MOOCs

[bigbadwolf]

MOOCs (Massive Open Online Courses) -- such as Udacity, Coursera, and so on -- have been discussed critically of late:

http://www.theawl.com/2013/01/venture-capitals-massive-terrible-idea-for-the-future-of-college

http://cacm.acm.org/magazines/2012/11/156587-will-moocs-destroy-academia/fulltext

http://iansommerville.com/techstuff/moocs-and-the-future-of-universities/

and here is a review of an introductory statistics course at Udacity:

http://www.angrymath.com/2012/09/udacity-statistics-101.html

In brief, here is my overall assessment: the course is amazingly, shockingly awful. It is poorly structured; it evidences an almost complete lack of planning for the lectures; it routinely fails to properly define or use standard terms or notation; it necessitates occasional massive gaps where “magic” happens; and it results in nonstandard computations that would not be accepted in normal statistical work. In surveying the course, some nights I personally got seriously depressed at the notion that this might be standard fare for the college lectures encountered by most students during their academic careers.

 

[Daniel Duffy]

You can see a course with its own product lifecycle, versions and revisions. Students give feedback, you improve your style, you add new material etc. etc.

I would reckon that a gestation period of > 3 years is needed before all the glitches have been removed.

The real acid test is the quality of the exercises in a course.

I have not looked at these much but how many lecturers have the flair and depth given by Gil Strang of MIT.

 

[bigbadwolf]

You can see a course with its own product lifecycle, versions and revisions. Students give feedback, you improve your style, you add new material etc. etc.

If there are thousands of students, the feedback is probably going to be something numerical (a series of questions that can be responded to on a spectrum of 1 to 5), which will then probably be statistically condensed ("the average response to this question was 2.78"). This is probably not a solid base on which to make concrete improvements to a course (assuming the lecturer even wants to).

I would reckon that a gestation period of > 3 years is needed before all the glitches have been removed.

They may not be glitches but inevitable shortcomings of the "product."

The real acid test is the quality of the exercises in a course.

There's more involved. Some of which is intangible or deliberately not measured by the various metrics.

I have not looked at these much but how many lecturers have the flair and depth given by Gil Strang of MIT.

Strang developed as a teacher probably because of decades of face-to-face teaching and lecturing experience.

The best teaching occurs on a face-to-face and one-to-one basis. And it's more of a dialogue and conversation rather than "teaching." With the mediation of technology, some of the spontaneity, the "meeting of minds," goes out the window. With a one-to-many ratio, the teaching assumes a more lecture-like quality, becomes more "talking at" rather than "conversing with," and becomes more one-size-fits-all.

 

[mfegrad]

this is somewhat disconcerting. i was hoping this would take off, mainly because a) it would help kill off the metastasis that is the college bureaucracy at this point and b) it would be a great thing for continuing education/brushing up on old material.

without having read the article, aside from the snippet, i wonder if the author is familiar with the current level of teaching in undergraduate institutions and how it compares.

 

[bigbadwolf]

without having read the article, aside from the snippet, i wonder if the author is familiar with the current level of teaching in undergraduate institutions and how it compares.

An excerpt from the second link:

The bitter truth, however, is that academic pedagogy has never been very good. It is well established that a professorial soliloquy is an ineffective way of teaching. We do know what works and what does not work when it comes to teaching. Much has been written in the last few years about "active learning," "peer learning," "flipping the lecture," and the like, yet much of academic teaching still consists of professors monologuing to large classes. We could undoubtedly improve our teaching, but MOOCs are not the answer to our pedagogical shortcomings.

And from the third:

Moshe’s statement that ‘a professorial soliloquy is an ineffective way of teaching’ – is sadly true for perhaps the majority of university teachers. It is a particular problem in situations where much of the teaching is done by uninterested graduate students who simply see this as a way to fund their PhDs – once a uniquely North American approach, sadly now spreading elsewhere. The tenure system in the US and the research assessment system elsewhere means that faculty get the message that research takes priority and really holds back those creative people who want to do the best job that they can as a teacher.

The argument isn't that undergrad education isn't gravely wanting (it generally is), but that online education -- touted both as a panacea for poor teaching and as a low-cost alternative -- is even worse.

Techno-fetishism in education goes back decades -- the idea of teaching machines goes back, if memory serves, to the 1940s.

 

[Trev]

They have their place. I've been filling up the next 6 months with a big stack of classes through Coursera primarily and my intention is to lay a good foundation for the coming school year, as well as get back into the learning mindset after decaying in a job for the past 6 years. For my purposes, it's great to have a structured environment to work through multiple subjects at once, rather than staring at a pile of textbooks, and the cost & convenience are obviously a big factor. The three I'm taking right now all have solid & passionate instructors and good materials, but I have quit several past classes early after finding the instructors or materials to be of poor quality. It's a bit hit or miss and you can tell which classes have cobbled together something just to be a part of the "disruption" and which ones have actually put some effort in.

Most classes tend to be more on the introductory side, so replacing a masters degree with Coursera classes is a pipe dream, but preparing for one on the other hand...

 

[Daniel Duffy]

Here is a set of interviews with Cornelius Lanczos.

One interview concerns teaching and writing.

