Recently I wandered around the South by Southwest ed-tech conference, listening to excited chatter about how digital technology would revolutionize learning. I think valuable change is coming, but I was struck by the lack of discussion about what I see as a key problem: Almost no one who is involved in creating learning materials or large-scale educational experiences relies on the evidence from learning science.
We are missing a job category: Where are our talented, creative, user-centric “learning engineers” — professionals who understand the research about learning, test it, and apply it to help more students learn more effectively?
Jobs are becoming more and more cognitively complex, while simpler work is disappearing. (Even that old standby, cab driving, may one day be at risk from driverless cars from Google!) Our learning environments need to do a better job of helping more people of all ages master the complex skills now needed in many occupations.
I am not suggesting that all subject-matter experts (meaning faculty members) need to become learning engineers, although some might. However, students and faculty members alike would benefit from increased collaboration between faculty members and learning experts — specialists who would respect each other’s expertise — rather than relying on a single craftsman in the classroom, which is often the case in higher education today.
Education technology has enormous potential to help. While often expensive upfront, it has the chance to make learning more affordable, reliable, available, data-rich, and personalized. The technology within new learning environments — for example, an interactive simulation offered as part of a well-designed MOOC — is available 24/7, and can provide patient, repeated, and varied practice with supportive feedback that does not embarrass learners.
In the future, these environments may follow learners across their life spans, filling gaps from their past while allowing faculty members to provide the coaching, feedback, and motivation that is possible only with human interaction.
Unfortunately, technology has only a chance to help — there is no guarantee. While we hope that only the best instructors are engaged with technology, imagine your worst college professor. In the old days, that person damaged just a few hundred students per year. Thanks to video on demand and other wonders of technology, today that person might damage a few hundred thousand students — a weapon of mass destruction. Not exactly a win for technology and learning.
Technology is not the problem. As Richard E. Clark suggested in his book Learning From Media: Arguments, Analysis, and Evidence, education technology serves only as a delivery vehicle. All technologies can deliver effective or ineffective instruction. The key question is what you ask students to do and how you help them do it, not what tools you use.
After decades of experimental work by cognitive scientists and others, we now know a lot about how people learn. Neurons do not follow Moore’s law, the prediction by Gordon Moore in the 1960s that semiconductors would double in capacity every two years. Since our brains’ cognitive machinery does not change year after year, the good news is that investing in learning science will have long-lasting benefits.
Science, however, is not enough. It’s never enough for real-world problems.
Consider the tens of thousands of chemical engineers working in the United States. Anyone building a modern pharmaceutical factory needs them. You trust them to get the safety and regulatory issues right, and to use modern chemistry.
Indeed, most of the design processes leading to the conveniences of modern life benefit deeply from mediation between science and its application to real-world problems. Physicians, too, can be seen as “engineers” who use their knowledge of human biological science to tackle various medical problems within the constraints of medical care, economics, regulations, and other factors.
So where are the learning engineers? The sad truth is, we don’t have an equivalent corps of professionals who are applying learning science at our colleges, schools, and other institutions of learning. There are plenty of hard-working, well-meaning professionals out there, but most of them are essentially using their intuition and personal experience with learning rather than applying existing science and generating data to help more students and professors succeed.
Not applying learning science leads us into trouble:
- We make assumptions about learning that don’t match the facts. For example, we talk about the need to understand various “learning styles,” yet meta-analyses over decades show no practical benefits from bucketing minds into style categories, compared with well-designed single instruction.
- Students, faculty members, and administrators seem reluctant to question educational suppliers (of software, textbooks, and other materials) who do not deliver good evidence that their products or services solve learning problems.
- Colleges rarely run controlled trials, commonly used in medicine, to compare one approach to learning with another. Sometimes there are ethical concerns with such an approach: If you think a particular teaching method is good, it would be wrong to withhold it, and if it’s not good, it would be wrong to use it widely. Yet many other fields recognize that a promising discovery does not necessarily lead to large-scale benefits — you need to test assumptions. Oddly, in higher education it is unremarkable to change a course with no evidence (by adding a new reading list or teaching practice, for example), while experimenting with a group of courses to test an idea seems controversial. Kaplan University, where I work, runs dozens of controlled trials to make sure we know if an approach or intervention makes a difference before we adopt it.
- We don’t do a good job measuring what students learn. For example, a chemistry professor creating a test problem about Boyle’s law (the mathematical connection between pressure, volume, and temperature for gases) may, without realizing it, formulate an item that tests reading ability more than comprehension of the concept.
So what are we to do? To get started, several recent books provide very approachable syntheses of learning science: E-Learning and the Science of Instruction, by Ruth C. Clark and Richard E. Mayer; Why Don’t Students Like School?, by Daniel T. Willingham; and Talent Is Overrated: What Really Separates World-Class Performers From Everybody Else, by Geoff Colvin. I’ll add to the list a volume that I wrote with Frederick M. Hess, Breakthrough Leadership in the Digital Age: Using Learning Science to Reboot Schooling.
Just being exposed to information is never enough. To learn, instructional-design and teaching professionals need the same things their students do: We have to provide explicit practice and coaching on applying the science about learning for everyone involved in instruction. At Kaplan Inc., we have developed a training program for our more than 100 instructional designers, to help them apply learning science to solve practical learning problems. They might, for example, decide against using a fancy 3-D video game to teach a particular concept once they see research that found that a simpler tool makes the point more effectively.
It is not simple to go from reading the science to putting it to work, day in and day out.
We also need decision makers in higher education — especially those who buy learning materials and educational-technology offerings — to ask harder questions. For example: What learning science underpins this offering? Is there learning science behind a particular professional-development activity as well? Do you have valid and reliable data showing that a new product works better than what we’re using? Will you conduct a pilot program to demonstrate that it works better? How are you using data to improve the learner and staff experience?
New technologies offer a real opportunity to revolutionize learning. The capacity for efficient, accessible, reliable delivery of learning and the generation of more data about learning than we’ve ever had before are huge assets. However, the challenge is to use these technologies correctly.
Whether in the classroom, at home, or at work, we owe it to learners, employers, and families to do a better job at “learning engineering” than we’ve done so far.
