Perspective: Data: A Wider View
As we enter an age where data seems to be everywhere, both educators and education researchers are becoming aware of its power. Yet our current view of data is highly limited. We have only begun to conceive of the possibilities that multiple forms of data can offer for teaching and learning. We will need imagination to envision all the novel ways data can empower education and fuel innovation. To fully understand how data stand ready to transform teaching and learning we must think broadly and look far ahead.
For decades the Concord Consortium has been beating the drum about the importance of data. In recent years, the rhythm has intensified. Our newest projects and our vision for the future reflect this, and they do so in some surprising ways.
Data games for learning
One intriguing view of data comes from gameplay. The use of games for education is a growing field with significant promise for STEM learning. Games provide a strong means of motivation and engagement, and align with many STEM learning goals. The data generated by players as they interact with games offers promise for researchers as they seek to better understand how learning through games takes place. We’re interested in all of these, but a newly funded project, Data Science Games, is making use of the data generated as students play digital games in a novel and creative way.
When students play a data science game, their gameplay actions generate data, but the data takes on more significance than in other gaming examples because it becomes essential to the game itself. To succeed at a data science game, students must visualize, understand, and properly apply the data their game playing has generated in order to “level up” and progress within the game. As they visualize and analyze the data, planning and plotting new, evolving strategies, students learn the fundamentals of data science.
Guiding teaching and learning via data
In other Concord Consortium research, we use the data generated by students as they explore virtual environments to better understand and aid the process of learning. Our SmartCAD project expands on this important line of work by improving and advancing the data analytics behind our successful Energy3D virtual CAD software. SmartCAD will develop a full-featured environment outfitted with sophisticated metrics for monitoring and understanding the process of engineering design in real time. This is an essential step in understanding what has until now been largely a “black box.” Building on highly promising past work with the same environment, we are now able to visualize process data in powerful ways. The fine-grained data will shed light on design iteration and, in particular, on how data can be leveraged to provide feedback to students during the design process.
This work is an example of our view of learning analytics. Fostering learning is the ultimate goal, not merely examining the analytics themselves. Another new project, GeniGUIDE—Guiding Understanding via Information from Digital Environments—is expanding this concept by integrating our game-based genetics work with robust intelligent tutoring technology. Rather than keep the tutoring system’s data limited to within the software’s feedback, this project explicitly acknowledges what is often a central mantra of ours: the most intelligent tutor is the one standing at the front of the classroom. By gathering the rich data streams generated as students breed dragons and attempt genetic challenges in our Geniverse environment, we can make use of such data to advise the classroom teacher and to connect individual students via direct feedback. By forging new ground for one of data’s most important uses—aiding and guiding teaching and learning—we’re blazing trails that we hope the field will follow in meaningful ways.
Discovering data amid the fabric of learning
With so much data streaming in from existing technologies and with new examples emerging every day, it’s easy to think that we have more than enough information about learning to go around. However, capturing the most critical and substantive interactions during the teaching and learning process—the discourse and conversation among students, teachers, and mentors—remains elusive. Data related to spoken language is the center of an incredible amount of effort within educational research, but these efforts are enormously time-consuming and subjective—as anyone who has ever coded video clips or performed detailed classroom observations is painfully aware. And even the most careful efforts in this direction miss huge parts of the teaching and learning picture.
The potential for capturing, processing, and gathering meaning from spoken language data via technology is increasing at an exponential rate, but its current development is missing a vital part of the picture—its use in education. Often-finicky interactions with Siri or Google products form most people’s image of modern spoken language technology. But these represent only the faintest view of the potential of today’s cutting-edge speech technology and its capabilities for automatic speech recognition and natural language processing. Many lesser-known technologies allow the extraction of highly accurate measures such as word counts or meaning-filled data from prosody (variations in tonality of speech, which can indicate emotion or stress). Spoken language technology stands at a powerful tipping point, and the central goal of one of our new projects is to rally attention to its potential for the education research community. We are at the forefront of a sweeping new era in educational research, in which spoken language technology will open up grand new horizons. In doing so, we hope to spark a revolution that can make wide use of the spoken interactions that have been at the core of educational exchanges for millennia.
Taking the long view of learning data
Through these and other initiatives, we are embarking on new directions and widening our current view of what data is and how it can be used for education. Our future vision, of course, far exceeds even our current work. In the spirit of forward thinking, we are thrilled to welcome another visionary, Janet Kolodner, to our team as Chief Learning Scientist. Her vision for the future is well known to many. She will bring our work into yet-uncharted territory in the area of learning science and technology. This will include work with project-driven learning, in which teachers and learners grapple with challenges that can extend for months, placing new demands upon software tools and platforms that must be able to make the related data discoverable and relevant for learning over time.
The future is exciting and by nature unpredictable. We know for certain, however, that it will be filled with data, whose expanding definitions we may have yet to understand. As we enter that future, we challenge others—as we’re challenging ourselves—to consider a new view of data, its definitions, and its significance across all STEM education.
Chad Dorsey (cdorsey@concord.org) is President of the Concord Consortium.