Sunday, August 12, 2012

Capacity focus, 55: Daphne Koller of Coursera (and other university-led online learning initiatives, etc.) on pointing the way to the possibilities for web based distance education; but, how can such wonderful alternatives pay their way?

 (NB: Two Sigma/Digital learning transformation series
  1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 built up from this)

Daphne Koller of Coursera has given a recent TED talk in which she highlights the revolutionary possibilities for online learning:


(The 20+ minute video is well worth watching and pondering. Observe how in one course the numbers moved from being a popular on-campus class size of 400 to a global phenomenon of 100,000. Note also, Udacity, edX, MIT OpenCourseWare and Khan Academy.)

I am particularly impressed by one slide she showed, building on some 1980's research by famed Educator Benjamin Bloom (the proponent of the well-known education taxonomies), on the effects of shifting from standard lectures to mastery learning approaches (master each topic then move on . . . ) and onward to the dramatic impact of 1:1 tutorials by a skilled person:

Bloom's Two-sigma problem, where a change in approach can have a dramatic impact on typical learning outcomes but the economics to do this using the traditional approaches and technologies are prohibitive. But, notice, what it would be like to have a 98% pass rate instead of maybe a 60 - 70% as "typical" pass rates
 In effect, credibly, moving to the mastery approach would improve learning by one sigma, leading to an acceptable degree of mastery by 84% of students, and moving to the tutorial approach would create acceptable performance by 98% of students for a typical course. 

In other words, we are talking about changing C students into A students simply by shifting how the course is taught, and almost all outright "bomb-out students" would now "pass." That is, the typical grade patterns resulting from education approaches we now use are not measuring real potential -- aka "brightness," but instead the degree of ineffectiveness of the approach. The trap here, is affordability: we could not possibly devote enough tutoring resources to teach each student 1:1, though if say parents could be involved and peers could be involved, we could probably make a big difference. (BTW, this is one reason why home schooling and private tutoring are so often so effective. And if we can figure out a way to turn "every student" into his or her own teacher, that, too, would have a dramatic effect.)

It is worth noting, from Wikipedia's article on the Bloom Two-sigma problem, a tableau of typical results on research on learning methods, factors and achievement outcomes:

Effect of selected alterable variables on student achievement.
Adapted from,[6] Walberg (1984).
Object of change process Alterable variable Effect size [+ sigma value] Percentile equivalent
Teacher Tutorial instruction 2.00 98
Teacher Reinforcement 1.2
Learner Feedback-corrective (Mastery Learning) 1.00 84
Teacher Cues and explanations 1.00
Teacher, Learner Student classroom participation 1.00
Learner Student time on task 1.00
Learner Improved reading/study skills 1.00
Home environment / peer group Cooperative learning 0.80 79
Teacher Homework (graded) 0.80
Teacher Classroom morale 0.60 73
Learner Initial cognitive prerequisites 0.60
Home environment / peer group Home environment intervention 0.50 69

The highlighted block of key interventions suggests that we can get much of the effect of 1:1 tutoring through a more interactive, rich feedback approach that reinforces learning step by step and provides opportunity for mastering and for remediation as necessary to fill in gaps. Indeed, the logical inference on the effect of 1:1 tutoring is that it is not so much an independent cause of the dramatic improvement, but that it trends naturally to promote several of the factors below it in the table, which leads to a dramatic cumulative effect. 

What web based, rich media, interactive approaches can therefore credibly do, is to bring to bear several of these factors, potentially leading to drastically improved learning outcomes.

That means we need to go this way. 

Barriers to doing such, we must simply bulldoze out of the way, the payoff for success is too high and the risk of doing nothing is too high.

And, given that others are already going this way and may well be pushing all sorts of ideological agendas -- often unconsciously, sometimes, deliberately --  with their "education" initiatives,  we had better act before we are acted on.

But, how can such pay its way? 

(And by extension how can the proposed AACCS pay its way? [Cf. also here.])

The Wiki article on Coursera shows an interesting set of possibilities for the so-called business model. Here is the Business Plan section:

Business plan

Coursera and participating schools each meet their own expenses, which are substantial on both sides. Any revenue stream will be divided with schools getting a small percentage of revenue and 20% of gross profits.[7][8] Schedule 1 of the contract between Coursera and participating universities[9] contains a "brainstorming" list of possibilities.[8]
Initial suggestions included:[8][9]
  • Certification: The student pays a fee[clarification needed] to the school which issues certification of completion or adequate performance in the course which Coursera makes accessible in a verifiable format.
  • Secure assessment: Coursera, for a fee, provides testing and verification of identity at physical locations.
  • Sale of information to potential employers: for a fee, and with student permission, access to a database containing information about students and courses they have taken is sold to enterprises.
  • Assessment of competency: for a fee paid by a potential employer or educational institution Coursera would evaluate the competency of a student.
  • Tutoring or evaluation of progress: for a fee an employee or contractor of Coursera provides personal attention, tutoring or evaluating a student's work.
  • Licensing or sale of the learning platform and courses to employers or schools for continuing education or course work, for example, at a community college.
  • Sponsorship: For a fee, firms or foundation would sponsor courses. Only "non-intrusive" advisement of the sponsorship is contemplated.
  • Tuition: after a free trial period tuition would be charged for full access to a course and materials. Another possibility is use of the platform and materials by on-campus, or on-line, students enrolled in the course at the sponsoring institution who are already paying full tuition; in which case a small fee would be paid to Coursera by or on behalf of each student.
As of July, 2012 certification and sale of information to potential employers were being actively explored.[8]
It is also worth clipping Wiki on the Khan Academy model, which shows how a deep-pockets plus broad/shallow endowment model (which BTW supports media entities such as TBN etc) can also make for a viable web based edu entity:
 In late 2004, Khan began tutoring his cousin Nadia in mathematics using Yahoo!'s Doodle notepad. When other relatives and friends sought similar help, he decided it would be more practical to distribute the tutorials on YouTube.[5][7] Their popularity there and the testimonials of appreciative students prompted Khan to quit his job in finance as a hedge fund analyst at Connective Capital Management in 2009, and focus on the tutorials (then released under the moniker "Khan Academy") full-time.[7] Bill Gates once said that "I'd say we've moved about 160 IQ points from the hedge fund category to the teaching-many-people-in-a-leveraged-way category. It was a good day his wife let him quit his job."[8]

The project is funded by donations. Khan Academy is a 501(c)(3) not-for-profit organization,[2] now with significant backing from the Bill & Melinda Gates Foundation and Google. Several people have made US$10,000 contributions; Ann and John Doerr gave $100,000; total revenue is about $150,000 in donations. Additionally, it earned $2,000 per month from ads on the website in 2010 until Khan Academy ceased to accept advertising.[9] In 2010, Google announced it would give the Khan Academy $2 million for creating more courses and for translating the core library into the world’s most widely spoken languages, as part of their Project 10100.[10]

Khan Academy has eclipsed MIT's OpenCourseWare (OCW) in terms of videos viewed. Its YouTube channel has more than 172 million total views, compared to MIT's 38 million. It also has twice as many subscribers, at more than 369,000.[11][12][13]
 A sampler will help:




Indeed, it is worth going on further, to see how Khan's approach works technically (noting along the way the "it's not rocket science" approach, and the way he emphasises open source software . . . some of which I think I am going to take a look at! [Stay tuned . . . ]):
The Khan Academy started with Khan remotely tutoring one of his cousins interactively using Yahoo Doodle images. Based on feedback from his cousin, additional cousins began to take advantage of the interactive, remote tutoring. In order to make better use of his and their time, Khan transitioned to making YouTube video tutorials.[14]

Drawings are now made with a Wacom tablet and the free natural drawing application SmoothDraw 3, and recorded with screen capture software from Camtasia Studio.[15]


All videos (hosted via YouTube) are available through Khan Academy's own website, which also contains many other features such as progress tracking, practice exercises, and a variety of tools for teachers in public schools. Logging into the site can be done via a Google or a Facebook account for those who do not want to create a separate Khan Academy account.


Khan chose to avoid the standard format of a person standing by a whiteboard, deciding instead to present the learning concepts as if "popping out of a darkened universe and into one's mind with a voice out of nowhere" in a way akin to sitting next to someone and working out a problem on a sheet of paper: "If you're watching a guy do a problem [while] thinking out loud, I think people find that more valuable and not as daunting".[16]

Offline versions of the videos have been distributed by not-for-profit groups to rural areas in Asia, Latin America, and Africa.[5][17] While the current content is mainly concerned with pre-college mathematics and physics, Khan's long-term goal is to provide "tens of thousands of videos in pretty much every subject" and to create "the world's first free, world-class virtual school where anyone can learn anything".[5]


Khan Academy also provides a web-based exercise system that generates problems for students based on skill level and performance. The exercise software are available as open source under the MIT license.[18]

Khan believes his academy points an opportunity to overhaul the traditional classroom by using software to create tests, grade assignments, highlight the challenges of certain students, and encourage those doing well to help struggling classmates.[7] The tutorials are touted as helpful because, among other factors, they can be paused by students, while a classroom lecture cannot be.[19]


The success of his low-tech, conversational tutorials—Khan's face never appears, and viewers see only his unadorned step-by-step doodles and diagrams on an electronic blackboard—suggests an educational transformation that de-emphasizes lecture-based classroom interactions.[20]
Notice, this largely seems to be a one-man effort, at least at the first!

Also, let us observe how it is beginning to snowball, with additional supporters and senior grade teaching staff (what Americans mean when they say "faculty") coming in, with a community-sourcing model also being developed:
Not-for-profit partner organizations are making the content available outside YouTube. The Lewis Center for Educational Research, which is affiliated with NASA, is bringing the content into community colleges and charter schools around the United States. World Possible is creating offline snapshots of the content to distribute in rural, developing regions with limited or no access to the Internet.[5][25]

Khan has stated a vision of turning the academy into a charter school:
This could be the DNA for a physical school where students spend 20 percent of their day watching videos and doing self-paced exercises and the rest of the day building robots or painting pictures or composing music or whatever.[9]
A November 2011 grant of $5 million from Ireland-based The O'Sullivan Foundation, founded by Avego MD and cloud computing pioneer Sean O'Sullivan, will be directed to three initiatives:
  1. Expanding the teaching faculty
  2. Extending content through crowd-sourced contributions following a Wikipedia-style model
  3. Developing curricula to help users blend the content with physical teaching through STEM learning
Recent teaching appointees as a result of the grant include Dr. Steven Zucker, formerly of Pratt Institute, and Dr. Beth Harris, from the Museum of Modern Art in New York, to produce art and history content. YouTube video creators Vi Hart and Brit Cruise have also joined the teaching faculty.[26]

A series of summer school camps are planned to start in Northern California from June 2012 to test curricula for real-world schools.[27]
 Udacity's model is a course-based, vid lecture one that is closer to Coursera:

Video lectures with closed captioning in conjunction with integrated quizzes and follow-up homework promote a "learn by doing" model.[22] Each lecture includes built in quizzes to help students understand concepts and reinforce ideas. Programming classes use the Python language; programming assignments are graded by automated grading programs on the Udacity servers.

Enrollment and certificates

Students can enroll in one or more classes before the due date of the first homework assignment. Upon completing the course, students will receive a certificate of completion, signed by the instructors.[22]
 I notice how Python seems to be catching on too.

The edX model is different again, and it seems to in part be a way to generate an observational base for research on education -- so one should be careful about how much one may be monitored. The initiative seeks to . . . :
offer university-level courses from a wide range of disciplines online to a worldwide audience at no charge. The two institutions [MIT and Harvard] have each contributed $30 million to the nonprofit project. Scheduled for launch in Fall 2012, edX will incorporate and build on MITx, a similar project launched by MIT in December 2011[1]. The "learning platform" will be developed as open source software and made available to other institutions of higher learning which wish to make similar offerings; there are plans to allow other schools to offer courses on the edX website also. Plans are to create online learning software that moves beyond videos of lectures to interactive experience.[2] For a modest fee certificates of successful completion will be offered but not college credit.
While we are at it, let us notice the free high school textbooks initiative from South Africa. (The Grade 10 intro Physical Sci book looks great for an intro to Physical Sci precursor to Physics and Chemistry, and comes with vids etc. A good model. Curriki is also worth a look.)

H'mm, that venerable institution, the textbook is in the cross-hairs:
Eric Frank, the co-founder of Flat World Knowledge, argues that there is a huge financial opportunity in outflanking the traditional textbook makers. His company homes in on colleges and gives away a free online version of some textbooks. Students can then pay $30 for a black-and-white version to be printed on demand or $60 for a color version, or they can buy an audio copy.

About 55 percent of students buy a book, Mr. Frank said, adding that the leading calculus book from a traditional publisher costs more than $200.

Publishers have started de-emphasizing the textbook in favor of selling a package of supporting materials like teaching aids and training. And companies like Houghton Mifflin have created internal start-ups to embrace technology and capture for themselves some of the emerging online business.

They are responding in much the same way traditional software makers did when open-source arrived, by trying to bundle subscription services around a core product that has been undercut . . . 
And, it goes on and on, even before we factor in the 7" Android tablet in a folio for ~ US$ 120 that should be coming along in a year or so now, say, like:

Such a device can carry a virtual classroom, teleconferencing and collaboration device, text- and- work- books, in-built software and interface digital instruments, office productivity software, software development environments and -- with a digital library to back it up -- a library, anywhere, anytime.

In short, the online educational revolution is already in progress. Will we surf on the wave or will we be swamped by it? That, is our choice. END