What does a contract with Coursera look like?

by on July 29, 2012 at 3:33 am in Economics, Education | Permalink

The contract [with University of Michigan] reveals that even Coursera isn’t yet sure how it will bring in revenue. A section at the end of the agreement, titled “Possible Company Monetization Strategies,” lists eight potential business models, including having companies sponsor courses. That means students taking a free course from Stanford University may eventually be barraged by banner ads or promotional messages. But the universities have the opportunity to veto any revenue-generating idea on a course-by-course basis, so very little is set in stone.

And this stunner:

When and if money does come in, the universities will get 6 to 15 percent of the revenue, depending on how long they offer the course (and thus how long Coursera has to profit from it). The institutions will also get 20 percent of the gross profits, after accounting for costs and previous revenue paid. That means the company gets the vast majority of the cash flow.

The full story is here, and for the pointer I thank my colleague Debra Lattanzi.

dan1111 July 29, 2012 at 4:09 am

Here Lattanzi offers first impressions from taking a course: http://livingethnography.wordpress.com/2012/07/25/stalking-distance-ed/

The kicker: all of the grading is by other students in the class! (That is how you can have one professor and 50,000 students).

This model is in an untenable position with relation to the existing system: The cost of running a single traditional university course is hundreds or even thousands of dollars per student; the true cost of this model is pennies per student. Something has to give: either this will collapse because it is worthless, or it will destroy the existing system by making it worthless.

This model would spell doom for most professors, also. Lecturing would become like professional sports or acting: only an elite few hundred would be needed to teach the entire nation.

Roy July 29, 2012 at 6:50 am

I’ve always thought it was funny how in sciences the lab section where the vast majority of interactive teaching is conducted are almost always taught with almost no supervision by grad students.

Andrew' July 29, 2012 at 10:16 am

This is viewed as a negative and a dirty little secret. Maybe it is really a positive. Maybe we need more grad student teaching and maybe they should get credit for it rather than the pretense that they don’t exist and professors do all the “real” teaching.

Keith Williams July 29, 2012 at 2:15 pm

Yes. Among the many things that I have proposed: formally partnering grad students with advanced undergrads, and partnering senior undergrads with new undergrads, etc. This structure would magnify the advising and teaching capacity, and introduce a sort of buddy system of mentorship. Just another of my ideas that is too far out there for such a conservative institution, apparently ;)

In my case, I worked in an experimental physics lab literally from day 1 as an undergrad, and that was perhaps the most important thing in my entire college education… comparable in value only to my study abroad experience.

Andrew' July 29, 2012 at 5:27 pm

It’s not conservatism, it’s territory protection of their monetization strat. Grad students teach, for sure, and then only 15% or so of GRADUATES score tenure track positions. Something is obviously inconsistent with that and what people think is going on.

Non Papa July 30, 2012 at 12:45 am

There are a couple reasons why it will be hard to squeeze more labor out of graduate students:

1) Graduate students have little incentive to teach. Many grad students are facing the most competitive labor market for tenure-track positions ever — the percentage of newly-minted PhDs who find TT jobs in this first year on the market is under 20% in some fields. The doctrine of “publish or perish” is more important now than ever, especially because candidates distinguish themselves based on their research, NOT their pedagogical ability. Teaching is a huge time-sink when grad students need to do everything they can to get published; many students are already doing grunt work as research assistants in order to curry favor and get their name slapped on a paper somewhere.

2) The future of the TA labor market looks grim. There are two factors here: the increasing push towards unionization of grad students at public universities and the unwillingness of schools to add more PhD lines. The unionization push has been in the works (mostly on the state level) since 2000 but it looks like the NLRB is now poised to allow collective bargaining across the country; I am not sure what the exact consequences of TA unionization will be, but I’m pretty sure it won’t be more teaching hours. There’s also no reason to expect that budget-conscious administrators will respond to greater demand for TAs by opening the floodgates to new grad students. Most deans I have spoken with talk about consolidating enrollment, especially in programs that don’t bring in tuition or research dollars to offset the cost of instruction (read: humanities and social sciences).

