Tuesday, April 30, 2013

Amherst College says "no" to edX, but how about Google MOOCs?

The Amherst College faculty voted 70 to 36 with five abstentions against joining edX, MIT's MOOC consortium, but they did not dismiss the idea of MOOCs. They passed a motion proposing that Amherst explore online teaching technology outside the context of the major MOOC groups.

The major MOOC platforms in the US -- edX, Udacity and Coursera -- are trying to build their brands by offering courses from "elite" universities. Most faculty senates will not have to consider the question of partnering with these organizations because they will not be invited in.

Furthermore, few universities have Amherst's $1.64 billion endowment, -- they could not afford edX even if they were invited to become a partner. EdX charges $250,000 to initiate a course and $50,000 each additional time the course is offered. They also take a cut of any revenue the course generates.

Does that mean that all but rich, elite universities are shut out of the MOOC game?


A university can develop and experiment with MOOCs or using a MOOC platform for smaller online classes or to supplement or flip face-to-face classes using open source software. Stanford and edX have combined their MOOC platform software in an open source project and Google also has an open source MOOC platform, Course Builder.

Amherst, or any other university, can explore online teaching by becoming familiar with these open source programs and providing a MOOC platform for use by their faculty. Many professors would ignore the resource, but others would begin using it. The university would gain experience with MOOCs and the way a MOOC platform can be used in a smaller class.

Even better, a university or university system could host a MOOC platform and offer accounts to faculty on multiple campuses. From there it is a small step to a public MOOC platform.

But, a public MOOC service would involve significant expense for servers, bandwidth and staff. How might it be funded? One could argue that such a platform should be funded by government as part of its education mission. Failing that, the usual advertising-based or "freemium" business models might support an open MOOC service.

How about Google? Google is fond of "moonshots" like self-driving cars, and any Internet use benefits their advertising business. For example, they may (or may not) be planning to expand their Google Fiber project and become nationwide Internet service providers. If the ISP venture broke even, Google would still make money by increasing the good-will value of their brand, improving their advertising quality and increasing advertising revenue.

Does that apply to MOOCs as well? Google has their own MOOC software, which they have field tested in running their own MOOC. They also have experience running very large services like Gmail and Google Docs, which are free for individuals and generate revenue from organizations.

How about adding Google MOOCs to the list?

Blaise Pascal argued that a rational person should believe in God because the cost of not doing so and being wrong was infinite while the cost of doing so and being wrong was small. The same goes for MOOCs. If MOOCs are the Next Big Thing, Amherst will be ready. If not, they will not have lost much. They may not even have to invest in their own MOOC platform -- their faculty might just open free accounts at Google MOOCs.  Think of it as YouTube for classes.


Update 5/1/2013

Inside Higher Ed ran an article on the edX discussion within the Amherst faculty (http://bit.ly/10s3Z4R).  It gives the pros and cons as seen by a number of faculty members.  It's based on an internal Amherst report that is not publicly available.  Also has quite a few comments.

Monday, April 29, 2013

Charlie Rose roundtable discussion on online education and MOOCs

Charlie Rose moderated a worthwhile round-table discussion about online education and MOOCs with Anant Agarwal, CEO of edX; Amy Gutmann, president of the University of Pennsylvania; Joel Klein, former New York City Schools chancellor and CEO of Amplify and Tom Friedman of the New York Times.

The panelists are all MOOC enthusiasts and the discussion is uncritically optimistic. They tout the upside and brush difficulties under the rug, but don't let that put you off.

As the panelists point out, today's MOOCs and Internet technology are still version 1.0. Regardless of today's MOOCs, they, like the Wright Brothers first airplane, are attracting attention, capital and smart people. I think MOOCs (and modular courseware in general) have kicked off a wave of innovation that will transform Internet-based teaching. We are beginning to depart from traditional online teaching that was modeled on the classroom and lecture hall.

(The roundtable video is 38 minutes long. If you would rather just listen to the .mp3 audio while working out at the gym or washing dishes, you can download it here).

Wednesday, April 24, 2013

The tactics and cost of the US telecommunication cartel

For a quick (six minute) summary of the cartel's tactics and their cost to the nation, listen to Brooke Gladstone's interview of David Johnston, author of the book "The Fine Print: How big companies use plain English to rob you blind."

Gladstone is co-host of OnTheMedia, an award winning NPR program that covers the Internet and other media.

(I've been ranting about the excesses of the telecommunication cartel for several years).

Grading the MOOC University

Maria Montero sent me a link to A. J. Jacobs' student-centered review of MOOCs, which appeared in the New York Times. Jacobs enrolled in 11 highly diverse MOOCs and finished 2 -- a typical browse/complete ratio.

Jacobs was also interviewed on NPR's Talk of the Nation (transcript and audio recording).

Jacobs "graded" MOOCs as follows:

The professors: B+
Convenience: A
Teacher-to-student interaction: D
Student-to-student interaction: B-
Assignments: B-
Overall experience: B

Jacobs' conclusion:

As these online universities gain traction, and start counting for actual college course credit, they’ll most likely have enormous real-world impact. They’ll help in getting jobs and creating business ideas. They might just live up to their hype. For millions of people around the globe with few resources, MOOCs may even be life-changing.

As for whether MOOCs will ever totally replace colleges made of brick, mortar and ivy, however, count me as a skeptic. A campus still has advantages for those lucky enough to afford the tuition — networking being one. (Even dropouts like Mark Zuckerberg made key social connections at Harvard.) And an online college will never crack Playboy’s venerable annual list of top party schools.

Friday, April 19, 2013

Google Fiber plus ubiquitous, free WiFi -- an offer your mayor cannot refuse

Google Fiber started in Kansas City. Based on that experience, they expanded to nearby Olathe, Kansas and hi-tech cities Austin, Texas and Provo, Utah will be next.

There has been a lot of speculation about Google's intention. Is Google Fiber just a proof of concept designed to spur the incumbent ISPs on? Were they picking hi-tech cities hoping to see some futuristic application development? Or, were they planning to become a nationwide ISP?

We got a hint when Eric Schmidt said Google Fiber was a "real business" and we read estimates of the cost of the Kansas City network and of a nationwide build-out.

Jason Calacanis has no doubt about Google's intention. His latest blog post is entitled "Google's Fiber Takeover Plan Expands: Will Kill Cable & Carriers."

You should read it for yourself, but I want to focus on one claim he makes:

More importantly, every Google Fiber home will have a public wifi component. In order to get Google Fiber, you’re going to have to agree to put a router in that lets anyone use a portion of your bandwidth.

That’s not announced, but it’s gonna happen.
His contention is bolstered by the fact that Google is an investor in Fon. I wrote about Fon a few years ago in a post about people owning their own Internet infrastructure and providing decentralized backhaul for Wifi. Fon gave users free, open WiFi routers, but there were two problems.

For Fon to succeed, the open modems had to be ubiquitous. I was an original "Fonista," but there were only a few others in my part of Los Angeles. The second problem was backhaul speed. I had a slow DSL connection at the time. If a lot of neighbors and passersby had connected to my Fon router, it would have impacted my connectivity.

Those problems disappear if Fon is tied to Google Fiber. Google Fiber is such a good deal that it would become ubiquitous and, if you have gigabit service, you won't notice the load imposed by WiFi users.

Google needs cooperation with cities if Google Fiber is to succeed. One "carrot" they have been offering is free connectivity for community sites like libraries and hospitals. What if they sweeten the pot with ubiquitous WiFi? That is an offer the mayor cannot refuse.

Google has another asset. Recall that the decision to start in Kansas City was based on a proposal from the city. I don't know how many proposals Google received or what the cities offered, but Topeka joked about changing the city name to Google and Kansas City offered significant subsidies. I bet Google got some sweet offers.  I know they got a sweet sales-lead list.

If Jason Calacanis is right, and I do hope he is, this is the beginning of the end of business as usual for wired and wireless ISPs (aka cell phone companies).


Update 4/26/2013

Time Warner Cable says they will offer free Wi-Fi in Austin (http://bit.ly/Y1vSUX).  It seems they got Google's message.  That being said -- I wonder whether they will be able to deliver.  What will be their backhaul strategy?

