Leaders in education rarely agree on the best approach to setting standards, designing curricula, and administering schools. However, there’s a growing consensus that learning is not a one-size-fits-all activity, but that it is both highly complex and highly personal. The optimal teaching approach for a specific student at any given time depends on the individual’s personality, state of mind, environment, past experience with the subject, plus a myriad of other external and internal factors.
It’s a formidable challenge. How do you address these factors, on a mass scale, to help each student discover their unique talents and realize their full potential?
The good news is there’s a tremendous, though largely untapped, resource that can provide a wealth of information: The numerous interactions, breakthroughs, and setbacks students experience every day. Digital learning tools, monitoring technology, and big data analysis can uncover insights on what really works and what doesn’t for each unique situation. Teachers and students can then apply that information to educate and learn in the most effective way possible.
Getting Beyond Annual Reports
Evaluating school and student performance data is nothing new. For decades, administrators have assessed schools based on standardized tests and graduation rates. But these standards only show a narrow view of student performance, shaped largely by how the tests are structured by those that develop them. This measure also fails to understand how individual circumstances adversely impacts the population of students who just don’t perform well under the pressure of standardized testing. In cases where test assessments indicate a problem, there’s no clear path to pinpoint what the issue is or how to resolve it.
Advanced data analysis tools are now helping to paint a clearer picture. Cloud-based software suites help school systems bring real-time data together from multiple sources to identify problems in schools earlier. By looking at a more comprehensive dataset and applying learning from past analysis, administrators can catch negative trends in schools early and work to right the ship.
For example, in 2011, Spokane, Washington, Public Schools hired a consultant to conduct a study of two graduating classes, bringing together years of varied data collected by the data analytics software used by the school district. Looking at data patterns spanning from elementary to high school, they found the strongest predictors of students dropping out of school in the future were (1) unexcused absences in elementary school and (2) failing classes in high school. Equipped with this knowledge, administrators were able to zero in on these and other patterns to help identify drop out risk and provide resources for future classes.
The data analysis tools used in the Spokane study also profile individual students daily, generating a dashboard with data on attendance, grades, test scores, behavior and more. With auto-generated and easy-to-understand reports and visualizations, administrators can be alerted in real-time when students may be at risk.
Analyzing How Students Learn… and How They Don’t
These measurements are useful, but they’re only the tip of the iceberg. The real potential of leveraging big data in education is making sense of what’s happening when students are actively learning as well as when they’re struggling.
The rise of digital education, such as learning software, games, and massive open online courses (MOOCs), has yielded a wealth of data on how students progress through lessons and solve problems. For example, one popular program uses algorithms to develop a custom curriculum for every student and teacher, based on their individual progress.
When “smart schools” integrate technology like personal tablets and laptops into their classrooms, they’re able to keep a highly-detailed record of each student’s progress — where they’re thriving, where they’re struggling, and what activities work best for them. In aggregate, this growing dataset will yield more and more insights on the best curriculum for individual students in different subjects.
Digitizing the Physical World
Digital learning tools have a big blind spot: The data doesn’t tell you what happens when the student walks away from the screen. Cutting-edge smart schools are looking to fill in the gaps, gathering data on what’s happening in the classroom.
For example, one startup school with locations in San Francisco, Palo Alto, and New York City, has brought audio recorders and cameras with fisheye lenses into every classroom. The footage lets teachers review important moments, such as critical breakthroughs or a class going off the rails. But the future holds the possibility of gathering billions of moments from thousands of schools, then using advanced data analysis to find patterns of success and failure. The system could give teachers real-time guidance, based on this extensive analysis.
This vision builds on successes from other industries. For example, some professional basketball leagues use motion-tracking algorithms to gather data on what every player is doing, 25 times a second, yielding insights down to the optimal number of times to dribble before taking a shot. Similarly, retail stores use motion-tracking analysis of shopper patterns to optimize their store layouts.
As smart schools advance, the analysis could get down to the biological level. Biometric devices could keep tabs on engagement, anxiety, and boredom, to help see what’s working in the classroom and what isn’t. Some academic researchers have developed experimental tutoring software with a digital tutor avatar that adapts to the student’s emotional state, indicated by analysis of expression, heart rate, pupil dilation, and more, gleaned from a video feed. The avatar then offers support, depending on whether the student is showing signs of confusion, stress, boredom, or engagement.
This vision worries many educators, psychologists and parents alike. The prospect of training cameras and biometric sensors on students all day while tracking their every digital move raises serious privacy concerns, and critics say the continual surveillance could take a heavy toll on students, or normalize an unhealthy level of surveillance. Additionally, some worry with increased reliance on educational technology, schools will lose sight of the valuable connection between teachers and students, and the intuition that comes from years of experience.
As new technology makes it possible to gather greater volumes of data, and advanced software makes it possible to draw deeper insights, schools must weigh the benefits against the risks and evaluate what is truly in the best interest of students. How can educators realize the benefits of smart school innovation without losing the strengths of traditional education or compromising student privacy?