TED Talks: Daphne Koller: What we’re learning from online education

Daphne Koller, one of the two founders of Coursera, describes some of the key features of the Coursera MOOCs, and the lessons she has learned to date about teaching and learning from these courses. The video is well worth watching, just for this.

However I’m probably going to suffer the same kind of fate of the Russian female punk band, Pussy Riot, by spitting on the altar of MOOCs, but this TED talk captures for me all that is both right and wrong about the MOOCs being promoted by the elite US universities.

Let me start by saying that I actually applaud Daphne Koller and her colleagues for developing massive open online MOOCs. Any attempt to make the knowledge of some of the world’s leading experts available to anyone free of charge is an excellent endeavour. If only it stopped there.

What I object to is the hubris and misleading claims that are evident in this TED video. As someone once said about one of Sigmund Freud’s lectures, what is new is not true, and what is true is not new.

Myth 1: MOOCs increase access to higher education in developing countries

She starts by using the example of students being trampled to death trying to get into the line for the very few places left open by the campus-based University of Johannesburg in South Africa. This is a particularly maladroit example. Yes, there is a desperate shortage of conventional university places in South Africa. But South Africa has probably the oldest distance and open teaching university in the world, UNISA, currently with over 160,000 students. Just providing not for credit open online learning from the USA will not solve South Africa’s access problems (especially as most of those seeking university places do not have home Internet access). Indeed, to suggest that Coursera is an alternative to conventional university education takes the pressure off governments such as South Africa’s to find their own, indigenous solutions to access to higher education.

If Stanford or MIT gave credit for these courses to students from South Africa who succeeded in the exams, and then awarded them full degrees, then that might be different. But these elite universities continue to treat MOOCs as a philanthropic form of continuing education, and until these institutions are willing to award credit and degrees for this type of program, we have to believe that they think that this is a second class form of education suitable only for the unwashed masses.

Myth 2: new pedagogy

Second, the teaching methods used by most of the Coursera courses so far are based on a very old and outdated behaviourist pedagogy, relying primarily on information transmission, computer marked assignments and peer assessment.  Behaviourist pedagogy has its value, especially where there are right and wrong answers, facts or procedures that must be learned, or students lack higher level cognitive processing skills. In other words it works reasonably well for certain levels of training. But it is extremely difficult if not impossible to teach higher order skills of critical thinking, creative thinking, and original thinking using behaviourist pedagogy, the very skills that are needed in a knowledge-based society. (It should be noted that the ‘Canadian’ MOOCs of Stephen Downes, George Siemens and Dave Cormier do not suffer from this fault).

Third, and this is the most enraging part of the presentation for me, Daphne Koller talks as if she invented online learning, and that nothing was known beforehand about works and doesn’t work in online learning. So she has discovered that students learn better if they are active, so there are lots of tests and activities in the courses. It is better to break up monolithic one hour lectures into smaller, more digestible chunks. Both these strategies in fact date back to the UK Open University print packages forty years ago and it has been standard practice to incorporate such strategies in most online learning since it began on a serious scale 20 years ago.

Her comparisons are all with the weaknesses of lecture-based teaching. For this we should perhaps be thankful but again this is not new – online educators have been making this point again for over 20 years. And now Coursera is creating local or online study groups: again standard practice in other forms of online learning.

Myth 3: big data will improve teaching

One example used in the video was how computer-tracking of student activities can identify weaknesses in the teaching. The example was over 2,000 students giving the same wrong answer to a multiple choice question. In other words, Coursera is using trial and error as a form of teaching: try something, and if it doesn’t work, correct it the next time round. However, if they followed good design principles from the outset – for instance working with an instructional designer who could spot such errors or pre-testing material before it goes out to hundreds of thousands of guinea pig students – many of these ‘errors’ in teaching would be avoided in the first place. It is far, far better to avoid errors in teaching than to try to correct them afterwards: unlearning is much harder. With massive numbers of online students, the negative impact is equally massive.

Myth 4: Computers personalize learning

No, they don’t. They allow students alternative routes through material and they allow automated feedback but they do not provide a sense of being treated as an individual. This can be done in online learning, but it needs online intervention and presence in the form of discussion, encouragement, and an understanding of an individual student’s needs. The TED lecture omitted any discussion of completion rates. Again, this should not be the measure of MOOCs, but if you are going to argue that this form of teaching is superior to other forms of online learning, then discussion of completion rates becomes valid.

Daphne Koller’s final comment is telling:

‘We should spend less time at universities filling our students’ minds with content by lecturing at them, and more time igniting their creativity … by actually talking with them.’ 

However, that requires the presence of a teacher, either in the class or online.


I am sad having to write this. Daphne Koller gave a good lecture. Even these MOOCs are valuable, because, coming from elite universities, they have woken up the media in particular, and brought online learning to the attention of the public. I believe MOOCs have great potential for higher education: but not these MOOCs. And please, is it too much to ask for a little humility? (Probably, from so-called elite institutions).

Lastly, be careful what you wish for. Underlying all this is a fundamental question: is online learning best left to computer scientists or to teachers (or even students)? I know where I stand on this. What about you?

Some other reading

See also Lloyd Armstrong’s excellent post: Coursera and MITx: sustaining or disruptive? Changing Higher Education, August 6

What should we do about MOOCs? The Board Discusses

Is online learning really cracking open the public post-secondary system? 

Thoughts on the pedagogy of Coursera-style MOOCs 

Qué está bien y qué está mal en el estilo MOOC- Coursera (Spanish translation):




  1. Tony,

    I think you misunderstand the value of classes such as the Probabilistic Graphical Models class that Dr. Koller teaches via Coursera when you write, rather dismissively it seems, that it is an example of “behaviourist pedagogy” which “has its value, especially where there are right and wrong answers, facts or procedures that must be learned, or students lack higher level cognitive processing skills.”

    To the contrary, classes such as PGM are tremendously valuable for, yes, transferring knowledge to students who *do* have so-called “higher level cognitive processing skills” in abundance, i.e., are smart.

    Your stance seems to echo a kind of paternalism I’ve been seeing in *certain* of the criticisms of both Coursera and Khan Academy, namely, the nothing can be “educational” unless it involves properly trained educators bestowing critical thinking skills upon students, who are apparently considered to be uniformly helpless until their betters help them. (I don’t know this attitude arises from some kind of overlap between professionally trained educators and social-sector types, or what.)

    At any rate, that kind of criticism tends to miss the point that there are a large number of hungry, curious, bright students in the world who would very much just like some appropriately structured information about a particular subject, please, thank you very much; and who will digest that information on their own–some of them right away, some of them years later when for the first time they really have to use a concept they previously “learned”.

    These resources help meet that need, and ought not be damned with faint praise for doing so.


    • I agree with your point that “that there are a large number of hungry, curious, bright students in the world who would very much just like some appropriately structured information about a particular subject”, but what I don’t entirely like is the way Coursera mimics university courses. I would like Coursera courses and seminars to evolve in a way that universities can’t, allowing people opportunities to learn but not marching them through in lock step. Of course people can use Coursera any way they want and that might be enough to overcome my objection. I would like to see more self-organization, little groups forming organically around sub-topics of interest etc. My opinions here are only starting to emerge. I originally was a big fan of Coursera. I still believe it has a big up-side, but I am less enthusiastic than in the past.

    • Even so late, I have to reply Richard. I am sorry, but this “large number of hungry, curious, bright students …” are just a few ones (approx. 5% of the world population). And it’s true they don’t need any help (neither MOOCs) to learn. But we have to think on the rest: the big majority, the ones that do need help to learn more and better.

      • What a bunch of cry babies. Hello free education, material, quizzes peer grading. Its not easy to personalize. You must realize there must be tradeoffs!!! Its common sense. It is not the same as traditional. Complain al, you want im still learning a lot. Talk about whining.

  2. The thing I like about MOOCs is that they are confronting myths of traditional education as well as the myths of nowadays e-learning theory. I think that creators of MOOCs have a lot to learn from veterans of e-learning (like Tony Bates), and that e-learning veterans can learn from MOOCs, challenging their/our own myths, primarily the myth that good e-learning course can not be cheap and with many participants, bringing constructivist theory to defend this statement. It is my oppinion that through e-learning quality education can get cheaper and more accessible if we deal with some issues in e-learning like:
    – Reuse of content: It is not only number of participants and related activities that contributes to price when “engaging” e-learning course, but also the price of content-making. Reusability of e-learning resources is now only a theory, and there is need in practice to really reuse e-learning resources, and to make them it in a way that they can be easily localizable to many languages. Free MOOCs put that pressure on nowadays expensive e-learning offerings.
    – Traditional teacher/mentor role defined in e-learning theory: This role can change as to be able to cope with many participants of the course, still offering “engaging” e-learning course, according to constructivist theory e.g. part of the mentor role to be replaced with activities of peer-learning, properly tracked with mechanisms of e-learn system, and analyzed by e-learning system before handled to mentor for action.
    – Offering e-learning brands and certifications (with certification systems) recognizable throughout the world as “e-learning Ivy League”: This valued certifications could actually compete with existing “real” universities’ certificates and diplomas in the real world, when e-learning is used and missused for selling diplomas more than traditional universities.

    • Yes Katarina
      Ivy League online is the only good online for the world .
      Ivy League online certificates will compete ( and win ) ” real ” universities ‘ certificates, diplomas and degrees in the real world . When time comes ivy league onlkine degrees will be presented too.

  3. Thanks for drawing our attention to this. I think blaming the “elite universities” for the uninformed discourse about online learning is probably unfair. Blaming the “technologists” might be closer to the mark. There is a tendency, exacerbated and complicated by the drive to privatize and market everything, for computer scientists to approach long-established fields as if they were being discovered for the first time. And as if they were the first to explore the territory. Whatever TED was to begin with, it seems to have become largely a place to pitch and promote business ideas – and Coursera is a business, with a business model very similar to Google and Facebook’s.

    That, it seems to me, is the critical difference between today’s “MOOC movement” and, say, the OU. The OU was driven by a political and social policy of increasing access to higher education, developing human capital and knowledge – democratising it you could say. And it was provided for collectively, from government and government revenues. The MOOC’s, it seems, are to be developed by private companies, with the aim of making profits (in ways that are apparently “TBD”). Discourse about “access”, the developing world, reducing costs and so on, should really be interpreted as marketing. Once a profitable business model is discovered (by trial and error) those groups and uses which are not profitable will be left out (again).

    On the other hand, taking a technological focus on the development of distance and online learning systems, one might say that MOOC’s differ from their predecessors in being based on internet technologies, and ask what, if any, difference that makes? It seems to me that the most significant one is that interaction through such technologies leaves a “digital trace” (in a way for example that watching a TV broadcast does not) – hence the appeal to “big data”. What we (should) have learned from search and social-networking is that this data is indeed of considerable value, and that we should not leave it in the exclusive control of private, profit-seeking, entities.

  4. Greetings, Tony and others

    Yes, Daphne Koller gave a good, well illustrated TED Talk. The narrative, like the visuals, is convincing. However, you have raised several important points. Although the provision of free lectures and machine testing is worth applauding for what it does offer, we clearly cannot take everything she says at face value.

    I suppose the elite universities and their collective projects can be expected to suffer from the failing that many established and successful institutions and individuals exhibit — the assumption that the they have the answer and the whole world is asking the same question. I am reminded of the famous One Laptop Per Child initiative that delivered a packaged technological fix to problems that had been identified from a distance without bothering to ask or involve the very people who were targeted to be saved (http://goo.gl/8lZ6M & http://goo.gl/frTN2). Sometimes, it is the soft, or social technologies, rather than the hard technologies, that need to be addressed. But, as we know, hard is easy and soft is hard.

  5. I had a very similar reaction to Koller’s talk, which I blogged about yesterday, along with Peter Norvig’s TED talk, George Siemens recent talk at EDUCAUSE, and my ongoing experiences in Coursera’s “World Music” class. I too am irritated and even insulted by the presumption that a) “Online education” and “MOOCs” are more or less the same thing, and b) Koller and colleagues invented it. But in a way, it isn’t that surprising. In a mid-July CHE article explaining how Coursera is planning on making money (I blogged about that one on July 20), Trace A. Urdan, who is a Wells Fargo analyst who specializes in education related companies, said “These are two of the most arrogant types of institutions—Silicon Valley companies intersecting with these elite academic programs,” he says. “Neither of them considers that anyone else has come to this place before they’ve arrived. They say, We’re here now, so now it’s sort of legitimate and for real.”

