Skip to content

If You Believe in MOOCs, You Are Assuming Too Much

2013 January 23

There has been quite a bit of hype about MOOCs lately.  If you haven’t been following the craze – and it is definitely a craze, just as much as Big Data is – a MOOC is a Massively Open Online Course.    I’ve written about them before – in this post, I provide a checklist to let instructors know if a MOOC can replace them.  MOOCs are marked by several key features, at least as they are implemented now:

  1. Anyone can enroll in a MOOC.  They are massively open in that anyone in the world can take the course for free.
  2. They are online courses.  The only way to get information to the masses at a very low per-head cost is to do so online.
  3. They rely on peer involvement to facilitate.  When taking a MOOC, you will almost certainly never meet, speak to, or email the instructor.  Help for assignments is typically provided by other people in the class.

MOOC fervor is getting even more intense, with the New York Times declaring 2012 the year of the MOOC and some universities granting credit for completion of MOOC courses.  It seems all is coming up roses for MOOC providers like edX and Coursera.  Proponents of MOOCs have taken on a certain fervor, promoting the idea that MOOCs are the “solution” to college – that all students will one day be enrolled in single massive MOOC for each course they need, and the current-day idea of “college” will be rendered obsolete.

I say: not so fast.  We heard similar rhetoric when online courses in general were introduced, and while many students do opt for online courses instead of in-person courses, most don’t.  The predicted mass exodus of students from in-person classrooms to the glorious Internet did not happen.  That’s because the in-person classroom still offers a lot of benefits that are difficult to provide online – at least, not without a lot of time and dedication on the part of the instructor.

In my research, I’ve even come across stories about the introduction of correspondence courses with the same sort of rhetoric as we see with MOOCs.  Correspondence courses were presented as the solution to over-crowded classrooms, and a way to get more people to go to college who otherwise would not have.  After all, why go to a college classroom when you can complete a course by mail?  As it turns out – there are a lot of reasons.  And there are a lot of reasons why you wouldn’t want to rely on MOOCs for your college education either.

I’ve noticed that proponents of MOOCs tend to not be in higher education.  This is a critical problem of perspective.  They make claims like, “We saw the Internet completely flip the music industry on its head; this must happen to higher education as well.”  Because of this perspective, they make several assumptions that upon scrutiny turn out to be unwise.

  1. Given: MOOCs rely on automated grading to evaluate student progress.
    Unwise assumption: If one class can be graded automatically, all classes can be graded automatically.

    Regardless of your personal stance on the value of grades for learning, they do provide a benchmark for progress (albeit an imperfect one).  Universities use grades to rank student comprehension of course material; a student with an “A” should come out of a course with greater skill than student with a “B”.  MOOCs can only grade questions where there is a single correct answer or where an algorithm can be used to derive all correct answers.  That means algebra is easy to grade; architectural design is not.  The most problematic area is writing.  Although an algorithm can be used to automatically grade written material, those algorithms, on the whole, are awful.  At best, they assess sentence structure, spelling, and complexity.  They cannot evaluate if your writing made any sense, which is really the only aspect of the writing most instructors actually care about (except perhaps English instructors!).  Artificial intelligence is not yet sophisticated enough to evaluate writing quality or logic, and it will probably be decades before we even start to be able to do this.  Think of it this way: how long have we been trying to create accurate translation software that actually converts the meaning of one language into another as effectively as a human translator can?  Hint: IT’S BEEN A LONG TIME.

  2. Given: MOOCs are effective at teaching basic skills that can be assessed automatically.
    Unwise assumption: The skills that can be assessed automatically are the ones we want to teach at the college level.

    I completely agree that MOOCs can effectively replace some basic skills training currently taught in colleges (e.g. spelling and grammar, basic mathematics, etc.).  But it does not follow that MOOCs can effectively replace all training currently taught in colleges.  I will even go so far as to say that the introduction of MOOCs is an excellent opportunity for universities to do away with all those courses.  Let me be clear: if course material can be taught and assessed effectively in a MOOC, it should probably no longer be a college course.  It should be a required prerequisite before the student even gets to college.  That way, time can be devoted in college to the skills that we actually want students to gain in a liberal arts environment: creativity, problem solving, and critical thinking.  If a student is paying for college credits to learn how to add and subtract effectively, they are wasting their money on college after already having been failed by K12.  We should not be spending instructional resources at the college level teaching these skills if a MOOC can deliver them equally effectively.

  3. Given: MOOCs will create a marketplace for learning skills without the need for degree programs.  People can pick the knowledge they want to gain and gain only that.
    Unwise assumption: Credentialing is not valuable.

    A common perspective is that MOOCs will make degrees obsolete – that somehow having a worldwide marketplace of courses that provide specific skills will be sufficient to prepare people for the careers they want.  But this assumes that the credential – possession of a degree – does not provide useful information in that worldwide marketplace.  But it does.  Possession of a college degrees acts as a signal to potential employers: “This person was able to organize their schedule, choose a program of study, and stick to it for multiple years at a sufficient degree of success for this institution to proclaim that they have earned a degree and for us to support that proclamation.”  This is why people have judgments about the quality of college someone attends – there is a belief that a more rigorous college will provide a more rigorous education, and the degree signals that.  While I agree that many employers put a little too much faith in what those degrees imply, this doesn’t mean that the degrees themselves are useless.  Let’s not throw the baby out with the bathwater.  A college degree indicates a substantial multi-year commitment and a major life accomplishment for the degree holder.  Let’s not minimize that accomplishment by saying that taking a selection of MOOC courses would be equally impressive.  Also important is that college exposes students to a much wider range of knowledge than they would otherwise be exposed to in high school.  I’ve had many students who took my Industrial/Organizational Psychology course to fulfill a requirement only to discover that they had a real passion for the topic.  Without a broad liberal arts education, they would never have discovered this about themselves.  In a MOOC-centered world, those students would never have known that passion.

  4. Given: MOOCs rely on discussion forums to facilitate learning.
    Unwise assumption: Discussion forums can effectively police themselves.

    One of the biggest challenges for an online instructor is managing a discussion forum.  Forums are used as a constructivist instructional strategy: the goal is for the students to actively think through the material, apply it to their own lives, and be able to explain their application of class concepts to others.  There are generally two ways such forums go.  First version: the instructor makes some regular assignments to “post in the discussion forum” or “reply to another student in the discussion once per week”.  A MOOC can do this too.  But these strategies rarely work very well; students do the bare minimum, usually resulting in a flood of posts the day that the assignment is due, and rarely do students demonstrate the application skills described above.  Second version: the instructor is an active participant in discussion, working to correct misconceptions, redirect conversations in productive directions, and generally applies their expertise to improve student engagement with the material.  The current sophistication level of artificial intelligence means that a MOOC cannot do this.  They must rely instead on the “wisdom of the crowd” – that a critical mass of other people in the classroom will 1) be interested in helping others, 2) have sufficient expertise to answer the question being asked and 3) happen to browse the forum at the moment a response is needed.  While this might work for a relatively common learning topic (like Introductory Algebra), when we get to relatively uncommon fields (like Advanced Organic Chemistry), those numbers are going to drop dramatically.

  5. Given: MOOCs rely on the support of others currently taking the course to provide the social context for a course.
    Unwise assumption: Individual attention from a mentor/leader figure is unnecessary for effective learning.

    Unfortunately, some college courses are taught by instructors that don’t really care about their students.  They are there because their are required to be there, riding  the tenure train for the long-haul.  But this is not the norm in most universities.  Most faculty care deeply about the success and learning of their students individually – they want every student to succeed.  A common misconception is that faculty want students to fail.  This is simply not true.  Instructors want students to strive and push themselves.  Instructors want to provide a sufficiently challenging course so that the most advanced students need to push themselves just as much as the struggling students do.  This is realized in a variety of ways.  Some provide a plethora of comments on student papers with the hopes that those students will read those comments and think about how they can improve for next time.  Some instructors schedule one-on-one meetings to talk about course progress and identify areas needing extra attention.  Some instructors schedule review days and feel out the classroom for areas of weak understanding so that they can target those areas with supplemental materials.  None of these things are currently possible with a MOOC, and all are valuable.  Little is more motivating than one-on-one attention, delivered by a mentor/leader figure that you know cares about you and your success.  In a MOOC, you are surrounded by other learners, and yet you are all alone.

In summary, I think MOOCs are an important development for higher education, but they are not poised to majorly reform anything.  At best, they can supplement and bolster existing efforts by replacing remedial and entry level courses, leaving the more advanced topics for the traditional college classroom, online or otherwise.  As MOOCs improve, this may change – but for now, this is the best they can provide.

Previous Post:
Next Post:
4 Responses leave one →
  1. Anonymous Coward permalink
    November 17, 2013

    I think you are to harsh on MOOCs, because you listen to the hype and take it at face value. But I think they are going to revolutionize and democratize knowledge, just not in the ways the MBAs at Coursera, edX or elsewhere wants you to believe …
    1 . You are right : not everything can be graded automatically, and even if the course material can be, the grading is only as good as the designer made it. And AI has the same problem …
    2. You are right again, but that is where I think people (even the one behind the MOOCs in the first place) are mistaken, why try to make the college or university learning experience fit in the MOOC delivery system? I had the best experience with some (yes, not all the courses manage to pull it of either) courses on Udacity.
    Why? Because they tried to adapt to the new medium!
    In the courses I undertook on Coursera, I felt I did not have the time to dedicate to learn things that you needed to undertake the course in the first place, because I do not have the time …
    I guess you need different MOOCs styles for different people
    3. I think that is the most misunderstood part, certificates are not degrees, but I think degrees could become more like TOEIC or TOEFL (where you go to take the test somewhere) in the future, not associated to one University and dissociated from the actual learning experience, again this could vary depending on subjects and ease of standardization, but we do not red to standardize every thing …
    4. You are right, I think that on that point progress is needed but, I ad a lot of good experience un Udacity’s forum (yes I am a fan of their interface) but the forum on Open2Study where not engaging (and courses there while interesting are just bait to lure students to Australia and that feels …)
    5. I do not agree, as you said and in my experience even at university or in school, a lot of teachers just do not care about the individual for various reasons, you are promoting best cases as if they were the norm in traditional learning … I can count the truly great and inspiring teachers I have encountered on one hand and one of them was in a video course on Youtube before the advent of MOOCs … The force of MOOCs is that the whole world can now learn from some of these very good teachers, but like in real life, there are plenty of courses with problems there too …

  2. November 18, 2013

    Get here today from your MOOC link on Slashdot (congrats). I agree with your general criticism of MOOCs, and likewise predict that in the long run they won’t make any fundamental impact. To counter the previous poster, I’ve investigated a Udacity course in my specialty (introductory statistics), by the founder Sebastian Thrun, and found it be absolutely dreadful (link).

    Here’s the place I think you should modify: In item #2, the claim is made that “MOOCs can effectively replace some basic skills training currently taught in colleges (e.g. spelling and grammar, basic mathematics, etc.)… the introduction of MOOCs is an excellent opportunity for universities to do away with all those courses… We should not be spending instructional resources at the college level teaching these skills if a MOOC can deliver them equally effectively.” But the unwise assumption here is the “if a MOOC can deliver them” clause.

    As someone who teaches both remedial and credit-bearing college math courses, I can say that the remedial classes are much more challenging for the students taking them. The weaknesses of the students are far deeper and more numerous; the need for personal attention is more intensely desperate; the failure rate is much higher. These are students who specifically lack the basic skills it takes to get online, manage a technical interface, organize their time, discipline their studies etc. At our school the remedial classes are kept to much smaller numbers than credit-bearing classes (like 25:35). So my argument is that MOOCs are actually much worse for remedial courses than anything else. MOOCs seem much more complimentary to fairly high-skilled students who are self-learners already, etc., etc.

    You can argue separately about whether remedial students should be kept out of college classes in the first place; a fair argument. But MOOCs are even less likely to solve this problem than any other one that you list. (Case studies: link1, link2.).

    • November 18, 2013

      That’s a really interesting point about remedial courses. I agree that they are more challenging for students taking them, but because in most of these cases there is a “right” answer, it means automatic feedback can be provided readily, and this is central to remedial education. In the real world, this would be an instructor watching someone solve a problem and stopping them before they get themselves too far down the wrong path. As a practical example, when I wrote my introductory statistics textbook, I developed an app that automatically generates datasets, runs stats on those datasets, and produces an interpretation of the results in terms of both statistical significance and effect size. MOOCs have the potential to take that to an even finer level – for example, in calculations, stopping the learner as soon as s/he’s calculated something incorrectly and gently nudging him/her back to the correct answer. It would be like having a real expert standing over your shoulder helping out when things went wrong. We’re not there yet – but I can see it happening in the not-too-distant future. I complete agree that MOOCs as they are now are not suitable for remedial education – but I predict that is where the first major breakthroughs in effectiveness will occur.

Trackbacks and Pingbacks

  1. Article Impression: If You Believe in MOOCs, You Are Assuming Too Much | Learn Programming

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS