Sunday, January 29, 2012

The hope of Udacity?

I've read a little commentary on this and I'm not sure what to make of it. I think it could be a good thing, but I also think that we have a history in the U.S. of too-uncritically embracing "the new" in education because it is NEW rather than because it is GOOD. And that sometimes what must be done to keep quality in something high is overlooked and the "oooh, shiny" aspect is emphasized.

In short: Sebastian Thrun, who was a tenured prof at Stanford, ran an enormous (and enormously popular, apparently) online course in artificial intelligence.

There were 160,000 students in the course.

Yes, you heard that right. All of the grading, all of the exams and such, were entirely automated.

Part of the...I guess you'd say, controversy? Was that Stanford would not "credential" students who took the course on line. (I presume that means they wound up taking it for no transferrable credit). (Another aspect: the course was apparently free...which is kind of amazing. And also is probably why Stanford couldn't grant credit for it.)

Thrun has since left Stanford, and wants to start up an online university (called Udacity, and I admit, the name makes me want to throw up a little, but I hate most cutesy brand names). Apparently Udacity is going to be free, which makes me wonder a little bit: how will they pay their "content providers" (or professors). How will they pay people who do grading in the courses that can't be 100% machine-graded? Will they sell ads, like some online publications do? Will they expect the people writing for it and working with it to do so pro bono? None of the articles I've seen address how Udacity will pay for itself. (Which makes me...not exactly suspicious, but concerned.)

I don't know. I admit, I'm somewhat leery of online teaching. Don't get me wrong: when it's done the right way, it can be very good. It can provide a wonderful benefit for students who work long hours and only have "school time" at hours that conventional programs would not be open. It's probably a Godsend for people with certain disabilities that makes leaving the house difficult. Or for people living in remote areas.

I can see it working well, for example, for a small-ish, writing-intensive class, where students submit their work to their prof, discuss in online fora, maybe even critique each other's work.

But in the sciences there are some challenges. Labs, for example: unless you expect students to come in to a campus once a week (or once a month for a long day), you can't really do labs the way you'd normally do them. Or, at least, I couldn' field labs, for example, are geared to specific sites and require multiple people working together as a group. And in some cases, safety issues would prevent students from doing certain labs at home. And, say what you will, I DON'T think online simulations of stuff are an acceptable substitute...maybe for a few lab exercises, yes, but I think manipulating "real" stuff and dealing with the "real" challenges that sometimes come up in field settings (or even in the lab) are valuable. And having students together in a lab room - I have seen that the students learn from each other, and that one group's problem provides a jumping-off point for discussion or troubleshooting by the other students. And I'm not sure how that would work online.

The other issue that faces online courses is cheating. Especially for classes, like, say, Introduction to Biology for Majors, like I teach. It's an intensive course, the students are expected to master it before they move I can see a motivation to cheat. (I've heard others who teach online classes say, "You just have to assume they're taking tests with their textbook open, and write the tests that much harder.") I know some classes use on-site proctored exams, which would maybe improve conditions, but my campus does not. (Every time we've raised the concern about exam security in online classes, the only response we get is, " know, 80% of students in conventional classes report cheating." Which, to me, is not a response: for one thing, no citation for this is EVER given, so it could be a made-up statistic (80% is depressingly high). Also, it's not clear if that 80% means "most of them tried it once, got busted, and never tried it again" - it would be a lot harder to catch someone cheating online).

(As I said before, maybe online classes are best for the kind where students are mainly evaluated based on papers, rather than exams)

My other concern with online teaching, is that the quality can vary widely. (Well, that's also true of in-person teaching). But it does seem that there's less quality-control in online courses...we've had a few people come in as transfers with online basic-level classes who didn't know jack, despite earning decent grades.

My biggest concern, though, is the uncritical, unquestioning acceptance, the making a one-size-fits-all solution out of any "new" educational theory/process/tool/whatever and telling faculty that they WILL use this tool. The idea of "adapt or die" (except, in this case, the adaptation may not actually confer an advantage).

The other thing that concerns me is that I could see some universities looking at this model, seeing giant dollar signs, and going "We'll get faculty to teach thousands of students! We'll be ROLLING in tuition money!" without really thinking about the challenges or costs of setting up such a system. (I've seen it happen, on a smaller scale, with regular on line classes...we had someone nearly quit because TPTB decided, two days before classes began, to double the enrollment in one of his classes without really telling him.) The problem is that some people are not really aware that teaching online is still TEACHING. That you still have some of the issues to deal with that you have in a face-to-face class (plus the added issue of "I missed the quiz and now it's time-locked, can you unlock it for me" and technical problems on the part of the students). If you're doing anything OTHER than purely machine-graded automated testing, the amount of time required is going to scale upward with the class size. (Oh, how I wish TPTB realized that for in-person classes....)

The other issue with the way the original huge AI class was taught was the lack of grades/credentials. I can see highly motivated people who just want to learn taking classes like these for personal enrichment....but without grades or some kind of final evaluation, you can't tell (as an employer, or grad school, or whatever) if the person who took the class really learned from it or if they slacked their way through it. (And for the autodidactic types: there are already many things out there to feed that noted in the comments on the Joanne Jacobs article - those "Learning Company" CDs and dvds, and, even simpler technology, books.) So I don't know how "new" it is allowing "adult learners" to learn about something they're interested in purely for fun.

I think the "credentialing" will be the biggest issue going forward: how do you "prove" someone learned something? I wouldn't go to a doctor who took all his or her coursework online for no grade; I wouldn't hire a technician for my lab without some kind of clear evidence that he or she knew what he or she was going to be doing. And I don't think most employers will have the time or patience or funds to apply a battery of tests to EVERY new potential hire to weed out the ones who don't know anything. (And as much as Thrun seemed to hate the idea of "weeder classes" - well, you DO need to decide who is good and who is not-good at something, and employ only the good ones.)

A lot of people are either heralding or damning Udacity for "bringing on the death of the university." I think my reaction is "not so fast, amigo..."

No comments: