Tuesday, October 21, 2014

Big Data for Poor Students

WASHINGTON, DC – Countries need skilled and talented people to generate the innovations that underpin long-term economic growth. This is as true in developed as it is in developing economies. But it will not happen without investment in education and training. If we are to end poverty, reduce unemployment, and stem rising economic inequality, we must find new, better, and cheaper ways to teach – and on a vast scale.

This goal may seem to be beyond even wealthier countries’ means; but the intelligent collection, analysis, and use of educational data could make a big difference. And, fortunately, we live in an age in which information technology gives us the right tools to broaden access to high-quality, affordable education. Big data – high-volume, complex data sets that businesses use to analyze and predict consumer behavior – can provide teachers and companies with unprecedented amounts of information about student learning patterns, helping schools to personalize instruction in increasingly sophisticated ways.

The World Bank Group and its private-sector lending arm, the International Finance Corporation (IFC), are trying to harness this potential to support national education systems. A recently launched initiative, called the Systems Approach for Better Education Results (SABER), collects and shares comparative data on educational policies and institutions from countries around the world.

In the private sector, the ability to collect information about teacher-student interaction, and interaction between students and learning systems, can have a profound impact. In Kenya, for example, Bridge International Academies is using adaptive learning on a large scale. An IFC client founded by three American entrepreneurs, Bridge runs 259 nursery and primary schools, with monthly tuition averaging $6. It is a massive learning laboratory for students and educators alike.

Bridge tests different approaches to teaching standard skills and concepts by deploying two versions of a lesson at the same time in a large number of classrooms. The lessons are delivered by teachers from standardized, scripted plans, via tablets that also track how long the teachers spend on each lesson. Exam results are recorded on the teacher’s tablet, with more than 250,000 scores logged every 21 days. From these data, Bridge’s evaluation team determines which lesson is most effective and distributes that lesson throughout the rest of the Academy’s network.

We know that a host of issues can cause a student’s performance to decline – scorching summer heat in classrooms without air conditioning, problems at home, or poor-quality teachers, to name a few. But when one gathers results on a large scale, variables flatten out, and the important differences emerge. That is the great value of big data.

Another case is SABIS, a provider of K-12 education in the United States, Europe, Asia, the Middle East, and North Africa. SABIS mines large data sets to ensure high standards and enhance academic performance for more than 63,000 students. Continuous tracking of annual student academic performance yields more than 14 million data points that are used to shape instruction, achieve learning objectives, and ensure consistency across the company’s network of schools in 15 countries.

Knewton, an adaptive learning platform that personalizes digital courses using predictive analytics, is another company at the forefront of the data revolution. With tailored content and instruction, even classrooms without private-school resources can provide individual learning. As a result, teachers spend their time in the most effective way possible – solving problems with students – instead of delivering undifferentiated lessons.

These benefits do not come without risk. We are only beginning to grapple with how big data’s tremendous potential for learning can be harnessed while protecting students’ privacy. In some cases, data-collection technology is outpacing our ability to decide how it should be collected, stored, and shared. No matter how rigorously data are secured, there is still a need for a clear licensing structure for its use. In many developing countries, there are no regulations for data privacy at all.

The interface between data and education holds the promise of new educational products for improved learning, with large potential benefits, especially for the poor. To realize those benefits – and to do so responsibly – we must ensure that data collection is neither excessive nor inappropriate, and that it supports learning. The private sector, governments, and institutions such as the World Bank Group need to formulate rules for how critical information on student performance is gathered, shared, and used. Parents and students deserve no less.

Hide Comments Hide Comments Read Comments (5)

Please login or register to post a comment

  1. CommentedNathan Coppedge

    Of course, it would help if large-scale corporations did more to protect students' mental health. Students that can handle any degree of stress are over-represented in society. There is a hidden potential to make better use of classrooms through creativity and stimulation-based learning, which not only has the potential to reduce long-term stress, but also promotes those abilities which are most beneficial for generating ideas which might influence long-term economics. If physical tasks are to be done by robots, it is time that intelligence were measured not only in terms of calculus and statistics, but also in terms of personality, development, and especially creativity. Although the products sometimes look like pottery or screen-printing, this technical focus is really an extension of the calculus focus. It should be realized that some tasks are performed without measurable results, and it is these results which sometimes prove rewarding in the long-term. In my view, this new focus is important, and it involves not only a focus on perfectibiilty, but also a significant component of enjoyment. These two factors can be used to offset eachother, producing an environment more conducive to paying the student, and thus to getting more quality results. It is high time the global organizations recognized that human capital has a lot to do with who we really are, what we really want to do, and the products of sincerity and our own innate or learned sense of perfection and communicability. It is time to foster the growth that individuals want for themselves---sort of like how to design your own video game, or how to interpret and design new environments. This doesn't have to involve soft stuff, because it doesn't really involve more resources. It just involves a little appreciation of what resources can do, and what it means for human interface and environment. I think it also involves perfectibility, and catering to the preferences of the student. This doesn't seem like a gamble to me, it seems like the only way to go. There are always people rising to the occasion. Why not make it easier, and get a little bit of something from everyone? After all, the limitation on standard testing is a combination of what it permits the student to accomplish, and whether the student can take it to a further level. With sufficient information, the same can be achieved on a more remarkable level by asking something from everyone, and then simply observing who (perhaps more than one person) has results in the long-term. It is high time quality, human-interface oriented learning entered the classroom, which may perhaps be a workplace.

  2. CommentedLinda Boudreau

    Great post, thanks for sharing. There is a lot of great work being done with data analysis and data linkage tools for the future of education with P20 and SLDS initiatives. Linking K-12 data with college and career data will certainly have a positive, significant impact on student achievement.

    Linda Boudreau
    Data Ladder

      CommentedLinda Boudreau

      One more thought: a single comprehensive and timely education data standard that ensures usability is key to the success of these programs. States need to take advantage of the data analysis that is being made available, turning education data into actionable information that can transform student lives.

      Linda Boudreau

  3. CommentedVal Samonis

    Well, unprecedented interaction opportunities brought about the Internet is the crux of the new approach to education that actually mimics the real world as the real world becomes first of all digital world.

    Val Samonis
    Vilnius U

  4. CommentedEric Nordin

    One of the other main concerns regarding big data and education is that such data focuses almost exclusively on quantitative measures, which is fine in certain circumstances. Education, however, is as much of an art as a science – no two students learn the in same manner. Whatever knowledge we can glean from large data samples may fail to include the nuances and complexities of the learning process. Given that many pedagogical thermos are focusing on student-oriented learning as well as peer-to-peer learning, big data most likely will not be a panacea for issues regarding educating students in emerging economies, but only copy half-measured solutions applied in developed economies.