MOOCs: Where’s the Lit Review?

One of the purposes of research is to establish a foundation of prior knowledge for future experiments to engage and extrapolate before proposing a new design that will further the field.  This is important; without an understanding of what came before, research runs the risk of reinventing the wheel, or even (worse yet) coming up with something more rudimentary than the wheel.

In my days of teaching creative writing, it used to be quite the stressor to get smart, motivated teenagers to take notes of their plots and characters.  These were students used to doing everything right and being able to beat the system just with what was stored in their heads.  I explained that creative writing was not about beating a system, and the more complex a story and a group of characters became, the more important it was to create a system where you could record those complexities so you could return to it as the story developed.  Some listened right away and got to work.  Some needed trial and error before coming to me so we could devise strategies.  Some never listened and became increasingly frustrated.  In the end, it was more likely for someone from the first or second group to have a coherent, rich story than someone from the third group.

I think about this as I read more literature on the history of MOOCs as described by the MOOC creators.  MOOC developers don’t spend a lot of time writing about this, which is concerning because just like any learning model, MOOCs deserve a thorough literature review prior to implementation (if a middle school teacher has to link a lesson on Where the Red Fern Grows to the saturation of various competencies and skills as dictated by the state department of education, an organization pining to offer courses for more than enrichment should be held to a similar standard).  I don’t necessarily blame them for not doing this on their own; popular media has not pressed the issue (and would rather run the narrative that MOOCs are ordained learning models that will democratize/economize/save education), and the MOOC craze has exploded to the point that I don’t expect Sebastian Thrun or Andrew Ng or Daphne Koller to sit down and comb through the prior research; they have other fish to fry.

But I’m doing it.  And I’m running into a problem.  First off, Thrun, Ng, and even Stanford’s Class2Go link their pedagogical practices to Salman Khan.  This isn’t uncommon or unexpected; Khan started his Academy in 2009, so it was fresh in the minds of these instructors and developers.  There is no written link as of yet between Khan and AI or Machine Learning, though Khan Academy’s movement toward data aggregation and personalized learning (which, as I will get into many times soon, sounds much better than it is) makes sense in an AI context. Khan recently published a pseudo-autobiography, the sort of book that takes a brief look at his history but focuses more on his vision for education.  And Khan puts a lot of stock into the idea that long lectures are not an effective learning model, so breaking up those lectures is paramount.  He cites some research, but not much, and in a number of cases cites the research as so (I am paraphrasing here): Yes, the research says this, but I/We/People figured this out on their own.  I just did this intuitively.  

First off, Khan is right — long lectures are not effective.  The problem is, lectures are not effective.  Or at least lectures are not an effective instrument in which to push a cornucopia of content onto a student.  Khan is drawing from research at the dawn of the cognitive revolution, a period in the late 60s and early 70s when psychologists (many of whom were working for the military) realized that people gained knowledge internally and a didactic exercise of that information was not going to be a reliable indicator of what was learned.  Research of the time looked at how people gain tacit knowledge, which is why future research looked at various learning styles and modalities for projecting information (whether it be auditory, literary, kinesthetic, visual, etc.).

This is what frustrated me about Anant Agarwal walking around a Future of Higher Education conference last month citing a paper from 1972.  There’s a ton of current research out there that builds on the cognitive revolution, realizing its importance but taking that focus to greater levels:  social learning theory, constructivism, connectivism.  Basic cognitive learning is part of it, but it’s not all of it, and we have 40+ years of research to prove that.  You wouldn’t know it from looking at the key players in the game today, though.  And that’s a problem.  There is a rich literature about distance education, online learning, learning theory and pedagogical models that notes the trials and errors since the cognitive revolution, trials and errors that sought to further the field:  PLATO, AllLearn, OpenCourseWare, Connectivism & Connective Knowledge.  MOOCs to this point have not shown an interest in bettering learning, but rather in promoting scale and access.  The approach from Khan Academy, xMOOCs and things like Sugata Mitra’s School in the Cloud is barren of contemporary pedagogy:  give students a computer and vetted content.  Why not give them a Kindle and a universal library pass?

I truly believe people like Khan, Ng, Koller, Thrun, and Mitra want to make the world a better place and see education as a catalyst, and believe in their systems.  I just hope someone in their camps takes a month to go to the library and read on the history.  Because as this phenomenon continues to explode, the $$$ backing it wants to get a return on its investment, and it doesn’t care whether the learning theory behind the model is outdated.