Higher education: Creative destruction | Hacker News Reader

Higher education: Creative destruction

economist.com

Interesting article, even though many of the 'facts' about MOOCs are largely based conjecture. There is little hard evidence that "MOOCs [...] have so far failed to live up to their promise. Largely because there is no formal system of accreditation, drop-out rates have been high."

From the research I did it comparing MOOCs to other online education offerings this is merely one of the contributing factors. IMHO the primary factor to the failure of MOOCs is that they are massive. Recent research, highlighted in the NYT, showed that the introduction style courses - the ones with 300-400 students - suffer from the same or similar problems. Once they took some of the poorest performers out of the lecture halls and put them in classes of ~30 they outperformed the students in the large classes.

One way of dealing with the teaching assistance shortage during the course is community discussion forums. The best implementation of it I found are on Udacity and Coursera. Another excellent teaching aid is Community Teaching Assistants on some of the Coursera courses, e. g. Machine Learning by Andrew Ng.

In one of the highest followed and most completed course, which happens to be the first Machine Learning by Andrew Ng, the level of completion was a meager 12.5%. And in that Ng's course is an outlier as no other courses with similar sizes, with the exception of AI, have come anywhere close to that level of completion. The numbers actually show the opposite, a large course has a lower percentage of completion.

IMHO if course community were a large factor this high trafficked course with a high level of community discussion would most likely have exceeded this. As you say later ML courses have had much community discussion, yet they have had lower completion rates.

At the moment there is no substitute for real interaction with the docent or professor.

Reply on news.ycombinator.com