The End of College: Q&A with Kevin Carey

Kevin Carey is an American higher education writer and policy analyst. He is serving as Director of the Education Policy Program at the New America Foundation, a non-profit, non-partisan research organization based in Washington, D.C. We asked Kevin about his recent, provocative book, The End of College: Creating the Future of Learning and the University of Everywhere.


Before we get to the substance of the book, I’d like to ask about its reception. If you’re like most writers, you are surprised by some of the interpretations of your work. What responses did you not anticipate?

Most of the book takes place in the past and the present, telling the story of how American colleges and universities became what they are and how information technology might allow them to become something new. In the last two chapters, I present some ideas about what the future might bring. As I see it, some students will learn mostly online, particularly as human-computer interaction and digital learning environments improve. But many will continue to attend in-person institutions that provide direct relationships with peers, teachers, and mentors. Those institutions won’t, however, have to be colleges as we know them today. In other words, the future of higher education won’t be solitary, with everyone but an elite few learning alone in front of machines. Some of the criticism of the book has left out that last part, which is disappointing. Then again, it’s been out for less than a month, so maybe people haven’t gotten all the way to the end yet.

The future of higher education you present in “The End of College” places more responsibility on individual learners to design and follow-through on their own studies. Self-direction has long been an instructional ideal – we want students to take responsibility for their own learning. But there’s a fear that they simply won’t, at least not to the extent required in the scenario you’ve laid out in “The End of College”. 

People vary a great deal in their educational skills and motivational constructs. Some of them need a lot more support, or different kinds of support, than others. Existing colleges are often impersonal and cookie-cutter in how they treat students, and I think the future will provide more opportunities for customization, personalization, and variation in organizational philosophy. Traditional colleges are less diverse than they like to believe — they share a great deal of the same organizational DNA.

I also think there’s an important distinction between “design” and “follow-through.” Pretty much everyone needs guidance on the design of their studies. Effective pure auto-didacticism is rare. In many ways, that’s the very definition of education: creating an environment and set of relationships that deliberately guide people in the acquisition of skills, knowledge, and habits of mind. On the follow-through side, I think simply being better will accomplish a great deal. Good education challenges people to work hard and move out of their comfort zones, but doesn’t push them past the threshold of total frustration. That’s crucial and not easy to accomplish, which is exactly why people like the cognitive tutor designers at Carnegie Mellon are working to solve the problem.

So there’s no reason to believe the future of higher education will reduce the amount of necessary support that students receive. Quite the opposite.

It’s been suggested that your vision for higher education is best suited to well-supported, above average students – leading to a sharper class divide. Poorer students get MOOCs, and the 1% get 4 years at Yale. How do we ensure that this isn’t the case? 

Only the 1% get 4 years at Yale now. That’s the way it has always been, and always will be, as long going to Yale means living in a faux-Gothic building in New Haven, Connecticut. So the question is, can we radically expand the number of people of people who have access to the educational riches of Yale, and institutions like it? The answer is clearly “yes.” You can already access the lectures, courses, syllabi, assignments, and exams at Yale, for free. That problem has been solved. The question, then, is what additional educational resources does Yale offer, and can we provide them to many students at an affordable price? The existence of organizations like the Minerva Project, which is just as selective as Yale and also provides a rich liberal arts curriculum, for about one-third the tuition, suggest that we can. There can be many more organizations like that.

A common characteristic of discussions about change in higher education is the tendency for lines to be drawn between commentators and initiatives from within higher education (e.g. faculty) and those outside (e.g. think-tanks, entrepreneurs). Several of the people and initiatives you highlighted in the book are from outside higher ed proper (even some of the MOOC providers felt compelled to step outside of their institutions to bring something to the masses). And you, as an analyst at a think-tank, might also be considered an outsider. Is this divide significant? What might be its’ origins? 

It’s significant, but it can be overstated. Anant Agarwal left a great, prestigious job as the director of the MIT Computer Science and Artificial Intelligence Laboratory to head up edX. This involved moving from a building on the MIT campus to a building located one block away from the MIT campus. I interviewed Robert Lue, faculty director of HarvardX, in his office in the Harvard natural sciences building. He hasn’t gone anywhere. He said to me, “HarvardX is Harvard. It is us.” The West Coast MOOC providers have definitely stepped outside the academy into the world of for-profit, venture-backed companies, but that’s an incredibly porous border. The companies are located in Palo Alto and some of the founders still teach at Stanford.

That said, it’s not surprising that people who are currently employed by colleges are only going to go so far in agreeing with the thesis of a book titled “The End of College.”

In a review of your book, Don Heller wrote the following:
“In his claim that MOOCs and other online learning materials will replace colleges and universities, Carey also provides a very narrow view of the goals of higher education. A bachelor’s degree is more than just a collection of individual courses; college — when done right — also satisfies other developmental objectives, including extracurricular learning, developing interpersonal communication skills (of both the online and face-to-face variety), and instilling a sense of an individual’s role in a democratic society.”

This is a common criticism of online higher education. But it’s also odd given that online higher education didn’t set out to supplant all aspects of the traditional college experience – its’ focus is instruction. On the one hand, I recognize that meeting your spouse or having inebriated conversations with other students at the campus pub is important – it was for me. But isn’t also the case that these extracurricular dimensions of higher education are primarily accidents of history – rather than by design. Is it the case that our institutions set out to provide instruction, but these other dimensions then began to form around the core mission? 

Per my response above — I have a lot of respect for Don as a scholar of education but I think his critique ignores some of what I say at the end of the book. I quote from some of the relevant parts of the book here in explaining why. In short: there’s no reason why we can’t have the best of both worlds, online and in-person, for a lot less money than traditional colleges and universities charge today.

