You are here
As learning goes digital, big data can guide us
"Fear and trepidation" are stalking the educational landscape, according to Roy Pea, who stood at the podium with an awful green monster projected on the slide behind him. It's MOOC hysteria, showing at a school near you: "Something New in Shock-Thriller Education."
Well, not quite. But Pea, the David Jacks Professor of Education, told an audience at the recent annual meeting of the Educause Learning Initiative (ELI) that missing amid all the hand-wringing and worry over the advent of massive open online courses (MOOCs) and digital learning is any serious consideration of the linkages between course design and the learning sciences, and between learning sciences and analytics.
Those gaps, and an exploration of what "learning sciences" actually means, were the subject of his keynote speech, "Learning Sciences and Learning Analytics: Time for a Marriage," delivered to educators and educational technology specialists gathered in Denver. [Visit here to watch the lecture.]
Learning sciences is a field that emerged in the late 1980s that crosses various disciplines including psychology, education, computer science and anthropology. It aims to understand how learning takes place under different conditions, and over recent years it has particularly been devoted to studying digital innovations in pedagogy and instructional design. The Stanford Graduate School of Education, where Pea is on the faculty, has masters and doctoral programs that include learning sciences. At Northwestern University, Pea founded the world's first learning sciences PhD program, in 1992, and went on to direct it, before moving to Stanford in 2000.
"The pace is truly astounding," Pea remarked regarding MOOCs, but the founding model of the genre — epitomized by Khan Academy videos or lecture-and quiz-centered Coursera classes — needs to be developed much further, he said. Short videos and short-answer quizzes are not enough, though he acknowledged that peer assessment and more expansive pedagogies are beginning to appear in MOOCs.
Pea was a member of the U.S. Department of Education's National Education Technology Plan (NETP) Technology Working Group, which in 2010 drew up a new vision for K-12 and higher education powered by technology. That group's findings and recommendations formed a foundation for his address.
Above all, he said, digital learning must place students and learning at the center. The NETP called for a "learner-centered framework for an always-on networked world" that could empower students to take control of their own learning. Technology would give both teachers and students increased flexibility to switch tracks or supplement lessons to make them more tailored to students' interests, needs and goals, and it could embrace not only classroom experiences but also "lifelong and lifewide" learning.
How to do this? You need data, lots of it, and therein lies the great promise of digital learning. Information on students' work habits and the effectiveness of tests and feedback can be harvested, analyzed and compared, similarly to how private commercial entities use web data for insights into consumption. Doing this in an integrated way in the world of education is no easy matter, however. In fact, the NETP refers to the development of such a system as a "grand challenge."
The second part of the title of Pea's address refers to learning analytics, which Pea defined, citing colleague Professor Erik Duval of the University of Leuven, Belgium, as "collecting traces that learners leave behind and using those traces to improve learning." Those traces can be reduced to numbers. The marriage in the title, then, brings numbers together with design and innovation.
Chief among the field's priorities, Pea said, is to develop connected "learning maps" across courses and disciplines, that are based on empirical studies concerning progressions in learner competencies and coherent knowledge integration rather than disparate facts and procedures. Learning maps incorporating web technology can enable the tagging of learning resources, periodic assessments and the tracking of student performance. In other words, learning can be personalized as the maps are interoperable and open.
In higher education, unlike in K-12, learning maps are conspicuously absent, and their development is an "urgent priority," Pea said. He is not the only one to think this, he noted, but that is also part of the problem, because lots of individuals and nonprofits and educational technology companies are developing their own learning maps. It's a bit like web and hypertext technology in the 1980s, he said, when there were no common standards or criteria. So learning maps must be incorporated into higher education, and, to some extent, standardized for K-12.
In his talk, Pea profiled an alliance of educators representing districts, schools, nonprofits, for-profits and foundations formerly known as Shared Learning Collaborative (and now as inBloom), that is working on making personalized learning — the availability of open access, flexible, useful learning maps and recommended learning resources for every student's specific interests and needs — a reality throughout U.S. schools.
He followed that with a cautionary note from the National Science Foundation: "Of all the transformational catalysts brought by the Internet and the Web as technology infrastructures," the NSF said, "perhaps the most fundamental is that innovators and entrepreneurs can draw upon shared, interoperable services and platforms." Letting a hundred flowers bloom may be a good thing for a while, but at some point ¾ and Pea seemed to suggest that point is now ¾ it's time to come together and develop coordinated standards and platforms.
The marriage between learning sciences and learning analytics must not only place the learner at the center of things, Pea said, it must also design for engaged social learning. This is especially critical with MOOCs, where students are never in the actual presence of their teacher and instead participate through online communities, with the notable exception of those using 'flipped classroom' models. In these instances, students watch lectures and do web-based activities remotely and then engage in discussions with their professor in a physical classroom later.
Another priority is to develop broad competencies, beyond cognitive achievements to include the interpersonal competencies (such as collaboration) and intrapersonal competencies (such as persistence and metacognition), which all collectively contribute to what the National Research Council in a 2012 report has called "deeper learning" for transfer beyond classes and classrooms. And, finally, remaining priorities are to understand learner goals, to develop interdisciplinary teams and to use richer pedagogical models such as project-based learning, problem-based learning, complex system modeling, cognitive apprenticeships, knowledge-building communities, computer games, learning in virtual worlds and immersive and embodied learning, among others.
The NETP report, released by the Department of Education in 2010, described how online learning offers opportunities for integrating assessment and learning so that information on those processes can be gathered and then form part of a feedback system, with data fueling new learning. The group's findings were elaborated upon in a subsequent report on education data mining commissioned by the Department of Education.
Educause, the sponsor of the ELI meeting, is a nationwide nonprofit organization dedicated to information technology in higher education.
R. F. MacKay is communications manager for Stanford Online.