Chief Executive of the Association for Learning Technology, Seb Schmoller has been part of the story of learning technology since the term first appeared and has direct experience of implementing lifelong learning and online distance learning projects across the post-compulsory spectrum. He was kind enough to spend the time to share his thoughts with us on a topic of emerging importance.
Q: Is there a recent change of any kind in technology or technology use that particularly interests or excites you?
The thing that’s caught my attention over the past six months has been the set of free online courses in computer science organised by, or connected to, Stanford University. I’ve had the pleasure and benefit of being a student on one of the three courses, “Introduction to Artificial Intelligence”, which has recently finished. Having been involved for more than half my working life in online distance learning, I feel that something fundamental has been discovered about how to do online distance education in a way that learners feel is giving them personal and personalised instruction, without it actually being the case that there’s any direct contact between learners and teachers.
Note from Seb “Since the interview the start-up associated with the Introduction to Artificial Intelligence course has morphed into a private provider – udacity.com – with the intention of enrolling 0.5m students onto two free courses, CS101 – Building a Search Engine – and CS373 – Programming a Robotic Car – starting in February 2012.”
Q: Do you think their approach differs from other massive online open courses? Is there something particular that Stanford have done?
I’ve not taken part in any other massive online open courses but from what I understand from what Stephen Downs and George Siemens have written, they are much more actively participative and less formally structured; they don’t feel like enrolling on a course of study that’s got a beginning, a middle and an end, and has teachers who are leading figures in their discipline running it. That’s not to criticise massive online open courses for what they do, or to assert that people learn less in such a course than in a course following the “Introduction to AI” pattern; but in the latter case the materials are prepared in advance, the course follows a syllabus and there is systematic testing and assessment.
The AI course used basically the same syllabus as the course undertaken by Stanford’s face-to-face students, and it was taught in parallel with it. It’s using different materials – I assume that the Stanford course is taught in one to three hour lectures – but the assessments were the same and the subject matter was the same. (Post-interview note from Seb. It turns out that the availability of the online version of the AI course led to a big reduction in attendance at the face-to-face lectures for the Stanford version of the course. You can learn more about this from the video I point to from http://fm.schmoller.net/2012/01/sebastian-thruns-reflection-on-the-ai-course.html. Ignore the horrid introduction.)
Q: Tell me more about their approach to online materials and presentation.
The defining characteristic seems to be that they’ve opted to break the content down into very short videos – ranging between thirty seconds and about six minutes – and to intersperse them with frequent questions that keep you connected, test your understanding and prepare you for the next bit. The questions have plenty of ambiguity and they’re generally not ‘now we’ve told you this, we’ll test whether you’ve understood it’, questions.
These videos have surprisingly low production values; they have almost a “homely” tone. The producers haven’t faffed about with making them look smooth and, for me, that makes them much more appealing to use because it feels like there’s a guy on the other end telling me something, rather than there being some kind of impermeable membrane in the way, for example caused by a designer trying to give a jazzy representation of some diagram that a teacher has drawn by hand.
Rob Rambrusch, someone from New York I got to know during the course described it thus: ‘In trying to explain the appeal of the class, this was the closest I could come: “This class felt like sitting in a bar with a really smart friend who is explaining something you haven’t yet grasped but are about to.” The whole drawn-on-a-napkin feel of the class was responsible for much of its charm. The napkin was visible to 160,000 people but that didn’t detract from the personal nature of the learning experience.’
Essentially, the presentation relies on things that are technologically mature – YouTube, the web, multiple-choice questions – in a VLE platform that is largely a marking engine. As is often the case with technology innovation, people constantly have their minds on the leading edge and the things that are new and exciting instead of focusing on stuff that’s quite a way back from the leading edge that has just been gradually maturing in such a way that it is now capable of supporting something that is really dramatic.
Q: What are your thoughts on how might the approach might develop in the coming years?
I think it depends quite a lot on whether the approach is more widely applicable. I believe it is, but I don’t know.
Whether you could apply the same approach to non-numbers-based courses such as property law, say, that’s the first question. And the second question is whether it would work at lower academic level. I think it could do and I feel very motivated and interested to establish if a level-two mathematics or English course, for example, could be built in this way.
So I’ve signed up for the human-computer interaction course to experience a course where I’ve got some proper knowledge. I’d also like to see how it works when the teachers are not quite so much leading figures in their field as are Sebastian Thrun and Peter Norvig. My suspicion is that for many learners what matters is from whom you are learning rather than with which institution. Certainly from a future employer’s point of view from whom you have learned is important.
Q: What does this mean for established models of teaching and learning?
It’s very challenging for providers and I think it’s deeply threatening in the long run to the academy as we now know it. Some of the public reactions to the Stanford development struck me a complacent, with the argument put that online provision can never match face-to-face (true, but only up to a point); or that people’s tendency to cheat will hugely devalue the meaning of any certification of learning that is issued, and therefore we do not need to worry about this kind of provision.
Q: But what about cheating; how did Stanford try to pre-empt this kind of criticism?
There’s nothing to stop me and somebody much better than me comparing notes, and me submitting their work as my assignment. And I’m sure there was some of that going on. But I think much less than one might imagine. Right at the beginning there were rather strange references to the Stanford Honour Code, which is something that Stanford gets students to sign up to, which is basically that you don’t cheat, and you do not help each other with marked work. But that all dropped away; it just disappeared as an issue. I have no doubt that students on the course who were in the top decile or one-percentile with their overall marks (not I, I must stress) will find that the (non-Stanford) certificate they were issued with will be of value in the job market even if only as a way to differentiate themselves from other candidates..
Interestingly, for the AI course, a couple of universities in Germany offered fully proctored, sit-down opportunities to come and take the exams and get credit from those universities for the course. People were driving and flying to these two locations from all over Europe for the mid-term exam, and are probably are driving all over Europe as we speak, to go and do the finals.
Q: Did anything else catch your eye or seem to be a sign of things to come?
Yes; there’s a whole learning analytics angle to this kind of mass provision. Notice that two of the courses are machine learning and artificial intelligence; the Stanford people are world authorities on how to extract meaning from mass data and I think they will be extracting interesting meaning from that mass data in the coming months. Watch out, therefore for Peter Norvig’s coming TED talk on the subject.
Q: If you could wave a magic wand, which obstacle to maximising the positive outcome from this approach to delivering online courses would you remove and why?
The obstacle I’d remove is the current exclusive focus on institutional competitiveness. The Department of Business, Innovation and Skills has a business plan for higher education that has as one of its top-level aims to “Establish a higher education, science and research framework that promotes world-class competitiveness in teaching and research”. Obviously national economies, and the businesses within them, need to be competitive, and there are circumstances when competition drives innovation; but we also need openness, as many innovative and successful businesses will attest. In education and research we need plenty of open collaboration and sharing to get the kinds of things that I’ve been talking about off the ground. In particular, we need collaboration between providers if any of them are to aggregate learners in sufficient numbers to run provision at the scale that is needed to make approaches of the kind described in our discussion work. A model where everyone is fighting with each other to be the best and to corner the students is almost doomed to fail.