Coursera
Posted by Joseph Phillips, Jr. on Tuesday, October 8, 2013 at 5:25 PM PDT
At the Western Association of Collegiate Schools of Business meeting this week we had the opportunity to hear from Andrew Ng, co-founder of Coursera. It was fascinating to listen to the insights of one of the pioneers of MOOCs (Massive Open Online Courses), which have become a hot topic in higher education.
According to Ng, his goal is to expand access to higher education, and he does not believe that MOOCs will drive traditional providers out of business. He believes that Coursera's impact to date has been to bring non-consumers into the market. He notes that 80% of Coursera's students already have an undergraduate degree and are working full time, and in that mode are too busy to pursue a traditional degree program.
Ng explained that in the traditional model, what is constant is time on task, but what is variable is learning. In the mastery model used by MOOCs, that is flipped and what becomes constant is learning and time on task is the variable.
He also talked about their experience with peer grading, which is better than you might expect. The median peer grade is highly correlated with instructor grades, plus students like peer grading because it allows them to calibrate their performance against that of their peers in ways that cannot happen in the traditional model. Coursera's algorithms can also detect "tough" graders and adjust scores accordingly.
They have also had good experience with peer answers to student questions as opposed to instructor answers. Coursera is also finding that peers take more time to provide more thoughtful and detailed answers than instructors do!
Coursera is also having good luck with its verified completion certificate program. Even though they provide significant financial aid for these certificates, revenues are strong and helping to monetize the company. Moreover, they feel they have solved the verification problem with scanned photo IDs and keyboard stoke recognition patterns.
Coursera is collecting lots of data on student learning, single-handedly amassing more data than all of higher ed has collected from the beginning of time. Analyzing this data gives them important insights on human learning that will grow in value as more of it gets analyzed.
After his talk, Ng said that one thing becoming clear is that we will have to "flip" the classroom to stay relevant. Faculty members or programs that do not do that will fall by the wayside. Flipping the classroom means putting content up on the web for students to view outside of class and using class time for more higher level work.
Ng noted that completion rates for Coursera courses are high if you only look at students who complete the first assignment (45%), but for all students runs at a much lower 7%.
Asked if Coursera had algorithms to identify the most talented students, Ng did not answer directly, but made it clear this was in the mix with some of his examples of what could be done.
Asked how he thought Coursera would be different in a few years, Ng responded that it was his hope that under-served groups would be able to receive financial aid to take a collection of Coursera courses (12 was the number he mentioned) that would allow them to find a job and enter the middle class in our society. It was clear that Ng sees the most important role for Coursera is to create access to education and economic opportunity.
While Ng sees a noble role for MOOCs, others are worried that Coursera will put professors out of a job. Ng's response is that MOOCs cannot replace the most important things that faculty do, which is to mentor and advise students. If anything, technology will allow faculty to reduce the low value activity of lectures and grading homework and allow for more time for high value activity such as mentoring.
It's tempting to accept Ng's benign description of MOOCs, but something tells me it will not be so easy for the rest of us. We will have to change to stay relevant. What will happen to traditional providers if the mastery model gains greater acceptance? What about the rising cost of higher education (not just what we charge, but what it costs to deliver in the traditional model)? Will the data advantage of MOOCs give them a significant advantage in improving their product? And if the faculty role changes to mentoring and advising, will we need as many faculty?
All the sudden the future does not look so sanguine!
According to Ng, his goal is to expand access to higher education, and he does not believe that MOOCs will drive traditional providers out of business. He believes that Coursera's impact to date has been to bring non-consumers into the market. He notes that 80% of Coursera's students already have an undergraduate degree and are working full time, and in that mode are too busy to pursue a traditional degree program.
Ng explained that in the traditional model, what is constant is time on task, but what is variable is learning. In the mastery model used by MOOCs, that is flipped and what becomes constant is learning and time on task is the variable.
He also talked about their experience with peer grading, which is better than you might expect. The median peer grade is highly correlated with instructor grades, plus students like peer grading because it allows them to calibrate their performance against that of their peers in ways that cannot happen in the traditional model. Coursera's algorithms can also detect "tough" graders and adjust scores accordingly.
They have also had good experience with peer answers to student questions as opposed to instructor answers. Coursera is also finding that peers take more time to provide more thoughtful and detailed answers than instructors do!
Coursera is also having good luck with its verified completion certificate program. Even though they provide significant financial aid for these certificates, revenues are strong and helping to monetize the company. Moreover, they feel they have solved the verification problem with scanned photo IDs and keyboard stoke recognition patterns.
Coursera is collecting lots of data on student learning, single-handedly amassing more data than all of higher ed has collected from the beginning of time. Analyzing this data gives them important insights on human learning that will grow in value as more of it gets analyzed.
After his talk, Ng said that one thing becoming clear is that we will have to "flip" the classroom to stay relevant. Faculty members or programs that do not do that will fall by the wayside. Flipping the classroom means putting content up on the web for students to view outside of class and using class time for more higher level work.
Ng noted that completion rates for Coursera courses are high if you only look at students who complete the first assignment (45%), but for all students runs at a much lower 7%.
Asked if Coursera had algorithms to identify the most talented students, Ng did not answer directly, but made it clear this was in the mix with some of his examples of what could be done.
Asked how he thought Coursera would be different in a few years, Ng responded that it was his hope that under-served groups would be able to receive financial aid to take a collection of Coursera courses (12 was the number he mentioned) that would allow them to find a job and enter the middle class in our society. It was clear that Ng sees the most important role for Coursera is to create access to education and economic opportunity.
While Ng sees a noble role for MOOCs, others are worried that Coursera will put professors out of a job. Ng's response is that MOOCs cannot replace the most important things that faculty do, which is to mentor and advise students. If anything, technology will allow faculty to reduce the low value activity of lectures and grading homework and allow for more time for high value activity such as mentoring.
It's tempting to accept Ng's benign description of MOOCs, but something tells me it will not be so easy for the rest of us. We will have to change to stay relevant. What will happen to traditional providers if the mastery model gains greater acceptance? What about the rising cost of higher education (not just what we charge, but what it costs to deliver in the traditional model)? Will the data advantage of MOOCs give them a significant advantage in improving their product? And if the faculty role changes to mentoring and advising, will we need as many faculty?
All the sudden the future does not look so sanguine!