Orange14 July 29, 2012 at 2:57 pm

ROFL! When I started grad school in chemistry at a Big 10 school back in 1970, we were thrust right into a teach assistantship position with no training at all! It was assumed that we would know what to do. I had responsibility for one discussion session (1 hour) and 2 four hour lab periods per week for freshman chemistry. What was most appalling was the lack of writing skills of the students, so much so that I had to give a couple of grammar tests and tell the students that poor writing would result in points being taken off. It was amazing how the quality of lab reports improved!

Kevin P. July 29, 2012 at 4:53 am

I think you’re missing a link, Dr. Cowen. Is the full story here? http://chronicle.com/article/How-an-Upstart-Company-Might/133065/

Keith Williams July 29, 2012 at 9:02 am

MOOCs target the introductory curriculum- the 101 classes that are typically very large. The irony is that these courses are already the least expensive for a university to provide (hundreds of students, one instructor; a few TAs). And most college students will have only a few of these “stadium” courses. Hence the impact of agreements such as those with Coursera appear to be very slight.

As for why these things are coming to pass, my own conviction is that we now have “two lane” traffic though American colleges: well prepared students bring ample AP credit and bypass intro curriculum, proceeding directly to advanced coursework. These “fast lane” students reach the more compelling material and may never see a stadium course or a MOOC. Meanwhile, students with less K-12 prep languish expensively in intro courses that frankly have no place at the college level. Thus it is the prep problem that has brought the MOOCs upon us.

The importance of the two lane supposition is that it shows how two different kinds of students at the same institution can have two very different outcomes- not just in college but also in the job market and as they pay off student loans. I wrote a bit more about this here: http://keithwms.blogspot.com/2012/07/americas-two-lane-higher-education.html

Andrew' July 29, 2012 at 9:57 am

Do we really need those small number classes? Or most departments for that matter? I think college doesn’t really do specialization because they can never approach the real specialization required by jobs. They do a faux specialization such as ultra high-level math as a proxy for the ability to learn specialties.

Keith Williams July 29, 2012 at 10:17 am

This question is really too big to answer in this format, but in brief, yes we definitely *do* need the smaller courses, particularly for S&T disciplines. That’s where students get the personal attention and build the mentor relationships that play such a large role in their career advancement.

Now, I am all for building bridges to good jobs through more practical coursework, internships, etc. I have been running myself ragged for years trying to convince certain bowtied administrators of the need for that. Many of our college administrations appear oblivious of their role in the cycle of investment and return, which of course includes employment post graduation. Too many schools are selling credentials that are, basically, expensive hood ornaments on a car with no destination.

The larger problem, as I described in the blog link above, is that American higher ed has allowed the valuable concept of liberal arts to be corrupted into remedial high school, which in turn takes time away from the more valuable experiences, at enormous cost to the students and their families. Personally, I see absolutely no reason to devote expensive faculty and facility with remedial high school coursework. But: for those few well-prepared students who can bypass the large courses, the college experience can focus on the high-value experiences. So let’s not throw the baby out with the bathwater.

Bottom line, if you see my point about fast versus slow/HOV lane higher ed, then I assert that you will better understand where the traditional/online debate is headed.

Rahul July 29, 2012 at 1:37 pm

I think the bigger problem is that most Professors treat teaching as a necessary evil. Given a choice many would prefer to only do research.

The rewards have become skewed: there’s little glory in being a great teacher.

Keith Williams July 29, 2012 at 2:07 pm

How is that a problem? Would you rather have students participate in research or sit in a classroom? Speaking from my own experience, I learned ~nothing from the latter and certainly didn’t establish the important relationships and communications skills in the classroom.

It’s unfortunate that some now equate college=classroom. Value for practical experience has eroded, particularly in S&T fields. And frankly, those in favor of MOOCs indeed have a lot of ammunition, when it comes to the early / remedial coursework. We are faced with a large influx of underprepared students, guided by the idea that everyone should get a college degree. Hence, students are greeted by stadium classes, and to top it off, they pay a lot for that impersonal experience.

Ted Craig July 29, 2012 at 9:16 pm

“It’s unfortunate that some now equate college=classroom.”

Try most people, especially those who are paying for universities to exist either through their children’s tuition payments or their tax dollars.

Keith Williams July 29, 2012 at 11:54 pm

Ted, I think you’ve taken the “college=classroom” logic and simply extended that to thinking that “college cost = instructional cost.” But if you consider that there is steady demand for more seats in larger classrooms and realize that faculty numbers and salaries have been stagnant, you should conclude that the instructional component is not what’s driving the inflation. Administrative and facilities expenses are driving up the cost, as well as (I allege) the attitude that all kids should go to college, which leads to the two-lane problem that I described somewhere above.

From my perspective, the best argument for MOOCs is actually not the savings on instructional costs, it’s to save on facilities expenses for the schools and for the students. It’s pretty inexpensive to put a bunch of kids in front of one instructor. It’s very expensive to house them on a campus.

Let me recommend “The Fall of the Faculty” by Ginsberg, which nicely expresses just how powerless many faculty feel in all of this. Here’s the amazon link, and if you care to, you can scroll down and see my little review.

http://www.amazon.com/The-Fall-Faculty-All-Administrative-University/product-reviews/019978244X/ref=dp_top_cm_cr_acr_txt?ie=UTF8&showViewpoints=1

Ted Craig July 30, 2012 at 8:38 am

You completely missed my point, Keith. I’m saying that to most people, their view of college is as an instructional institution, not a research institution.

Jeff Trawick July 29, 2012 at 2:00 pm

>MOOCs target the introductory curriculum- the 101 classes that are typically very large.
Two of the classes I took from Coursera were classic upper level undergraduate computer science classes: Design and Analysis of Algorithms (a bit more theoretical than the commonly taught algorithms class) and Compilers. The third is definitely not an introductory class, and is probably not a good candidate for a stadium class either. In Computer Science at least, Coursera has shown that the MOOC format can be used successfully with upper level material.

Keith Williams July 29, 2012 at 2:18 pm

Yes, the situation does vary among disciplines. Obviously, CS is far better suited to MOOCs than some disciplines.

Era July 29, 2012 at 9:21 am

that’s right!

Bill July 29, 2012 at 11:41 am

Waiting for the BakeOff

I would like to see some controlled experiment testing how well students do with various online learning courses….

So, controlling for IQ, ACT, SAT, etc., give several groups different online learning classes, say, for linear algebra, calc I, etc.

Then, give them a standardized test and see which course had the best results.

Bill July 29, 2012 at 11:43 am

…and, it goes without saying, also compare the results with bricks and mortar classes.

Corey July 29, 2012 at 1:01 pm

I’m surprised Western Govorners University doesn’t get brought up more. Getting an IT or business degree from WGU is a great example of a working 95% online school.

Tren Griffin July 29, 2012 at 1:57 pm

The University of Washington and others will supplement the video lectures with testing and quiz sections to provide certification (which is different from education). Others who can’t afford college will use real work deposited in places like GitHub as certification. Bricks and mortar college is fantastic but unaffordable for many. Self-education will get better and better over time and that is disruptive for an incumbent.

Bill July 29, 2012 at 2:54 pm

Students grading students.

Hmmm.

In international distribution, you will hire an independent testing organization to inspect the goods before they are shipped to assure that you have the quality you desire. Buyers and sellers often split the cost, with the buyer picking the inspection agency.

Today, we bundle teaching with testing. The teacher reads and scores the exams.

You could envision a model where someone teaches, and an independent testing organization tests and grades, with the student paying for the testing service if the student wants a grade.

I could also envision that the person correcting the test lives in India.

Orange14 July 29, 2012 at 3:28 pm

Warning – Rather Long Post
During my second year in grad school (Big 10 U) my thesis adviser was teaching the freshman chemistry class and had just written a textbook. He began a collaboration with a prof from the school of education to see if they could improve student performance by giving students multiple exams on each unit and taking the highest score. As I recall there were about 500 students in the course and four units per semester with three exams per unit. If you do the math you see that this means that 6000 exams have to be graded and the only way this can be done is through computerized scoring of multiple choice questions. Guess who was given the task of writing all the questions? Yup, we poor grad students. We started on the task in June prior to the September start of classes. Students would get randomized exams and there was a flexible schedule to take the exams so you didn’t have all 500 students taking the test at the right time.

I was skeptical of this whole venture as I didn’t see how it would be possible to generate enough test questions to make this truly random. When I sat down to write test questions there were only certain types of questions that could be written and one had to carefully anticipate how students would get the questions wrong so that the difference in answers was subtle enough that the correct answer would not be readily apparent (my admiration for those who write SAT test questions rose exponentially after this experience). I and my grad student colleagues quickly found that we could only spend about 1-2 hours a day writing questions as it was very labor intensive and of course took time away from our lab work which is why we were in grad school to begin with. Even though we thought we had a very good set of questions the experiment didn’t work.

The students were very adept at gaming the system. They would do very little studying prior to the first time they took the exam; if they got a top score they were done with the unit and didn’t need to take the exam again. They also quickly figured out there were very few different types of questions and they would quickly develop an infrastructure to share the types of questions so that they were studying against the test rather than actually learning chemistry. This approach would be similar to taking an SAT exam that consisted mainly of questions found in one of the test prep books! Even though we poor grad students alerted my professor to what we thought was going on, he and the ed school professor felt that the students were learning as the grades of this cohort were much higher than was seen in the past five years for freshman chemistry.

Fast forward to the next year. All these top grade chemistry students felt that if freshman chemistry was such a breeze, let’s continue on this track. The enrollment for organic chemistry (second year and a requirement for pre-med, pre-dental, and a bunch of other disciplines) was the highest ever. the outcome was the worst ever. 80% of the students either failed or received a ‘D’ on the first exam and things only got a little better from there. I remember the professor who was teaching organic chemistry that year coming in and asking my professor whether the students learned anything at all the previous year; they didn’t. The experiment ended up a huge failure and things went back to the previous model the next year.

You can count me as a skeptic of this whole effort. What will be worthwhile is making courses widely available to those of us old codgers who want to learn new tricks. I’ve looked at Zeke Emanuel’s health care course from U Penn and it’s pretty well done. I would like to take some macro- and micro-economics courses to better understand some of the current debates and so on. For this I would be happy to pay a nominal fee for access to the course material. This is where I see things going (much like the great courses CDs that have been around for some time now)

Bill July 29, 2012 at 3:41 pm

There are many uses of online beyond teaching the class or old coggers….

You could envision students taking the online free class prior to taking the class in college (sort of a AP class in highschool) with a high school teacher, and then colleges would know what they are getting;

You could envision students taking the Stanford online artificial intelligence class and Stanford picking the [Indian] winners to attend on a scholarship

Or, you could see how students do on a free online class and use the test results as a filtering feature for selecting students to attend your college;

Or, you could use free online classes to sell your textbooks.

Orange14 July 29, 2012 at 3:49 pm

all good points, especially the one on selling textbooks. I was surprised (though I shouldn’t have been) about seven years ago when we took our youngest daughter up to U of Rochester to start college. I thought I would browse the bookstore and see what new textbooks looked like. Of course the chemistry texts hadn’t changed much since freshman and organic are pretty much taught the same way as when I was in school. Price tags however were quite high: freshman chem was $110; organic was $165!

keith Williams July 29, 2012 at 4:07 pm

I especially like your idea of “students taking the online free class prior to taking the class in college.” This could be coupled with entrance exams to ensure that students don’t begin college with too little preparation. This might be an effective way to address the two-lane problem that I described.

On the whole, colleges should aim to *raise* their standards and set higher challenges for students. But that’s the opposite of what we see, and I say that not only because of stadium coursework, but also because of grade inflation, which started long, long before MOOCs.

Paul July 29, 2012 at 7:50 pm

It is worth noting that college professors seem to be the only ones making the argument that small classes and one on one interaction are essential to the college experience. And professors are the ones with the greatest to lose from any other point of view.

If this is true, how come more non-professors do not make this argument?

My personal experience was that class size was uncorrelated with what i learned or what value i received. The best two classes i took were of size 80 and 400; the latter was a physics class with a lab of size 20 taught by a TF more than capable of teaching undergrads.

Professors – it is true that your jobs are at risk, but you might be overestimating the importance of your one on one interactions.

Keith Williams July 29, 2012 at 9:06 pm

It is perfectly natural for teachers to present arguments for smaller class sizes: they know the most about effective teaching and have the most day-to-day interaction with students. I can tell you that a lot of what I do, as a teacher, is a hodgepodge of of mentorship, career and personal counseling, and homework assistance etc. It’s not just writing things in chalk in a classroom. Most of the work is outside the classroom.