    I more or less see that kind of “elite institution” arrogance on a daily basis. I’m a faculty member (teaching both online and in person) at Eastern Michigan University, which is an “opportunity-granting” sort of regional university that also just happens to be less than 10 miles from the University of Michigan. As far as I can tell, almost everyone at U of M either doesn’t know EMU exists, and those that do know we’re here don’t think of us being anything close to a “comparable institution.” So even though places like EMU have been involved in online teaching for years and years, it is not all surprising to me that folks at U of M (and Stanford, UPenn, Harvard, etc.) think that they’ve “discovered” this new thing.

  6. Tony,

    I was delighted and at the same time amused by your comments of the Koller presentation on TED. While the
    ‘hubris’ is a little discomforting I thought the technologies behind the effort would be helpful if shared with the institutions located in many of the countries where the trampled millions are found.

    Bottom line for me – every effort helps.


  7. Carol Edwards, BCIT, writes:

    I watched the video mentioned in the article above last night and my reaction was, “Isn’t it wonderful how brilliant and smooth a speaker she is; isn’t it sad how the Coursera courses bear almost no relationship to what she is discussing; isn’t it even sadder that no one in the audience has done a Coursera course, and so they are completely incapable of applying critical analysis to what she is saying.”

    The only thing the two Coursera courses I am doing have in common with her speech is that you can pause the video, rewind it or replay it. Other than that, nope … not the same … just traditional long lectures, with no so-called “personal” attention in the form of quizzes and tests of your knowledge as you move through the materials, which are presented EXACTLY LIKE a normal school lecture, except that the lecturer, rather than being physically present, is on video.
    In other words, what she says they are doing in the video is NOT what they are doing in reality. Why? I suspect because it is far too expensive to build any kind of interactivity into the massive amounts of video they now have online (43 subjects x minimum of 2 hours of video per unit x 10 units … you do the math of how massive their video data base is becoming). In other words, to keep costs down they have had to dump what little so-called “personalization” they used in their first course.

    Let me also put in a few extra comments using Tony Bates’ headings:

    Myth 1: MOOCs increase access to higher education in developing countries – try doing a 2 hour long video if you are an ESL student! Yes, there is translation software out there, but judging from the increasingly desperate pleas for help on the Forums, in broken English, they aren’t very effective at translating high level, complex sentences and ideas.

    Myth 2: new pedagogy – nope — just traditional lecture format, quizzes and assignments, delivered by a computer, and marked in a completely dissatisfying and irritating fashion by a computer or by a few individuals, who regardless of their skills / abilities / intelligence, have been arbitrarily defined as your “peers”. Combine that with no personal attention and you get a lot of desperate pleas for help on the forums, with copious numbers of often badly written responses, which a person unfamiliar with the subject matter has to sift through trying to figure out what is relevant, useful and accurate. Sure you might get an answer in 22 minutes, but the QUALITY of the answer is far more important than the RAPIDITY!

    Myth 3: big data will improve teaching – nope, this is just the same old, same old, with a high tech “spin” because it is delivered via a computer. It’s about processing large numbers of people as quickly and as cheaply as possible in the hopes that someone can figure out how to make a profit off a product labeled as “Education”.

    Myth 4: Computers personalize learning – nope! Saw no evidence of personalization in anything Coursera is doing. It is a massive funnel pushing large quantities of information down people’s throats and then testing/marking them in a largely useless fashion. Despite the high quality of the lecture videos and the obvious dedication of the Instructors/team involved, there has been no spark of creativity, just the mindless “filling our students’ minds with content by lecturing at them.” (to quote Ms. Koller). I am finding the experience to be stultifying, constricting, frustrating, irritating and dissatisfying – anything but inspirational.

    I will add one last, critical comment: Notice that Ms. Koller does not provide any data on the drop-out or failure rates from Coursera courses. I posit that, like the original AI course offered through Stanford, that they are getting, if they are lucky, a 1 in 8 “success” rate. Translation: “someone” actually made it through all 10 units, though Coursera are not entirely sure who that “someone” is, since Coursera can have no idea if the “someone” they enrolled actually did the assignments, quizzes and/or exams or if it was their friend, for example, or the collaboration of a group of friends, or if they paid someone to do the work for them.