If alternative credentials grow in stature and students gain access to more information about the actual learning gains that a particular provider can offer, are we setting in motion a highly competitive education market (i.e. informed consumers with choice)? Are our institutions ready for this? Will they allow it? 

Yes, No, and It won’t be up to them. We need a competitive market in providing students a great education. We really don’t have one now. Educational institutions are complicated and hard to understand and most consumers are naive. There’s a dangerous conflation of price and quality. Existing institutions are theoretically in a great position to take advantage of this opportunity, because they employ a lot of smart, knowledgeable educators and have brand names that people trust. But many are deeply committed to not changing, are they are likely to struggle. They’ll fight opening up the market to new credentials, for obvious reasons of self-interest, but that’s not a fight that can be won forever. If you’re battling against better information, your days are numbered.


Keith Hampson is Managing Director, Acrobatiq, a Carnegie Mellon University venture.

University of the People as Disruptive Innovation

Founded in 2009, University of the People (“UoPeople”) holds the distinction of being the first tuition-free, non-profit online university. To date, UoPeople has admitted more than 2000 students from over 150 countries, including the Sudan, Indonesia, and Haiti. Students enrol in either Business or Computer Science, for Associates or Bachelor’s programs. Most of the academic labour is supplied by volunteers, for which there is apparently no shortage. Open content is used and delivered using open source technology. Students rely heavily on peer-to-peer learning.

Screenshot 2015-01-15 16.43.50

In higher ed circles, discussions about of UoPeople typically frame the initiative in philanthropic terms: it’s grand and generous, but quite removed from higher education proper.

Certainly, generosity came to mind when I spoke with Shai Reshef (pictured above); the Founder and sponsor is modest and plain-spoken, and clearly has a sincere desire to help students.

But UoPeople can also be understood as a highly resourceful example of business model innovation; it bears a striking resemblance to the textbook definition of disruptive innovation (before the term was stretched beyond recognition):

  • it combines technology with a new business model to offer a stripped-down, feature-poor version of an existing service (higher education);
  • it seeks to serve “non-consumers”, rather than offer an alternative that attract students that have other, convenient and affordable options for getting a college education.

It’s true that the UoPeople relies on volunteers, many of whom are only able to share their expertise freely because they have jobs with universities. (The Provost is from Columbia, a Dean from NYU.) But Reshef confirms that the operation is sustainable in its current form and, in time, the anticipated surplus of funds could be used to hire a certain percentage of staff at market rates. It’s also important to remember how relatively large the “free economy” is in the education sector. (For more on the “free economy” concept, see Chris Anderson’s original riff here).

More than a philanthropic effort, the UoPeople may also be understood as an example of the different forms that higher education can take to meet the needs of increasingly global higher education system – particularly in the developing world.

Learn more about the University of the People.

Keith Hampson, PhD is Managing Director, Client Innovation at Acrobatiq LLC (A Carnegie Mellon University venture)

Sharing What Works in Online Higher Education (2 of 2)

This post is Part 2. Part 1 can be found here

The Growing Chasm in the Online Higher Education Market (2 of 2)

One of the key characteristics that distinguishes faster growing, more scalable, and increasingly high-quality online universities (described in “The Growing Chasm”) is the systematic use of knowledge about what works in online instruction and what doesn’t. This handful of US institutions tend to capture more data about student learning, learn from it, and act on it.chairAndnewspapers

As simple as this process sounds, it’s difficult to implement in our traditional colleges and universities, where course design and development is typically a very decentralized activity. Instruction is determined on a course-by-course basis and there’s rarely a systematic, robust process in place for identifying and sharing knowledge about what drives student-learning outcomes most effectively. I recall, as a new member of faculty, being surprised to learn that my colleagues had little to no knowledge of how their other colleagues in the department ran their courses, for example.

This need not be the case, though. The required technology now exists and if used creatively in conjunction with basic change management practices, it is possible to increase the volume and quality of information sharing, even within the most decentralized institutions.

Initiatives to drive instructional innovation are most likely to be led by central support units responsible supporting online courses and programs. It’s in these departments that the pressure to ensure quality and bring about change is felt greatest.

The core elements include:

  1. Get buy-in
  2. Define the current state of affairs
  3. Make new approaches clear and easy to adopt
  4. Distribute tools to measure student learning

A Few Details . . .

Pitch it. Put together a clear and compelling description of your plan. Use it to get feedback and solicit buy-in. (It won’t hurt if you get buy-in from people with influence, but passion goes a long way, too.)

Catalogue it. Take an inventory of the instructional practices currently being used. (Maintain full confidentiality of Instructors). Share this inventory, once organized for simple review, with all stakeholders. There will be some surprises. And many educators will benefit from this tactic, alone – given the current dearth of information. “What are others doing?”

Package it. Ask a team consisting of representatives to select 8-10 interesting instructional strategies that they believe would be of value to others within the institution. Showcase these examples, using events, a dedicated website, external conferences, and other means available. Be an insufferable promoter. Reconfigure these 8-10 instructional strategies so that they can be easily understood and copied. Vagueness, here, is your enemy. Be clear, simple and above all, concrete. As Dan and Chip Heath suggest: If you want people to eat a more healthy mix of foods, don’t tell them to “eat low fat foods”, tell them to “buy skim milk.”

Assess it. Implement learning analytics that each instructor can use to measure student learning. If the goal is to have students learn more and more quickly, then the analytics must actually measure student learning, not merely track their behaviour; I call this “engagement analytics”. The type information collected in engagement analytics often includes:

  • Number of page views (per page)
  • Contributions by students to discussion threads
  • Which students (and what percentage of the total cohort) have completed the assignments
  • Number of logins

Engagement analytics do not necessarily measure learning, per se. What’s measured is student activity, which may or may not signal actual learning. For example, engagement analytics is often used to track student page views. The student’s presence on that particular page within the course site tells us that the student has been exposed to that part of the curriculum. But it doesn’t tell us whether the student understands the curriculum. In fact, it may be that the student inadvertently left the browser window open while searching the Internet.

Learning analytics, on the other hand, measure the student’s actual learning state; what students know, what they don’t know, and why. It’s this kind of information that’s needed if individual educators are going to imagine new and better ways to stimulate learning. See figure A. Information that can be captured by learning analytics include:
  • What aspects of the course did the student master?
  • Which students are struggling, and with which concepts, topics and problems?
  • What misconceptions about the curriculum are leading to poor performance?
  • What topics require more attention or better presentation?

Let us know which of these tactics you’ve tried and which has served you well.

The Growing Chasm in the Online Higher Education Market (1 of 2)

A chasm is beginning to appear between institutions of higher education that offer online programs. The divide is the result of the different strategies taken for designing, sourcing and managing online education programs.

A small number of institutions in the U.S. have adopted methods for producing and supporting online courses that have the potential, if not the likelihood, to improve learning outcomes, increase the speed with the institution improves the quality of teaching and learning, increase value (quality/cost). If present trends continue, these institutions could reconfigure the deeply embedded hierarchy that organizes higher education.

A couple of scenarios

An acquaintance of mine, currently an Assistant Professor at a mid-size university, was asked in mid-July by her institution to create and deliver a new online course for the Fall (September) semester. In the time available, she had to define the new curriculum, determine the instructional tactics to be used, collect existing resources, and create new materials, including assessments.

Throughout the process, she worked alone. Although an instructional designer was on-hand, the staff member had little time and offered not much more than a checklist of best practices. The Instructor’s budget for the course development? Nil.

Her experience contrasts sharply with practices at a handful US universities. These institutions typically focus, sometimes exclusively, on online education, offer open-admissions, and have centralized management of teaching and learning. Consider this depiction; a composite of a few institutions I’ve had a chance to investigate:

An academic department – after conducting a thorough, regularly scheduled review of learning outcomes – determines that a full rework of a key program is required. Starting what will be a twelve-month process, the department conducts a deeper analysis of the current program, consulting with student support staff, faculty, academic leadership, and industry advisors – to define the overarching set of objectives and instructional strategies for the revamped program.

A team is assigned to the project, including specialists in learning analytics, subject matter experts, managers of assessment systems, faculty, teaching assistants, student support staff, and technology managers.

The institution’s team identifies a number of things they want to offer their students that can be done more efficiently by forming partnerships other universities, consortia, and vendors, so as to complement internal strengths. The course development process ultimately involves more than a dozen people, three external organizations, and costs more than 100k per course, when including internal labour costs. Following the first year of the new programs’ delivery, a review is conducted to identify where refinements are needed.

Not an inconsequential impact

There are a number of issues of note:

All things being equal, this handful of institutions will offer students higher quality education. By bringing the right mix of talent, resources, funds, and processes together, the institution has a much better chance of providing students with a well-conceived, thoughtfully-executed, and well-resourced learning experience.

The institutions have considerably greater ability to scale-up learning to meet demand. They can build new courses and programs more quickly, and with greater assurance that each will meet institutional standards for quality.

These institutions pay considerably more attention to the results of their instructional strategies. Internal reviews are common, and many are now turning to analytics to generate even more detailed and extensive insights into what’s working and what isn’t. This knowledge provides the basis for better decision-making, which in-turn can provide progressively better learning experiences for students.

This last quality needs to be underlined.

Knowledge about how to design and support learning in higher education held by individual faculty – whether online or not – is rarely systematically shared with the institution. Teaching is approached as individual pursuit. Indeed, faculty members can work in the same department as other academics for several years without ever seeing each other teach. Each Instructor operates individually. Strictly speaking, this isn’t by design: it’s a by-product of the traditional organizational structure of the institution and conventions of the academic occupation. But the effect of this characteristic is that it limits the flow of knowledge across the institution about effective teaching.  It fits nicely the centuries-old conventions of the occupation, it may ultimately limit the breadth and depth of the knowledge that is brought to bear on each course within the institution.

These upstart universities see knowledge about teaching and learning as the domain of the institution. The institution, not the individual educator, captures, interprets and applies knowledge about how best to serve students.  Knowledge is applied on an institutional level, not on a course-by-course, instructor-by-instructor level.

Of course, the downside of this approach is the potential to suppress the kinds of innovations that can arise from radical decentralization – letting a “hundred flowers bloom”, if you will.

But supporters of this more centralized approach contend that the benefits of a collective, institutional approach to knowledge building and sharing may be greater at this point in the evolution of online education. Higher quality learning, they argue, requires a more deliberate and disciplined approach. At times, I can appreciate this perspective: conference presentations about “how to teach online” offered in 2014 have striking resemblance to those we heard in 2001. We don’t seem to be making significant headway by placing the burden of course design and delivery primarily on the backs of under-resourced individual Instructors.

In the next post, I’ll consider the implications of this chasm on the credibility and status of online higher education institutions. Please feel free to contact me directly > Email Me


Keith Hampson, PhD is Managing Director, Client Innovation at Acrobatiq LLC (A Carnegie Mellon University venture)

Stuck in the 90s: Online Course Design in Traditional Higher Education

When I entered the field of online higher education back in the Dark Ages of 1997, the first 56.6 kbps modems were just starting to appear in homes. Online enrolment was a fraction of what it is today. The push to grow online learning was largely faculty-led; few university Strategic Plans included more than a passing reference to online education.

Today, of course, we’re accustomed to reading breathless editorials in mass-market publications like the New York Times about the disruptive power of online learning. University Presidents proclaim that online education is a “game changer”. Clay Shirky explains that technology is bringing an end to business-as-usual for higher ed, just as it has for newspapers and the music industry.

Yet, in the midst of all this talk of revolution, one aspect of online higher education has remained virtually unchanged since the 1990s: the way that the majority of traditional colleges and universities go about designing, creating, and financing in-house online course development.

Now, as in 1997, individual instructors assume the bulk of the responsibility for course design and development. Support is now often available from an instructional designer and technical staff, but their impact is limited by workplace conventions that encourage faculty to work alone (and staff to let them). Funds for course development are similarly constrained, due to the conventional notion/practice that course materials should be built for use only within the institution from which they came.

While this model of course development may work for classroom education (where the organizational and financial model originated), it places severe limitations on the kinds of digital learning experiences that committed educators can make available to their online students.

Development of more sophisticated forms of digital learning such as personalized instruction driven by analytics, immersive gaming, or the use of rich media, to name but a few possibilities, require a team of specialists, longer development schedules, and considerably more funding than is available in the current approach. Placing the burden on lone educators with minuscule (or non-existent) funding — who are not hired for their strengths in instructional media development — is both illogical and unfair.  But more to the point, it’s a lost opportunity to leverage high-quality course design to drive improvements in learning outcomes.

As a result, students across North America are frequently presented with online courses consisting of repurposed classroom PowerPoint slides, home-made graphics, and an incoherent pastiche of free content from the Net — each element developed for different purposes and pitched at different levels. Worst of all, these online learning experiences are being developed without deep knowledge of the science of how people actually learn most effectively — knowledge, ironically, that universities themselves have generated.

By simply transferring the existing roles, responsibilities and financial model to the online context allowed institutions to quickly “put courses online”. It made the migration to online education relatively painless for the institution, but adherence to this approach has ultimately limited our capacity to offer our students a more thoughtfully crafted, rigorously developed online course experience.

The current approach to developing online courses took hold not because we thought it was the best way to create a great online learning experience for students, but for the simple fact that it fit with the institution’s existing structure, processes, and culture — all of which are derived from the deeply engrained logic of the classroom model. Nor is the endurance of the classroom model due to the fact that we tried new approaches and found them wanting. There have been great breakthroughs in using new course design strategies by a handful of innovative organizations. Extensive evaluations of Carnegie Mellon University’s “Open Learning Initiative” – from which my organization springs – regularly found that their approach significantly reduced the amount of time it took students to complete courses, while still maintaining or even improving learning outcomes.

Of course, not all aspects of online course design require a team of specialists, a longer development time, and more funding. Some things can be done quickly, cheaply and by individuals with focused skill sets. But those types of instructional materials are not the problem — we’ve got that well in-hand. The challenge now is to find ways for our colleges and universities to build and access those types of instructional experiences that take fuller advantage of the Net’s unique properties to combine data, interaction, media and rigorous instructional design.

We know, for example, that students benefit from having regular opportunities to apply their knowledge as they are learn and then to receive immediate, meaningful feedback on their efforts. But, of course, the current approach makes it all but impossible for lone instructors to produce hundreds of learning activities for each course. And it’s even less feasible for the instructor to provide feedback to each and every student in real-time. Similarly, we know that students come to us with very different cognitive and educational baggage; educators have little opportunity to identify and respond to these differences.

When built with a deep understanding of how students learn, technology can meet both of these needs. We can build online courses that provide students with hundreds of opportunities to test their knowledge. Using scientifically-based learning analytics, we can provide each learner with immediate, context-specific feedback. We can build software that constantly responds to each student’s cognitive and educational differences and serves up activities that address these differences.

But these kinds of learning tools take time, money and talent to create. They can’t be put together a few weeks before classes start.  If we’re going to take online education seriously, we need to take course design seriously. We need to start imagining new models for building, acquiring and sharing instructional media.


Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @Acrobatiq

Consortia as a Driver of Innovation

In the late 90s — during the “early years” of online higher education — many colleges and universities didn’t have the internal resources required to build, support and market online education. Some institutions saw fit to join online consortia; by pooling limited resources, each institution gained access to the resources they needed. Many of these early initiatives are still operating.

Access and/or Innovation

In a recent review we conducted of online higher ed consortia, we found that the majority of consortia, and the vast majority of those that started ten or more years ago, are designed primarily to increase access. That is, these initiatives define success by the number of online courses created and/or supported by the consortia, and the number of students enrolled in these courses. More courses, means more access.

Access is obviously important. However, for a number of reasons, we believe that a recasting of the consortia model would be beneficial. Given the state of online education, the focus needs now to shift from ensuring institutions can launch and support online courses, to stimulating innovation and improving quality.

Time to Focus on Innovation

First, and most obviously, the needs of member institutions have changed and consequently consortia need to change, as well. Over the last decade-and-a-half, most colleges and universities have significantly augmented their internal capacity to develop, support and market online education. The LMS is now near universal. The majority of university leaders see online education as fundamental to institutional strategy, and far more instructors have experience teaching online.

moving beyond the basics

As internal capacity of member institutions increase, the functions that can’t be done well (or at all) within each member institution change too. Although this may seem so obvious as to be not worth mentioning, our review suggests otherwise. Many consortia we reviewed continue to provide only the basic requirements of creating and supporting online courses. One consortium, for example, simply assigned a single instructional designer to work with a lone instructor from the member institution to develop an online course. No meaningful quality standards are employed, the instructor isn’t even paid for the course development. Fewer and fewer institutions need these basic services. It isn’t surprising that our review found that institutions that have set more ambitious goals for online education are less interested in participating in consortia.

focus on innovation

Our review suggested that more consortia should focus less on providing basic, increasingly common, services and more on helping institutions test and scale more ambitious online learning strategies that can improve outcomes and drive down costs. If the fundamental value proposition of consortia is that it enables member institutions to do what they can’t done alone, then the initiative should be deliberately and systematically focusing on those functions that are anything but “basic”. Services that fall into the category of “ambitious” in 2014 might include the development of rich media, the use of learning analytics, and the development of competency-based programs.

why consortia

Consortia align particularly well with three trends in online higher education:

  1. A slow migration to the software model of course development, in which upfront costs for course development are relatively high, but maintenance and distribution costs are marginal. By pooling resources, consortia can accommodate higher upfront costs and then coordinate distribution at scale.
  1. Growing use of analytics to inform and personalize learning. The more data is shared and compared across institutions, the greater its value. Again, consortia are well positioned to facilitate the proper movement of data-generated insights across institutions.
  1. Online education will continue to demand new, increasingly complex skills and knowledge that are not readily available within each institution. Consortia can serve as a central, shared source of talent and technology across individual institutions.
defining ROI

Consortia need to define and then share clearer and more concrete objectives with member institutions. In particular, it would be useful for consortia to provide members with more robust assessments of the initiative’s ROI. If success is defined by the consortium (as noted above) by the number of online courses and students that are supported by the consortia, then members should be able to assess whether the cost of running the consortia is greater than the actual increase in enrolment and number of courses. ROI is always difficult in education, but consortia — given their frequently tenuous financial stability — may be less inclined to produce this kind of information. Member institutions should demand it.

built to change

Lastly, consortia must be built to change. If, as suggested, the basic purpose and value proposition of the consortia is to do what member institutions can’t do separately, then the services offered must change as technology, costs, and objectives change. Again, this may seem obvious. But consortia struggle with change like other organizations. Nevertheless, the value proposition of consortia requires that they continually adjust their services to meet changing conditions.


Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @Acrobatiq

Services offered by consortia to member institutions include the following:

  • Course registration and course registration systems
  • Help desk (technology and/or administrative) for students
  • Professional development for instructors
  • Learning management systems
  • Video conferencing (hardware, software and support)
  • Webinar hosting and management (hardware, software and support)
  • Sharing of online courses between institutions
  • Instructional quality assessment and rubrics
  • Development of new applications
  • Multimedia development (instructional material)
  • Market research services
  • Instructor training on educational technology
  • Instructional design
  • Tutoring services (student online/phone)
  • Learning object repositories
  • Project management/coordination
  • Marketing / clearinghouse of members courses and programs

Taking Course Design Seriously (Finally)

When I first entered the field of online education back in 1997, the first 56.6 kps modems were appearing in homes. Most institutions had no more than a handful of online courses. Universities treated educators with an interest in “distance education” with either suspicion or indifference. Online education was largely a faculty-led effort; few university Strategic Plans included more than a passing reference to online education.

shutterstock_103492922Obviously, much has changed. But one aspect of online higher education remains largely intact: the way that traditional colleges and universities go about designing, creating, and financing in-house online course development.

Now, as in 1997, individual instructors assume the bulk of the responsibility for course design and development. Support from an instructional designer and technical staff is available, but their impact is limited by availability and the conventions of academic work. Funds for course development are similarly constrained, due to the limited revenue that can be generated from offering a single course at a single institution.

This “cottage industry” approach took hold not because we thought it was the best way to create a great online learning experience for students, but because it fit with the institution’s existing organization and processes — one based on the classroom model. Mirroring the classroom model made the shift to online relatively painless.

As a consequence, though, the quality of the online courses produced within our traditional colleges and universities falls to professionals who were not hired on the basis of their knowledge of graphic design, information architecture, programming, or the learning sciences — the very qualities required to consistently create great online courses. As has long been the case for classroom education, the overarching, but implicit assumption is that putting course design in the hands of people with deep subject matter knowledge translates into a good learning experience for students. There’s little evidence that this is the case in the classroom, and it’s less true for online education, where a whole new host of skills and knowledge are required.

By simply transferring the existing roles, responsibilities and financial model to the online context, allowed institutions to quickly “put courses online”. But it also all but ensured that these institutions are unable to produce more sophisticated and ambitious online courses that support better learning outcomes or reduce costs. Ironically, it restricts the use of the very instructional techniques, resources, and new business models that research claims can improve the value of higher education — research produced by our universities.

Examples of more ambitious online course design include:

  • Courses that offer students hundreds of opportunities to test and apply their knowledge and skills. And an equal number of moments of feedback that let them know of their progress;
  • High-production value media, including illustrations, animations, audio and video, that explain difficult concepts and process in clear and powerful ways;
  • Software that adapts to student input, enabling the optimal sequence of learning activities;
  • Learning analytics, with instructional activities, that provides the student, educator, and institution with detailed explanations of a student’s relative strengths and weaknesses;
  • Instructional methodologies tied to new business models that have the capacity — unlike the current model — to drive down costs by leveraging economies of scale.

These tactics and others like them accomplish the obvious: they take advantage of the unique economics of the Internet and leverage the medium’s ground-breaking capacity to combine media, interactivity, social interaction, and data; the very things we dreamt of doing back in the 1990s, but have failed to accomplish.

Simply getting courses online is no longer sufficient, and it can no longer serve as evidence of a strong institutional commitment to online learning — as Penn’s Robert Zemsky argued a full decade ago. Institutions need now to turn their attention to finding ways to build or acquire instructional content that truly takes advantage of technology to improve learning. There’s no better place to start than course design.


Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @Acrobatiq

Consortia typically offer a range of services for member institutions:

  • Course registration and course registration systems
  • Help desk (technology and/or administrative) for students (e.g. BC Campus)
  • Professional development for instructors (e.g. Campus Saskatchewan)
  • Learning management systems (e.g. eCampus Alberta)
  • Video conferencing (hardware, software and support)
  • Webinar hosting and management (hardware, software and support)
  • Sharing of online courses between institutions
  • Instructional quality assessment and rubrics
  • Development of new applications
  • Multimedia development (instructional material)
  • Market research services
  • Instructor training on educational technology
  • Instructional design
  • Tutoring services (student online/phone)
  • Learning object repositories
  • Project management/coordination
  • Marketing / clearinghouse of members courses and programs

Teaching a Child to Ride a Bike via Lecture

Well done. A great (because it’s simple) explanation of active learning techniques.

Hand-Picked Selections> Week of September 08 2014

“Worth Reading” is a hand-picked weekly collection of new, not-so-new articles and downright old ideas, events and other items for higher education professionals.


UMUC: The Future of Learning An animated promotional video from UMUC outlines what its vision of the future of online higher education. The vision includes, notably,

  • competency-based assessment
  • use of mentors as the primary contact for the student
  • direct assessment
  • prior learning assessment
  • student portfolios
  • open educational resources mixed with university licensed materials;
  • extensive predictive analytics


So Bill Gates Has This Idea for a History Class …

New York Times Magazine cover story (no less) on the “Big History Project”.

“As Gates was working his way through the series, he stumbled upon a set of DVDs titled “Big History” — an unusual college course taught by a jovial, gesticulating professor from Australia named David Christian. Unlike the previous DVDs, “Big History” did not confine itself to any particular topic, or even to a single academic discipline. Instead, it put forward a synthesis of history, biology, chemistry, astronomy and other disparate fields, which Christian wove together into nothing less than a unifying narrative of life on earth.”


Not insignificant news. A major university, Purdue, is interesting a competency-based program.

“The national interest in competency-based education, also called direct assessment, comes on the heels of U.S Department of Education guidelines released last year for institutions wanting to provide federal student aid to enrollees in such programs. In July, the U.S. House of Representatives also passed legislation that further enables institutions offering competency-based degrees to participate in federal student aid programs.”


Article in The Atlantic that encourages us to look past the low completion rates of MOOCs, and to focus on the significant volume of learning taking place.
” . . . focusing on the tiny fraction of students who complete a MOOC is misguided. The more important number is the 60 percent engagement rate. Students may not finish a MOOC with a certificate of accomplishment, but the courses nonetheless meet the educational goals of millions.”

A Norwegian government policy document that outlines the use of MOOCs for credit. Thin edge of the wedge. (Thanks to Kris Olds for pointing this out.)


Economist article that points to research detailing rising grades at elite US institution.
“In 1950, Mr Rojstaczer estimates, Harvard’s average grade was a C-plus. An article from 2013 in the Harvard Crimson, a student newspaper, revealed that the median grade had soared to A-minus: the most commonly awarded grade is an A. The students may be much cleverer than before: the Ivies are no longer gentlemen’s clubs for rich knuckleheads. But most probably, their marks mean less.”

Online Consortia: Course Sharing (Part of a Series)


We heard a couple of months back about Unizin launching a consortium of large state universities to share course content, content systems, and analytics. More recently we learnt about Cal State Online’s decision to slow down its already tepid push into a system-wide collaboration to offer online courses. These initiatives join a long list of efforts to expand online education by sharing resources across institutions.

The logic of building collaborations is infallible: joining forces can potentially bring down costs, reduce risk, increase access to better resources, stimulate innovation, ward off competition, and more. But there are more a few failed efforts to ward off any naive assumptions that academic collaborations — particularly those that concern shared courses — are fool proof. There are more than 50 consortia in North America and almost as many types. There are right ways to do it and wrong.

Having had a chance to review online consortia recently for a client, I want now to share a few observations in a handful of posts, beginning with “known obstacles” of course sharing initiatives. These obstacles are not insurmountable. But like any undertaking, it helps to know where the potholes are before you set on your journey.

one . . .

Most online consortia that include course sharing agreements were created in the 1990s; a time when many institutions had yet to make substantial investments in online learning  (e.g. technology, student support systems, professional development for faculty) and/or had yet to develop a sufficient level of confidence to carry out the functions required. As of 2014, the vast majority of universities have internal resources in place to support online education. The capacity to scale-back these investments is often difficult, owing to labour agreements and established practices. Nor do I suspect many institutions would chose to scale back at this point — given the growing strategic importance of online education to institutions. This, possibly more than any other factor,  could dampen enthusiasm among institutions for participation in shared course delivery models.

two . . .

Many course sharing initiatives target large enrolment, foundational courses because, first, these courses appear to offer the greatest possible savings and, second, because the curriculum is thought to be relatively generic. But of course these high enrolment courses also often generate higher net revenue for universities than other courses and activities. Revenue from these courses is regularly redirected to other university activities and program areas that are less “profitable” (i.e. “cost-shifting”). Consequently, initiatives that facilitate students enrolling in foundational courses at other institutions may appear to administrators as a threat to a strong and stable source of revenue.

three . . .

Owing to concerns about intellectual property, some university faculty and instructors have voiced concern about distributing their instructional content outside of their home institution. While open educational resources is often presented as a solution, in practice this approach continues to face opposition, and seems to have greater traction for the distribution of academic research papers (e.g. open journals) than instructional resources.

four . . .