Also, I think you have a misperception that teachers are worried about money. I find that quite humorous. Look, anybody trying to make serious money is not going to go into teaching in the first place. I make a pittance for my teaching compared to what I make for other work. Teaching is a service. Very few people make bank on teaching, even if they’re good at it.

Let me turn your question around: who’s making the argument for larger class sizes? Anybody with actual teaching experience? Don’t assume that the people advocating MOOCs have a clue about financial practicalities; scroll back up to the title of this thread. Also bear in mind that some for-profit schools are in deep financial trouble; see for example the recent news from DeVry. Because of financial pressures, there is a lot of experimenting going on and nobody knows for sure where all of this leads.

I do believe there is a very strong case for online coursework, but it is *not* because of larger class sizes per se; rather, it is because the portal can link diverse students and non-students alike, and improve overall access to information. Sure, I definitely see a place for *some* large classes, but again, why do so many assume that the college experience is all about the classroom? Hardly. Some of the best lessons are learned through practical experience… doing those things that can’t be streamed over the internet, e.g. hands-on research.

Non Papa July 30, 2012 at 12:09 am

The benefits of small classes and one-on-one interaction are not actually in dispute. There is a ton of academic literature out there that shows a significant negative correlation between class size and student achievement as measured by grades and exam scores. The problem is how to achieve it. Sure, some pedagogically-minded professors might hope that state legislatures will swoop down, throw bags of money at their university systems, and turn every regional public institution into a small liberal arts college. I think most progressive-minded faculty realize that that’s not going to happen, and that’s where online pedagogy comes in.

Online education still has a reputation as a sterilized, one-size-fits-all environment. Fortunately, all of the innovation in online platforms is leading towards a smaller, more intimate virtual classroom. Look at the products offered by the best-known learning management systems (e.g., Blackboard); they’re moving towards Skype-like videoconference classes, voice authoring (e.g., for language practice), application sharing, virtual whiteboards accessible by the entire class, etc. There’s still a long ways to go as far as personalizing online pedagogy, but most providers (and, slowly but surely, educators) are headed in the right direction.

Barkley Rosser July 30, 2012 at 3:56 pm

This fad is the latest manifestation of the higher ed bubble, and indeed it is cost savings that is driving it. How bad this is will only become clear when every Tom, Dick, and Podunk U. gets on the bandwagon. How strong the pressure for this is was shown at U.Va. where the firing of the prez was strongly related to her perceived slowness to get on the bandwagon, although it appears those BOV members were not on top of what was going on.

This stuff is just a higher tech version of the old correspondence courses, which were also supposed to replace classroom teaching. Yeah.

Dana July 30, 2012 at 8:34 pm

“This fad is the latest manifestation of the higher ed bubble, and indeed it is cost savings that is driving it. How bad this is will only become clear when every Tom, Dick, and Podunk U. gets on the bandwagon.”
It’s difficult for me to see where the cost savings come from, except perhaps in the physical infrastructure of the uni. You don’t need buildings if everything happens in cyberspace. But it seems…unlikely that schools will scale back construction much. The real benefits seem to be in new sources of revenue, but I’m not sure those will ever materialize in the numbers bandied about by online ed boosters. Non Papa’s comment above yours seem instructive on this point: all kinds of technology will make possible a learning experience more or less as good as the classroom. But can you really teach more students via skype-like videoconferences, virtual whiteboards, blah blah, etc etc? I don’t think so, not without *putting in more time*. So really this is about extracting more labor from the academic workforce. Your last comment, on a higher tech version of correspondence classes, is apt. There’s been effusive praise of making lectures available online and reserving classroom time for discussion/individual tutoring. It’s not surprising to me that this is successful–turns out, students learn more when you double instruction time! Why didn’t we think of that sooner? Of course we did–you could do it with a VHS tape of a lecture. But now you can wring twice as much work out of people and assign the gains to technological advancement, even though we’ve been able to do precisely the same thing for at least 30 years now.

Comments on this entry are closed.

Previous post:

Next post: