The notion of making courses available to people to watch at their leisure is not a new idea: this has been done on television and in the “Open University” forms going back 40 years. The UK was a pioneer in this area. And you may have heard of the DVDs one can buy from companies like the Learning Company, now renamed The Great Courses, which by the way, markets a hundred different history courses that are widely used and have many admirers.

MOOCs are an extension of this idea; they leverage certain developments in online technology, and they also incorporate a large social network in which people who are taking the courses can interact. This means that the courses are offered somewhat synchronously—that is, they happen in real time, though you don’t have to download them or look at them on a particular day. But if you want to enter the social network and comment on what’s going on, you need to keep up with a course on a more or less weekly basis.

Let me deal with a couple of widespread myths or misunderstandings about MOOCs. First, it is very important to keep two different conceptions in mind. One conception is of a pure online course in which the teaching is done only through the online format and the accompanying social networks. An example is the Coursera platform, currently available for free. Other platforms are not available for free. A great many professions use these—if you ask your doctor what he used to pass his recertification, he may give you an answer that has something to do with online education.

But there is another concept which marries the online vehicle with classroom instruction, in some sort of hybrid form. That is highly unusual, and requires an unusual commitment on the part of the students (and this is what we did at the University of Virginia).

It’s important to keep these two different concepts in mind. They have different audiences and different purposes. The reason I mention all that is because the primary reason the University of Virginia spent money on my course was as an experiment in this second model. We were working on different ways to try to crack the problem of how best to deliver survey classes, in which the large lecture-hall format has certain strengths and weaknesses, and retain in­person instruction. So my course was kind of a test bed in which to play that out, and it had the spillover effect of creating quite a lot of course material that you could make available for personal enrichment to an audience around the world, for free. A lot of people have different concepts in mind when they speak of MOOCs, but everyone should understand why we spent money on this.

Another widespread myth and misconception about MOOCs is that universities see this as a way of saving a lot of money. Not my university. These courses are not cheap. Any university that tries to do this in a serious way quickly discovers that if you want to do it well it is not cheap. There is a significant upfront investment, and it requires constant attention and work. So it’s not so much that this is a gigantic money saver; it’s actually devised to deliver certain kinds of instruction in a way that allows you to take advantage of certain possibilities, depending on what your conclusions are about this experiment.

So let me tell you about my personal experience with MOOCs. You’ll often see people advertise their registration numbers for their courses, which are huge. These numbers are meaningless. Anyone who tells you they mean something is not giving you the full truth. People will sign up for courses that they have no intention of actually using and never actually even look at. So the only interesting number is how many people actually try out the course. Did they watch as much as five minutes of the course to see if they’re interested in it? In my case that number was somewhere in the 30,000 range worldwide. And we have some decent data on where those people were around the world.

Of that 30,000, only about 15,000 decided to give the course a serious try (I think my numbers are better than the Coursera average). About 10,000, more or less, stayed with it for the whole duration and watched all or most of the 92 video segments—and 92 videos is a big­time commitment. About half of them bothered to sign up for a certificate. About 5,000 were online auditors who took the course for their own enrichment. About 2,500 people were downloading and watching offline, not taking quizzes or participating. But they are still part of it.

Teaching people all around the world was the most gratifying teaching experience I have had in my whole career. And a lot of online instructors have experienced this. I touched a lot of people’s lives, and they responded in a number of ways—e­mails, flowers, handwritten letters, testimonials, messages to my bosses, to the board of visitors, all unsolicited. It’s the kind of stuff any teacher would appreciate.

It was an incredibly gratifying experience, but in a way the most interesting part of it was what I learned about flipping my residential class. In some ways I think that how I reinvented my residential course, leveraging this material, is the most interesting part of this experiment.

The standard model for a history course we all know: a professor lectures to a large and passive group of students. In my case, it was 120 students. In the revised model, I created 92 taped lectures, with me sitting at my desk as if you had come to my office. In the standard model, I’m giving presentations in which I’m extemporaneously using classroom technology. I’m trying to juggle managing my PowerPoint or whatever I’m doing on the screen. Hopefully I’m not messing it up in the middle of my lecture. I’m doing it extemporaneously and hoping it all holds together.

In the revised model, I can prepare a much more elaborate integration of media, including writing on my slide, zooming in on maps or parts of paintings, animating maps, and I can prepare all that in advance and try in multiple takes to get it right (I do not use a script, but obviously these are not done in one take).

In the standard model, your lectures have to fit into two time boxes. Whatever content you have has to be grouped into two 50-­minute segments. In the revised model, I can create lectures of varying lengths depending on how much time I want to spend, which has varied from 100 minutes a week to 200 minutes a week because I wasn’t bound by time boxes. And I can cut up my segments in a way that makes narrative sense.

In the standard model, students get one chance to listen and take notes on your lecture. In the revised model, they can play the lecture again and again, stop, pause, or, if they find you really boring, speed up or skip. They can slow down for note taking. They can freeze on a map or chart if they want to linger and puzzle it out a bit more.

In the standard model I have to clump my readings into a weekly clump or a biweekly clump. And it feels like homework. Students may or may not do the readings. In the revised model, my reading assignments were geared to the segment being presented. And because they are watching the presentation at home and doing the reading at home, it can become a seamless process of reading and viewing together. If you think about this for a while, it is actually quite a remarkable change. You’ve changed the way students cognitively take in this material.

In the standard model, how often do you quiz students on lecture material? You probably test them only two or three times per semester. In the revised model, I test their recall and understanding in every single one of the presentations. Ninety­-two times. Psychologists will tell you that this kind of testing and recall exercise has really important cognitive benefits.

In the standard model, who does the follow-­up explanation on the big lecture? Your grad students. In the new model students get the follow-­up explanation from me. And I never met with my students in a classroom of 120. I broke them that into tutorial segments no larger than 60, and we’d have meaningful discussions. So I’m doing the TA sections. But then, what did the the TAs do? I used the TA time to create a whole new dimension to the course that never existed before: history labs, where TAs created a whole new body of primary sources on 10 different cities around the world, focused on the development of those communities over time, and then related the themes that emerged to the macro themes in the big lectures.

In conclusion, this is an experiment, but it has two important features. One is the online­only feature, and there the question is whether personal contact with the professor is important. A lot of people found the purely online course personally enriching. And that’s important; we care about the students don’t we? And how many people are getting a chance to learn. The potential of MOOCs is the ability to meet the vast body of students who want this education and are not between 18 and 22 years old, which is most of the world. It lets us transform the whole notion of who is the audience for university education.

Then there’s the second concept, the experiment in flipping classrooms, and whether you can use these tools to do innovative classroom design. I’m highly satisfied with this experiment. And I will repeat it this spring.

—Philip Zelikow is the White Burkett Miller Professor of History and associate dean for the Graduate School of Arts and Sciences at the University of Virginia.