Re-evaluating Online Universities

Jan 11, 2014

Here’s an article I read a couple days ago about the founder of Udacity and his struggle with very low success rates for online courses.

http://www.fastcompany.com/3021473/udacity-sebastian-thrun-uphill-climb

To compensate for a high dropout rate, Udacity has made a shift to focusing on corporate training. Company sponsored courses on the practical tools and frameworks that are used today. These courses are generally taught by corporate “Technology Evangelists”, or some such title.

I have a problem with this for a couple reasons:

  1. I don’t think the dropout rate is a problem
  2. I really like theoretical, foundational courses
  3. Great teachers make great courses

The dropout rate is awful if you’re comparing to conventional universities. But no conventional university could see 160,000 enrollments from people across the globe with nothing in common but internet access. There is no barrier to entrance so there is no harm in trying a couple classes and dropping out.

Why is this comparison even being made?

Let’s compare it to Wikipedia instead. I wonder what the “completion rate” is for the average longish article on Wikipedia. I would guess that it’s not much better. It doesn’t mean that the system is broken, it simply means that the subject matter may be wrong for that particular person. It may be too trivial, too hard or just not where they are interested in investing their time.

Some will dive in deep, and quit when they have found something they want to pursue in more depth. Some will just skim the content to get a sense of what the subject is all about.

There is nothing wrong with this.

In university, there is a lot of inertia keeping you within the course you have chosen. It’s hard to change course. In a free online course, there is absolutely nothing holding you back from moving on if your time is not being spent optimally.

The second point is maybe more of a personal preference. If I want to learn an API or framework or language, I’m quite comfortable reading the docs. Pretty much anything worth making one of these courses about will have a comprehensive tutorial as well as clear and current API docs. Taking a course by some “Evangelist” just seems like pointless hand-holding.

My favorite Udacity course was CS212 with Peter Norvig https://www.udacity.com/course/cs212. The focus was on HOW to think about the problems you tackle. The choice of subject matter and language was secondary to the underlying concepts and the mental process to work through them. I did learn some python tricks, and more than I need to know about the rules of poker. But that wasn’t what was being taught. It was about choosing good data representations and designing clear functions and abstractions.

And this leads me to the third point: the instructor. Programming can be a pretty isolated endeavor, so the chance to get a glimpse into the mind of a really amazing programmer is a rare thing.

It may sound weird, but it’s really important to take note of how good programmers talk. How they frame their questions. Their pace. The ideas that flow naturally as well as the concepts they step through one idea at a time. Each is different, but they all have something valuable to offer. This is another type of deep learning that is hard to find - even at a good university.

I really hope to see more of this type of thing in the future. Online courses where we can soak in the lessons (and character) of great minds at our own pace.

tags: Programming