Wednesday, April 17, 2013

Lies, damn lies and open data -- Reinhart and Rogoff challenged

This is one of several cool XKCD cartoons on
statistics. See http://bit.ly/XGLLju for others.
In 2010, economists Carmen Reinhart and Kenneth Rogoff published "Growth in a Time of Debt." They analyzed national income statistics from 44 nations over a span of as much as 200 years. Their dataset incorporated over 3,700 annual observations.

Their main conclusions were:
  • Above 90 percent debt/GDP, median growth rate falls by one percent, and average growth falls considerably more.
  • In emerging markets, "When external debt reaches 60 percent of GDP, annual growth declines by about two percent; for higher levels, growth rates are roughly cut in half."
  • For the advanced nations as a group, there is no apparent contemporaneous link between inflation and public debt levels, but in emerging markets inflation rises sharply as debt increases.
The study was influential. It has 450 Google Scholar citations and was widely quoted by Paul Ryan and other politicians arguing for austerity.

But, a new paper by University of Massachuets economists Thomas Herndon, Michael Ash and Robert Pollin analyzes the same data and comes to different conclusions.

This does not mean Reinhart and Rogoff are charlatans or Ryan and other politicians are liars. Ryan (hapilly) accepted Reinhart and Rogoff's conclusions as fact and Reinhart and Rogoff built a number of questionable assumptions into the spreadsheet they used to analyze their data. They were also careless, evidently not catching a spreadsheet coding error.

But, there are many ways to analyze any multivariate dataset. Data analysis is like the proverbial blind men describing an elephant while touching different parts. A study like this does not give The Answer, it gives an answer.

The best way I can think of to cope with such complexity is to let many people analyze the same data. As the open source software aphorism states "given enough eyeballs, all bugs are shallow."

While I may not agree with the assumptions they made in their analysis, I applaud Reinhart and Rogoff for publishing their data and spreadsheet. You can download them and do your own analysis.


Updated, 4/21/2013

Paul Krugman has weighed in on the Reinhart and Rogoff affair, commenting on its practical political and economic influence in a column on "The Excel Depression" (http://nyti.ms/14IrRd5). I typically agree with what Krugman writes, including this column, but its title is a bit over the top.


Updated 5/31/2013

Reinhart and Rogoff have come in for more criticism -- does debt cause slow growth or vice versa? See details of the data analysis in this paper by University of Massachusetts economist Arindrajit Dube.

One might also compare recent performance of the Japanese and pro-austerity European economies.

Monday, April 08, 2013

Presidential science initiatives (including the Internet)

President Obama has called for initiatives to map the human brain and bring an asteroid to Earth for study.

What a breath of fresh air to see a president who is thinking beyond today's simple minded political jousting -- even if the proposed budgets are just a down-payment for planning and design.  It is reminiscent of President Kennedy telling Congress:

I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the moon and returning him safely to Earth.
(I get kind of misty typing that quote).

Of course there have been other noteworthy presidential efforts -- check the "Progressive Professor's" list of the top 15 presidential science investments, starting with Thomas Jefferson's sponsoring of Lewis and Clark.

Although no single president can take the credit, I would add a sixteenth to the professor's list -- the funding of the research and development that led to the Internet. The US taxpayers got quite a bargain -- this is what it cost us:

One can argue that this $124 million was only for research and development and we should also count the money spent on government procurement -- most notably SAGE, a networked early warning system for manned bomber attacks (video), which cost an estimated $8 billion. (The SAGE video is a bit off topic for this post, but it highlights technical progress beautifully .  Even if you include SAGE and prototypes like MIT project Whirlwind, the networking investment has paid off handsomely.


Update 4/9/2013

Prominent scientist thinks the brain map should have more funding.  Harvard professor George Church, a veteran of the human genome project and an advocate for this study says the project should start small and have funding from several agencies. be ambitious and focus on developing new technologies from the start.  Lots more ... this is a recommended article.

Update 8/13/2014

For Vint Cerf's description of the ARPA/Internet project, see this post.

