Wednesday, July 11, 2012

Khan Academy and online learning

Two math professors critique the Khan Academy in the video below. See also the excerpted Chronicle article.

An earlier post on online learning and the future of higher education.

Chronicle of Higher Education: ... Khan Academy is a collection of video lectures that give demonstrations of mechanical processes. When it comes to this purpose, KA videos are, on the average, pretty good. Sal Khan is the main reason; he is approachable and has a knack for making mechanical processes seem understandable. Of course, his videos are not perfect. He tends to ramble a lot and get sidetracked; he doesn’t use visuals as effectively as he could; he’s often sloppy and sometimes downright wrong with his math; and he sometimes omits topics from his subjects that really need to be there (LU decomposition in linear algebra, for example). But on balance, KA is a great resource for the niche in which it was designed to work: giving demonstrations of mechanical processes. ... 
This is not to say that Khan Academy can’t play a useful role in learning calculus or some other subject. I don’t deny that mechanical skill is important for getting to the higher-level cognitive tasks. But mechanical skill is a proper subset of the set of all tasks a student needs to master in order to really learn a subject. And a lecture, when well done, can teach novice learners how to think like expert learners; but in my experience with Khan Academy videos, this isn’t what happens — the videos are demos on how to finish mathematics exercises, with little modeling of the higher-level thinking skills that are so important for using mathematics in the real world. So the kinds of learning objectives that Khan Academy videos focus on are important — but they’re not enough.
I tried out a couple of Khan Academy videos on my kids recently and I thought they were reasonably effective. Khan is not as precise as a real math professor but he gets the message across.

Small nitpick: he referred to negative numbers as "smaller" than positive or less negative numbers (e.g., -100 is "smaller" than 1), which is I think confusing and even a bit misleading. I think he should have used "less than" rather than "smaller". If you are familiar with complex numbers then you'll probably tend to think in terms of a magnitude (big vs small) and an orientation (with negative numbers along the theta = pi direction), so that -100 is not really small (in magnitude) relative to 1. Presumably Khan learned complex analysis at some point in his education (although maybe not, he went to MIT ;-) IIRC he started out wanting to do physics.


jacob said...


Farnam Street posted an interesting video discussing the effectiveness of teaching through video. While quite controversial, it's worth watching.

One of the main arguments seems to be that online videos are "too" effective to stiumulate learning and they should be more confusing.


Richard Seiter said...

The Khan videos might not be perfect, but the point I think a lot of the critics are missing is that they fill a niche quite nicely. I think some of the math instructors going ballistic would do well to ask themselves why Khan seems to be gaining such traction. Is it perhaps because their way of communicating math is not reaching a substantial subset of students? I think the talk of Khan Academy replacing educators wholesale is overblown, but I think it has a valuable role to play. The lack of precision can be a problem (Steve has a legitimate nitpick, I found some of those in the video less compelling). But, as someone who has a problem with digressing during explanations (to add precision usually) I see the benefit of some simplification (even if it is not precise). Also, how precise is the average elementary/high school teacher? Not to mention the average friend/parent who the student might otherwise consult for help? The MST 3000 parody style is amusing, but I worry about criticism like that making "the perfect the enemy of the good".

Steve, I think you would have enjoyed the T-shirt I saw recently (you probably know about this already. I think it was a from a 2005 prank/hack, but I was amused to see it in the wild). Front side "MIT" (to all appearances one of the standard MIT shirts), back side
''Because not everyone can go to Caltech". (full disclosure, I went to MIT, but was admitted to both)

Yan Shen said...

I like the fact that his videos cover a wide range of topics. For the rest of us non-genius laymen, I think the KA videos serve a fairly good purpose. I remember watching a couple of videos related to the subprime mortgage crisis and also a few related to how the foreign exchange markets functioned.
Probably those videos weren't very technically precise, but I think they give the average person an okay understanding of the big picture.

By the way Steve, it really seems like things have calmed down on this blog. The quality of both the commentary and the commentators seems to have improved dramatically in recent weeks! Now, I know that correlation doesn't necessarily imply causation, but I do wonder whether or not anything I did played a role in helping to bring about this Pax Infoprocana... ;)

Stephen S said...

My take on Khan is that it's fine for people who need to be proficient and then forget everything. It's the same with all videos. If you want to _really_ know a subject, you need to commit yourself to _reading_ about the subject, and constantly challenge yourself with problems. Shortcuts won't produce mastery; listening to other people talk will give you a false sense of mastery, inhibiting mastery. Simple as that.

Richard Seiter said...

Thanks for the video link. One thing that drives me crazy about most of the discussions about learning/teaching effectiveness is that they fail to acknowledge that different people learn (best) in different ways. Can anyone help me understand why this is not a larger part of the discussion?

David Coughlin said...

I really only learn when I see the same thing different ways [at least, for complicated stuff].

David Versace said...

RWCG posted the penultimate view of the entire online college/signaling theory issue.
I’ve basically given up on trying to follow along with the Bryan Caplan-led discussion on Econlog over whether college is ‘signaling’. But I realized that I don’t even know who to root for.
Look, the middle-and-above classes have a natural, understandable desire to bequeath an Advantage Stamp to their children with their wealth. (What else to do with wealth?) Naturally and understandably, they wouldn’t want to spend $200k’ish on the thing if the Advantage given by the Stamp depended solely on the actual talents, brains, and efforts of their kid. What good would that be? No, to have any value, the Stamp has to work on some baseline level, independently of the particulars of their actual kid. So it seems axiomatic that college is, at least in part, about ‘signaling’. If college didn’t have at least some signaling component the middle-and-above classes would just demand some other wasteful service or credential which did. They will go out and purchase some Thing that differentiates their kids from the others. It doesn’t have to be college but if it’s not they’re going to waste their money on something else.
This dynamic may be what underlies the ‘occupy’ resentment, with its focus on student loans and lack of opportunities (read: six-figure high-status white-collar opportunities) for graduates; as the value of education has been diluted by federally subsidizing universal college attendance, many are finding the Advantage Stamp they thought they’d purchased (with borrowed money) to be not so valuable as they’d been led to believe. The frustration is understandable and it’s fair to point to the misguided campaign to universalize college as a culprit.
But let’s say the Caplanites win this debate and halt that project. The demand for the Advantage Stamp would not go away. Either college is what currently serves this function, or something would have to take its place (and would be equally unshy in clamoring for public funding to purchase it). Would that thing be better for society than the four-year-summer-camps we’ve already set up? I’m not optimistic. At least college as an institution has some history going for it. Do we really want to find out what sort of Alternative Advantage Stamp will be cooked up (and then demand federal funding) by people like, say, trust-fundee Edward T. Hall III in Occupy laboratories?

Emil 'Deleet' Kirkegaard said...

There is also duolingo for learning languages. Usually from an english perspective. More are being added. I hope they add Esperanto. Thus having a more gamified tool than

Bobdisqus said...


Blog Archive