The efforts to boost the nascent field of learning analytics could bring about a sea change in education, making it possible to personalize — on a massive scale — students’ learning by their individual interests and needs, according to a comprehensive report, involving experts from academia, business, nonprofits, foundations and government . It was written by Roy Pea, the David Jacks Professor of Education and Learning Sciences at the Stanford Graduate School of Education.
The research and associated report, conducted under the auspices of the Learning Analytics Workgroup, describe how education data could transform how students are taught; how teachers are prepared and further developed; how education research is conducted; how education-related information is used and managed; and how foundations’ funds are allocated. The report points to the ways that business, health and other sectors are beginning to capitalize on the exponential growth of data, but underscores that education lags behind them — and that a major hurdle is its lack of expertise.
“Data science, as a distinct professional specialization, is in its infancy,” Pea observes. “What we are calling for is an even newer specialization, education data science,” which is the backbone of learning analytics. “Technology has run ahead of the readiness and human capital in [this] emerging field,” he adds. “Demand is ahead of supply and will continue to be without a systematic effort at capacity building in the form of training programs and field building.”
The report, titled Building the Field of Learning Analytics for Personalized Learning at Scale, draws on a series of meetings held over the past three years by the Workgroup, which consists of 37 leaders in the field from companies, universities, government, foundations and nonprofit organizations. The report incorporates research and ideas presented in 11 specially-commissioned white papers, and builds upon interviews with hundreds of teachers and school administrators. The work was made possible by support from the Bill & Melinda Gates Foundation and the MacArthur Foundation. Pea served as the project’s principal investigator.
As part of its analysis of the challenges and opportunities that education data science poses to different stakeholders, the report examines the shortage of specialists in learning analytics in higher education. It suggests “bringing current education faculty — especially those who study psychometrics and educational measurement — into learning analytics,” along with reaching out to faculty in not only computer science and statistics but also bioinformatics, business intelligence, particle physics and other fields that do advanced work with large data sets. “Graduates from computer science, data science, learning and educational sciences, computational statistics, computational linguistics and other areas are all potential fits for learning analytics postdoctoral training,” Pea says.
In addition to schools of education partnering with data scientists from a variety of disciplines, the report emphasizes that K-12 teachers must be instructed in data literacy so that its fluent use can be integrated into their daily practice.
The report goes on to address the need to establish a range of “success metrics for learners” to suit the needs of different education stakeholders. Such a step, in turn, demands developing an evolving “learner model” that would incorporate, for any learner, data beyond those that measure cognitive abilities, including indicators of student interactions during learning activities, student mindset, learning media genre preference, and perseverance and persistance (sometimes called grit), among other things.
A major goal, Pea writes is to “create predictive learner models that get the greatest percentage of learners to competency in the shortest time at the lowest cost.” He offers a concrete example: “Learning analytics systems presumably will allow researchers and educators to identify early warning indicators when learners struggle with key developmental phases like prealgebraic thinking prior to their enrollment in early algebra classes.”
“Overall, the motivating question for the field is how to develop personalized learning systems,” Pea writes. “For which learners does a learning intervention work or not, under what conditions, and why?”
The report further examines what’s needed to establish a learning analytics infrastructure that is functional and allows data to be easily shared, while protecting privacy. “It will need to incorporate visualization and data reporting systems like learning dashboards, which will be accessible to decision makers (including teachers, learners, administrators, and policy makers),” Pea says.
In its final section, the report offers a detailed road map that policy makers, foundation leaders and researchers can pursue to foster the field of learning analytics. It cites a number of recent advances — the development of the Common Core standards, more sophisticated measures of effective teaching, growth in data mining and analytics, personalized and blended learning models, and digitally-born learning innovations, to name a few — toward realizing the promise of personalized learning pathways.
“Philanthropic foundations and government granting agencies are waiting in the wings to determine if we can draft a plan that is worthwhile,” Pea writes. “The learning analytics community needs to step forward with a plan to address the challenges and opportunities discussed in this report.”
Subscribe to our monthly newsletter.