Saturday, March 9, 2013

My first MOOCs

My university recently invited George Mehaffy, Vice President for Academic Leadership and Change of the American Association of State Colleges and Universities to speak to us about  the “[Perils and Promise in a New Age].” What can higher education expect in the next 10 years?  Despite protests of a few colleagues who feel we are doing just fine, Mehaffy left me a bit shaken.  He pointed out, and not too gently, that what happened to journalism, to the music industry, to news and publishing, to photography, and to telecommunications ...  is happening next to higher education.  There are several factors at play.  The public is fatigued by taxes and has grown skeptical about the value of or need for a traditional higher education.   Higher education has not kept pace with rapid social and technological change.  And the cost of college, he said, is at the “tipping point.” There is a hunger, outside, for change.

One big threat comes in the form of new technology: Massive Open Online Courses, affectionately known as "MOOCs."  In 2010 about a third of U.S. college students were taking at least one course that was entirely or almost entirely online.  Many of these have 30 students or so and fairly traditional in their content.  But many are MOOCS.  One Stanford course on Artificial Intelligence famously drew 150,000.  The dropout rate is high but maybe 10,000 students completed it.  

MOOCs are now organized in consortia, the most successful of which has been Coursera, less
than a year old, with 2.8 million users.  Last month it announced that it will double its number of partner institutions.  The problem for those of us in the industry is this: these courses are free -- for now -- and are likely to be very inexpensive in the long run.  

In 2012 alone private U.S. companies spent $400,000,000, Mehaffy warned us, “to take away your jobs.”   “If you’re not worried, you’re not paying attention.” MOOCs tend to be created by [top institutions], those with deep pockets:  – Brown, Rice, Ohio State, Princeton, Rutgers, Stanford, Northwestern, etc.  And they offer a lot of classes.  Flip through this site [MOOC-List] and you’ll see hundreds and hundreds, by title, by recent addition, by what’s about to start, by category, by university, by length of course, by estimated hours of effort weekly and even by “cloud tags” which show that math, health, and programming  courses are among the most popular at the moment.  There are nine full majors and more universities contributing all the time.

I find it unsettling that I hadn’t been fully aware of the magnitude of this force, or of its possible implications.  But something else troubles me more.  We try to give our students the skills to be nimble, to think creatively, and critically, and to adapt to the unpredicted changes they are likely to face.   But as an institution I’m not sure that higher education itself is able to change.  In Teaching Naked, Bowen wrote “College teachers in general have no formal preparation for teaching, so they teach as they were taught, going back in an unbroken chain to the founding of Bologna, Paris, and Oxford universities in the 11th and 12th centuries.”  I do sense the truth in this; I sometimes feel that we aspire toward conservatism, on the academic and also on the administrative side of the academy.

Now I wouldn’t exactly call myself an innovator because I don’t often invent things but I am an early adopter.  I bought my first computer in 1985, was emailing in 1989, wrote a departmental web page before graphical browsers were available.   I first taught GIS in 1989, way before it was popular, and I moved all my courses into Podium years before Powerpoint even existed.  I started a wildly successful social network called "Bughouse" on my campus a half decade before Facebook and I made a detailed 3D model of campus in 2007, shortly after Google made that tool available.  My whiteboard, made of a hacked Wii remote, plumbing parts, and free software which I learned about on TED, cost me less than $30.  All this time I didn’t invent anything, I just used the cool new tools as they emerged.   

So after the Mehaffy talk my curiosity was piqued and soon I was signed up for two MOOCs.  I chose Introductions to Genetics and Evolution, by Mohamed Noor (Associate Chair of Biology at Duke University, editor of the Journal Evolution, and past president of The American Genetic Association) ; and Game Theory, by three faculty members at Stanford: two computer scientists Professor Yoav Shoham, and Associate Professor Kevin Leyton-Brown, Stanford PhD and now of the University of British Columbia; and a Stanford Professor of Economics: Matthew Jackson.  They are not slouches.

But they are online courses for Pete's sake.  Surely at least one would be too easy or boring, poorly done, a bad fit, or otherwise unsatisfying.  Maybe I just wouldn’t’ have time to do them both, so I would drop the one I disliked the most and see the other one through, thick or thin.   And so on January 4 2013, we began. I and 20,000 classmates in one course and on Jan. 7, 40,000 in the other.

These are both topics I am drawn to.  I’ve been studying evolution for most of a decade in my spare time,but I’ve never mapped a genome before or learned why disease alleles tend to be recessive and not dominant. Nor have I calculated the rate at which genetic drift will permanently fix or eliminate a non-selective trait, or have even wondered why it does.  And I knew game theory well through just one game: prisoner’s dilemma, which  I credit (in its iterative form) for a great deal of altruism in nature and which I’ve written about in a [previous blog].  I suspect natural systems and animal behavior incorporate games of other sorts as well.

Not surprisingly, maybe, but I ended up liking them both.  There went my weekends; each took about 5-6 hours a week, but even so I found the pace and structure to be motivating.  I’d like to master this material, I want to see how a MOOC is run, and (gosh darn it) I’d  like to get that certificate of completion.  Fast forward: I passed Game Theory last week (along with 6,000 [15%] of my original classmates) and I’m two weeks from the final for Genetics.  I’m eager to see what Professor Noor has to say about sympatric speciation in week 9; it's a topic that has [vexed me for years], but the course takes a Spring Break, of all things, and I have to wait.

I’ll describe the course structure.  In Genetics it was always the unflappable Mohamed Noor, an excellent lecturer whom I’ve somehow come to think of as my friend.  For Game Theory there were three quick and competent but more humorless professors.  In both courses -- and maybe this is a Coursera thing -- the lecturer is in one corner, highlighting Powerpoints with a stylus.  Pretty straightforward and effective, I thought.  Some lectures were interrupted by a short (ungraded) multiple choice question which returned “Correct!” or “Incorrect, try again!,” and then was explained by the talking head immediately afterward.  There were other optional, ungraded, weekly quizzes for those who were inclined. 

I liked Matt Jackson's manner, and occasionally Leyton-Brown suggested pausing the video to think, which I found amusing, and I did it of course.  One can bump the timebar back to see something again and again, or watch the whole thing twice or download it.  Most consistent was the level of instruction which I would rate very good to excellent (half a star was lost by some obtuse lectures in Game Theory, a few ambiguous exam questions and at one point someone had the audacity to introduce what he called an essential concept but then skipped over it  completely because “I’m just “giving you a taste.”  The nerve!

Podcasts were about 15 minutes long and there were 7 or so each week, available a week in advance for the eager beavers.  Each week ended with an untimed but graded quiz: about 10 tough questions with a Sunday night deadline.   A discussion board was monitored by some graduate assistants but mainly was used for peer-to-peer collaboration, and that was at a fairly high level, I'll note.  There was an active thread for each question and while no one ever gave away an answer (they’d be dropped from the course if they did) my classmates did bump my thinking back in the right direction more than once. 

After the deadline passes you see your grade on the quiz, and receive a brief explanation for wrong answers.  Then comes the midterm or final, with a four hour time limit from the moment you open the exam.

So what do you get when the course is over?  Besides the knowledge, in Game Theory, if you earn over 70% you get a Certificate of Completion; over 90% is “With Distinction.”  For Genetics, it’s a Statement of Accomplishment signed by the instructor. 

That’s all very fine and good and it doesn’t affect traditional institutions.  But in the first week of Genetics I received an email: "Exciting news -- we'll be the first class offering the new Signature Track option! ... this is an exciting step in the evolution of Coursera and MOOCs in general."  And within a month, another: "Exciting news -- Introduction to Genetics is also one of five Coursera classes approved for college credit recommendation!"   The recommendation comes from the American Council on Education (ACE). This is the official explanation of the  Signature Track and college credit options:


How Signature Track works:  The Signature Track will give you the chance to link your work in this class to your identity and earn a Verified Certificate from Duke University and Coursera.
Your Work, Your Identity: Link your coursework to your real identity using your photo ID and unique typing pattern.   Earn a Verified Certificate: Earn official recognition from Duke University and Coursera for your accomplishment with a verifiable electronic certificate. Share Your Success: Share your electronic course records with employers, educational institutions, or anyone else through a unique, secure URL.
Price: The regular price of joining the Signature Track for this class is $90, but as part of the first launch you'll be able to join for an introductory price of $49.
How to receive ACE college credit recommendation:  Follow these additional steps after joining Signature Track to receive a college credit recommendation:  Sign-up for and take course’s Credit Exam: Take an online proctored exam for $69 to receive an ACE CREDIT college credit recommendation.  Send your ACE CREDIT transcript to your school once you meet the passing criteria for credit recommendations.
This class has been evaluated and recommended for 2.0 semester hours of Introduction to Biology or General Science college credit by the American Council on Education’s College Credit Recommendation Service (ACE CREDIT).
I didn't pay, but I did look at the [lists the colleges and universities] which have agreed to consider these credits (which would have cost me$59 each). There are 1,244. Now, enrollments are already suffering from demographic declines. If the ACE accreditation becomes fully legitimate -- able to insure that the courses are top quality and that students do their own work -- two more bad things might happen.
  1. There will be no option but to accept the credits because univerisies not doing so will be hurt the worst.  Traditional institutions may become vehicles for polishing off the major.
     
  2. Or employers will use the ACE approval directly.  If employers can trust that an employee knows X, Y, and Z (and that would include not only history and math but critical thinking, creative writing, public speaking, teamwork, diversity, etc), it's possible that the era of the bachelor's degree will be over.
So how does Coursera know it’s really you taking the course, and not a ringer?  In my courses there were checkboxes before every quiz and exam, to the effect of “I certify that I am me.”  Ha Ha.   But notice the comment about the “unique typing pattern” in the description of the Signature track, above.  You get a paragraph to type and the speed and cadence of your strokes is a unique identifier.  That and a webcam -- cool, huh.   For the ACE Credit, they simply proctor the exam the old fashioned way: a person, a driver's licence, and a certified terminal.

Not long ago most MOOCs were of the calculating sort, like these two I have enrolled in. But Coursera now offers the gamut: sustainability, ancient Greeks, ADHD, self-knowledge, business strategy, food and economics, writing, rhetoric, immigration and citizenship, philosophy and so on. And in a [recent article] in The NY Times I learned of a new economic model for course delivery -- I call it the Netflix model. [Lynda.com] offers 90,000 training videos in design (photography, video, web design, animation, audio production ...) -- 1,659 full courses so far. All that, for $25/mo.

Now what about quality and rigor.  After the first few weeks of casual engagement I realized that if I was going to pass these classes I’d needed to get serious and take notes like … well like a college course.  So I started a composition book for each class and took about as many notes as I would in a face-to-face class.  Judge for yourself the difficulty of the exam: here is a question cut from an end-of-the-week quiz for Game Theory:
War Game
Two opposed armies are poised to seize an island. Each army can either "attack" or "not-attack". Also, Army 1 is either "weak" or "strong" with probability p and (1−p), respectively. Army 2 is always "weak". Army's 1 type is known only to its general. An army can capture the island either by attacking when its opponent does not or by attacking when its rival does if it is strong and its rival is weak. If two armies of equal strength both attack, neither captures the island.
The payoffs are as follows.  The island is worth M if captured.   An army has a "cost" of fighting, which is equal to s>0 if it is strong and w>0 if it is weak (where s<w<M).  There is no cost of attacking if its rival does not attack. These payoffs are pictured in the payoff matrices below:
Weak
1 \ 2AttackNot-attack
Attack-w,-wM,0
Not-attack0,M0,0
with probability p
Strong
1 \ 2
Attack
Not-attack
Attack
M-s,-w
M,0
Not-attack
0,M
0,0
with probability 1−p. When p=1/2, which is a pure strategy Bayesian equilibrium (there could be other equilibria that are not listed as one of the options): (1's type - 1's strategy; 2's strategy)
a) (Weak - Not-Attack, Strong - Attack; Attack
b) (Weak - Not-Attack, Strong - Attack; Not-Attack);
c) (Weak - Attack, Strong - Attack; Attack);
d) It does not exist.
It took a lot of scratch paper but I calculated it correctly and with confidence.  This could be a graduate level course, I often thought and others on the discussion board felt the same.  

So what else?  Coursera helps arrange what they call “meet ups” where you can actually meet with classmates in your city.  Today there are apparently 2,334 “Coursera Communities” around the world.  In Chicago there was no meetup for either of my courses but 9 people in Foundations in Business Strategy are getting together, as I draft this sentence -- at a Panera Bread on Diversey. Other groups have recently met at the Harold Washington Library, the Cultural Center, Intelligencia, Starbucks, and at a private studio. There is also a "google hangout," on youtube where a handful of selected students from around the world get a chance for a totally unscripted conference-like chat with the professor and his T.A.  I saw a bit and found it stimulating; delightful would be the word.  Of course it was recorded and is now at [this link].

Now there are clear benefits to humankind if education beyond high school becomes cheap or free, and ubiquitous.  Great gaps will be spanned; knowledge -- even new knowledge -- will spread efficiently, quickly.  From a macro perspective it's hard to see MOOCs as bad.  But, like 16th century scribes facing the Gutenberg press; like newspapers facing online news -- institutions may be hurt.

The best question, probably is what's next?  First, what will they come up with next.  Foolproof ID systems:  Response-based learning paths?  Real-time curricular updates?  Location specific course material?  Courses that learn?  And second, how will -- or will, or when will -- the brick-and-mortar universities respond? I'll share a little proposal about what institutions like mine might do, another day.   

But for me, "what's next" is easy.  I broke the news to my wife Nancy today at a strategic moment.  She’s been a little concerned about the amount of time this has taken and what it means for things that might need to get done around the house, and so on.  We were in line at our local coffee shop, a public place, and I mentioned “...well I finished the one and in two weeks I’ll wrap up the Genetics course.”  
“You sound disappointed,” she said. 

“I am a little.
  but at least I did sign up for two more.”

(They are Surviving Disruptive Technologies, from the University of Maryland
And A Beginner’s Guide to Irrational Behavior, Duke University.
They start in 15 days.  I can’t wait!)