Home > math education > Good news for professors: online courses suck

Good news for professors: online courses suck

February 19, 2013

If this New York Times editorial is correct, and it certainly passes the smell test, students are not well-served by online courses but are by so-called “hybrid” courses, where there’s a bit of online stuff and also a bit of one-on-one time. From the editorial:

The research has shown over and over again that community college students who enroll in online courses are significantly more likely to fail or withdraw than those in traditional classes, which means that they spend hard-earned tuition dollars and get nothing in return. Worse still, low-performing students who may be just barely hanging on in traditional classes tend to fall even further behind in online courses.

This is important news for math departments, at least in the medium term (i.e. until machine learners figure out how to successfully simulate one-on-one interactions), because it means they won’t be replacing calculus class with a computer. And as every mathematician should know, calculus is the bread and butter of math departments.

Categories: math education
  1. February 19, 2013 at 8:56 am

    If I had to guess, I’d say that we are in the ‘toys for hobbyists and enthusiasts’ stage of online education. This is similar to the stage of personal computers in the late 1970s, or the stage of the WWW in the early 1990s. In the long term (decades), online education will be a big game changer.

    The chalk n’ talk of the traditional lecture has been basically unchanged for centuries, it’s a rather inefficient way to teach, and disruption is coming.


  2. Steve Stein
    February 19, 2013 at 9:04 am

    One would think that Calculus would be among the easiest of topics to cover in an online course. Has MIT done an evaluation of their open courseware offerings for 18.01 and its variants?


  3. JSE
    February 19, 2013 at 9:19 am

    I’m not sure it’s relevant to professors’ future job prospects whether online courses suck. The logic of disruptive innovation is that an inferior product can (some people would say must!) disrupt its competition if it is free/cheap, convenient, and scalable.


    • Richard Séguin
      February 19, 2013 at 2:14 pm

      A perfect example of disruptive innovation is the MP3/iTunes phenomena. Many young people now view music as a cheap commodity and have no idea how much better recorded music can sound from a quality LP, CD, SACD, or high resolution music file. The superior technology is suffering in the market place as a result.

      (One exception is that SACD seems to be doing well in the classical music world.)


    • dg
      February 19, 2013 at 3:43 pm

      While this argument is tempting, it doesnt really explain the fact that already people are willing (in fact, more than willing) to pay more money to go to [insert your favorite second-tier private college or university here] than to go to [insert your favorite third-tier state school here]. Clearly some people perceive that the educational benefits are worth the money. Or at least they believe that other people perceive that the educational benefits are worth the money.

      Tangentially, one of my biggest pet peeves with all discussion of MOOCs is how they lump all face-to-face courses together. There are many many many college faculty who have given up on the ‘chalk and talk’ model in their own classrooms, and it seems clear to me that this trend will only continue.


  4. grwww
    February 19, 2013 at 9:26 am

    Computer based demonstrations of the meat of calculus still don’t seem to flourish. There doesn’t appear to be anyone focused on a step by step learning toolset, which can adequately analyze a students comprehension of the subject matter.

    I went through advance math class programs in Jr. High and Sr. High school. It wasn’t until I was in college writing software to draw a graphical representation of a circle that I finally “got” how the unit circle worked, really. I was never taught that sin and cos are ratios, really. Instead, they were always magic “functions” that returned magic “numbers”.

    I needed to see cos(x) ^2 + sin(x)^2 = 1^2, not cos(x)^2 + sin(x)^2 = 1. There were lots of other things about how trig was not taught which also created problems.

    But calculus was even worse, due to a number of things, some of which were my issues with time due to my major. But in the end, I still don’t think that there is enough “information” conveyed in calc learning material to really ground the student in the facts and the concepts those facts are applied to and why things are so grand when you finally know how to use calculus.


  5. February 19, 2013 at 9:50 am

    I am certain that I have made this point here before. I teach service calculus courses regularly. I also tutor undergraduates in my university’s math tutoring center. Many students are weak on fundamentals. I am certain that excellent students can learn from MOOCs, just as excellent students can learn by reading a textbook alone. However, most students need one-on-one interaction to push them, to keep them involved, to show them the shortcuts.


  6. February 19, 2013 at 10:03 am

    Steve Stein mentions OCW and 18.01. I coordinate 18.01 twice, and I contributed the first 18.01 materials to OCW. I think 18.01 is a perfect illustration of my point. Although MIT has excellent students, those (few) students who are not math proficient end up in 18.01 — the lowest level math course MIT offers. Those students need one-on-one interaction, and happily MIT provides the resources those students need. I spoke with an older faculty who said it would be better if there were even more one-on-one interaction / smaller class size for those student (typical class size is about 150 students). I do not think MOOCs would work for those students.


