Skip to content

Online Education and Jazz

Alex Tabarrok writes:

There is something special, magical, and “almost sacred” about the live teaching experience. I agree that this is true for teaching at its best but it’s also irrelevant. It’s even more true that there is something special, magical and almost sacred about the live musical experience. . . .

Mark Edmundson makes the analogy between teaching and music explicit:

Every memorable class is a bit like a jazz composition.

Quite right but every non-memorable class is also a bit like a jazz composition, namely one that was expensive, took an hour to drive to (15 minutes just to find parking) and at the end of the day wasn’t very memorable. The correct conclusion to draw from the analogy between live teaching and live music is that at their best both are great but both are also costly and inefficient ways of delivering most teaching and most musical experiences.

Excellent points (and Tabarrok has additional good points that I haven’t quoted).

We’re not all the way there yet: We have the technology to record a live class for repeated consumption but it’s still not so easy to do, certainly not as easy as pushing a button on a tape recorder to get a rough if imperfect capture of a music performance. It would’ve been cool if my class this fall had been video recorded, but it never happened.

One difference between music and lectures is that, for most people, the only substitute for live music is broadcast or recorded music. In contrast, there’s already a substitute for live lectures, and that’s textbooks. It may well be that in the near future the good lectures will be recorded and students will be required to watch them before they come to class. This will be a great thing. Already, though, students are supposed to read the book before coming to class, and they don’t always do it! One reason I never post my slides before class is that I’m worried that students would then read the slides as a substitute for reading the book. After all, reading the slides is easier. So now I fear the future where students will watch the video instead of reading.

To put it another way, Alex writes, “Recording allows for repeated listening and study. Indeed, one might say that only with recording, can one truly hear.” I think that’s right, but remember that students can read the textbook over and over again too, in fact physical (paper-based) textbooks are random access and I find much easier to go find things in, compared to video.

Another way of saying all this is that we have different goals when doing the following two things:

1. Attending a lecture, reading a textbook, or watching a lecture video.

2. Attending a concert or listening to broadcast or recorded music.

In general, people are doing 1 to learn, and they’re doing 2 for enjoyment. So I think we should expect some differences. Yes, some people listen to music carefully and repeatedly, but that’s still different from taking a class, I think. To put it another way, if you’re a music student listening to music, that’s like a university student taking a class. But lots and lots of people listen to music just to hear it, not to study it, while not so many people watch college classes just for fun.

This is not to disagree with any of Alex’s points but rather to explore them. I agree with everything that Alex wrote, and I also agree that it would be good to move toward a future where instructors do very little lecturing and instead spend class time helping students working on problems in small groups. That’s how they do things in high school, as I recall (just with textbooks rather than videos). When discussing teaching strategy with our young instructors here, I always advise them to go with the book rather than spending long hours preparing supplementary reading material (which, I assure them, the students will never read). I also tell them not to prepare lectures, not to lecture at all but just to go with the textbook and spend class time working out problems (this is particularly appropriate and easy to do in an introductory class). They usually don’t follow my advice, but I think that’s not because they think their lectures are so great but rather because things are set up now so that it’s much less effort to stand up and lecture than to arrange problem-solving and discussion sections. That’s what happens with me: if I’m not well-enough organized, lecturing is a default style. Which I don’t like, but that’s how it often happens, cos I haven’t prepared enough problems ahead of time for students to work on and discuss in class.


  1. […] Addendum: Andrew Gelman comments. […]

  2. Travis says:

    I have the same concern about whether students will watch videos before coming to class, but maybe they don’t read the textbook because they know the same material will be covered in lecture. If it’s clear the class meetings aren’t lectures to introduce new material and basic knowledge will be assumed (like going into a literature class having already read the novel), maybe they would be more motivated to keep up out of class? Maybe that’s hoping a lot…

    • Clifton says:

      I agree with Travis. I do not “cover” material in class nor do I use ppt. I assign the readings and in class ask the students to explain what the readings mean, to them, the the subject area, to the field, and to real life. It makes for a vibrant process with much valuable discussion.

  3. Paul says:

    If only the writers of textbooks could experience what it is like to read a textbook for someone in the bottom 99% of IQ. I’m finishing a PhD, and I still find textbooks quite a chore to read. They are dense, tedious and soporofic. Mathematical steps are incomprehensibly skipped or their purpose is unclear.

    Watching a video of a derivation is far, far easier than reading one in a textbook. A lot of things are said and not written down, which add clarity to the solution. The instructor can use his fingers to point and explain the meaning of different parts of an expression in a way that would be extremely tedious in a textbook.

    I find a textbook most useful *after* I have seen an instructor teach a topic.

    Maybe the very smart people who write textbooks also find textbooks to be easy to read, and an efficient and enjoyable way of learning new things. But I doubt it is true for most of us.

    • Rahul says:

      I definitely wasn’t in the smart kid cohort yet I always found a good textbook far superior to most instructors I had. The worst classes were the ones with no text and instructor notes and handouts alone.

      When you select a good text you are essentially learning from the best in the field. I’d need to be mighty lucky to always get an instructor of the quality of the author of the best text. Notes and handouts often seem half-baked, unpolished and error prone. Purely probablistically, the chance of errors being caught and fixed is far higher in the 3rd edition of a text that sold 20,000 copies worldwide than some in-house pack of notes.

      Another horror was textbooks that got used not on their merits but because the author belonged to our department, was an emeritus or actually teaching the class.

    • Millsy says:

      Many texts are poor. Some are very good. The better ones tend to come in at the higher level, as those with higher levels of knowledge write them. I think this is the disconnect. Elementary statistics books often have incorrect or misleading or useless information in them.

      At the very least, I would love if statistics/econometrics books like Casella & Berger would refrain from the use of:

      “it’s easy to show that…”

      Not to suck up here, but this is one of the reasons I like ARM so much. It has both the expertise and the ability to begin at a relatively low level. One could pick it up right after an introductory statistics course that teaches through two-way ANOVA, assuming they used R for the introductory stuff. Unfortunately, I don’t teach beyond introductory statistics at the MA level at this point so I have not been able to use it for a course extensively.

  4. […] Quite right but every non-memorable class is also a bit like a jazz composition, namely one that was expensive, took an hour to drive to (15 minutes just to find parking) and at the end of the day wasn't very memorable.  […]

  5. Matt says:

    I concur with Paul – having acccess to video or audio lectures, books, AND a live teacher to ask questions of, is incomparably better than any single one of these, or even any two. Books are great to give me information quickly when I already have a conceptual framework – learning a new programming language when I already understand the basic concepts of programming. But a college or graduate level text can be tough sledding if it assumes understanding of a concept that you don’t already possess. That’s where interaction with a teacher is indispensible.

    Recorded lectures are a great way to introduce a new topic – I have taken many of the Yale on-line courses, not for mastery, but to introduce me to new topics that I previously knew very little about. The recording can cover the “dumb” introductory aspects of the material that all new students need, without much difficulty, and I can listen or watch while doing menial taks (like washing dishes). I know far more than the average person about finance after listening to Rober Schiller’s lectures, even though I did not master all the material in the course. (I don’t work in finance or any related field!) If I want to continue, I can interact in a more deliberate and focused manner with written material or a live teacher. It’s a waste of an expertt professor’s time to teach intro level stuff over and over again – that’s an optimal use of recordings.

    My point is that each medium can complement the others without replacing them.

    • Andrew says:

      I agree. It would be great if there were a video accompaniment to Bayesian Data Analysis. In fact, when I teach the class, I tie the lectures very closely to the book, starting with Section 1.1 and going forward from there. So the lectures (or, even better, the video-lectures) would enhance, not replace, the book.

  6. Citywalker says:

    Many people who extol the video lecture as a better way to learn haven’t ever had to provide any sort of performance of their understanding: a homework assignment, a test, to apply the concepts and accomplish something. On-line teaching has its uses, but are we in classic bubble phenomenon? Also, there’s a limit to how much of it any one person can consume without going into a totally passive state. Imagine having 5 courses worth of video to watch at once.

    By the way, MOOCs were predicted in the science fiction book, The Forever War, by Joe Haldeman, published in 1974. The protagonist returns to Earth after two years on a military mission to another galaxy to find that about 26 years have elapsed on Earth. Education has changed:

    “Sir, is it eighteen years public school in very country? Where do they find enough schools?”
    “Oh, most people take the last five or six years at home or in a community center, via holoscreen. The UN has forty or fifty information channels, giving instruction twenty-four hours a day.”

    Here’s the rub: MOOCs are a feature of a dystopian world.

    • BMGM says:

      Since I was laid off as a government scientist nearly a year ago, I’ve taken several MOOCs. I’m particularly fond of the edX platform where short lectures are interspersed with quizzes and exercises that .extend. the learning. The best MOOCs are not passive entertainment vehicles. They are deeply engaging learning platforms.

      BTW, I have a BA in math and a PhD in physics. I always wanted to learn more statistics, so I took Coursera’s Stat 1 with Andrew Conway, then Data Analysis with Roger Peng and then edX’s biostatistics. If I had known statistics was so much fun, I would have gone to grad school in stats instead of physics.

      I’m now taking edX’s Poverty (Economics) class. Some of the MOOC participants whine that the class is too hard and contains too much statistics. A year ago, I would have been lost, too. (Statistical methods commonly used for physics are not the same set used by social scientists or in medicine.) But, thanks to MOOCs in the prior paragraph, I understand all the lectures and research journal articles we have had to read for homework.

      A well-designed MOOC can be wonderful for a motivated and well-prepared student.

  7. Bill Harris says:

    I see two threads. First, I agree with the importance of texts and of reading them before class. When I teach, I assign the reading and the homework and then cover the material after the homework is submitted. When I do it well, the class session is more about small group problem solving than about covering the material in the text, but I’ll admit I don’t always do it perfectly well.

    One point on reading that Paul’s comment sparked: I learned, later than I should, that reading technical texts is not at all about reading; it’s about working things out in parallel. Read a bit of text, grab paper and pencil or computer and try out the ideas, redo the proof, and work out similar problems. One can do all the things we might hope the professor or video to do by ourselves and then come to class armed with those questions where we truly were stuck, despite our efforts. That’s hard, and I admit to wishing I could read a technical book like a novel at least sometimes. Still, the reason for an education is to learn how to learn for the rest of our lives, right? (And I generally like reading (and doing) over watching because it’s easier for me to control the timing.)

    Second, when comparing MOOCs and online classes to music recordings, think about what you wish for. It’s my impression that recordings have improved the productivity of music delivery, but that has also changed the economic dynamics in the business.

    Finally, there is an alternative to listening to recorded music and listening to live music: it’s called making music yourself. Is that in the same category as particpating in a MOOC, attending a live course, or working through the text or the concepts oneself?

  8. Has anyone thought much about the idea that recording a lecture changes it. For example, the audience might be affected by the knowledge that things are being recorded, and the lecturer might too. I have a suspicion that the best lectures and lecture discussions couldn’t be recorded, in some sense. I guess some kinds of Jazz enthusiasts would say the same thing about concerts. But I am making a serious point about interactive lectures with an active audience and open discussion. Recording might be bad.

    • Andrew says:


      I’ve been recorded on occasion and it doesn’t change a thing. After a minute or so I forget about it entirely.

      On the other hand, if I’m doing a lecture expressly prepared for recording, then I might plan it differently. What works with a live audience doesn’t necessarily work so well for people outside the room.

      • Millsy says:

        We are currently brainstorming about the best way to do this in our department. I think the discussion aspect suffers a bit if a lecture is being recorded. Technology is beginning to allow those watching to participate, but only if it is live of course. Going off on a tangent during a lecture due to a few questions from those in class could interfere with those following along online, especially if they cannot interact as well. So do you script it and do it twice (one for live classroom and a scripted one for online with a separate, later discussion)? Have them all together at once to improve the efficiency of deliverance?

        Just questions I don’t know the answer to.

  9. […] Andrew Gelman’s reaction to an excellent post by Alex Tabarrok: [W]e have different goals when doing the following two things: […]

  10. Jed says:

    Seems like a lot of influential people discovered textbooks and public libraries last year. Ivy League graduates have been especially generous in their praise. It’s surprising they didn’t notice sooner — books aren’t exactly a recent innovation. If they had, the long-term decline in library funding might already have been reversed. Still, with so many prominent voices speaking out at the same time, you can be sure that something really big is about to happen. I’m sure before long the Republicans will have library cards and textbook subsidies for everyone.

    Then we can get rid of those inefficient Pell grants.

  11. jrkrideau says:

    Somewhat off-topic but an interesting paper on student learning which I hope to finish reading in a day or two.

    Dunlosky J, Rawson KA, Marsh EJ, Nathan MJ, Willingham DT. Improving Students’ Learning With Effective Learning Techniques Promising Directions From Cognitive and Educational Psychology. Psychological Science in the Public Interest. 2013 Jan 1;14(1):4–58.

  12. Live lectures, textbooks, or recorded lectures? “What works?” is an empirical question to which ongoing experiments will provide more useful answers than will any accreditation agency. Perhaps students vary enough that there is no single correct answer.

    Markets excell at discovering and serving specialized customers. A legislature concerned for the well-being of students and taxpayers would mandate that all schools which receive taxpayer funds accept credit-by-exam for all courses required for graduation. Unlimited credit by exam would stimulate demand for a variety of instructional services. Demand, whether for genuine education or for credentials, will evoke suppliers. Not all students need a PhD professor to read their Russian History textbook to them. All your legislature (or Congress) needs to do to save $ millions ($ billions) per year AND improve instructional methods is to license qualified institutions to offer a standard course of exams. Let competition between Sylvan Learning Centers, the Kumon Institute, and the University of Phoenix drive the cost of a high school diploma or college degree down to the cost of books and grading exams.

    Schools teach, test, and certify. A legal regime that allows schools to test their own students creates a serious conflict of interest. If it is fraud for a physician to charge for the treatment of a healthy patient and if it is fraud for a mechanic to charge for the repair of a functional motor, then it is fraud for a teacher, school, of government to charge students or taxpayers for the instruction of a student who does not need our help. If the US K-PhD school system is NOT an employment program for employees of tax-subsidized schools, why cannot any student take, at any age and at any time of year, an exit exam for all courses required for graduation? What possible “public goods” argument justifies charging taxpayers and students for something that (potentially) costs taxpayers next to nothing and that students could obtain (nearly) free?

    Credit by exam would bust this racket.

  13. This whole discussion, which is happening across much of the web I frequent, deeply frustrates me. From my perspective, it’s … irrelevant and insane?

    My perspective is as someone whose primary educational experience was in the context of an undergraduate institution among the top-five student:instructor ratios, where there were no lectures and every class was a seminar of one to (at most) two dozen students. And this experience was preceded by several attempts at a typical large university education where most of the lower-level classes were lectures attended by several dozen to hundreds of students, and often taught by harried TAs or disinterested profs.

    I’m sorry, but I saw almost no learning going on in those lectures. Questions were rare … they are, after all, problematic with that many students. And the students in those courses at those large universities aren’t that interested in asking questions, anyway. Many of them aren’t even paying attention.

    In that context, insisting that students show up for these classes, as opposed to various distance-learning methods, including online, seems to me to be perverse. Or, to be frank, it seems to me to be a bunch of offered rationales that are pedagogically unconvincing but make a great deal more sense as frantic attempts to justify a status quo with many vested interests.

    Mind you, I am the last person who would ever claim that any sort of distance-learning could substitute for the pedagogical interaction of the small, communal classroom, seminar-style or otherwise. But that’s the exception at the undergraduate level and especially the exception with regard to most of the courses that are being offered online.

    Another problem with this discussion is that it presupposes as the standard for the lecture to be that engaging, “jazz” performance by a talented academic lecturer. And putting aside for the moment whether it makes sense to deny that this experience couldn’t largely be preserved as a recording, the simple fact is that such lecturers are a tiny exception to the rule of the pedagogical void of simple, bored regurgitation of textbook material, with slides. I’d be willing to fight to preserve these wonderful lectures if they were the norm, but they’re not.

    From my perspective, what’s threatening to many academics and other interested parties about online education is that it makes it pretty clear just how truly low the American standard actually is for a university classroom experience. It’s not at all what we like to think education is, it’s not like all the stories we tell about inspiring teachers and exciting classrooms, and we recoil at it. But, really, what we’re recoiling from is our subconscious awareness that for most undergraduate university students the classroom is pretty much never like our ideal and online education is sort of a capitulation to a reality that we don’t want to admit.

Where can you find the best CBD products? CBD gummies made with vegan ingredients and CBD oils that are lab tested and 100% organic? Click here.