be to develop a successful user interface. It should be highly graphical and easy to understand and use. It should draw users to a depth sufficient for meaningful learning while at the same time maintaining interest, pace, and playability. Users report playing SimCity many times to experience its wide variety of different scenarios, exogenous events, and patterns of consequences. SimU should elicit a similar degree of interest.

SimU might open by inviting the user (you) to choose among institutional types and control. Do you want to lead a private research university, a public comprehensive, a private liberal arts college? In what year would you like to begin play? What would you like to name it? You might choose to “grow your own” generic institution or load one of a handful of pre-scripted scenarios that present a specific institution in the throes of a specific dilemma based on actual case studies. The scenarios would define “victory goals” that you must achieve to “win.” (In regular game mode, you will be free to define what success means yourself as you hone your own goals over time. ) Versions of the game might be developed that would allow tailoring to match more closely your specific real-life institution (e.g., through the loading of custom data sets).

Play might open with a panoramic view of the campus: a map with icons representing various organizational units and functions that then segues to a close-up of “Old Main,” your administrative headquarters. Double-clicking on it reveals the interior of an office (yours) complete with desk, file drawers, computer (for e-mail and information display), perhaps a door to a conference room in which meetings could be conducted and a window overlooking the campus that reveals significant changes in various aspects of the campus environment (e.g., dilapidated buildings if maintenance is deferred for too long, fewer students milling around if enrollment declines substantially, hostile faculty if you have not recently appeased them). You, by the way, are the president/senior administrator of this institution and have been blessed (burdened?) with an uncannily high degree of omnipotence.

Clickable icons to the side of the screen could represent schools and departments, offices for managing various functions (e.g., admissions, fund raising, buildings and grounds department), athletic facilities, dormitories, etc. Clicking on an icon would provide information about and/or encounters with the people or activity–analogous to searching out reports and management by walking around. Figure 1 provides a sampling of the kinds of activities and reports that might be included in the simulation.

Figure 1. Sample Activities, Decisions, and Reports Operating Units Academic departments Student services and student life Admissions and financial aid Institutional advancement Alumni relations and public affairs Libraries Information technology support Intercollegiate athletics Finance and administration Plant operations and maintenance Dormitories and food service Financial Decisions & Reports Operating and capital budgets Tuition rate and financial aid policy Research overhead rate Sources and uses of operating funds Faculty and staff salaries Faculty early retirement buyouts Endowment asset allocation and investment return Debt issuance and retirement Balance sheet operating surplus/deficit Other Actions & Reports Admissions selectivity and yield Enrollment by degree level and major Attainment rates & times to degree Course offerings Teaching method mix Course availability Class size distributions Teaching loads Sponsored research volume Publications record Faculty awards, prizes, etc. Popular prestige ratings Academic prestige ratings Faculty age distributions Faculty hiring & retention Staff additions and layoffs Staff turnover rates

At this point you might proceed in one of two ways. You might provide a set of presidential goals: in effect, a “platform” that calls out the priority you attach to various stakeholders and outcomes. Or, you might begin by playing with the goals programmed into your chosen scenario. These goals would influence certain aspects of simulated behavior. Moreover, they would provide the institutional performance benchmarks needed to define what it means to “win” the game. Alternatively, you might choose to simply explore the simulated world. Rather than trying to “win,” you would be occupied with observing the intuitive and sometimes surprisingly counterintuitive consequences of various inputs you and others (driven by the underlying game engine) make over time.

Semester by semester, time passes as you observe (and seek to modify) outcomes like faculty gains and losses, shifts in applicant pool and graduation rates, growth or decline in externally sponsored research, crumbling buildings and infrastructure, and accreting or eroding financial health. Conditions permitting, you might raise or borrow money and construct new facilities. You may at any time review data in your office or by walking around, call meetings, or change certain policies. (Nothing will happen at the interface while you are engaged in one of these activities, but computations will continue in the background.)

You might visit a particular department and perhaps try to influence faculty behavior: e.g., numbers and types of courses, teaching loads, submission of proposals for sponsored research, involvement with students outside of class. Such efforts might or might not be successful, depending on the institution’s incentive-reward environment (which would stem in part from your own prior decisions) and other circumstances. Even if successful, they might exact a price in faculty morale–however, the opportunity to exert influence would allow you to seek changes at the academic working level that might otherwise appear out of reach.

Three kinds of pre-programmed events punctuate the passage of time:

• Scheduled events marking milestones or providing periodic information for which no response is required: e.g., quarterly and fiscal-year financial reports, key athletic outcomes, admission of the next freshman class, commencement. Simulated time continues.
• Scheduled events for which a response is required: e.g., submission of the annual tuition recommendation, the operating budget, and the capital budget; and the Board’s annual presidential performance evaluation and your acceptance or disagreement with it. Simulated time halts while you prepare the budget or react to the performance evaluation.
• Unscheduled events arising exogenously or because of some condition within the simulation: e.g., a stock-market crash, a dean or professor pleads a case or airs a grievance, a Faculty Senate action or student protest, a fire or a safety problem. Simulated time may continue or halt depending on the event.

By combining these events with the user-initiated ones, the SimU program should be able to provide a simulation that is sufficiently rich to realistically represent the essentials of university leadership and capture user interest. Most of the databases needed to specify the model already exist, and a growing number of research findings are available. Indeed, pulling together the information needed to build the model will be a valuable exercise in its own right. A companion handbook and strategy guidecould provide background, help focus play, and draw out lessons contained in the simulation.

SimU Actors

We have already described how you (as institutional leader) might interact with the SimU simulation. But in a university, the administration’s word is not exactly law. Much of what happens in SimU would result from the actions of various constituencies–simulated actors and stakeholders that operate inside or outside the university. Your actions would influence constituency behavior, but not control it.

A list of potential constituencies follows. Some are depicted as individuals while others represent aggregations of individuals. Some would appear in the simulations for all institutional types, others would apply to one or two types only. At this point we do not know how practical it will be to include all the following constituencies in the game’s initial version.

Internal constituencies

• The governing board might provide financial oversight and offer evaluations of presidential performance .
• School deans might be simulated as independent agents who have independent objectives and sets policy.
• Faculty in each department might be simulated as a set of cohorts with age-rank characteristics and probabilities for promotion, departure, and retirement.
• Department chairs might decide about course offerings, teaching loads, and research emphasis–in effect, representing the aggregate view of departmental faculty.
• The faculty senate might represent the view of the faculty taken as a whole.
• Students might be simulated in terms of admission cohorts, degree levels, and majors, each with course-taking, graduation rate, satisfaction, and similar characteristics.
• The student senate might represent aggregate student views.
• Non-academic operating units–e.g., support services, administration, operations and maintenance–might be described by production functions relating the quality and quality of outputs to budget allocations (see the list in Figure 1 for a more complete list).
• The non-faculty workforce (staff) might be portrayed as a small number of groups–e.g., professional and administrative, clerical, operations and maintenance–whose numbers would grow or ebb according to budget allocations; staff morale and efficiency might depend on workload in relation to numbers, and on compensation level.
• Prospective students might be simulated in terms of application and matriculation rates by market segment; “market research” data might be used to convey attitudes and predict behavior.
• Research sponsors might be simulated on a discipline-by-discipline basis, with each discipline characterized by the level of total funding and the intensity of competition.
• Alumni and potential donors might be simulated according to their interest in one or another department or in the whole institution; gift-giving might depend on department/institution performance and prestige.
• The media might be simulated as a single constituency, with media actions being illustrated with newspaper clippings or television stories.
• Public opinion also might be simulated as a single constituency; public opinion might drive regulatory decisions, and influence state funding decisions in game sessions dealing with public institutions.

Exogenous factors also might affect SimU’s fortunes. Economy-wide inflation and family income growth might drive up cost and mediate the effect of tuition increases on admissions yield and public opinion. Demographics might affect student demand. Governmental funding decisions might drive sponsored research, and for public institutions, state appropriations. A natural disaster might disrupt campus operations. Technological change might restructure cost functions and engender new competition that challenges market shares in education or research. Presidential actions throughout the simulation would determine how well the institution weathers the storms and captures and opportunities.

Issues to be Addressed

SimU would address at least four kinds of issues. The interaction of player decisions with data and response functions built into the model would determine how a college or university evolves and whether the president’s goals are achieved. Gaining insights about these issues and learning to analyze them in systems terms would constitute one of SimU’s most important benefits. The issues include:

1. Capital investment vs. spending for current operations (spending vs. saving): policies governing financial capital (endowment, reserves), physical capital (facilities, equipment), and spending for operations. Most institutions bias decisions toward spending for current purposes, especially salaries. The simulation would address the consequences of such imbalances.
2. Operating budget allocations: decisions to spend more on one field or activity than another; determination of cross-subsidies between fields and activities. Spending on certain fields may be seen as more or less consistent with the school’s mission, and fields will vary in their ability to generate enrollments and sponsored research dollars. Spending on academic support services may improve educational and research quality and competitiveness, institutional-support investments (e.g. G&A and O&M) may improve infrastructure and efficiency, institutional-advancement investments may increase giving levels, and so on.
3. Transactions with customers and stakeholders: student applications, admissions, and yields; sponsored research finding; gift acquisition; and, for public institutions, state appropriations. Outcomes may be affected by quality and prestige, net prices (e.g., tuition minus average financial aid, the effective research overhead rate), and “marketing” expenditures (e.g., for admissions and institutional advancement), as well as uncontrollable factors.
4. Academic department actions, which produce the institution’s instruction and research outputs. The range of simulated action might include: the profile of courses as represented by teaching method mix (lectures, seminars, labs), course level, and degree of specialization; faculty teaching loads; and the degree of emphasis placed on research. Considerable attention would be placed on departmental actions because such actions constitute the central focus of academic production.

Success Criteria and Performance Measures

Both game designers and players will have to address basic questions dealing with the university’s or college’s fundamental mission. Should the mission stress the preservation and transmission of knowledge (teaching) or the generation of knowledge (research)? Should the mission cater to the few or virtually everybody (the elite or the mainstream)? Should the institution focus its mission or should it try to serve a broad set of constituencies?

The SimU simulation would be rich enough to permit a large number of performance measures to be reported, but it would not dictate what players should pay attention to. Indeed, much of the data available as a byproduct of the behavioral simulations would not be displayed unless the player searches it out by clicking on the appropriate icons. While certain success criteria would be defined–either by the player or as part of the chosen scenario–the player will retain a great deal of latitude.

Much of SimU’s value will come from discussions about values, performance measures, and the functions programmed into the game. These discussions would be stimulated but not brought to closure by the software and supporting data. The players themselves would have to supply the missing pieces, but the game would supply two crucial elements.

1. First, the game would provide a specific set of stimuli for discussion–a context within which to explore one’s own values and understandings and, depending on the circumstances, to compare them to those of one’s colleagues.
2. Second, the game would enforce the disciplines of conservation and causality. Money allocated to one priority is not available for another. Actions and failures to act have consequences that must be considered when trying to satisfy one or another constituency. All constituencies cannot be satisfied to the full extent of their desires, especially when exogenous forces infringe on the institution’s market power or freedom of action.

These are important lessons in their own right, and their application in the context of discussions about values would add yet another important benefit. Without consideration of conservation and causality, discussions about values become unbounded, and the university is urged once again to be all things to all people.

Conclusion

The motivation for using SimU is to understand better how a university works. What performance measures should be considered? How do decisions made by the administration, the faculty, and other agents affect the performance measures? Why can’t the university simultaneously maximize the agendas of all its stakeholders? Some participants will challenge the theories used in the simulation, but the very act of challenging requires the formulation of an alternative hypothesis–which can be analyzed and compared with assumptions and data used in the model. SimU also should be fun to play, since learning depends on engagement and engagement will be stimulated and sustained if the activity is intrinsically interesting.

Faculty, staff, students, and trustees must develop more coherent and realistic perspectives about their institutions. Working with a simulation game can build experience and broaden perspectives. Gaming can help all stakeholders understand issues at the level of the institution–and from viewpoints of other stakeholders–and see the issues through less parochial eyes. Even experienced managers find that playing a sophisticated game expands their horizons and motivates broader discussion of management issues.

Ultimately, we would hope that development of SimU would bring three benefits:

1. New knowledge: advances in fundamental understanding of how a university functions will come from facing for the first time the challenge of modeling the whole of a university.
2. Education of a broad group of stakeholders: given a reasonably sound simulator, a rather large number of stakeholders, numbering in the tens of thousands, may enhance significantly their understanding of the university as a system by “playing the game.”
3. Development of new management tools for universities: while SimU would be generic, it could prove the concept of university simulators and stimulate the subsequent development of more detailed, realistic simulators appropriate for specific institutions or classes of institutions.

The authors understand the difficulty of considering the university as a complex system. But because complexity lies at the heart of the university’s current problems, we feel it is important to address the issue head-on. Even the limited models that are practical using today’s knowledge can begin to capture the dynamics and the interactions of the parts. At a minimum, they can help organize the data that will be needed to simulate a university in finer grain, and they will lead to better definition of parameters, variables, and outcomes. But the real payoff–achievable, we believe, with today’s technology–will be to move higher education’s many constituencies toward more shared understanding of how the academy works.

Notes

¹ This paper was motivated in part by a session on simulation an gaming, led by John Hiles of Thinking Tools, Inc., at last fall’s Stanford Forum for Higher Education Futures (Annapolis, November 1995).