    • February 19, 2013 at 10:09 am

      First, I do think that MIT is a perfect test case in the sense that, if it doesn’t work for that level of smarts and motivation then it has no chances (but if it does work for them, then the jury’s still out for normal students).

      Second, though, I don’t see a natural obstacle for algorithms to eventually take the place of one-on-one prerequisite exploration and instruction. After all, you probably have a rote set of things you ask when a student has difficulty with a certain idea – first you check if they understand the basics of such and such, and if they don’t, you go back further to something else, but if they do then you move on one step forward, etc.

      As soon as someone’s written that down they can encode that. It’s a question of developing something complex and good. It will happen.


      • Moeen
        February 19, 2013 at 12:39 pm

        As to your second point: I don’t think it’s quite that simple. Different students will learn and understand the same concept differently. I may have to explain the same idea from different perspectives to find ones that click to the student, and you don’t know in advance which ones will. It really depends on how individuals learn concepts and I don’t know if you can encode that. What worries me is that MOOCs will encourage a “one size fits all” approach, which has serious drawbacks.


        • February 20, 2013 at 6:49 am

          But you don’t know in advance which will work either! You have to try different things. Computers can be trained to do that too.


  7. Cynicism
    February 19, 2013 at 10:30 am

    Jack Bennett (@32000days) :
    The chalk n’ talk of the traditional lecture has been basically unchanged for centuries, it’s a rather inefficient way to teach, and disruption is coming.

    This is a rather strong statement. Would you care to back it up? By what measure is the traditional lecture inefficient and why is that a valid measure?


  8. grwww
    February 19, 2013 at 10:38 am

    If anyone leaves the room without 100% conveyance of the information, then the “Chalk and Talk” has failed as “The Tool”. It is well known, that different people have different learning styles. Many people need to see the same information presented in different ways, so that the see the “facts” in two contexts, and thus work out what is “Fixed” and what can be “Variable”. Demanding that some people can go to class once, and others have to go to class 2 or even 3 times, to “learn”, is where the issue lies. We should help people identify their learning style, and provide a single classroom experience that makes it possible for them to learn in that way. Certainly, there are educators who have varied learning styles as well, and they could easily provide the correct formulation of the materials for a style because they understand how that works for them.

    It’s the separation of the “information” from the “how to learn it” that makes the big difference. Facts on paper are just that. Until they can be consumed, they have no value to the student.


  9. Dan L
    February 19, 2013 at 11:32 am

    For average unmotivated students, most of the value of a traditional class is to force them to show up somewhere at an appointed time, try to pay attention to something for one hour, do assignments in a timely manner, and study for exams at regular intervals. Ironically, for such students, MOOCs will succeed by becoming more rigid, even though the main advantage of MOOCs for well-prepared, motivated students is their flexibility.


    • February 19, 2013 at 11:41 am

      If I understand your point, clickers would serve the same purpose. Clickers are cheap, they are already widely used, and clicker classes are much more flexible than MOOCs.


  10. February 19, 2013 at 11:43 am

    I am such an online learning partisan. It was a little hard for me to read. But I’ll try to be rational and fair.

    Having an online class with only 25 people is the worst of both the online and offline worlds. Online classes shouldn’t just be offline classes with video.

    You want to build massive communities where students can help each other in an active forum. You want to get large data sets, so that you can see where the common problems are and design followup lectures to address the big, common misunderstandings.

    I really distrust a single community college running an online course. If you could get a hundred community colleges together, find the three most passionate teachers of some given subject, and have them each exert as much effort as they would teaching 5 classes (which is still a savings of teaching effort!), then run an online class with many students each from those hundreds of community colleges, I’d be more concerned.

    Online classes shatter the proximity requirement, which allows a whole new level of scale. If you aren’t benefiting from that scale, you are wasting your time.

    It’s kind of like saying that search engines don’t work because it’s better to have a librarian help you rank and sort a few hundred documents of some weird corpus than having a machine do it.


  11. NotRelevant
    February 19, 2013 at 1:09 pm

    They say 90% dropped the free Stanford artificial intelligence class, but 0% paid a tuition fee. There was no cost for quitting. When the course costs real money then the drop rate will decrease, I promise. But one of the two major points was the attrition rate, and so half their argument is already gone.

    People are just now learning how to teach stuff over the internet. In a sense, the professors of today are like those who conceptualized automobiles as “horseless carriages.” They haven’t begun to understand the potential of the thing.


  12. February 19, 2013 at 1:14 pm

    Why is Calculus the bread and butter of math departments?


    • February 19, 2013 at 1:20 pm

      Good question!

      Because the vast majority of courses in a given math department consist of calculus I, II, III, and IV. In some schools, the math department has lost calculus – which is to say, chemistry, economics, physics, etc. each teach their own version of calculus for their majors. And in those schools the math departments are generally very very small.


      • Thads
        February 20, 2013 at 6:45 am

        Good question!

        Many have suggested that we offer other foundational courses as alternatives to calculus: say, a little discrete math, or a little probability. The intellectual diversity would be appealing. However, economics departments want calculus, and medical schools want something uniform, so calculus it is. On the bright side, calculus is intellectually deep and of great interest from both pure and applied points of view.


  13. Zathras
    February 19, 2013 at 1:19 pm

    “…calculus is the bread and butter of math departments.”

    It’s not the bread and butter of community college math departments. The bread and butter of community college math department is development math, i.e., math needed to get into college algebra, which in turn is 2-3 steps behind Calculus. It’s really shocking when you look at the course schedule of a community college and calculate the ratio of developmental math class sections to college algebra sections. This ratio is usually over 10:1. Doing the same with Calculus sections adds a digit to this.


  14. John
    February 19, 2013 at 5:12 pm

    I suspect we will eventually see a more hybrid model. I find the short video segments and on-line quizzes that are the staple of Coursera courses to be superior to a long lecture for teaching the unit operations that are so necessary for learning. As Salman Kahn learned, his nephews preferred him on video rather than in person because they could pause/rewind/use at their own schedules. These are also useful for refreshers.

    That said, there is no substitute for good recitation section where one works problems and hammers out misconceptions. Discussion forums just aren’t as good for this. Peer grading on a complicated assignment can be a nightmare, as we are discovering in my current Coursera class. The prof has great intentions (and is working very hard on the class) but simply designed an assignment that – in my opinion – is way too complicated for on-line peer grading. MOOCs are a learning experiment for all involved. And I think they will lead to positive changes in brick and mortar based education.

    The other aspect that will be hard to replace is exams proctored by humans with good ID verification. Credentials from online courses will just be too suspect.

    As always, the discussion I find here is enlightening.


  15. February 19, 2013 at 9:35 pm

    It’s as old as Skinner’s first suggestion for “evil teaching machines”. Some teachers think the machines will inevitably replace humans; Some administrators actually think that’s possible. Machines can be made to do lots of things, but they can’t BE human. Interested in thoughts on this:


  16. bgpratt
    February 19, 2013 at 11:33 pm

    It’s got plenty of room for improvement, but I have found Jim Fowler’s (Ohio State University) Coursera offering outstanding — his approach is really worth noting – and the “Mooculus” system that presents increasingly more difficult problems as you answer correctly, and gets easier as you struggle – including hints, etc. shows lots of potential. Courses like this show the possibilities and I would be surprised if this category doesn’t innovate its way upstream to compete head-on with in-person University formats – my University in person calc class was absolutely awful by comparison. To be honest, I wish I’d had Fowler in person – but online is good enough.


    • February 24, 2013 at 1:35 pm

      I’m glad you’ve enjoyed mooculus, bgpratt; I’m certainly very excited about the potential.


  17. February 21, 2013 at 11:15 am

    I would only add that, used correctly, the new technology will free up instructors to focus on the parts of teaching that only humans can do. This should be a boon to instructors. Online education is in no more danger of displacing teachers than calculators are of replacing understanding of mathematics. It’s already easy to see how Coursera et al can monetize by developing models that take advantage of economies of scale, in order to bring excellent instruction to millions at a very low cost per student. We could be on the verge of an education explosion.


  18. Spikels
    March 1, 2013 at 2:53 am

    This editorial did not pass my smell test! I did not find it very convincing. The low completion rate is due to people trying out the free class and is not comparable to either paid online or offline courses. The article then jumps to studies the effectiveness of online courses developed in-house at community colleges. These are not the best-of-breed approach that is clearly the future of online education – best teachers, best course design and best software platform. Will it solve every problem with education – no. But it is rapidly becoming a viable alternative for many students.

    I suggest anyone interested in this issue take a class on Coursera and see how it actually works and form your own opinion. I think MathBabe should consider putting her Data Science class on one of the MOOCs. She could teach more students in one class than teaching at Columbia for the rest of her life!


  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: