The Hyped Success of MOOCs

I tried to find an experience report written by one of the teachers, because I vaguely remember reading possibly reading one, but could not find any through google scholar. So, I don’t know about the logistics of running the course yet. How much time it takes to develop the materials and supporting technology, the best ways to ensure participation, how to combating the high rate of dropouts and lurkers, how to prevent poser problems when giving credentials, etc. are all still unanswered.

The following are mostly quotes from the above search, but I’ve inlined some of my own comments and thoughts.

Of the initial 104,000 people in Ng’s machine-learning course, only 46,000 completed any assignments, and only 25,000 did a majority of the homework.
More curiously, when a Minnesota company that does weather for businesses had an open position in the spring of 2012 for a mathematical scientist, more than 25% of the resumes listed Ng’s course.
— How Coursera, A Free Online Education Service, Will School Us All

Norvig and Ng were only a week ahead of the class, and help online ‘office hours’ where they answered questions that had been through a voting/ranking process. Author used Norvig+Thrun’s class for the content of his own AI course at Uni of Mass. Lowell, using their classtime for conversation about material. The fall 2011 course for matriculated Stanford students included programming assignments, and the online course did not. But the new Udacity courses do include programming.
— Will Massive Open Online Courses Change How We Teach?

The 2011 graduate level AI-Stanford class, CS221: Introduction to Artificial Intelligence, taught by Sebastian Thrun and Peter Norvig was a massive open online course with 160,000 registered enrolees of which 20,000 completed all coursework. The course included feedback on progress and a statement of accomplishment. The curriculum drew from that used in Stanford’s introductory Artificial Intelligence course. The instructors offered similar materials, assignments, and exams. An additional 200 registered for the course on campus, but a few weeks into the semester, attendance at Stanford reduced to about 30, as those who had the option of seeing their professors in person decided they preferred the online videos.
The AI courses had many computer-programming experts in the field. The range of participants went from junior-high school students and humanities majors to middle-aged middle school science teachers and more than 50 retirees. This paper also comments on issues with Lurkers and Dropouts, using statistics from 3 online offerings of courses about MOOCs.
— MOOCs and the AI-Stanford like Courses: Two Successful and Distinct Course Formats for Massive Open Online Courses

The 2011 graduate level AI-Stanford class, CS221: Introduction to Used an overhead video cam to talk and record writing. “It’s like listening to a friend explain something to me at a coffee shop”. Typical video is never more than 6 minutes, and pauses for a quiz question at the end to simulate one-on-one tutoring. Questions are more conceptual and open-ended than calculative, and students answer before proceeding to the next video. Videos only available for a week and tied with a discussion board. Everyone works on the same thing in the same time frame, so no procrastination. Online discussion was mostly organized by the students themselves across many networks (FB, Google+, Reddit, Aiqus, etc). 160,000 enrolled about 20,000 finished all the homework and 1/2 watched at least one video per week. Norvig is more interested in the electronic data they can gather about what makes for effective learning. I’ve seen Kahn academy doing a great job at that already.
— Peter Norvig’s 6 minuted TED Talk “The 100,000-student classroom”.

Required volunteer army of 2,000 to translate into 44 languages.
Stanford students preferred videos, in spite of the fact they pay $33K/yr and can see famous prof in person.
Collected very powerful student testimonials: including one from Afghanistan who spent the day running from rocket and mortar attacks to make in to a safe zone at an air field and spend his only hour of internet access on the course that night.
Old “weeder class” habits die hard: Prof gets used to asking really hard questions, letting students fail, and coming to their rescue to present themselves as very smart. Grades are the failure of the Edu system, it’s much better to give students many chances to master the material, and not intimidate with tough questions. There is also no reason to stay on a tight schedule, with recorded videos students can spend time on material until they reach mastery.
Videos feel more like a ‘personal tutor’ combating the ‘anonymous in a large class’ syndrome.
—Sebastian Thrun Aims to Revolutionize University Education With Udacity

It’s too easy to cheat, and the incentive will be to do so as soon as employers recognize the certificates and grades. Star students can’t shine, because the teacher can’t write 900 reference personalized letters for the 1% of high achievers. Employers avoid weird people, especially those who avoid authority. Computers can’t grade everything. Money can substitute for ability, because you can hire tutors.
— Why Online Education Won’t Replace College — Yet