MIHNEA MOLDOVEANU on creative destruction in higher education
AN UNPRECEDENTED — and massively overdue — wave of innovation in the higher education industry is about to be unleashed, and it will bring unprecedented disruption to the field.
The waves of digitalization of content, connectivity and interactions that have disrupted the media, retail, travel, entertainment, publishing, manufacturing and financial industries are about to strike the higher education industry, presenting a massive opportunity for the redesign of a field whose practices have remained unchanged since the early 1000s.
This is not a typo: Early Renaissance paintings depicting classrooms and historical accounts of learning practices both indicate that the basic choreography of content, context, learner-teacher interactions, and structured drilling and quizzing as a prerequisite to certification have not changed for more than 1,000 years. The lecture-problems-recitation-exam format — canonized by repeated and unquestioned practice in early modern Europe and North America — has formed the basis on which learners are sorted, measured, incentivized, evaluated and ‘taught’.
Remarkably, these practices have persisted in spite of a century’s worth of empirical evidence — in cognitive and applied psychology, in educational practice, and more recently in artificial intelligence — that there are faster, better, cheaper ways of helping learners acquire new skills than those that populate current college and university classrooms and labs. Spaced learning, variable-delay reinforcement-based learning, socialized learning, hyper-resolution feedback, problem-based learning — amongst others — present modern-day educators with building blocks for the redesign of the learning experiences of learners in ways that increase the efficacy and efficiency of both skill acquisition and skill transfer — i.e. the application of a skill outside of the context in which it is acquired.
In spite of the accumulating and accelerating evidence for the sub-optimality of current pedagogical practices, innovation in the ~$3 trillion + (2017 USD) higher education field has been slow, spotty and segregated. The behavioural blueprints of learning experiences — courses, classes, recitations, tutorials, quizzes, problem sets, essays, exams — have yet to change in ways that resemble the restructuration of everyday experiences in the music, retail, publishing, travel, communications or financial industries.
The explanations usually offered for this painful factoid draw on the macro- and micro-incentives of research-active academics and departments that use teaching--
driven revenue to subsidize research activities whose outcomes are the ones that ‘count’ and the institutional forces of research-centric universities that align in the direction of minimizing the logistical unpredictability that innovation waves trigger. They point to the sociology and social, cognitive and developmental psychology of homo academicus — a creature better suited and predisposed to speaking about a phenomenon (say, innovation, usually taken to mean innovation in a different field) than to practising it, to analysis of innovative options rather than to the prerequisite action and to representing rather than to intervening. Or, they take the ‘tough-mindedly realist’ position that higher education is a filtering and evaluation process of students for employers, wherein learning and development are desirable but rare and accidental by-products. Of course, this is precisely the sort of (quasi)-causal explanation whose proliferation causally contributes to the perpetuation of the status quo.
In the face of such massive synergistic forces, how could it not be that the practice of teaching and learning lags behind insights and empirical findings by a good century?
But, this explanation is both incomplete in factual base and erroneous in inference. The last 10 years have seen massive innovation in the field. MIT’S 20-year-old Open Course-Ware initiative and Stanford’s 30-year-old commitment to continuous, remote learning have morphed and proliferated into a massive, open-learning ‘exoskeleton’ which under the guises of EDX, Coursera and Udacity bring state-of-the-art content to millions of users while, at the same time, making it possible for dedicated instructors to learn how to teach from one another.
Curricular innovation in professional programs — notably schools of business and medicine — has been on the rise since the early 2000s, responding to new demands for quintessentially-human and executive skills from evermore-savvy recruiters, whose own in-house training programs have also grown in sophistication and size (witness a tenfold increase, from 200 to 3,000, of ‘corporate uni- versities’ between 2004 and 2015).
Responding to the need for contextualized learning that combines conceptualization and technical skills with the practical know-how provided by context, leading-edge Engineering programs — such as the Olin College of Engineering — have redesigned their learning vehicles ‘from scratch’, and from first principles, to maximize on the still-elusive objective of skill transfer from classroom to ‘life’ — and the life-world of organizations, in particular.
Alongside positive evidence for curricular and institutional innovation, there is no evidence that course-level innovation happens less frequently than does innovation in any other field — including those considered to be considerably less inert than academia. New techniques for polling learners, drawing them into the socialized and disciplined dialogue of the classroom, and making them co-accountable for the efficiency of the learning production function of their program are finding their ways into graduate and undergraduate courses alike. The current and burgeoning wave of investment in ‘Edtech’ — educational technologies meant to increase the effectiveness of learning through personalization of content and context — suggests that pedagogical innovation is alive.
Alive, yes; but, why does all of this innovation not translate into a radical transformation of learning practices across the field? Why does the industry increasingly appear to live up to Peter Drucker’s indictment of it as the ‘largest burden on the backs of taxpayers’, stimulating increasingly-shrill calls for radical technology-based transformation?
A large part of the answer lies in plain sight. Local, desynchronized, segregated innovation needs an open, integrative platform to generate both internal momentum and an industry-wide transformation. Advances in telecommunications — we are now working on 5G systems — provide a telling example. Innovations in the physical and medium access control layers (layers 1 and 2 of the OSI hierarchy) had been frequent and significant ever since the construction of the first digital modems in the 1970s. But, until the
person a liz the the most effective forms of learning are personalized to the learner, socialized to her learning group, and contextualized to her work and life.
IEEE standards and the standard setting process evolved to the point where innovations from companies large and small could be synchronized and harmonized into networklevel blueprints for coding and modulation — i.e. until the IEEE created an innovation platform — we could not even contemplate sending Youtube videos over handheld devices. What IEEE’S platforms enabled for telecommunications, open-source repositories and platforms like Github enabled for the development of algorithmic building blocks that took us from Web 1.0 to the current Web 2.5 — and open AI initiatives are promising to do for innovations in self-refining algorithms.
In the educational field, the Learning Management Engine (LME) provides the equivalent innovation platform that promises to aggregate and integrate across isolated innovations in learning and instructional design. It provides a locus of innovation that allows both learners and instructors to learn about the best ways to learn, and to teach while at the same time learning.
A recent, large-scale study jointly undertaken by the Rotman School of Management and Harvard Business School has identified the massive gaps in skills learned and skills transferred that besets the higher education field, showing that the most effective forms of learning are personalized to the learner, socialized to her learning group, and contextualized to her work and life environment.
That is precisely what a learner-centric LME will do: It will allow instructors to collaboratively and interactively design content and learning experiences adaptive to the preferences, backgrounds, cognitive and affective styles of learners, by interfacing to platforms and applications used in recruitment, admissions and alumnae/i relations that track learner backgrounds, interests and employment patterns, while at the same time allowing instructors to do quick A/B testing of content and learning experience designs.
Data analytics — proprietary and closed in current systems, but open in the learner-centric LME — will allow for continuous tracking of learner profiles and learning-oriented behaviours and learning outcomes, and for in-depth understanding of what-works-for-whom- and-when when it comes to the design of learner-instructor and learnerlearner interactions.
Unlike current LMES, which do not allow for instantaneous in-band transfers of data between the core engine and other learning-enhancing applications, the learnercentric LME will enable instructors and learners alike to use and share learning apps in the same learning environment, thus deepening collaboration among instructors and programs and tapping into the burgeoning ecosystem of ED Tech applications that is currently ‘waiting on the sidelines’ and being only sporadically used.
With the flexible allocation of decision rights to learners and instructors and the free flow of data and content across programs and schools, a real ‘learning innovation ecosystem’ will be enabled. Higher education is a densely and tightly coupled network of activities and tasks, which include selecting and motivating learners, informing and testing them, connecting them to instructors, content and other learners — all while heeding the metronome of the academic year and program guidelines. If we think of an
LME as a network of user behaviours enjoined and engendered — not just web pages and apps — it becomes clear that its network must have a similar level of complexity to that of the system it is serving, if it is to function as an innovation hub in which instructors can learn from one another — and from their learners.
Learner-centric LMES will enable a flexible allocation and re-allocation of decision rights over the learning process: Whereas current LMES give the preponderance of authority over class constitution, allowable content sharing, analytics, co-horting, apps deployment and interfacing to administrators and developers, a learner-centric LME will allow instructors and learners to collaborate in the design of the learning experience itself, by selecting the additional applications, data analytics, testing protocols and class designs that best fit their learning and instructional objectives. Even without any of the improvements in the learning production function promised — and likely over-promised — by pundits and gadflies in the ‘new AI’ movement, proven, reliable techniques like collaborative filtering can be deployed in the learner-centric LME to produce the social multiplier of learning efficacy that by now has been amply documented in empirical research.
Why Are We Not There Yet?
Once articulated, a learner-centric LME seems oddly obvious as a large piece of the solution to filling the ‘innovation hole’ of higher education. Why are we not there yet, despite having par-coursed four generations of learning-management systems and engines—and of the ratification of a new standard (Learning Technologies Interoperability) designed to assure the very kind of openness to applications and analytics we currently lack?
A quick look at the evolution of the LMS/LME industry — currently sitting at about US$3.6 billion/year and projected to grow to $7 billion/year by 2020 — gives us the requisite hints. The industry is heavily concentrated around entrenched providers of admin-centric LMS platforms that are not interoperable, closed to state-of-the-art analytics engines, and closed to learning applications that are originating in the Web 2.0+ environment of socialized, networkbased learning.
Although some of them started from open source platforms ( Canvas, Moodle), they developed interstitial modules for data transfer and interoperability that make their current instantiations de facto closed, which allows them to charge universities richly for analytics on their own