    My conclusions match Tony Bates’ ones – and I will go a step further – though MOOC’s are valuable because they are bringing attention to online learning, in the absence of a computer such as in Star Trek, with the ability to respond intelligently, rapidly and flexibly to student needs, and to identify student identity, MOOC’s will prove to be a dead-end as an instructional method.

  8. Tony
    How can you make such statesment

    ” We have to believe that
    they think that is a second class form of education suitable only for unwashed masses ”

    I do not think that MITx is a second class education.
    I hope you will take your words back .

    • Two points: this article is about Coursera, not MITx.

      The only time in this post that I refer to MITx is its decision not to award MIT credits to MITx programs. If it does not recognize its own courses for transfer into a degree, but will allow its own students who have been admitted to campus-based programs to take them for credit, then logically the decision-makers in the institution must see the MITx students as inferior to the ones who take their campus-based credit courses.

      Just because an elite institution offers a course or program does not necessarily of itself make it better than any other program. There is a distinction between world class content and world class teaching. MIT has world class content. It hires the top experts in a particular subject. This does not necessarily mean though they are the best teachers – some are, some aren’t.

      In an ‘open’ online environment, the teaching becomes apparent. MITx to date has not been using best practice in online learning, nor, other than making the courses open to all, has it shown improved or innovative teaching methods. This may well come, but it’s not there yet. As a result many more students struggle with MOOCs than is necessary.

  9. Tony

    MITx provides a certificate of mastering the course if one pass a very stri,ct exam.
    MITx has a very long term strategy, please do not take their name with MOOCs and Coursera.

    Coursera is for profit.
    MITx is not .
    Coursera does not provide certificate,
    MITx does
    Coursera does not even have a business plan
    MITX has a 20 years long range plan

  10. http://www.academicearth.org

    is in the market for 5-6 years .
    It is non-profit. But did not advertise themselves widely.
    They have more than 150 courses mostly from elite schools. All free.

    We did not call them MOOC.

    Now Coursera is here. Almost same universities but more second tier universties .
    Now Coursera is for profit.But they market themselves as free.,But they do not have a business plan yet.
    They do not know how to finance the project
    They really market themselves. Everybody talk about them.

    That is the reason I do not like the concept of marketing.

    On the other side there is MITx + Harvardx + Berkeleyx
    Excellent program . Provide a certificate. Will charge a small fee
    Still people talk about Coursera . Amazing . I do not understand people .

  11. I agree with the article that Daphne is hyping Coursera to a level that that is not borne out in practice. Also, Coursera itself does not practice what it preaches. Rather, I interpret her talk as to WHAT is the potential of MOOC if done right.

    Regarding private vs public participation in education, public participation may work at small scales like in a few developed countries but at large scales like China and India it does not even begin to scratch the surface. Public education must set up a benchmark that private institutes can aspire to initially and then better.

    Govt supported higher education suffers from all the problems that plague the government itself – politics, delayed decision making, lack of funds, trying to satisfy whims of all parties etc. India is a perfect example. Vote bank politics has messed up higher education beyond repair in public universities. That’s one of the reasons most Indians love US education which is like a breath of fresh air.

    I believe that Coursera will do well in developing countries where the system does not work and assuming language barriers can be overcome.

    Regarding India, I think students will love coursera as it solves the lack of good faculty problem rampant in all colleges. I think the degree does not matter as much as knowledge as anything is an improvement over current conditions.

  12. Hey guys. I’m actually having a completely different experience with Coursera from most of the rest of you. I’m currently in the middle of Keven Werbach’s Gamification course, and it does what it says on the tin: not only is he an extremely engaging lecturer, but also the videos stop to ask you questions, and if you re-take the quizzes, they give you a chance to really learn the subject matter. Yes – you’re right – it’s still being lectured AT, but my other option is to read a book on the subject. I’m happy to do that, but this is even better.
    Regarding the subject of credit (and I believe there was something in there about ‘unwashed masses’ but I’m not going to scroll back up to look at it): I already have a degree from an Ivy-League university, then a masters degree from another well-regarded institution. At this point in my life what I want from a course is not another degree, but the knowledge behind it that I can use to further my career. It wouldn’t surprise me if a lot of other people taking these courses don’t have other options and are happy simply to get the information, even if it doesn’t come with a credential.
    I realise that the second part of my response could apply to any on-line learning platform, but I wanted to share that I’m really happy so far with my Coursera experience.

    • Very clearly spoken. Extremely encouraging for a beginner like me. I too have my degrees and am looking only for a place to master specific domains. The structured content seems very appealing and the design of study is extremely convenient and easy-going.

  13. Tony,

    Have you taken a Coursera course? From your comments I would imagine that maybe you have not.

    My experience is great: I’m finishing Mathematical Thinking by Stanford’s Keith Devlin and it has promoted higher cognitive skills (the whole course is about that!) and it has been an extremely engaging experience. The fact that there are lots of forums and peer assessment takes us to a connectivist approach where we are all learning form others. We all share resources, there are lots of open ended problems and the automated quizzes are quite useful (even if they reflect a behaviorist approach). Of course, there is room to improve, but I think that your points of view do not reflect the richness of Coursera offerings.

  14. I was at first very enthusiastic about Coursera. It was like being set loose in a Candy store — interesting course blurbs and flashy introductory videos. I read constantly and am very ideational and therefore was attracted to Coursera. I liked the structure of a course format as a change of pace. After a while I found it was a little like running on a thread mill however. I became more and more interested in finishing courses rather than learning. In some respects, I found the courses were dulling my imagination even as I was learning more and more facts. I am not especially interested in typical history (I am more interested in philosophy, literature, and intellectual history), but I found the history courses assembled by the University of Virginia to be excellent. There were some peculiarities about some of the other courses. Some of the grading criteria were very strange to say the least. A course by the University of Edinburgh gave a quiz on seven different topics and the lowest of your quiz scores was issued as your grade for the course. So if you received a 87,5% (or worse) on one quiz and 100% on the seven other units, your grade would be the lowest score. None of the other courses were like this so this was peculiar, I Of course the grades in these course are all fake grades just as current the certificates don’t have any real currency, but it is almost addicting to get caught up in earning them, The Coursera model has potential but I don’t think can in the near future compete with universities. Also since professors are for the most part the instructors what incentive would there be to compete with the university model? I suppose only time will tell as stranger things have happened and MOOCs may morph in unexpected ways, I would like to see more course offerings that allowed individuals to do them at their own pace. This would perhaps allow people to focus more on learning rather than consuming the courses as commodities.

  15. I just discovered Coursera today, and I am impressed with the attempt to challenge students with a variety of offerings. I wish they had offered Religion/Theology classes online. Their low or no-cost means to additional education reminds me a little of the Free Universities of the 1960’s in the US: Anyone can teach, anyone can learn, where classes were free or low-cost. However, my first question for Coursera is: Are the instructors paid low wages for an impossible task?

    Next, in a world jammed with online possibilities from an endless line of “educational” providers, beware what you apply for. Anything that looks this good needs to be thoroughly examined before dropping ANY money, because according to Coursera, there are select classes that are not completely free.

    Finally, classes are not small. They are huge, involving thousands of students for one instructor. The idea of the personal touch is a joke. And the ratio of students to instructors reminds me of a company called K12 that is making promises it can’t keep, while raking in huge sums of money for online classes aimed at acquiring public school students as clients.
    The old saw is still true: If it looks too good to be true, it is. Walk away.
    Wherever there is an opportunity offered by companies like Coursera, money is somewhere in the mix.

  16. Your words provided me with an excellent view to review her video. Especially your critical summary of the child being trampled and their solution being online education for those with no means of accessing it.

    Nevertheless, the main issue with the MOOCs is the lack of a piece of paper that states your a pro. Similarly, I avoid internal training from my corporation because what good is it externally. Therefore, I spend my time else where–on education I pay for from an accredited University. The pedagogy issues can be resolved with MOOCs, as they can with any form of education. However, without a certificate,diploma, degree how transferable is this education in the hunt for employment.

    I disagree with “Daphne Koller talks as if she invented online learning”, it’s a sales pitch!


Please enter your comment!
Please enter your name here