by Salman Khan
Such idylls aside, the hard truth is that in terms of learning, the great majority of summer hours are wasted. Kids watch TV or play video games while waiting for their parents to come home from work. Some kids read books; most don’t. As for actual academic study, how could that possibly happen? Last year’s textbooks have been given back. Teachers are away. Feedback is unavailable. Buildings are locked. Brains are in suspended animation.
How, then, should the school of the future approach the question of summer vacation?
My preferred scenario would be to trade it for a perpetual school experience where vacations can be taken whenever there is a need for one—not too different from what happens in companies. If students are working in multiage groups, all at their own pace, there is no longer an artificial stopping point when you transition to the “next” grade. If your family wants to travel to Europe or you have people coming over for the holidays or you want to start a business, it’s not a big deal. Just take the time off. You can’t “miss” class, because you’re working at your own pace. Even better, you can still do a lot of learning while on the road now that you have access to self-paced videos and exercises. The same flexibility would apply to teachers. Because of the multiteacher environment, teachers could stagger vacations during the year. No one would be asked to give up a restorative break or time for travel, but these would happen without the need of shutting down the entire system.
But okay, I’m a pragmatist and I realize that summer vacation—one of the most sacred of the education establishment’s sacred cows—probably isn’t going away anytime soon in the majority of schools. Fortunately, computer-based, self-paced mastery learning can solve many of the problems that summer vacation creates.
First of all, Internet-based lessons such as those offered by Khan Academy are always available. The Internet doesn’t close for the summer! Motivated kids can continue advancing and reviewing. Their minds stay active, their neurons keep firing.
This leaves the question of teacher help and feedback. As we’ve seen in the discussion of the Los Altos pilot program, the Academy has, with the help of experienced classroom teachers, developed a sophisticated feedback dashboard to give teachers real-time information regarding students’ progress and difficulties. The dashboard doesn’t need a school building in order to be accessed; there is no reason why teachers couldn’t continue to monitor student work and serve as online tutors during the summer. This would basically be an updated version of “summer school,” though with far lower cost than the current version—and with far greater mobility available to both students and teachers.
The Future of Transcripts
Given that we live in a competitive and increasingly interconnected world, and given that there will always be more applicants than slots available for the finest schools, how do we decide who gets to go to Harvard or Oxford or Heidelberg, or, for that matter, the leading universities in Taipei, Bologna, or São Paulo?
Given that there are not enough resources available to give every person extensive postgraduate training in his or her first choice of career, how do we decide who gets to be a doctor or an architect or an engineer?
Given that the most desirable job openings will always be met by a multiplicity of candidates, how do we decide who gets the gig or the promotion? Who should become the leader whose skill and character will in turn affect the livelihoods and morale of many other people?
These are bedeviling questions. They’ve always been bedeviling questions, and have become even more so as school applications are less and less limited by national borders and as corporations scour the entire globe for the best minds, the most creative thinkers, and the most motivated workers. How do you compare one applicant to another when they grew up in different cultures, speaking different languages, with widely divergent economic standing and, in turn, the various opportunities or lack thereof that go with wealth or poverty? How can you settle the question of which academic or personal criteria really matter as predictors of success? In the name of fairness and also practicality, how can you be confident that you’re comparing apples to apples?
Conventional education has done a woefully inadequate job of even asking these questions, let alone answering them.
How do conventional schools appraise their students? The first way, of course, is by letter grades. Could anything be less precise, less meaningful, or more capricious? As everybody knows, all schools have “easy markers” and “hard markers.” If standards can vary widely on either side of a hallway or a row of lockers, how much less uniform will they be from state to state or nation to nation? Yet letter grades are where the rankings start. Combined into that serious- and objective-sounding statistic called the grade point average, letter grades take on a seeming legitimacy and a determinative power well in excess of their reliability. If the individual grades are hazy and subjective, why should we imagine that their composite is precise and scientific? GPA is at best a blunt instrument. True, it can provide a general idea of whether a kid showed up, engaged with school, and played the game. But it is sheer blindness and folly to imagine that GPA alone tells you much about a student’s intelligence or creativity. Does someone with a 3.6 necessarily have more to offer to the world than someone with a 3.2? I wouldn’t bet on it.
Then there are the standardized tests to which students are subjected from third grade straight on through to grad school. As I’ve said, I am not antitesting; I believe that well-conceived, well-designed, and fairly administered tests constitute one of our few real sources of reliable and relatively objective data regarding students’ preparedness. But note that I say preparedness, not potential. Well-designed tests can give a pretty solid idea of what a student has learned, but only a very approximate picture of what she can learn. To put it in a slightly different way, tests tend to measure quantities of information (and sometimes knowledge) rather than quality of minds—not to mention character. Besides, for all their attempts to appear precise and comprehensive, test scores seldom identify truly notable ability. If you’re the admissions director at Caltech or in charge of hiring engineers at Apple, you’re going to see a heck of a lot of candidates who had perfect scores on their math SATs. They are all going to be fairly smart people, but the scores tell you little about who is truly unique.
Tacitly acknowledging the inadequacy of grades and testing as measures of ability or worthiness, many schools and employers also use extracurricular activities, third-party recommendations, and applicant-written essays as part of the selection process. In principle, this is a good thing, as it moves beyond snapshot moments and seeks insight into applicants as flesh-and-blood individuals. The obvious problem, though, is that the game is rigged in favor of those who understand how to work the system. These tend to be families that are already educated, well connected, or wealthy. Children of doctors, professors, and engineers have access to people whom they can do research under. Students who have parents, siblings, or cousins who have attended selective education programs get coaching on how to optimize their chances. A kid whose family friends include CEOs and legislators will tend to have more articulate and impressive recommenders than a kid who comes from a blue-collar family. Does any of this say anything significant about the applicant himself? Even on the so-called personal essays, students from wealthy or particularly ambitious families sometimes have help from well-paid counselors and consultants… who give them tips on how to sound sincere! Good luck to the overworked admissions officer who has to thread her way between the heartfelt and the cleverly bogus.
How then, in my thought experiment school of the future, would I appraise both the performance and the potential of my students?
First, I would eliminate letter grades altogether. In a system based on mastery learning, there is no need and no place for them. Students advance only when they demonstrate clear proficiency with a concept, as measured either by the ten-in-a-row heuristic or some future refinement of it. Since no one is pushed ahead (or left behind) until proficiency
is reached, the only possible grade would be an A. To paraphrase Garrison Keillor, all the kids would be way above average, so grades would be pointless.
In pursuit of the elusive ideal of comparing apples to apples, I would keep some version of standardized testing, though with several significant changes. I would alter the content of the tests from year to year far more than is currently done, include richer tasks, and attempt to incorporate an open-ended design component; this would limit the appeal of test-prep factories and in turn lessen the unfair advantage held by kids from wealthy families. The emphasis of the exam would also not be a one-time snapshot, but something that could and should be retaken after refining one’s skills (more affluent students already treat the SAT this way). And in recognition of the hard truth that standardized tests will never be perfect, I would put far less emphasis on them than is currently the case.
Instead, I would propose, as the centerpieces of student appraisal, two things: a running, multiyear narrative not only of what a student has learned but how she learned it; and a portfolio of a student’s creative work.
As we saw in the discussion of the Los Altos pilot program, readily available technology gives us the ability to track students’ progress, work habits, and problem-solving methods in unprecedented detail. The software needed to do so can be customized to the particular needs of any school, and is becoming more sophisticated all the time. The simplest part of the available feedback is quantitative: How far did a student go in math? How many concepts did he master in a given length of time? Is she above or below the median level for her age?
While this information is important, the far more interesting element of the feedback is qualitative. This is where tremendous progress remains to be made—a very exciting prospect for the very near future. Aside from counting concepts and measuring time, what can we infer from a student’s efforts at Khan Academy or some other version of computer-based education? What can we learn about his or her work ethic, persistence, resilience—elements of character that are at least as important as sheer intelligence as predictors of success? Johnny gets stuck. Does he flee the frustration by putting in less time, or does he dig in and work harder until he’s figured it out? Sally goes through a patch where her progress is labored and slow. Does she bounce back from it or give in to discouragement and waning confidence? As a seventh grader, Mo seems disengaged and puts in very little time at lessons. As a ninth grader he’s spending hours on biology; what does this say about his growing maturity and his possible gift for a particular field?
Clearly, this sort of information, if carefully interpreted, gives us a far more three-dimensional picture of a student than does a bunch of letter grades and numerical scores; it gives us a picture not just of a test-taker, but a learner.
I can also envision a category of data that would track a characteristic currently altogether ignored in student assessment but hugely desirable either on a college campus or in a workplace: the ability and willingness to help others.
The large and mixed-age classes I envision would be learning environments in which an important part would be played by peer-to-peer tutoring. And part of the running narrative of every student’s educational career should make reference to that, should record and honor not only the time and effort put in on one’s own behalf but also the work done on behalf of other people. Software could easily be developed to track this, and I believe the data would be valuable. A generous student will grow into a generous colleague. Someone who communicates well in school will likely communicate well in life. People who are skilled at explaining concepts to others probably understand them deeply.
If I were an admissions officer or a personnel director, I would love to have some insight into candidates’ tendencies in terms of their willingness to help, to give, to pursue not only their own goals but the general good of a community or team. A multiyear, data-based narrative—privacy-protected, of course, and made available only to people of the student’s choosing—would present a compelling and many-sided preview of how an applicant would be likely to function and contribute in the world.
This brings me to the idea of the “creative portfolio” as a central part of a student’s “transcript.” Everyone is beginning to recognize that curiosity and creativity are more important attributes than a mere facility for a particular subject; yet except for narrowly defined art schools, few institutions even consider an applicant’s creative output. This is doubly wrong. First, it implies that only “art” is creative—a view that is provincial and limiting. Science, engineering, and entrepreneurship are equally creative. Second, if we fail to take a serious look at what students have created on their own, above and beyond lessons and tests, we miss an opportunity to appreciate what is truly special about them. More than any data, grades, or assessment, someone’s actual creative product is the best testament of his or her ability to create from scratch, to make a solution out of an open-ended problem.
Serving the Underserved
Let me remind you of the mission statement that has guided Khan Academy since day one: To provide a free, world-class education for anyone, anywhere.
Admittedly, this is a rather grandiose ambition. It probably springs at least in part from the fact that I myself am the child of immigrants and I have seen with my own eyes places like Bangladesh, India, and Pakistan, where the inadequacy and unfair distribution of educational opportunity is a scandal and a tragedy (and pre-Katrina New Orleans wasn’t much better). But if my internationalist perspective is partly a function of personal history and emotion, it is also a matter of simple practicality. We live on a small planet, in a world that, as Thomas Friedman put it, is “hot, flat, and crowded.” A problem in one place—whether that problem is a financial crisis, a political revolution, or a new virus of either the electronic or biological variety—quickly becomes a problem everywhere. Lack of education, and the poverty, hopelessness, and unrest that tend to go with it, are therefore not local issues but global ones. The world needs all the trained minds and bright futures it can get, and it needs them everywhere.
As a parent myself, I completely understand the human tendency to regard one’s own kids as the most precious in the universe. To every mother and every father, of course they are; biology takes care of that. But there is a somewhat dangerous corollary to this natural parental love. Sometimes it seems that, both as individuals and as societies, we think it’s okay to be selfish as long as it’s on behalf of the kids. Clearly, there’s a hypocrisy here; we’re still serving the interests of our own DNA and our own narrow clan. We give ourselves a free pass on something that is emotionally right but morally wrong. As long as our kids are getting educated, we won’t worry about the kids a block, or a nation, or a continent away. But are we really doing our kids a favor by taking this isolationist, me-first position? I don’t think so. I think we’re condemning them to live in a world of broadening inequality and increasing instability. The better way to help our kids is to help all kids.
I believe that computer-based, self-paced learning offers an incredible opportunity toward leveling the playing field all around the world. Contrary to many people’s assumptions, it can be delivered very cheaply. It can be deployed in thousands of communities where tens of millions of kids currently have no educational access at all. If computer-based learning has the power to transform education in the developed world, it is potentially even a bigger game-changer in the developing world. Consider an analogy with cell phones. Cell phones have changed life everywhere, but they have positively revolutionized it in the developing world. Why? Because the developing world had so few landlines. For most people there, cell phones aren’t just an add-on, they are it. As with telephones, so with education—the more egregiously that people were underserved before, the more revolutionary an improvement they will experience.
To be sure, there are daunting challenges in bringing any sort of education to the world’s poorest and most badly administered places. I do not claim to be an expert in on-th
e-ground conditions in Africa or Borneo or remote towns in the Andes. But I do know something about the Indian subcontinent, which can, I think, serve as a sort of poster child for the kinds of difficulties to be faced.
In many rural areas, even the most basic prerequisites for education are often missing. Child malnutrition is a huge problem; it’s hard to learn on an empty stomach or with maladies that sap strength and concentration. School buildings are few and far between; money for traditional school supplies is hard to find. The skill sets of village kids tend to vary far more widely than even those of the most underserved in the developed world; one twelve-year-old may have somehow managed to attain the same grade level as his middle-class peers in the United States or Europe, while another hasn’t even learned to read.
And the list of difficulties goes on. There is a terrible shortage of teachers—and an even bigger shortage of those qualified to teach relatively advanced subjects like trigonometry or physics. Because of vast distances, bad roads, poor communications networks, and administrators who are lax, corrupt, or simply overwhelmed, there is virtually no oversight of school performance, or even teacher attendance. The World Bank estimates that 25 percent of teachers in government primary schools don’t come to work on any given day, and only 50 percent of those who do actually teach.2 Nor is there any reliable way to monitor student work and progress. Is education happening at all in many rural villages? Often it’s impossible to know.
These are realities that educators of any stripe must grapple with. But for various reasons I am convinced that software-based, self-paced learning has the best chance of thriving in these sorts of circumstances.
Why? Let’s start with cost. If school districts in poor countries can’t even afford secondhand textbooks, pencils, and blackboard erasers, how can they possibly afford up-to-the-minute video lessons? The answer is that the lessons, in their most basic form, could be delivered virtually for free.