YouTube U. Beats YouSnooze U.

Randy Lyhus for The Chronicle

October 31, 2010

Every day during the academic year, tens of thousands of students across the country sit passively in 300-person lecture halls listening to 90-minute lectures on freshman or sophomore-level calculus, chemistry, or biology (and this isn't even counting the students who have decided to punt the lecture altogether). Some students take notes to keep up. Most are lost or bored or both, trying their best to stay awake. Professors stare at a sea of blank faces while delivering a lecture not much different from the ones they have delivered in each of the past 10 years.

Students go back to their dorms to work on problem sets in a vacuum. They fight through 1,000-page, 10-pound tomes to get at the nuggets of information they really need or can comprehend. Many give up and copy from their peers.

This cycle continues for several weeks, until just before the midterm or final exam, when students cram everything they should have learned into one or two sleepless nights. Regardless of whether they can prove proficiency in 70, 80, or 90 percent of the material, they are "passed" to the next class, which builds on 100 percent of what they should have learned. Fast-forward six months, and students are lucky to retain even 10 percent of what was "covered."

IN THE RIGHT COLUMN: More on Online Learning
BROWSE THE FULL ISSUE: News, Commentary, and Data
BUY A COPY: Digital and Print Editions at the Chronicle Store

This is a grand exercise in labeling and filtering students with arbitrary grades rather than teaching them. It is a hugely inefficient use of time and resources, but no one wants to notice, because it is the way things have always been done.

Perhaps the worst artifact of this system is that most students end up mastering nothing. What is the 5 percent that even the A student, with a 95 percent, doesn't know? The question becomes scarier when considering the B or C student. How can they even hope to understand 100 percent of a more advanced class? Is there any point in studying differential equations if you don't have an intuitive understanding of basic calculus? Is there any point to taking biochemistry when you have less than perfect understanding of first-year biology and chemistry?

There are some college experiences that don't fit this mold. Many seminars and advanced courses are based on hands-on projects and small-scale discussions with professors. Those are undoubtedly valuable. But core classes tend not to be taught that way. The very classes that should establish a student's base understanding of a subject are taught like assembly lines—lecture, problem set, exam—with no quality control. Sure, the product's quality is graded, but nothing is done about defective understanding as the student is pushed down the line.

It's not the professors' fault. Many are excellent teachers and do amazing things within the framework they have to work with. The problem is the system itself. The best calculus teacher on the planet will make little headway with even innately intelligent students if they have weaknesses in algebra. The current system does not allow for addressing those gaps, so both professor and students settle for superficial coverage of the material. Students don't retain anything because they didn't intuitively understand it to begin with.

Apologists for the lack of retention and intuition by students argue that what­ really matters is that they are being taught "how to think." The reality is that because students have ever larger gaps in their knowledge as they progress, they learn to get by through pattern matching and memorizing. They learn to fake understanding, not to think.

In my former career, in the hedge-fund world, I encountered summa cum laude graduates of Ivy League colleges—some were CFO's or CEO's of large enterprises—who couldn't see or solve an eighth-grade algebra problem (for instance, realizing that one can figure out marginal gross margin given the revenue and the total gross margin in two different periods). I have interviewed economics majors from those same elite universities who could confidently repeat words or thoughts they've heard from professors or pundits, but could not explain how the Federal Reserve actually operates or how to analytically decide whether it makes more sense to buy or rent a home. I'm sorry, but these are superbright students, with scores and grades in the 99.9th percentile, who simply weren't taught to think.

There was nothing practical that anyone could do about this broken "learning" model until recently. But we can now deliver on-demand content to any student for nearly zero incremental cost. The video content can be paused and repeated as needed. Content producers can get real-time data on use, including student attention and efficacy. Students can focus on exactly what they need to know. They don't have to be embarrassed to fill in remedial gaps. They don't need to take notes. Crucially, the lectures can be given by superb communicators, with a deep, intuitive understanding of the material.

Why aren't students watching lectures on their own, at their own pace, in their dorms? Why aren't we using the 300-person gathering at 10 a.m. every Tuesday and Thursday as an opportunity for active peer-to-peer instruction rather than a passive, one-size-fits-all lecture?

The only argument I've seen made in favor of the 30- to 300-person lecture today is that it gives students an opportunity to ask questions of a live professor. The reality, however, is that the great majority of students are too self-conscious or considerate of their peers' time to raise their hands. Those who are bold enough to do so are often more likely to derail the professor's pacing than to ask something relevant to more than 5 or 10 percent of the students in the room. With on-demand video, however, the very act of watching a 10-minute video is equivalent to the student's saying, "I don't understand this, tell me more about it" without fear of being judged by peers or the professor, or of wasting everyone's time. Being able to directly learn from the professor actually becomes much more likely when the lecture happens on the student's own time, with on-demand video. Then the professor is freed to be an active participant in an interactive, peer-to-peer problem-solving powwow in the classroom.

Textbooks are even more archaic. The publisher or author has no data on which parts of a book are being used and when. They have no idea whether the changes they make in successive editions make the books more effective. Students get no real feedback on the exercises they do (the solutions to some problems in the back of the book are not real feedback). It is obvious to anyone but the most naïve observers that the focus of the multibillion-dollar textbook industry is on making each new version heavier and more expensive, while making previous editions obsolete with strategically renumbered pages and revised problems, in order to extract maximum revenue from students. Calculus hasn't really changed since Newton and Leibniz, but the powers that be have us believing that a new version of their freshman calculus book is necessary every three to five years.

My not-for-profit Khan Academy, which has recently gotten support from Google and the Bill & Melinda Gates Foundation, has a free, self-paced application that generates exercises for students dynamically. It is being developed as an open-source project and is already used by several tens of thousands of students. Over the next two years, we intend to have unlimited exercises covering every major math concept through calculus, and then to continue even further.

We collect data on when a student does a problem, how long it takes, and what happens before and after a video is viewed. And we can present this information in real time to the student, professor, parent, or administrator. We provide immediate feedback on proficiency and give step-by-step explanations of every problem. Most important, students continue to get exercises until they correctly answer 10 questions in a row (not once they have answered 70 or 80 or 90 percent of the questions correctly).

We are developing other applications to create repositories of teacher- and student-generated questions (on which we would collect the same metadata and ratings on quality, difficulty, and importance). Between these apps and our 2,000-plus and growing on-demand video library, which is being used by more than a million students a month, there is a genuine opportunity for educational institutions to rethink the system so that it is both more effective and more economical. We have decided to do this as a not-for-profit venture, so that our goal of optimizing learning never conflicts with profit maximization (which leads to the type of behavior we see in publishers).

The college experience is a valuable one from the perspective of students being immersed in a (hopefully) intellectual residential campus, but the paradigm of lecture, homework, and snapshot-assessment is broken. Ten years from today, students will be learning at their own pace, with all relevant data being collected on how to optimize their learning and the content itself. Grades and transcripts will be replaced with real-time reports and analytics on what a student actually knows and doesn't know.

The classroom will be a place for active interaction, not passive listening and daydreaming. The role of the teacher will be that of a mentor or coach as opposed to a lecturer, test writer, and grader. The institutions that will remain relevant will be those that leverage this paradigm, not fight it.

This is not just another prediction. It is merely a conservative extrapolation of what is already happening. The technology exists, and it is free and open-source. It is already being used regularly by a larger population of students than on any physical campus in the world—and that population is growing severalfold a year. When it is complemented with the human experience that can be had only on a physical campus with leading researchers and thinkers, we should see a new renaissance in learning.

Salman Khan is executive director of the Khan Academy (