Update 4/27/2015

In this video of his keynote presentation at the SCI Institute, Alan Kay describes the ARPA funded research that led to the development of the Internet and the personal computer. The talk begins with the visions of people like Vannevar Bush and JCR Licklider and runs through the years at the Xerox Palo Alto Research Center (PARC).

Kaye concludes that the work at PARC cost Xerox about $40 million in 2012 dollars -- funds for around 30 researchers for four years. He estimates that Xerox made 300 times its that through sales of laser printers, even though they failed to capitalize on other PARC inventions. He estimates that the payoff to the US and world from all the PARC inventions was around $33 trillion by 2012.

The entire talk is over and hour, followed by questions and answers. The portion in which he talks about PARC and estimates the return on investment, begins at the 1 hour, 1 minute point.

Update 10/22/2015

The US National Economic Council and Office of Science and Technology Policy have published a Strategy for American Innovation.

It is a long document with both strategic and specific points. This is one paragraph from a section on the Federal Government’s Foundational Role:
When U.S. companies develop a breakthrough product like a smartphone, it is appropriate to celebrate American firms and workers. But it is also important to recognize the value of the decades of Federal investment in R&D that provided these new products’ scientific and technological foundations, such as the Internet, the Global Positioning System, speech recognition, electronic design automation for advanced microprocessors, and artificial intelligence for virtual assistants. Although companies must ultimately invest a great deal to commercialize emerging technologies, the new insights, early prototypes, and the first markets for them are often supported by the Federal Government. Absent Federal investment, many new products would not ever reach the market, let alone reach world-changing scale.
The smartphone example is taken from Mariana Mazzucato's “The Entrepreneurial State," and provides a good example of the role of the Federal Government in research and procurement of early, limited-market products.

Saturday, April 06, 2013

Combined Course2go and edX platforms will lead to hosting services

When MIT announced the creation of edX, they promised to open source their MOOC teaching platform.  Last September, both Stanford and Google open sourced their MOOC platforms, Course2go and Course Builder.

Stanford and edX have announced that they are merging their platforms.  Stanford's Course2go is now in maintenance mode and its features will be moved to edX by June.  I wonder what Google plans to do.

This is good news -- it means schools will be able to offer MOOCs without using the services of a commercial firm like Udacity or Coursera.  A school or a system like the California State University could host MOOCs using the edX platform.

The next step will be hosted services open to any teacher or anyone else who wants to teach a class -- MOOC or other.  A thousand flowers will bloom.  How will this affect course management systems like Blackboard and Moodle?


Update, May 30 2013

The Stanford/edX MOOC platform source code will be posted tomorrow.

We have a few hints: developers have the source code for the edX learning platform, including code for its Learning Management System Studio, a course authoring tool; xBlock, an application programming interface (API) for integrating third-party learning objects; and machine grading API's.

That will be the second major open source MOOC platform -- Google's Course Builder is already open source (http://bit.ly/ZmhRyy).

Will this lead to the democratization of MOOC and MOOC module development? Are you planning to test either of these platforms? Will Google or edX make hosted versions available at some point in the future? Will someone else?

Friday, April 05, 2013

Is there a place for online training companies in the MOOC discussion?

My daughter teaches an upper division finance course, in which spreadsheets are an important tool. In order to take her class, the students must have passed two classes -- an introductory information systems course, which includes learning to use Excel, and a prerequisite finance class.

She tells me that a number of her students do not understand that you can enter a formula into a spreadsheet cell. Others may understand that they can enter formulas, but lack basic arithmetic skills like understanding percents and rates or converting from one unit of measure to another.

It turns out that they do gain rudimentary skill with Word and PowerPoint in the introductory course (or before they get to college), but we are doing a bad job of teaching rudimentary spreadsheet skills.

What can we do? One approach is to put off the introduction to spreadsheets until they are needed in an accounting or finance class, and present them there, but that would lead to curricular redundancy.

How about a short course on spreadsheets that focused on just the features needed for subsequent classes? That sounds like a good idea to me, but it is less than a typical college course -- perhaps one unit. That would be administratively awkward and, more important, we do not seem to be capable of teaching spreadsheet skills to a significant number of our students.

How about a spreadsheet MOOC? That would pay off in terms of economy of scale, and we should be able to spend the resources on it to become really good spreadsheet teachers. We could even make the course modular so a given school could elect just the spreadsheet features they wanted their students to master.

Cool, let's do a MOOC, but who should teach it? Today's MOOCs are offered by university professors, but would it make more sense to turn to an organization with expertise in training?

I Googled around and found two organizations with Excel training online, Udemy and Lynda.com. Would they be better candidates for teaching a spreadsheet MOOC than a university? Would they be willing to sell wholesale to a university -- giving us reduced prices for, say, 100 students and tailoring the course to the features we want covered?

Udemey charges $99 for unlimited lifetime access to their 16-section Excel course. What would they charge for, say, 100 students who had access to only 8 of the sections for one semester?

Lynda.com charges $25 per month for access to all of their training material. What would they charge for, say, 100 students who only had access to their Excel course for three months?

Universities around the world are beginning to offer MOOCs and we are thinking about giving credit for completion of MOOCs offered at other universities. Perhaps we should also think about giving credit for or requiring completion of courses taught by training companies.

Can we grade and give feedback on college essays automatically?

I am skeptical, but a New York Times article says the folks at edX will be doing just that using "artificial intelligence." They also say students will be able to improve their essays using feedback from the system. EdX promises to open the grading platform to others -- presumably as a Web service.

EdX has hired Vik Paruchuri to work on the service, which he developed for an automated essay grading contest sponsored by the Hewlett Foundation. The contest results are reported in a paper by the contest organizers, which concludes that automated essay scores correlate well with those of human readers. (Before you settle for that, read this critique of those results by MIT professor Les C. Perelman).

The Times article does not contain a link to a service and I can't find one on the edX web site, so I will remain skeptical. Grammar checking? For sure. Meaningful feedback? Show me the API or URL.

Could we afford a full-time teacher for one course?

We've heard suggestions that MOOCs might enable us to take courses offered by all-star professors at elite universities. Why take a course from a professor at a local state university when you could take a course from a renowned Harvard scholar?

But, would that be a good use of resources?  Albert Einstein was too busy to teach a physics MOOC.

Government subsidizes education for the benefit to the society and economy.  If we assume that MOOCs work well for some courses and some students, government should be willing to foot part of the bill.

In an earlier post, I did a back-of-the envelope calculation speculating that a MOOC might generate annual revenue of, say $200,000. In my estimate, most of that revenue would come from students seeking entertainment and enrichment, not certification.

What if the government would match that, allocating another $200,000 for a MOOC. At that level of funding, could we cover the overhead -- management and administration, production, software refinement, community facilitation, hosting, bandwidth, etc. -- and have enough surplus to hire an outstanding, motivated teacher and perhaps a TA to spend full time presenting and improving a single course?

If you are a good teacher at a state university, community college or high school, think of how well you could teach a course you were interested in if that one course were your full time job and you had support staff.

Thursday, April 04, 2013

MOOC monetization -- a free sample strategy

The high cost of education has been one of the drivers of interest in MOOCs, and nearly all MOOCs have been free up to this point. We are in the "invest to get users" stage.

I've heard speculation on ways to monetize MOOCs, and would like suggest one I've not encountered -- the free sample model.

We are starting to get some data on MOOCs. For example, Katy Jordan pulled together completion rate data on 26 of the MOOCs that have been offered by Coursera, EDx and Udacity:

MOOC critics point to these low completion rates as evidence that MOOCs are a Bad Idea, but those large enrollment numbers are still encouraging, so let's look further. Here is a plot of the number of views of the first and last videos each week in the Bioelectricity course at Duke University:

As you see, it is long-tail, y=1/x type graph. One way to look at the long tail is to conclude that the majority of enrollees were dissatisfied and the course was a failure. But an alternative explanation is that the folks were not really enrollees, they were browsers, and, like shoppers in any situation, most decide not to buy.

We can also think of their motivation. Maybe some were not interested in "completing" the course and getting "credit" -- they wanted something less complete, less formal. We have data on that from the Duke course as well. As shown here, fun and enjoyment was the strongest motivator.

Similar motivation was reported by students in a programming class. Sixty four percent of the students said they were interested in the course out of personal interest and curiosity, 33% for professional advancement and only 3 percent for university studies.

These three studies are small and limited, but perhaps we should be thinking about MOOCs as entertainment and lifelong learning as well as for formal education and vocational training, which brings us to free sample pricing. We can view enrollment over time as follows:

The spike in the first few weeks is mostly browsers -- people who want to see what the course is like and what the workload is. They would be getting a free sample.

After a few weeks, the enrollment decline slows and we are left with the students who are interested in completing the course. As we have seen, some are seeking formal credit and others entertainment and the joy of lifelong learning. Those people would presumably be willing to pay for the course. How much?

San Jose State University is currently running a trial in which students will get credit for Udacity courses for $150. The results of that pilot study are not yet in, but $150 is less than the University price.

How about the non-credit students? Let's say a typical MOOC runs 10 weeks and the first three weeks are free. What would people pay for seven weeks of entertainment or lifelong learning? Would it be as much as a movie? A hardback book? A steak dinner?

We can noodle around with a simple spreadsheet and get some napkin-sketch revenue estimates. For example, let's assume a for-credit price of $50 for a certificate of completion and a non-credit price of $20. The median number of students completing the MOOCs shown above was 2,777. Assuming that ten percent of those are for-credit students and the MOOC can be offered four times a year, we get annual revenue of $214,000. If we use the mean completion rate of 4,447, the revenue estimate increases to $343,000.

Note that under these assumptions, the bulk of the projected revenue is from the entertainment students, not for credit students.

If an organization offers several courses, are numbers like that sufficient to cover the overhead of management and administration, production, software refinement, community facilitation, hosting and bandwidth, etc. and pay a teacher who is full time on a single course?

A question for any good teacher reading this -- how effective a teacher would you be if you only taught one course and spent full time delivering and improving it?


Update, 4/17/2013

Coursera revealed that they had revenue of $220,000 in the first quarter after starting to charge for verified completion certificates (http://bit.ly/12nd8Rb).  The income was from students paying between $30 and $100 per certificate.  Coursera co-founder Daphne Koller said prices were "around" $50.

This gives us a first cut estimate of what someone hoping to get job or school credit for a course might be willing to pay.  No doubt, the going price for certified completion will be higher than the price for un-certified completion for entertainment and curiosity.

(Coursera uers who pay for certification have to submit a photo ID of themselves to the company and are also tracked based on their “unique typing pattern” to ensure that people who take tests or turn in assignments are who they say they are).

Update 5/23/2013

I posited a simple characterization of people signing up for MOOCs as either "browsing" or "enrolling." Phil Hill presents a more nuanced characterization in his blog post on blog student types, classifying students as no-shows, observers, drop-ins, passive participants and active participants. Details are presented in the paper called Deconstructing Disengagement: Analyzing Learner Subpopulations in Massive Open Online Courses.

Update 11/27/2013

This post in the Chronicle of Higher Education argues that MOOCs have failed in higher education, and are better suited to vocational training and lifelong learning.

Update 12/1/2013

Prominent MOOC provider Udacity has shifted emphasis from college credit courses toward vocational training and lifelong learning.

Update 12/9/2013

A University of Pennsylvania study of a million enrollees in 16 MOOCs shows they have relatively few active users and low completion rates.

  • Course completion rates are very low, averaging 4% across all courses and ranging from 2% to 14% depending on the course and measurement of completion.
  • Across the 16 courses, completion rates are somewhat higher, on average, for courses with lower workloads for students and fewer homework assignments (about 6% versus 2.5%).
  • Variations in completion rates based on other course characteristics (e.g., course length, availability of live chat) were not statistically significant.
  • The total number of individuals accessing a course varied considerably across courses, ranging from more than 110,000 for “Introduction to Operations Management” to about 13,000 for “Rationing and Allocating Scarce Medical Resources.”
  • Across all courses, about half of those who registered viewed at least one lecture within their selected course. The share of registrants viewing at least one lecture ranged from a low of 27% for “Rationing and Allocating Scarce Medical Resources” to a high of 68% for “Fundamentals of Pharmacology.”