Daphne Koller recently gave a TED talk about Coursera from her perspective as a co-founder. She painted a very encouraging picture of the results they were seeing, making the platform look very good. And in many cases, she had some data to back it up.
For instance, she was able to show a visualization of the kinds of wrong answers students were giving for a particular assignment, and discussed how that helps instructors in discovering and repairing student misconceptions. The data also points to the effectiveness of interrupting video lectures with questions that student must answer correctly before moving on.
But the glowing review of how well students are learning with Coursera isn't the whole story. There are still issues, ranging from high drop-out rates to the inability to grade certain kinds of assignments.
Koller does address the latter problem in her talk. She admits that robo-graders cannot currently grade critical thinking in disciplines like the humanities (Coursera deals with this with peer-grading). But others have argued that robo-grading doesn't really work all that well even for subjects it is used for, like math.
Math educator and PhD student Dan Meyer doesn't believe math is a subject that can be automatically graded by computers, and is disappointed that ed-tech entrepreneurs don't seem to understand this:
[T]he message from Silicon Valley and the message from our best math classrooms contradict one another more often than they agree. On the one hand, Silicon Valley tells students, "Math is a series of simple, machine-readable tasks you watch someone else explain and then perform yourself." Our best classrooms tell students, "Math is something that requires the best of your senses and reasoning, something that requires you to make meaning of tasks that aren't always clearly defined, something that can make sense whether or not anyone is there to explain it to you."(I'd argue that the same goes for other subjects that are seemingly easy to grade, like computer science.)
I haven't yet tried any MOOCs personally (though there are a few I'd love to take someday). I am relying on others' experiences to get the big picture. For example, Lorraine Hopping Egan shared the following when I posted Koller's talk on Google+:
I took a swing at the Stanford course on game theory. The profs are super well-informed on the topic, of course, but not so great at creating videos that are clear, let alone interesting.(Context: "Mine was one experience—one course from one university— and I'm not a typical student (having graduated a few decades ago).")
After three weeks or so, I decided I was better off learning the material on my own (which I had done with a Great Courses series on the same topic—far better produced).
The benefit would have been more social interaction, study groups that worked (they fell apart, IME), and some challenges that are more interesting and creative. Not asking to be spoon-fed or entertained. But, given, the audience (tens of thousands) and the hunger for this knowledge, it seems to me there's a high market for well-produced graduate level classes.
Hack Education's Audrey Watters had this to say about her completed Coursera course after failing to finish a similar Udacity offering:
[When writing about Udacity] I said I “wasn’t sure” about the level of difficulty, the forums, or the “robot-grader.” I am sure about those those latter elements now: robot-graders can be incredibly frustrating, and forums can make for poor learning communities.But while there's work to do, the goal of making education accessible to everyone around the world is a good one. As Mark Guzdial points out, these online courses are experiments, and we need to do more of them to figure out what works. It will be worth it.