Enterprise IT, E-Learning, and Transformation: Prospects in Higher Education
Which way to Millinocket? "Well, you can go west to the next intersection," a Mainer explains,
". . . get onto the turnpike, go north through the toll gate at Augusta, 'til you come to that intersection . . . well, no. You keep right on this tar road; it changes to dirt now and again. Just keep the river on your left. You'll come to a crossroads and . . . let me see. Then again, you can take that scenic coastal route that the tourists use. And after you get to Bucksport . . . well, let me see now. Millinocket. Come to think of it, you can't get there from here!"
Skeptics have argued that transforming higher education, especially to decrease its cost, is like getting to Millinocket: We know where we want to be, but we can't get there from here.
Might information technology help find the way? Maybe. The Apple II was supposed to transform higher education. So were the PC and the Mac, and then the "3M" workstation (megapixel, megabyte, megaflop) for which MIT's Project Athena, Carnegie-Mellon's Andrew Project, and Brown's Scholar's Workstation were designed. Simulated laboratories, BITNET, and then the Internet, MUDs, Internet2, Wi-Fi, artificial intelligence, supercomputing — all these innovations were at some point seen as auguries of change. Each helped higher education grow, evolve, and gain efficiency and flexibility. Even simple technologies — e-mail, for example, and texting — changed higher education in ways hard to imagine before personal computers and the Internet came on the scene.
What about social media, streaming video, multiuser virtual environments, mobile devices, massive online open courses (MOOCs), and the cloud? Will they be evolutionary, or transformational? If higher education needs the latter, can we get there from here? To explore these questions, I will address the following topics:
- Structural challenges in higher education
- IT roles in higher education
- Driving and enabling change
This may help us figure out how to get to Millinocket.
Structural Challenges in Higher Education
We first need to consider three things about higher education:
- How students distribute across institutions
- How the economics of higher education constrain efficiency
- How society's expectations that higher education will be at once comprehensive, diverse, and cost-effective complicate efficiency
Scale and Distribution
According to the Integrated Postsecondary Education Data System (IPEDS), there are about 4,200 undergraduate degree-granting colleges and universities in the United States, enrolling about 18 million students. Yet the largest 20 percent of the institutions (840 of the 4,200) account for 70 percent of undergraduate enrollment, very nearly a Pareto (20/80) inequality (figure 1).
Figure 1. Enrollment vs. number of higher education institutions
That students and institutions distribute so unequally is one reason higher education adapts so slowly to change. Policies favoring most institutions may not favor most students, and vice versa. Politics complicate accordingly.
James Surowiecki recently wrote that higher education costs illustrate "Baumol's cost disease":
Some sectors of the economy, like manufacturing, have rising productivity . . . which leads to higher wages and rising living standards. But other sectors, like education, have a harder time increasing productivity . . . . The average student-teacher ratio in college is sixteen to one, just about what it was thirty years ago. In other words, teachers today aren't any more productive than they were in 1980. . . . Colleges can't pay 1980 salaries, and the only way they can pay 2011 salaries is by raising prices.
Demography and Baumol's disease may help explain why so much of higher education is under stress. Adding to this stress are constrained funding sources, including state appropriations, federal and state financial aid, research funding, endowment yields, and family income. If a recent Bain study is correct, perhaps one-third of all colleges and universities are headed for insolvency.
Society expects more of higher education each year. The introductory electrical engineering course that I took in 1966 barely mentioned solid-state electronics, for example, whereas the modern equivalent course covers complex integrated-circuit design. Today a well-educated American student is expected to have some familiarity not just with Western literature but also with Eastern, Islamic, and other literatures.
Content is only part of growing expectations. Graduates need greater reasoning and critical-thinking skills — for example, the ability to make judgments about news, data, arguments, and analysis, now that so much of it arrives unfiltered by editors, commentators, librarians, or other moderators. And a vastly more diverse student population — itself a major achievement of higher education, with important and positive consequences for our society — is by its very nature more difficult and expensive to educate, since learning styles and requirements are also diverse.
IT Roles in Higher Education
Based on data from IPEDS and the EDUCAUSE Core Data Service (CDS), colleges and universities devote about 5 percent of their overall expenditures to information technology, perhaps $21 billion in all, with some variation across institutional types. Using staffing as a metric, over half of this goes to "Enterprise IT," which includes core infrastructure, administrative systems, communications and other "commons" systems, and general support (see figure 2).
Figure 2. IT staffing as a metric
Perhaps the most striking attributes of IT progress have been its pervasiveness, convergence, integration, and migration. Today, not only are servers, networks, and end-user devices tightly interconnected and interdependent, but their components are also increasingly indistinguishable, as well as provided and operated by organizations other than one's own. Network switches are essentially servers, servers often comprise internal networks plus vast arrays of the same processors that drive end-user devices, and end-user devices readily tackle tasks — voice recognition, for example — that once required massive servers.
The locus of technology has shifted dramatically from the institution — be it home, workplace, school, or campus — to the mobile individual. The locus of control and responsibility is shifting accordingly: connectivity, content, services, even identification come from providers external to one's immediate location or affiliation — from the "cloud" — in a sharp departure from past practice.
Through what applications might information technology help higher education evolve? It is important to distinguish two different but overlapping roles that information technology might play. In the evolutionary category are four overlapping educational functions. Information technology can
- streamline administration,
- amplify and extend traditional pedagogies, mechanisms, and resources,
- make educational events and materials available outside the original context, and/or
- enable experience-based learning.
In the transformational category are two more-radical functions. Information technology can
- renew and redefine the social environment and/or
- replace the didactic classroom experience.
Paper-based registration, scheduling, and student-record systems managed separately have given way to highly integrated online student systems based on relational databases, enabling registrar offices to do more work with fewer staff. Decentralized, inconsistent distribution and access mechanisms for instructional and research materials — everything from reserve collections in the library to "course packs" at the local copy center to stacks of syllabi outside faculty offices — have given way to sophisticated library and learning management systems. And of course e-mail, the web, file sharing, and other messaging and information systems have fundamentally changed how students, faculty, and staff interact with one another and with intellectual property. Together, all of these elements of enterprise IT have transformed how colleges and universities operate organizationally.
In part because enterprise IT has made data mining and other sophisticated methodologies feasible, administrative analytics have become prominent. Purdue, for example, after using analytic and data-mining tools to analyze data on student progress, learned that reduced use of the learning management system (LMS) by students was an early indicator of academic trouble ahead, and it focused academic support resources accordingly.
Bringing analytics and big data to bear in this way not only can improve administration but also can empower students. Institutions can provide students with sophisticated analytic tools they can use to appraise their own choices and progress relative to that expected of them or relative to their peers' selections and experiences.
Amplify and Extend Traditional Pedagogies, Mechanisms, and Resources
Even though chairs outside faculty doors have given way to Blackboard, Sakai, or Moodle LMSs, and even though wearing PJs to early-morning lectures has given way to watching lectures from a dorm room over breakfast, campus-based higher education remains recognizable. In true distance education, this changes: students never (or rarely) set foot in a classroom. In 2008, 3.7 percent of students took all their coursework through distance education, and 20.4 percent took at least one class that way.
Extending traditional pedagogy across the network leaves unsolved how to extend traditional pedagogy across linguistic, cultural, stylistic, and time-of-day boundaries. Similarly, although network-based technologies enable materials to be available simultaneously in diverse arrangements and formats, thereby enabling adaptation to diverse learning styles and speeds, achieving this diversity can increase rather than reduce "teaching" effort. Also, rendering material electronically requires technical specifications, protocols, and associated software, which so far have proven frustratingly transient, not just for images and video but even for text.
Make Educational Events and Materials Available Outside the Original Context
As materials become electronic, barriers to broader distribution shrink. As barriers shrink, it becomes possible for materials to find new uses beyond their original intent. For example, MIT's OpenCourseWare (OCW) started as a publicly accessible repository of lecture notes, problem sets, and other content from MIT classes. It has since grown to include similar materials from scores of other institutions worldwide and most recently helped spawn the edX initiative, which provides not only materials but also mechanisms for demonstrating and certifying mastery of the content. Although some of these materials are used in traditional classes, albeit in institutions other than those creating them, much is simply used by individuals who want to learn independently.
Similarly, the newer Khan Academy has collected a broad array of instructional videos on diverse topics, some from classes and some prepared especially for Khan, and has made them available for anyone interested in learning the material. More recently, MOOCs have proliferated. Some of these courses simply involve widespread distribution of existing video, but some use materials designed explicitly for the medium.
It is one thing to distribute materials beyond traditional boundaries; it is quite another to build a business model around such materials or to decide which other components of higher education — guidance, curriculum, assessment, certification — are required for the extension to be worthwhile.
Enable Experience-Based Learning
As computers have become more powerful, simulations — which include many computer-based games — have become more complex and realistic. As the latter have moved to cloud-based servers, multiuser virtual environments have emerged. These go beyond simulation to replicate complex environments — indeed, sometimes they are complex environments. As social media like instant messaging, Facebook, LinkedIn, Second Life, and Twitter have enabled geographically scattered communities to grow and flourish, student experience-based learning has jumped beyond campus boundaries.
Communal experiences like these were impossible before the advent of powerful, inexpensive server clouds, ubiquitous networking, and graphically capable end-user devices. That they are possible now potentially transforms the notion of "campus" in important ways. At the very least, these applications shift traditional notions of authority. For example, for students, crowdsourcing among widely distributed peers already vies for respect with faculty expertise based on traditional credentials or with documented scholarship recorded in refereed books and journals and collected in libraries.
In simulation-based environments, educational material is in effect existential and irreproducible, something traditional records management cannot handle. Similarly, as information technology enables the personalization of educational experience, every student's education becomes different. How that experience is measured and recorded becomes a new problem accompanying a new opportunity. (In a similar case from the past, Hampshire College's pioneering, highly individualized curriculum and assessment mechanisms in the 1970s yielded 30-page transcripts comprising extensive descriptions of a student's performance in every class — transcripts so unmanageable and difficult to evaluate, I remember from my days serving on a Harvard graduate-school admissions committee, that they came to be almost counterproductive.)
Renew and Redefine the Social Environment
One channel for socialization in higher education is academic: spending time in the scholarly community and learning how to evaluate information and argue positions under its formal and informal rules. A second channel is cultural: being exposed to diversity in all its dimensions, be it involvement with different kinds of people, different rules of informal rhetoric and evidence, different social expectations, different kinds of food, or different perspectives on life.
Most students already live off campus — in 2007–08, only 14 percent of undergraduates lived in college-owned housing — and so are less intensively involved in extracurricular activities than residential students. Less emphasis on campuses can mean less socialization. Reversing this through technology requires more than Twitter and Facebook; it requires focused online mechanisms so that students who are scattered across the nation or the world can commune educationally with faculty and fellow students.
Technology thus should make it possible to expand rather than shrink the socialization effect of higher education. Today's students are as comfortable communicating and exchanging views electronically, especially through social media such as Facebook and Twitter, as they are interacting in person — perhaps even more so, in some cases. However, in many cases students handle their social media environment orthogonally to their academic environment, resisting efforts to combine the two.
Replace the Didactic Classroom Experience
This most controversial application of learning technology — using or sharing online instruction rather than hiring campus-based faculty — drives most discussion of how technology might transform higher education. This is especially true for disciplines and topics in which instructors convey what they know to students through classroom lectures, readings, and tutorials.
PLATO, now a commercial education provider, emerged from the University of Illinois in the 1960s as the first major example of computers replacing teachers. It has been followed by myriad attempts, some more successful than others, to create technology-based teaching mechanisms that automatically tailor instruction to how quickly students master material.
More typically, the classroom experience is not so much replaced as reengineered, by bringing information technology to bear on peri-classroom activities such as student support (e.g., transfer credits, counseling). The University of Texas's Course Transformation Program (CTP), for example, "is designed to improve student success in large, lower division gateway courses by . . . the identification, development, and adoption of appropriate evidence-based approaches to teaching and learning," especially in large, "gateway" courses such as introductory calculus. Carnegie Mellon's Open Learning Initiative (OLI) goes a step further, in that its courses are designed explicitly to replace classroom experience, but like CTP, it remains firmly grounded in a traditional institution.
Replacing the traditional face-to-face experience with technology, even in the CTP or OLI sense, is controversial. Some opposition stems from legitimate concerns that technology cannot replace certain kinds of interaction and that it is inappropriate for certain subject matter. But some opposition stems from a more basic antipathy to replacing labor with capital, thereby eliminating jobs.
Driving and Enabling Change
Information technology can help higher education evolve and transform itself to better tackle the pedagogical, sociological, economic, and financial challenges of the future. Yet obstacles loom. These require colleges and universities to contemplate difficult, fundamental changes in how they organize and educate.
Although the need for such changes has been apparent for some time, progress has been uneven. This may be because some of the necessary changes are, as Ken King put it in a note to me, "dangerous to discuss":
Libraries still add miles of books every year that are never used [and] institutions still teach introductory courses in large classrooms where there is very little faculty-student engagement. Students would learn as much by watching a video lecture, or better still, a video lecture from a faculty star at some other University . . . . Soon, I believe, massive amounts of capital will be invested in producing automated teaching factories.
That image of "automated teaching factories" triggers much instinctive resistance to transformational technologies. Its implicit threat to "miles of books every year" does the same in many libraries.
If information technologies are to transform higher education, we must exploit opportunities and address problems. Our goal should be not only to make these important entities more efficient than they are today, which is where evolutionary technologies are likely to play a central role, but also to make them better, which is the principal goal for transformational technologies. Achieving both greater efficiency and better outcomes through information technology requires a commitment to fundamental, unfettered thinking about the future both within and outside current institutions.
If information technology is to achieve its transformational potential in higher education, we must think differently about policy and practice in the following areas (which by no means constitute a complete list):
- Guidance and pedagogy
- Paths and swirl
- Borders and treaties
- Inputs and outputs
- Division of responsibility
- Privacy and analytics
- Complexity and permanence
- Who's who
Guidance and Pedagogy
As information technology makes experiential learning more feasible across a broad array of disciplines historically dependent on professors instructing students, the 19th-century German university model may lose ground to very different models based on direct, albeit often simulated, experience. This, in turn, entails faculty who advise and guide inquiry rather than (or, more likely, in addition to) provide direct instruction. Guidance calls for skills different from those needed for direct instruction, and it takes more time.
Students, in turn, must learn to tell the difference between valid and invalid experiences, reliable and unreliable sources, and accurate and inaccurate simulation. When computers first became prominent across higher education, many institutions implemented "computer literacy" classes and requirements, sometimes as part of the formal curriculum. Some of these evolved into classes on "information literacy" — well suited to students undertaking experiential learning (as was "bibliographic instruction" in libraries). For the most part, however, these classes disappeared rather than evolved, leaving students to fend for themselves. Preparing faculty and students for effective experiential learning remains a major challenge.
Paths and Swirl
My undergraduate career was at one institution, with credit for two or three summer courses taken elsewhere. My sister, two years younger, attended five colleges and earned credit from other institutions as well. Decades ago I was the norm, and she was the exception.
Today my sister would be closer to the norm: very few students stick with one institution from admission through degree. This raises important questions about who maintains and certifies a student's academic record, who has access to which parts of the record, and similar issues.
Although institutional registrars are the current custodians of student records, it may be time for a change. This is already happening for K–12 education, in the form of state longitudinal data systems (SLDSs). The steady increase in student swirl may make this change imperative and inevitable for higher education as well. If that is the case, then it would be good to begin work on the necessary transmission, validation, security, and privacy protocols.
Borders and Treaties
Many institutions think they can offset declining resources by enrolling distant students — especially, in the case of state-funded institutions, students in other states. Because of this, state authorization procedures — especially those that may apply to distant institutions — have become controversial. Some states have very simple authorization processes; some states have onerous authorization processes. The jurisdictional issues are unclear.
Uncertainty about jurisdiction works against progress in distance education. The likely solution involves reciprocity, whereby states with similar requirements recognize each other's authorizations. However, since this issue has implications beyond distance learning — for example, must an institution comply reciprocally with the data-breach requirements for every state from which its students come, or does compliance with its own requirements suffice? — it may take some time to resolve.
Inputs and Outputs
Higher education generally measures student attainment in two ways: through assessment of what students produce (e.g., graded assignments, achievement tests, class participation) and through the time students spend in class — even though strictly speaking, the latter measures input rather than output.
Although for the most part institutions choose their own assessment mechanisms, there is some standardization, such as the examinations required for entry into graduate schools or professions. On the time-in-class side, in contrast, standardization is the norm: few institutions depart from the "student hour" definitions promulgated originally by Harvard's president and then by the Carnegie Foundation.
It is one thing to track hours spent sitting in regular class meetings, or attending labs and sections, or even watching live video of lectures. It is quite another to track how much time students spend learning from online material, especially pre-recorded or self-paced material. We could possibly track how students click through pages or perhaps even use a camera to record their "attendance" and attention, but none of this would be reliable. Far more likely, online education will drive attainment measures away from attendance and back toward direct assessment of competence or mastery.
It is useful to remember that what drove the development of the time-based Carnegie unit in the first place was inconsistency in direct assessment mechanisms across different colleges and universities. We need to be careful not to circle back. Collaboration on standard assessment mechanisms should work — but this too has been controversial, as the history of the College Entrance Examination Board, the American College Test, and the Educational Testing Service make clear.
Division of Responsibility
We can imagine a future where students readily satisfy one institution's degree requirements with coursework taken elsewhere. Several institutions were early pioneers in this direction, organizing around competency-based degrees based on self-paced learning and life experience. These include Empire State College and, more recently, Western Governors University.
Further down this road, we can imagine a transformed future in which some institutions admit students, prescribe curriculum, certify progress, and grant degrees — but have no instructional faculty and do not offer courses. This, in turn, might spawn purely instructional institutions and even purely instructional institutions that operate entirely online — institutions that would have every incentive to use technology in novel, effective, efficient, non-Baumol ways.
Boundaries and finances immediately become an issue. Who, for example, decides whether a course taken at A is good enough to be credited toward a degree from B — especially if B has no faculty in the course's discipline? Does the degree-granting institution pay the instructing institution, or vice versa, or neither? If hundreds of institutions use the same online course taught by a faculty member at C, how does C recover its costs — and does the faculty member share in C's revenue?
Privacy and Analytics
I studied hard for one of my first undergraduate quizzes. It was immensely difficult, and I didn't finish all the problems. When I got the quiz back, I had scored something like 12 out of 100. I was briefly devastated, but then I saw on the board that the class average had been about 15, with a range from near zero to a high of about 30. So I was below average but, I was relieved to learn, not an outlier. The simple feedback I had received — the overall score distribution, the "curve" — was immensely helpful to my understanding of where I stood. With similar data from my other classes, it helped me allocate my study time. Information empowered me.
Today, student information systems enable institutions to provide much more sophisticated data to students, thereby empowering them even more. For example, if I were a student taking that quiz today, I might be able to find out not only how I'd scored compared to others in the class but also, perhaps, to others with similar concurrent enrollments or to others with similar entering test scores or to others in my fraternity or dormitory.
If the student information system was managed poorly, I might even be able to figure out exactly how well my friends did. Down the path beyond access to data lies invasion of privacy. Empowering students through access to data is a good thing; invading privacy through access to data is not. Analytic systems must be designed to strike the right balance.
Partly because of unease about privacy, institutions are reluctant to release student data — even anonymized — to outside entities. This makes it difficult for outside entities to assess student progress or instructional efficiency across different institutions and thereby to recommend or make choices regarding which institutions should receive support and how much. In K–12, as noted above, SLDSs have begun to collect student data from schools across states, largely to deal with swirl, and the same is beginning to happen in higher education. As SLDSs expand into higher education, they will encounter tension between analytic desiderata and privacy protections.
Without thoughtful, modern policies and practices to balance analytic and privacy issues, it will be difficult to empower students through sophisticated analytics. It will also be difficult to empower state and other coordinating and funding entities whose decisions would be improved by analysis of student data.
Complexity and Permanence
Rendering material electronically requires technical specifications, protocols, and associated software applications. So far these have proven frustratingly transient, not just for images and video but even for text. For example, I wrote my 1977 Harvard dissertation entirely on the university's mainframe computer, using a variant of the RUNOFF text-processing program developed in the mid-1960s at MIT. But Harvard refused to accept the computer file as my dissertation or, for that matter, the smudge-prone output from the computer line printer. After considerable negotiation, the university accepted copies of the computer printout Xeroxed onto archival paper. Today the paper fused-toner copies survive and are eminently readable (as is the microform copy at ProQuest); my computer version became unreadable long ago.
As pedagogical materials evolve into born-digital form, the risk of impermanence grows. The issue is especially pronounced for non-text materials and for materials protected by some kind of digital rights management (DRM) or otherwise encrypted. In pre-digital times, the standard way to guarantee permanence was by depositing copies in one or more libraries or archives. But this worked well only for standardized media, namely printing, writing, or drawing on paper. Unlike material on paper, electronic materials require maintenance and updating, which are often beyond the technical and financial capacity of existing repositories. Effective, efficient, robust digital archives remain an unattainable goal.
Especially as distance learning replaces classroom instruction — even within the traditional campus-based program — identity management becomes important. Two issues dominate: assessment and financial aid. It is clearly important that the individual who is assessed at a distance be the same individual whose achievement is to be documented and certified. Likewise, it is clearly important that fraud by individuals who create fictional identities, use them to obtain grants and loans, and then abscond with the funds be discouraged if not prevented. To address these and other identity issues, we need reasonable ways to verify the identity of remote students and then to secure the online mechanisms that can be used to authenticate them thereafter.
A national ID system of some sort might work, but it would be very controversial. Perhaps colleges and universities could piggyback on the mechanisms that banks use to identify distant customers. For example, I have an account at ING but have never set foot in an ING facility; ING trusted Chase's verification of who I was. Or perhaps campuses could act as identification sites for other, distant campuses.
Getting to Millinocket
These days, drivers headed for Millinocket rarely pull over to ask directions from farmers. Instead, they rely on the geographic location and information systems built into their cars, phones, or computers, which in turn rely on network connectivity to keep maps and traffic reports up-to-date. Tourists find their way to Millinocket, but their reliance on GPS and GIS insulates them from the diversity of the local residents they pass along the road, much as interstate highways have standardized cross-country travel. The gain from geographic information technology is not without cost.
The same is true for administrative and learning technologies in higher education, whether focused on organizational efficiency or intellectual outcomes. As we choose new practices and policies, we will both gain and lose.
Effective progress can result only if we explore and understand the technologies and their applications, decide how these relate to structure and goals, identify obstacles and remedies, choose our tradeoffs, and figure out how to get there from here.
This essay draws on an earlier paper prepared for a meeting sponsored jointly by EDUCAUSE and the Center for American Progress.