Anecdote: You realise 3 times that you don't understand:

1. as a student
2. when you give a lecture
3. when you write a book

http://www.maths.manchester.ac.uk/lanczos/

 

[Shantanu Kumar]

I have done bunch (around 4-5) courses in total. I would say Coursera has better quality over Udacity.
But all the courses are not of same quality on Coursera also. I really enjoyed Neural Network for Machine Learning from Prof. Hinton. On the other hand Computational Investing from UofW wasn't any good. So I would say quality of teaching varies from course to course.

 

[Shantanu Kumar]

Five Courses Receive College Credit Recommendations

 

[bigbadwolf]

Another stake through the heart of the undead vampire known as MOOCs:

This is not a good sign for online education: 72 percent of professors who have taught Massive Open Online Courses (MOOCs) don’t believe that students should get official college credit, even if they did well in the class. More importantly, these are the professors who voluntarily took time to teach online courses, which means the actual number of professors who discount the quality of MOOCs is probably much (much) higher.

http://techcrunch.com/2013/03/22/72...ses-dont-think-their-students-deserve-credit/

 

[Barny]

I think the failure of MOOCs is fundamentally down to how people view college-level education. What is the point of a degree? It is mostly signalling these days. Not much human capital goes on. People only put up with their degree(s) because at the end of it they get a gold-ticket which they can put on their CV which will make people consider them for jobs they otherwise wouldn't. Nobody gets a degree for the joy of learning, since you can do it better and more efficiently elsewhere. This is the problem with MOOCs. Why would you sit through hours of listening to a man poorly explain to you subjects from their field which they deem important, only to sit some phony exam and get no signalling from it at the end?

You surely wouldn't. If you wanted to learn something, you sit down on your own with a stack of books, a pen and paper and google. This is how real learning happens.

The most valuable part of my degree was being pressured to study things I wouldn't otherwise have done because of the pressure of having to attain a degree. The lectures themselves were a waste of time - I didn't go to 90% of them and in some cases I didn't know what my lecturer looked like when I turned up to their exams.

So in short MOOCs capture all of the useless and bad side of university education and none of the good (discussions and questions with professors at the top of their game).

 

[bigbadwolf]

Agree with your post (and avoided lectures like the plague myself in my time). To the extent that education is a social process, I think it happens on a one-to-one basis between a teacher and a student, unmediated by technology and bureaucracy.

 

[IlyaKEightSix]

Ehhhh...I happen to think MOOCs have their place. It's just right now, they're sort of fragmented. Either you have a bunch of intro or for-fun courses, or you have some really niche advanced course (natural language processing that demands a bunch of Python background no other MOOC offers, or Probabilistic Graphical Models, which just tosses you into some obnoxious matlab material without a background either). So far, only Udacity has really tried to make some sort of coherent pathway, albeit most of their classes are either straight Python dev or web dev (I really, really enjoyed Dr. Norvig's and Dr. Thrun's classes).

I recently finished the data analysis course from JHU (only got certificate of accomplishment b/c I was in Israel and didn't pay attention to the peer-review rubric), and hey, it gave me a little bit of background to random forests and had me implement my first model ensembles. Not bad.

If Udacity can teach people to do machine learning with Python to at least be comparable to R, that'd be wonderful.

Of course, it'd be a dream come true if there'd be a MOOC on the R quantstrat package, or something a bit more in-depth that made Trevor Hastie's "Elements of Statistical Learning" more digestible.

 

[bigbadwolf]

http://moreorlessbunk.wordpress.com/2013/03/27/they-can-hire-one-half-the-professoriate/

 

[IlyaKEightSix]

http://moreorlessbunk.wordpress.com/2013/03/27/they-can-hire-one-half-the-professoriate/

Who benefits when the elite give away their knowledge for free?

The students, duh. I love how that blog posts laments that omg universities are going to vanish! Which IMO is an excellent thing. If universities weren't so widely available, and the only option for most people was to educate themselves on MOOCs, employers would have to adjust by not demanding obnoxious degrees (AKA just another way of maintaining the status quo), and would have to look for MOOC completions instead, which are free and optional, as opposed to expensive and mandatory.

 

[Polter]

http://www.cs.uni.edu/~wallingf/blog/archives/monthly/2013-02.html#e2013-02-17T12_16_01.htm

[Skinner] understands that it is not the *machine* that teaches, but the person that writes the teaching program. And he is better informed than almost the entire current educational press pool in that he states clearly that a "teaching machine" is really just a new kind of textbook. It's what a textbook looks like in an age where we write programs instead of paragraphs.

 

[Daniel Duffy]

Personally, I really enjoyed my undergraduate lectures in maths. But that was Gen X era :D

 

[Trev]

A strange paradox: these professors are simultaneously the most and least accessible teachers in history

http://www.nytimes.com/2013/04/21/o...-the-mooc-university.html?pagewanted=all&_r=0

 

[Daniel Duffy]

Professors are very busy people.

And at the end of that it is not The Society of Saint Vincent de Paul giving the course.

 

[bigbadwolf]

Professors are very busy people.

Well, why? A lot of the time they're busy chasing grant money. Or directing their "research." Or forced into educratic activity (meetings and so on). The ones who suffer are the students. That's why the quality of teaching at liberal arts colleges is ofttimes so superior to that at ranking research universities.

When I was an undergrad in England, the prof teaching us classical mechanics was mostly away in the USA. Likewise for the prof teaching us number theory. In both cases an ill-prepared graduate student took over. The undergrads got shafted.