Bill Gates and Mark Zuckerberg are backing it with hundreds of millions of dollars. States from Florida to Vermont have adopted supportive laws and policies. And school districts across the country are embracing this emerging education trend.
But as 鈥減ersonalized learning鈥 takes root, it鈥檚 also coming under greater scrutiny.
Leading researchers say their work does not support the most enthusiastic claims being made by personalized-learning supporters. Education experts are raising questions about implications for teaching and learning. Tech-industry critics are sounding alarms about Silicon Valley鈥檚 growing influence over public schools. And a small but vocal coalition of parents and activists from across the political spectrum deride the term 鈥減ersonalized learning鈥 as an Orwellian misnomer for replacing teachers with digital devices and data-mining software.
Such resistance is probably not surprising.
When any new educational strategy receives money and attention, questions arise. Like other efforts to improve U.S. schools, personalized learning is getting swept up in decades-long disagreements over how children learn, the proper role of teachers, and who gets to decide how public education is organized.
But personalized learning also faces some unique challenges. The biggest is lack of clarity around what the term actually means.
In general, personalized-learning models seek to adapt the pace of learning and the instructional strategies being used to best fit each individual child鈥檚 strengths, weaknesses, and interests. In the digital age, realizing these goals is often seen as dependent on technology鈥攖o help measure in real-time what each student knows; to develop 鈥渓earner profiles"; and to help match each child with customized learning experiences and 鈥減laylists.鈥
In practice, though, 鈥減ersonalized learning鈥 is used to describe everything from supplemental software programs to whole-school redesigns.
As a result, the term has become a blank slate on to which supporters and skeptics alike project their own hopes, fears, and beliefs.
To help K-12 educators and policymakers consider personalized learning from every angle, Education Week is taking a close look at the perspectives of critics like Tiffany Dunn, a veteran teacher in Kentucky鈥檚 101,000-student Jefferson County school system whose strongly held concerns broadly reflect the worries of other opponents.
鈥淭his whole thing is coming from the tech industry, which doesn鈥檛 understand that what kids need is someone to love them and get excited about them,鈥 Dunn said. 鈥淚鈥檓 not aware of any research that says sticking a child in front of a computer for hours on end does them any good.鈥
Do such arguments hold weight? How do proponents respond? What do third-party experts say?
What follows is a breakdown of the three main critiques leveled against personalized learning.
Argument#1: The Hype Outweighs the Research
No one has studied personalized learning more closely than the RAND Corporation.
And RAND is unambiguous about what its research shows.
鈥淭he evidence base is very weak at this point,鈥 said John F. Pane, a senior scientist and the group鈥檚 distinguished chair in education innovation.
Still, the hype around personalized learning has continued to grow. Take, for example, Facebook鈥檚 Mark Zuckerberg. Along with his wife, pediatrician Priscilla Chan, the founder of the giant social-networking company plans to use his Chan Zuckerberg Initiative to invest billions of dollars into personalized-learning efforts. Zuckerberg regularly articulates his goal of 鈥溾
Such lofty ambition is common throughout the ed-tech industry. Dozens of companies now tout a wide range of software, platforms, and apps as transformative tools for personalizing student learning.
69传媒 are buying in: 97 percent of district leaders surveyed by the Education Week Research Center last year indicated that their districts had invested in some form of personalized learning.
Proponents say it would be a mistake to dismiss all those efforts as misguided, or not based on any evidence at all.
For one thing, RAND has found early signs that some comprehensive, whole-school personalized learning models yield modest student achievement gains.
And at its heart, said Richard Culatta, the CEO of the International Society for Technology in Education, personalized learning is about giving students control over their own learning; differentiating instruction for each child; and providing real-time feedback. supports each of those strategies.
What鈥檚 new about contemporary versions of personalized learning, Culatta maintained, is the use of technology to bring those individual approaches together into a coherent whole. What鈥檚 happening now should be understood as research-and-development, with many experiments taking place at once, Culatta said.
How can K-12 educators and policymakers make sense of such a confusing, evolving landscape?
Louis Gomez has some ideas.
An education professor at the University of California, Los Angeles, Gomez studies the ways technology initiatives play out inside school systems.
The reality, he said, is that many schools purchase off-the-shelf software and call it 鈥減ersonalized learning,鈥 without being able to say what is supposed to change in the classroom. And even when schools do take a broader view, they often fail to recognize that success depends largely on decisions that educators and administrators make on the ground.
That鈥檚 in part because many versions of personalized learning seek to change or replace fundamental processes that shape the day-to-day life of schools鈥攅verything from how teachers prepare lessons to how students are graded.
Inevitably, Gomez said, that causes tension.
RAND鈥檚 research, for example, has consistently found that even in the best-supported personalized-learning schools, teachers frequently say there鈥檚 not enough time to truly tailor the learning experience to each child.
Does that mean that schools should ditch the personalized-learning experiments?
No, Gomez, Culatta, and the RAND researchers agree. There鈥檚 real reason to be excited.
But it鈥檚 OK to wait for more and better evidence, they said. If you do plunge in, be sure the initial stakes for failure are low, and build in opportunities to learn.
鈥淚 would not advise schools to dump massive resources into going fully into personalized learning,鈥 said Laura S. Hamilton, the associate director of RAND Education. 鈥淓xperiment with some new approaches that might be a good fit for your particular school or district, but monitor it very closely.鈥
Argument #2: Personalized Learning is Bad for Teachers and 69传媒
Every student is a unique combination of individual skills, abilities, and preferences, writes Todd Rose in , which has become a kind of bible for the personalized-learning movement. Therefore, trying to peg each child鈥檚 learning experiences to some notion of what is typical for most children is both misguided and harmful.
Instead, many personalized-learning proponents suggest, schools should leverage technology to expand the benefits of good one-to-one human tutoring. By pursuing that vision, personalized learning can help children develop the skills, purpose, and passion to not only learn content and skills, but to think critically and build relationships, said Diane Tavenner, the CEO of California鈥檚 Summit Public 69传媒 charter network.
鈥淗ere鈥檚 what I consistently hear,鈥 said Tavenner, describing her interactions with the thousands of educators across the country receiving training on Summit鈥檚 technology platform and instructional model (which were developed with support from Facebook and the Chan Zuckerberg Initiative.) 鈥淭hey say, 鈥楾his lets me be the teacher I鈥檝e always wanted to be.鈥欌
But many experts, from across the political and pedagogical spectrum, aren鈥檛 sure they buy it.
The have come from progressives like Alfie Kohn, the author of Schooling Beyond Measure.
鈥淚t鈥檚 behaviorism on a screen,鈥 Kohn said of personalized learning.
Here鈥檚 what he means:
First, many critics believe that personalized learning boils down to kids working alone on software, an approach they say ignores the crucial social aspects of learning and reduces teachers to the role of 鈥渇acilitators.鈥
Because so many different approaches get labeled as 鈥減ersonalized learning,鈥 it鈥檚 hard to say how often that really happens. In the schools RAND is studying, the researchers describe such practices as relatively rare.
But in August, the Silicon 69传媒 Fund, which provides millions of dollars to leading personalized-learning networks, released a learned during five years supporting the field. Among the concerns were 鈥渟chools that were often very quiet,鈥 because 鈥渟tudents were head-down, working on their computers at their exact level.鈥 Now, the fund says, most of the groups it supports are trying to reduce the time students spend alone on computers to 20 to 40 percent of the school day.
On a deeper level, Kohn and some other progressive educators believe that real learning happens when students are driven by their intrinsic curiosity to pursue answers to their own questions about the world. By contrast, Kohn said, much of what鈥檚 marketed as 鈥減ersonalized learning鈥 amounts to little more than breaking knowledge and ideas down into 鈥渋tty-bitty parts,鈥 then using extrinsic rewards to 鈥渕arch kids through a series of decontextualized skills they had no meaningful role in choosing.鈥
Some conservative education-policy experts raise similar concerns.
Michael Petrilli, for example, is the head of the Thomas B. Fordham Institute. Like Kohn, he worries that some versions of personalized learning encourage a 鈥溾 that 鈥渂reaks learning into little bits and scraps and bytes of disparate skills, disconnected from an inspiring, coherent whole.鈥 That tendency is particularly troublesome in the early grades and in subjects other than math, he said.
But while Kohn worries that personalized learning doesn鈥檛 give students enough control over their learning, Petrilli fears the opposite.
It鈥檚 already hard enough to ensure that all students are held to challenging standards and offered rigorous curriculum, he said. Personalized learning鈥檚 emphasis on offering different content and paths to each student could open the door for watered-down expectations, greater inequities, and more difficulty holding schools accountable.
Tavenner, of Summit Public 69传媒, described both critiques as 鈥渓egitimate"鈥攂ut only in those cases where personalized learning is solely about helping students efficiently master discrete information and skills.
That鈥檚 not the case at Summit, Tavenner maintained. She said the strongest personalized-learning models offer the best of what both conservatives and progressives want: high-quality standards and content for students, with opportunities to apply that knowledge via self-directed projects, all supplemented by human mentors and technology tools that help students keep track of their own learning.
Do such claims hold up to outside scrutiny?
There鈥檚 not yet any definitive answer. But Stanford University professor Larry Cuban, a long-time ed-tech skeptic, recently spent a year observing classrooms at Summit and other personalized-learning schools in Silicon Valley. Cuban said he was surprised to find 鈥渕eaningful, incremental improvements in how teachers organize and teach a lesson"鈥攂ut not the kind of revolutionary new model of education that Tavenner describes.
Overall, the personalized-learning field is still marked by significant unresolved pedagogical tensions, said Benjamin Riley, the executive director of the nonprofit Deans for Impact, which seeks to improve teacher preparation. Among the biggest: the appropriate role for software in the classroom, how much autonomy is best for student learning, and the challenge of maintaining high standards and social interaction when every student is pursuing his or her own path.
Too often, Riley said, proponents gloss over such concerns.
鈥淭here are tradeoffs here, and we at least need to acknowledge them,鈥 he said.
Argument #3: Big Tech + Big Data = Big Problems
Finally, many critics are worried that 鈥減ersonalized learning鈥 is cover for an aggressive push by the tech industry to turn K-12 education into a giant data-mining enterprise.
For parents and activists like Karen Effrem, that belief provokes intense fear and anger. 鈥淲e鈥檙e sacrificing our children鈥檚 privacy, and we鈥檙e allowing corporations to make potentially life-changing decisions about our kids, all for technology that doesn鈥檛 actually help them,鈥 said Effrem, the president of Education Liberty Watch, an advocacy organization that supports parents鈥 right to control their children鈥檚 education.
Evaluating such critiques can be difficult. The proprietary technical engines that drive the personalized-learning movement are largely black boxes, inaccessible to public inspection.
What is clear, though, is student data are the fuel that makes many of those systems go.
Just listen to Bharat Mediratta, a former Google engineer who helped launch personalized-learning pioneer AltSchool, which aims to harvest information about students via everything from classroom cameras to computer keystrokes.
鈥淔irst, we need to generate the big data,鈥 Mediratta told Education Week last year. 鈥淭hen, we start figuring out how to use it to transform education.鈥
Writ large, that philosophy has prompted worry about the sheer volume of information collected from children in school, especially amid the recent push to better understand students鈥 behaviors, feelings, and mindsets.
Security and privacy are also front-of-mind: Just this spring, for example, popular digital-learning platform Edmodo was hacked, resulting in the personal information of an estimated 77 million users being put up for sale on an unregulated part of the internet.
And many critics worry that algorithms are increasingly being used to make key decisions shaping children鈥檚 futures, without any real way for students and parents to understand how those choices are made, or challenge them for possible errors or biases.
Personalized-learning proponents generally acknowledge such fears as legitimate, suggesting they can be solved through dialogue and better regulation.
鈥淎s we in the tech world get closer to having a better picture of each student at any given point in time, all of these discussions about data and privacy need to continue,鈥 said Kristen DiCerbo, the vice president of education research at global publishing giant Pearson. 鈥淚t鈥檚 not too late.鈥
But some critics say that stance presumes that more data-driven personalization in education is both desirable and inevitable, so long as the wrinkles can be ironed out.
K-12 educators and policymakers should question the forces driving that narrative, said Audrey Watters, an independent researcher who maintains the popular Hack Education blog.
The push for more personalized learning isn鈥檛 happening in a vacuum, Watters maintained. The movement is emerging from the values and technologies that already define Silicon Valley. And it鈥檚 clear how the companies behind those systems operate.
鈥淲hen Facebook promises personalization,鈥 Watters said, 鈥渋t鈥檚 really about massive data collection.鈥
In the consumer sector, critics say, the downsides of that approach are only now becoming evident, after the technologies have already been deployed at enormous scale.
Facebook, for example, has recently come under fire for its role in creating 鈥,鈥 in which users only see information determined by algorithms to fit with their existing preferences. Critics also decry the ways the company has reportedly sought to identify and manipulate users鈥 emotions, including a that it had told some advertisers it could identify when teenagers feel 鈥渋nsecure鈥 and 鈥渨orthless.鈥
And Facebook is one of several tech giants currently in hot water over the way its platform was in the run-up to the recent presidential election. Mounting evidence suggests foreign actors targeted unwitting Americans with personalized ads and misinformation designed to influence their voting behavior鈥攁 problem made possible by the extensive data that Facebook collects from users, and by the lack of public visibility into the customized information that each user receives. In late September, Zuckerberg, the company鈥檚 CEO, for 鈥渢he ways my work was used to divide people rather than bring us together.鈥
Is it crazy to worry there could be similar unintended, unanticipated outcomes if personalized learning systems spread rapidly throughout K-12?
There are key differences between the consumer side of Silicon Valley and the work happening in education, according to Jim Shelton, the president of the Chan Zuckerberg Initiative鈥檚 education division. Most groups trying to bring personalized-learning technologies to public schools aren鈥檛 seeking exponential annual growth, Shelton said.
While CZI has given millions of dollars to AltSchool and its big-data-driven approach, the initiative is also supporting models like Summit, which says it avoids algorithmic recommendations and targeting in its platform. And as leaders like Zuckerberg bring their technologies, skills, and experiences into the K-12 arena, Shelton said, they face the burden of demonstrating that the products and approaches they support actually help children learn and grow.
Facebook and CZI 鈥渕ay be related, but they are not the same,鈥 Shelton said. 鈥淏eing deliberate, being humble, and taking a posture of learning are all ways we鈥檙e trying to be responsible as we do this work.鈥
Still, educators and the public would be foolish to not carefully consider the tech sector鈥檚 influence in public schools, especially given its recent stumbles, said Ben Williamson, a lecturer at the University of Stirling in the United Kingdom who studies big data in education.
Billions of public dollars are at stake, he said. So are big questions about the fundamental nature of schooling: How do we believe children learn? Who should decide what students need to know and get to experience? How will we determine what they鈥檝e learned?
鈥淲e need to open up a bigger debate about whether we really want Silicon Valley establishing this new model of data-driven schooling,鈥 Williamson said. 鈥淭hese are people whose vision for reforming public education puts their own industry in charge.鈥