t is becoming increasingly clear that the education and career pathways of the future will be marked by a thorough, ongoing—and, one hopes, effective, equitable, and ethical—application of data to student and worker outcomes with a degree of refinement never before thought possible.
While it is important to acknowledge that the COVID-19 crisis has accentuated the need for these transformational developments, they arise out of longer-term trends already well underway. After all, there is nothing new, per se, in applying data to the management of students and workers. Taylorism transformed the shop floors of early 20th-century industry, while Progressive education reformers practiced quantitative “scientific” education reforms (giving us the IQ and SAT tests, a decidedly ambivalent legacy). All of these developments applied aggregated information rendered possible by the proliferation of social science survey methods to categorize patterns in large demographic sets and then apply those categories toward individual outcomes.
The past few decades have witnessed a growing dedication of data accumulation to helping inform, guide, track, and empower students and workers in their educational and career paths. This has been enhanced by the ongoing digital transformation of our education and workforce systems, which has facilitated the capacity to provide more richly textured and “real-time” assessment of individual trajectories. To take one typical example, many colleges today provide automated guidance for course selection to students as they move through the requisites for a particular degree program.
It is crucially important, of course, to ensure wholesale implementation of processes for identifying and correcting the biases that often inhere in such large social data sets and that, consequently, emerge in the algorithmic outputs derived from that data. There is clear evidence of the impacts of discriminatory conditioning of data-driven systems in criminal sentencing and hiring, for example, perpetuating deep-rooted systemic inequities. Nothing could be more critical as we adopt these tools in education than to guard against perpetuating or deepening systemic patterns of bias. Fortunately, data can and should also be used to identify and correct for biased outcomes through regular auditing.
The use of big data and machine learning is part of a broader trend that has been discussed throughout our modern economy, from personalized medicine to internet algorithms that shape, say, an individual’s Google search results. Education has been a lagging indicator of these trends compared to fast-adopting sectors such as tech, notwithstanding some visionary and ambitious educational innovation.
One of the more notable applications in 21st century classrooms has been the drive to implement personalized learning programs that allow students to move through curricula at their own pace. These initiatives have often reminded me of my own early 1980s elementary school experience as part of an Individually Guided Education district experiment. Personalized learning at its core hearkens back at least to John Dewey and Maria Montessori, or even Rousseau’s Emile. Today’s iterations are largely distinguished from such predecessors by the digital tools that can enhance teachers’ capacity to design and monitor divergent pathways simultaneously.
At JFF, we have been developing systems for empowering and assessing individual career pathways for more than a decade. Yet there is a growing sense that we are approaching the limits of the tools at our disposal without having achieved satisfactory solutions to many questions, including:
- How demonstrable are the pathways completed and maintained by individuals in our programs?
- How do we know individuals are benefitting from programs in the long-term?
- Can we ensure that the pathways are relevant and customized to individuals’ needs?
At the same time, we are on the cusp of an inflection point, made possible by the new capacities inherent in big data and algorithms to address these and other critical questions.
Key to this development, is that data sets available to us (e.g., career outcomes data that draws on resumes and job profiles) are massing exponentially, while sophisticated AI programs are improving our capacity to mine this big data for pattern recognition. Overall, this is driving what some have termed a shift from personalized to precision delivery. In medicine, for instance, the availability of DNA sequencing is making possible “precision” medicine that is specific to a patient’s genetic profile, such as selecting an individual’s chemotherapy treatment based upon the DNA profile of that patient’s cancer cells.
While the education and workforce sectors have been slower to adopt these methods, a promising application can be seen in a recent white paper from the Strada Institute for the Future of Work on “the new geography of skills,” where the skill needs of industry within micro-regions are used for tailoring the development of regional education and training programs. Still another intriguing step in this direction can be seen in the Burning Glass program designed to identify skills of workers displaced by the COVID-19 impact and match them to temporary jobs—lifeboat jobs—in sectors that are hiring amid the economic downturn. To this end, Burning Glass utilized their “database of more than a billion job postings and resumes worldwide.”
It is possible to envision going further by designing career guidance systems that empower each worker with a set of career pathways informed by multiple data vectors—including regional geographies of in-demand skills but also anticipated skills trends, local patterns of career advancement, and a rich repository of an individual worker’s lifelong trajectory of learning and credentialing. When these recommendations are complemented by support from a well-trained career counselor, students and workers will have both the data and individualized support to make decisions that support long-term career growth and align with their interests and needs. This will be increasingly possible, effective, and necessary in the new skills economy that will place greater premium upon learning over a worker’s lifetime—and the ability to stack and combine credentials.
Already, there are multiple start-ups rushing into the market to provide repositories of individuals’ lifetime of transcripts and other markers of qualifications. The application of blockchain, coupled with likely breakthroughs in quantum encryption, will soon be able to provide users with a high degree of confidence in the security of such personal files. The Education Blockchain Initiative announced by the American Council on Education in February of 2020, which would utilize blockchain technology to facilitate a secure ecosystem of centralized data files on each individual’s education and skills that would be accessible to approved schools and employers, is a promising step in this direction.
It may even be possible soon to add other dimensions to such profiles, such as multi-year portfolios of work or longitudinal profiles of learning outcomes unique to each person. A student who, for instance, joins a MOOC to refresh job skills might grant the online program confidential access to her profile so that the MOOC system algorithms could provide content delivery tailored specifically to that student’s unique set of learning strengths, modalities, and career path. Further, these recommendations could consider the high-demand opportunities with long-term growth potential within an individual’s local labor market to build stronger connections between supply and demand.
That, indeed, would be precision learning and workforce development of profound import, a consummation of sorts to the more than century of striving toward the optimal application of data monitoring to ensure individualized pathways of unprecedented precision and nuance.
Sometimes in history, quantitative capacities build up over an extended period of time until they reach a tipping point that ushers in a qualitative paradigm revolution. We are on the cusp of such a moment, driven by the advent of big data and AI capacities that enable the transition from personalization to precision in an array of sectors, including—transformationally—education and career pathways.
The sheer power of this emergent capability is breathtaking. The upside potential for offering individuals agency and efficacy in their learning and work journeys is considerable. The downside potential is also far from trivial, challenging us in the years ahead to ensure that these capacities are applied with a commitment to individual empowerment and equitable access and outcomes.
a global affairs media network
Moving toward Precision Learning in Postsecondary and Workforce Education
August 15, 2020
The COVID-19 crisis has accentuated the need to use data to refine education and career pathways for students and workers, but that need arises out of longer-term trends already well underway.
I
t is becoming increasingly clear that the education and career pathways of the future will be marked by a thorough, ongoing—and, one hopes, effective, equitable, and ethical—application of data to student and worker outcomes with a degree of refinement never before thought possible.
While it is important to acknowledge that the COVID-19 crisis has accentuated the need for these transformational developments, they arise out of longer-term trends already well underway. After all, there is nothing new, per se, in applying data to the management of students and workers. Taylorism transformed the shop floors of early 20th-century industry, while Progressive education reformers practiced quantitative “scientific” education reforms (giving us the IQ and SAT tests, a decidedly ambivalent legacy). All of these developments applied aggregated information rendered possible by the proliferation of social science survey methods to categorize patterns in large demographic sets and then apply those categories toward individual outcomes.
The past few decades have witnessed a growing dedication of data accumulation to helping inform, guide, track, and empower students and workers in their educational and career paths. This has been enhanced by the ongoing digital transformation of our education and workforce systems, which has facilitated the capacity to provide more richly textured and “real-time” assessment of individual trajectories. To take one typical example, many colleges today provide automated guidance for course selection to students as they move through the requisites for a particular degree program.
It is crucially important, of course, to ensure wholesale implementation of processes for identifying and correcting the biases that often inhere in such large social data sets and that, consequently, emerge in the algorithmic outputs derived from that data. There is clear evidence of the impacts of discriminatory conditioning of data-driven systems in criminal sentencing and hiring, for example, perpetuating deep-rooted systemic inequities. Nothing could be more critical as we adopt these tools in education than to guard against perpetuating or deepening systemic patterns of bias. Fortunately, data can and should also be used to identify and correct for biased outcomes through regular auditing.
The use of big data and machine learning is part of a broader trend that has been discussed throughout our modern economy, from personalized medicine to internet algorithms that shape, say, an individual’s Google search results. Education has been a lagging indicator of these trends compared to fast-adopting sectors such as tech, notwithstanding some visionary and ambitious educational innovation.
One of the more notable applications in 21st century classrooms has been the drive to implement personalized learning programs that allow students to move through curricula at their own pace. These initiatives have often reminded me of my own early 1980s elementary school experience as part of an Individually Guided Education district experiment. Personalized learning at its core hearkens back at least to John Dewey and Maria Montessori, or even Rousseau’s Emile. Today’s iterations are largely distinguished from such predecessors by the digital tools that can enhance teachers’ capacity to design and monitor divergent pathways simultaneously.
At JFF, we have been developing systems for empowering and assessing individual career pathways for more than a decade. Yet there is a growing sense that we are approaching the limits of the tools at our disposal without having achieved satisfactory solutions to many questions, including:
- How demonstrable are the pathways completed and maintained by individuals in our programs?
- How do we know individuals are benefitting from programs in the long-term?
- Can we ensure that the pathways are relevant and customized to individuals’ needs?
At the same time, we are on the cusp of an inflection point, made possible by the new capacities inherent in big data and algorithms to address these and other critical questions.
Key to this development, is that data sets available to us (e.g., career outcomes data that draws on resumes and job profiles) are massing exponentially, while sophisticated AI programs are improving our capacity to mine this big data for pattern recognition. Overall, this is driving what some have termed a shift from personalized to precision delivery. In medicine, for instance, the availability of DNA sequencing is making possible “precision” medicine that is specific to a patient’s genetic profile, such as selecting an individual’s chemotherapy treatment based upon the DNA profile of that patient’s cancer cells.
While the education and workforce sectors have been slower to adopt these methods, a promising application can be seen in a recent white paper from the Strada Institute for the Future of Work on “the new geography of skills,” where the skill needs of industry within micro-regions are used for tailoring the development of regional education and training programs. Still another intriguing step in this direction can be seen in the Burning Glass program designed to identify skills of workers displaced by the COVID-19 impact and match them to temporary jobs—lifeboat jobs—in sectors that are hiring amid the economic downturn. To this end, Burning Glass utilized their “database of more than a billion job postings and resumes worldwide.”
It is possible to envision going further by designing career guidance systems that empower each worker with a set of career pathways informed by multiple data vectors—including regional geographies of in-demand skills but also anticipated skills trends, local patterns of career advancement, and a rich repository of an individual worker’s lifelong trajectory of learning and credentialing. When these recommendations are complemented by support from a well-trained career counselor, students and workers will have both the data and individualized support to make decisions that support long-term career growth and align with their interests and needs. This will be increasingly possible, effective, and necessary in the new skills economy that will place greater premium upon learning over a worker’s lifetime—and the ability to stack and combine credentials.
Already, there are multiple start-ups rushing into the market to provide repositories of individuals’ lifetime of transcripts and other markers of qualifications. The application of blockchain, coupled with likely breakthroughs in quantum encryption, will soon be able to provide users with a high degree of confidence in the security of such personal files. The Education Blockchain Initiative announced by the American Council on Education in February of 2020, which would utilize blockchain technology to facilitate a secure ecosystem of centralized data files on each individual’s education and skills that would be accessible to approved schools and employers, is a promising step in this direction.
It may even be possible soon to add other dimensions to such profiles, such as multi-year portfolios of work or longitudinal profiles of learning outcomes unique to each person. A student who, for instance, joins a MOOC to refresh job skills might grant the online program confidential access to her profile so that the MOOC system algorithms could provide content delivery tailored specifically to that student’s unique set of learning strengths, modalities, and career path. Further, these recommendations could consider the high-demand opportunities with long-term growth potential within an individual’s local labor market to build stronger connections between supply and demand.
That, indeed, would be precision learning and workforce development of profound import, a consummation of sorts to the more than century of striving toward the optimal application of data monitoring to ensure individualized pathways of unprecedented precision and nuance.
Sometimes in history, quantitative capacities build up over an extended period of time until they reach a tipping point that ushers in a qualitative paradigm revolution. We are on the cusp of such a moment, driven by the advent of big data and AI capacities that enable the transition from personalization to precision in an array of sectors, including—transformationally—education and career pathways.
The sheer power of this emergent capability is breathtaking. The upside potential for offering individuals agency and efficacy in their learning and work journeys is considerable. The downside potential is also far from trivial, challenging us in the years ahead to ensure that these capacities are applied with a commitment to individual empowerment and equitable access and outcomes.
