The Backlash Against Personalized Learning

Tuesday, December 11, 2018

Newsletter
Students at the Secondary School for Journalism in Brooklyn, New York walked out of class to protest it. Another New York City public school dumped it. And in Cheshire, Connecticut, the superintendent eliminated a “personalized learning” program after families complained that users received limited attention from teachers, gamed the system, faced data privacy violations, and experienced increased levels of anxiety.

These approaches rely on software to lead each student through lessons deemed appropriate for that student at that time, thus assisting or supplementing teachers who are feared to have a lesser capacity to individualize. “Individualized” instruction may be a better name for these approaches, but advocates have popularized the “personalized instruction” name, and we thus use it here.

All three of the above cases involved the Summit Learning Platform, which is currently used in more than 380 schools. Summit was built with assistance from Facebook engineers and promoted financial backing from company founder Mark Zuckerberg. As such, they are arguably impacted by the recent backlash against Facebook, which was sparked by revelations that the social media giant improperly shared data and permitted election meddling. (The National Education Policy Center deleted its Facebook account in March over these and other concerns.)

But is personalized learning more broadly facing a backlash?

Maybe. In October, for example, The New York Times ran a series of articles about efforts by affluent parents (including those in Silicon Valley) to limit students’ use of screens not only at home—where they are often used for entertainment—but at school. For example, the private Waldorf School of the Peninsula has attracted families of executives of tech companies such as eBay, Google, Apple and Yahoo with its computer-free approach.

In a policy brief for NEPC, Vanderbilt professor Noel Enyedy writes that “recent studies show little evidence for the effectiveness” of personalized learning programs that aim to use computers to tailor digital instruction to individual students. Such programs often merely translate problematic features of traditional learning into the digital context. For instance, Enyedy writes:

(T)he basic formula of both traditional and computerized instruction has been ‘I, we, you,’ where the teacher (or computer) tells the student something, followed by a worked-out example gone over together, and ending with independent student practice. Everything we know about teaching and learning tells us that this formula is flawed and not working.

Another challenge is that there’s no one standardized definition of, or approach to, personalized learning.

“The systems lumped together under the umbrella term of Personalized Instruction differ widely,” Enyedy writes:

In fact, there is so much variability in features and models for implementation that it is impossible to make reasonable claims about the efficacy of Personalized Instruction as a whole. Worse, when decision makers consider adopting a particular system, it is usually hard to tell whether available evidence applies to the specific system under consideration.

One major complaint about Summit Learning is that there is too much digital learning and not enough instructor intervention: One student told New York Magazine that she met with her math teacher for just a few minutes a month. Survey results suggest that teachers in schools that use personalized learning are less familiar with their students and their lives inside and outside of schools. Other complaints about Summit include:

Software glitches One student told The New York Post she was locked out of the software for two months.)
Concerns about data privacy (Summit shares data with 19 companies, including Amazon and Microsoft. An NEPC research brief notes that most school districts lack the resources to manage all their student data.)
Lessons that are easy to game and trick (By skipping lessons, for example, and making educated guesses on multiple choice questions.)
Prof. Enyedy’s brief concludes with a series of seven recommendations, including the following four:

Education policymakers should continue to invest in technology but should be wary of advocacy promoting computerized instruction to an extent that oversteps the current research.
Policymakers should encourage more partnerships among developers, educational researchers and teachers. Such partnerships have great potential to produce systematic and rigorous evidence of what works and what doesn’t.Administrators must ensure that investments in technological infrastructure and software licensing are accompanied by substantive professional development for teachers in order to provide them with skills that have not historically been in the teacher’s toolbox.
All stakeholders should refrain from assuming that Personalized Instruction is the only model for computers in the classroom and be open to investigating new models integrating technology into the learning process.