Course sharing initiatives typically involve moving students between institutions (rather than the courses moving between institutions). In these cases, institutions may question whether courses offered by other institutions are inconsistent with, or inferior to, the instructional practices and academic standards of their own institution.

Education is a positional good; it’s used to define the status of the student (and the faculty and institution). Not surprisingly, my review identified cases in which more prestigious institutions refused to participate in course sharing programs with less prestigious institutions. Initiatives that bring together institutions of similar status may produce higher rates of buy-in.

five . . .

Institutions with more robust and successful online learning operations tend to less interested in participating in consortia. As a result, the consortia will fail to benefit from the participation of institutions with the greatest capacity and interest in online education.

Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @Acrobatiq

It’s About Potential

The recent PBS News Hour series, Rethinking College, posed the question we’ve heard countless times: can “computers” truly provide a quality of education equal to in-person teaching? To create conflict, the broadcasters juxtapose proponents of online education with those that believe that “real” education only occurs when the student can see the whites of the instructor’s eyes.

Rethinking Education, PBS News Hour, 2014. 

What strikes me as peculiar about these debates, other than the tendency to ignore evidence about the effectiveness of online learning, is the tendency of both the critics and advocates to talk about online education as if it is a mature, established style of teaching and learning; that what currently constitutes online learning is the end of the road; that we can expect no more improvements.

This is odd because the most striking difference between in-person teaching and online as of 2014 is their relative potential to improve. Online education, unlike in-person education, is in its infancy. It will improve dramatically. The only questions are how quickly and who will lead the way.

While online education is still in that early, awkward stage, the classroom format long ago reached the plateau. This isn’t to say that in-person teaching isn’t often of great value — that’s not the point. The point is that our capacity to dramatically increase the value of in-person learning is relatively limited.

New technologies and new ways of doing things tend to follow a common (“s-curve”) trajectory: at first they offer little value, but in time — as conditions change and improvements are made — the value escalates rapidly. Eventually the rate of progress subsides, and the value of the once new approach settles into a plateau, at which point improvements are merely incremental, unless you significantly increase costs, which in turn, reduces value. (Value: a balance between quality and cost.)

In-person teaching is a mature format. At the plateau stage, the focus is on trying to extract more value. Examples of this include hiring more adjuncts and creating teaching and learning support offices. These tactics may help, but the improvements in value are predictably incremental.

By comparison, online education still has a tremendous amount of room to improve. As of 2014, most online courses in traditional, brick and mortar institutions in North America are still built and delivered according to a model borrowed from the classroom. Lone instructors without the necessary time, incentives, skill-sets and funds maintain ultimate responsibility for course design and development. Funding for course development is limited to what can be generated by offering the course at a single institution for a few semesters, until such time that it needs a revamp. It’s still built in a cottage-industry fashion, like the classroom. It’s not surprising that online education has not yet led to lower tuition for students. Costs haven’t declined. The value proposition to students is still largely based on convenience, rather than improved educational value.

But in time, new instructional strategies and technologies will join forces in new ways that will drive up quality and reduce costs. There’s already pockets of innovation on the fringes: The Open Learning Initiative and NCAT, albeit in different ways, have shown that new approaches can increase value, whether in reduced costs, improved learning outcomes or both. Innovative program designs from institutions like Western Governors University, Arizona State University, SNHU, and Rio Salado, and Capella University will provide templates for other institutions to follow and — as a result of their fast enrolment growth — stimulate healthy competition.

What the future of online education will look like isn’t actually all that mysterious. We’ve been talking about many of these instructional strategies and features since the late 1990s.

  • Provide students with hundreds of opportunities to practice and test their understanding of the material (rather than, say, a midterm and final);
  • Use high-production value instructional media that explains difficult concepts and processes through sophisticated illustrations, animation, video and audio;
  • Employ adaptive software that personalizes learning to meet students where they are at that moment;
  • Design courseware that is customizable so as to enable economies of scale, thereby driving down costs per student;
  • Draw on the best instructional design strategies that are based on the rich field of learning sciences;
  • Leverage learning analytics to provide learners with constant feedback so that they come to learn how they learn best.

Institutions that recognize the current trajectory of online education — and can find ways to make it part of their operations — will be well positioned for the coming decade. Those that assume that the future state of online learning will be similar to the status quo in 2014 may miss out on one of the few great opportunities to innovate in higher education.

Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @Acrobatiq 

Podcast Interview on “Meet Education”

Nick Di Nardo, host of “Meet Education” interviews Dr. Keith Hampson of Acrobatiq.


Scale in Online Higher Education

“Economies of scale are factors that cause the average cost of producing something to fall as the volume of its output increases.”

One of the common and more provocative scenarios for MOOCs involves using these courses at multiple institutions. In other words, build once (e.g. by Harvard), use often (multiple institutions). The benefit of this approach, at least potentially, is that by sharing content across multiple institutions we drive down the cost of offering online courses and, second, the relatively high volume of users makes it possible to increase the level of investment in each course. San Jose State University’s attempt to integrate the Udacity course serves as a high profile example.

This isn’t a new idea. People have been considering how scale might fit into higher education since the 1990s. It wasn’t until MOOCs that we had a concrete and well-known example to serve as focal point for this discussion.

Higher education has not put a great deal of effort into finding ways to reduce costs through scale. While other sectors seized the new economics of digital content storage and distribution, higher education has continued to produce and use digital instructional content locally; a digital cottage industry model, of sorts. For elite institutions or those aspiring to be elite, scale runs counter to one of the trappings of elite status: exclusivity. For others, the highly decentralized organization of the institutions makes institution-wide scale improbable. Another factor is concern amongst decision-makers about the impact of sharing courses on labour market value and autonomy.

“Let’s not kid ourselves; administrators at the CSU are beginning a process of replacing faculty with cheap online education.”

San Jose State University faculty responding to the call to use Michael Sandel’s MOOC as part of their program.

Just How Common Is It?

Scale is directly tied to the degree to which curriculum is common or shared across institutions.  This component of scale hasn’t, to my knowledge, been sufficiently addressed. If a course is to “scale” there must significant commonalities across institutions of higher education of what they teach and how. It is implied in the Coursera/Udacity/edX business model — but surely never stated outright — that undergraduate curriculum in the US (and beyond) is sufficiently common.

Commonality is the premise of the textbook industry, of course. Certain courses, particularly in the first and second year of programs,  can be satisfied by a single textbook across hundreds of institutions. US textbooks are used in other nations, in fact — only modified lightly to refer to local conditions. A highly provocative study by (what was then called) Coopers and Lybrand suggested that 25 individual courses constituted 80% of registrations.

The concept of The Long Tail might provide a structured approach to understanding the degree to which courses are common and how this might impact costs.

Anderson argues that consumers have historically purchased “hits” — not because they are indifferent to less popular fare, but because of a lack of choice. But the Internet is now removing the bottleneck between suppliers and consumers. And as search and distribution technologies improve, and costs continue to decrease, Anderson forecasts that the top sellers in a variety of markets will constitute a smaller share of total sales, and the number of different products available will increase dramatically (i.e. further flattening and lengthening of the distribution of sales). While the top-rated television shows often captured more than 50% of the audience in the 1960s, today they represent  only10% to 15% and declining. New business models like Netflix have only amplified the shift.Chris Anderson’s “The Long Tail” (2004) contends that the Internet has fundamentally changed the economics of producing and distributing digital products. “Shelf space” on the Internet is virtually infinite and increasingly inexpensive. It’s now financially feasible for vendors to sell a much wider variety of digital products, particularly books, films, and music and other media. Marketing strategy is shifting from a dependence on a relatively limited number of “hits” or “blockbusters”  (e.g. Top 40 radio; New York Times bestseller lists) to serving niches.

The Endurance of “Hits” and Digital Content in Higher Education

Several studies have suggested that Anderson’s original work overstated the degree to which people migrated away from a select number of “hits”. Anita Elberse’s analysis, Should You Invest in the Long Tail?  (Harvard Business Review), suggests that the market for “blockbusters” remains largely safe from the onslaught of multiplying niche markets. Despite the changing economics of content authoring and distribution that Anderson describes, the bulk of sales are still found in the “head” and the “tail” is remarkably flat.  Greater variety is not always matched by demand.

The degree to which Elberse’s argument invalidates the Long Tail theory is somewhat dependent on where we draw the line between the “head” and the “tail”; that is to say, what we think constitutes a “hit”. What’s most useful about her work is that it reminds us that there are important forces at play in each sector that give shape to the distribution of sales (head and tail).

The insights from Anderson, and the questions posed about these insights by Elberse, can help us understand and anticipate the factors that might the demand for a more diverse range of content in higher education? Do we have a preference or need for “hits” in higher education?

As a starting point for addressing these questions, I offer three issues that may effect the length of the “tail” of content in digital higher ed:

  • Quality (Re)assurance. Do we need assurances from others in the field about the quality of content, and from whom, exactly? There are conventions in place: In traditional textbook publishing, it is conventional to employ currently employed academics from well-known (preferably) institutions of higher education as authors. In OER, we find the use of simple rating systems, such as stars (one-to-five), to crowd-source evaluations. To what degree will the need for assurance from others limit the expansion of the “tail”?
  • Consistent and Coherent Curriculum. To what extent must the content be consistent with the curriculum within the institution and other institutions? Although not to the same degree as K12, higher education is a “system” in which students progress, step-by-step. There are levels into which content must fit. When a student moves from first year to second year, or transfers from one school to another, there is an assumption (hope?) that the first year accounting course at University A is roughly equivalent to the same course at University B.  The Bologna Process is relevant here.
  • Production Quality. How important is it to educators and students that the content that they use meet a minimum standard of production quality? That is to say, at what point does “home-made” content become a liability because it is either difficult to integrate into an LMS, “buggy” (in the case of content embedded in applications), or simply difficult to use for students and instructors? How far along the “tail” will content of sufficient quality be found?

As the variety of content increases in the coming years, educators, institutions and publishers may want to pay attention to these and other issues to determine how they go about creating, acquiring and distributing content. Although it’s too early to be certain, my suspicion is that like the markets of music, film, and books, the demand for “hits” in digital edu content will remain surprisingly strong.

:: ::

Note: A number of people have written about the relationship between The Long Tail and education. You’ll recognize, though, that some of them use the concept of the Long Tail to analyze the diversification of students. That is, the tail gets longer as more people participate in higher education.  Instead, I use the concept to analyze the diversity of educational content. Although the former approach may be of great value, my focus on educational content is more in line with Anderson’s original use of the concept.  

Note 2: The concept of “mass customization” overlaps with The Long Tail in important ways. For more on more customization, see “How Technology Can Drive the Next Wave of Mass Customization” (McKinsey)

Dr. Keith Hampson is Managing Director, Client Innovation at Acrobatiq, a Carnegie Mellon University venture born out of CMU’s long history in cognitive science, human-computer interaction, and software engineering. @keithhampson & @Acrobatiq


Dr. Keith Hampson :: Managing Director, Client Innovation, Acrobatiq LLC


Get every new post delivered to your Inbox.

Join 3,693 other followers

%d bloggers like this: