Monday, 29 April 2013

What makes a MOOC?

Since the New York time article labelling 2012 as the year of the MOOC, there has been a bit of a ‘MOOC bubble’ lately with a very loose labeling of open courseware , youtube videos or even TED talks as MOOCs. As Massive Open Online Courses become increasingly popular, the MOOC acronym is being used for anything remotely associated with education on the web.

Let’s break down the acronym:

  • Massive:

Far from being a byproduct of the MOOCs’ success, scale is required for two reasons:
- The investments (time, resources, money) for launching a MOOC only make sense if a high number of students are going to participate;
- Meaningful student interactions via a dynamic forum, student-led video conferences and physical meetups require a minimum sizeable student body.

  • Open:

Anybody with internet access should be able to join with no restriction of any kind.
It is up to the participants to ensure they have the skills and knowledge required to successfully complete the course.
Open access will also ensure a diverse student body (age, country, profession, educational background, etc.), making student interaction all the more valuable.

  • Online:

Even though MOOC students have found useful to organise offline meetups and even though online courses material can be used by brick and mortar institution as part of their curriculum, a MOOC should be entirely online (teachings, exams, student-professor interactions, etc.).

  • Course:

A MOOC is a course first. It has to be structured like a course, with specific learning objectives, online classes (as opposed to what has been coined ‘xMOOC’ where the course content is actually produced by the participants), homework to practice what has been covered in class and exams to assess the learning.

By definition, a MOOC isn’t a set of tweets, a ‘serious game’ or even a free language learning web app like Duolingo.

Write a comment below or share this post!

Friday, 26 April 2013

A review of Udacity

Udacity is a Massive Open Online Courses platform launched in February 2012 by Sebastian Thrun, one of the two professors behind the seminal Stanford AI Class.

Reflecting the key competencies of its founders, Udacity initially offered courses in computer science from beginner level (CS101, building a search engine) to more advanced courses in cryptography and Artificial Intelligence for robotics.

The specificities of Udacity:

Unlike other MOOC platforms, Udacity has chosen to develop its courses in house. The courses therefore tend to be all on a very similar format:

- 7 weeks long
- short videos (1 to 3 minutes) punctuated by quizzes
- weekly homework assignments
- weekly recorded office hours (in which the professor and teaching assistant answer the forum questions with the highest number of votes)
- last week devoted to a ‘field trip’ (interviews with outside experts to put the course in context)

Despite Sebastian Thrun openness to offering on Udacity any subject that can be taught online, the focus on scienctific subject remains (with the exception of Steve Blank’s course on lean startups).

Udacity is a for profit company and its plan has been early on to offer courses for free but make money by identifying and placing its best students. In this spirit, it has started to offer courses with corporate partners:

From early 2013, Udacity also started a partnership with San Jose State University to offer online courses for credit (for a $150 fee).

Udacity’s strong points:

  • Udacity’s strong focus on pedagogy really shows. With its CS101 course, anybody actually can learn how to code a search engine with Python from scratch in 6-7 weeks.
  • After their first run, Udacity’s are still available as asynchronous courses which is more convenient than one off courses on other platforms.
  • Udacity’s career service is run by an ex-google recruiter and leverages knowledge of students’ forums activity as well as academic performance to match them with companies.
  • With Steve Blank teaching a course on entrepreneurship and Steve Huffman teaching a web development course, Udacity’s professor could not be more prestigious.

Udacity’s not so strong points:

  • Udacity is in the unique position of building curriculums rather than just courses. It is still not quite there though. CS253 (web development) which was advertised to CS101 students has a natural next step, threw off quite a few by requiring a lot of knowledge never covered in the previous course.
  • The user interface which was revamped in early 2013 offers an experience that isn’t as polished as on other platforms.

Udacity is a leader of the MOOC platforms neither by number of courses offered nor by number of students enrolled but its integrated model is probably the one holding the most promise for charting the future of education.

Write a comment below or share this post!

Wednesday, 24 April 2013

Caltech MOOC report

The California Institute of Technology released today the short video below going back over the experience of 4 of its professors teaching massive open online courses:

The video goes behind the screen, illustrates what the online courses looked like (lectures, quizzes, forums, etc.), and goes back over some of the motivations traditional institutions like Caltech have to get involved with massive open online courses. The takeaway message here being that MOOCs are here to stay.

Write a comment below or share this post!

Monday, 22 April 2013

A bold suggestion - My essay for Coursera's Beginner's Guide to Irrational Behaviour

There are a few misconceptions about what following a MOOC entails so, in a spirit of openness, I have posted below my entire essay in which I raise the issue of dropout rates and make a bold suggestion about monetization...

You and I are blessed to be taking part in the wonderful adventure of the Massive Online Open Course (MOOC). For the first time, through this new phenomenon of free and open online education, the increasing proportion of the World population with a connection to the Internet can not only access raw data, information but can also access knowledge and learn how to learn.
Companies have raised millions of dollars from venture capitalists with the hope of figuring out how to make money in this new field while non-for profit endeavours like edX have also sprung up.
There are problems plaguing free and open online education however. The main problems have to do with monetization and student retention.

In the seminal AI class launched in September 2011, less than 20% of the students enrolled earned a certificate. Although course’s dropout rates may vary and data is hard to come by, a figure of 90% can be considered conservative. Duke University’s first MOOC, A Quantitative Approach to Bioelectricity saw 346 students take the final exam and 313 pass it, when 12,735 had registered (Inside Higher Ed, Measuring the MOOC Dropout Rate by Ry Richards).

I believe monetization and student retention to be closely linked issues and will focus on the latter in this essay. The dismal student retention is problematic because of the unrealised potential of individual dropouts as well as because of its implication for MOOCs to be truly taken seriously as a force to be reckoned with in the future of education.

Let’s now see if the existing research could shed some light on this issue of high dropout rates in MOOCs. What existing research could be relevant to this behavior?

First of all, registering for courses is presently free and as Professor Ariely explained (Video Lecture 2.8, 2:50), the psychological effect of “the price of free” is such that people fomrget rational comparison between two desirable products. For instance when invited to choose between a 25 cents chocolate truffle and a 1 cent piece of chocolate most people choose the truffle. However, when choosing between a 24 cents truffle and a free chocolate, the vast majority goes for the free option, even though the price difference hasn’t changed. It is fair to assume that something similar is going on with MOOCs, most people would think hard between an expensive college class and a relatively cheap online version of it but enrolling for free in a MOOC seems a no-brainer.

Secondly, as describe by Professor Ariely (Week 2 Office Hours, 2:30) as MOOCs are a new phenomenon for which the quality hasn’t yet been established, there is a significant risk of students inferring from the current price (free) that the quality must be very low. It is therefore likely that this phenomenon leads to students not perceiving the full quality of MOOCs and eventually dropping out.

Finally, another factor that might be contributing to the high dropout rate of MOOCs compared to traditional on campus university courses or even fee-based online courses is cognitive dissonance. Professor Ariely describes the Feltzinger experiment (Video Lecture 4.6, 2:30) in which after doing a tedious task, subjects who were well paid admitted to the task being boring whilst subjects who were paid poorly convinced themselves they had found some intellectual stimulation from it. It is similarly easier for students enrolled in a MOOC to drop out because of a lack of interest than it would be for students of expensive courses to admit to themselves that they paid for something they do not deem useful anymore.

What solution could we propose to reduce the dropout rates in MOOCs?

I believe the best strategy to be for the MOOCs organisers to charge for registration but guarantee students the option to get their money back at graduation.
This solution addresses the negative implications for perceived quality of the price of free whilst allowing poor students to still graduate at no cost.

Furthermore, I believe that the courses could be engineered to make the IKEA effect as described by Professor Ariely (Video Lecture 4.4, 3:10) come into play. As participants in the experiment valued their own origami more depending on the effort spent on it, students who will have gone all the way through the course will have the greatest appreciation for its value.

Making students feel part of the course through peer-grading, forum participation and other means should not only ensure that the graduating students have learnt a lot from the course but also that they feel they own it and place a sufficiently high value on it that the vast majority will not exercise the ‘money back’ option at graduation.

Write a comment below or share this post!

Friday, 19 April 2013

A review of Duolingo

Launched in June 2012, Duolingo's core idea is simple: you learn a language for free while helping to translate the web.

Where is the idea coming from?

Duolingo is the brainchild of Luis von Ahn. Luis invented CAPTCHA, the blurry snippets of texts used by websites to tell computers and humans apart.

As a tool for preventing algorithm to signup to web services, CAPTCHAs were pretty effective. Taken collectively, they were also a big waste of time. Aware of this, Luis made amendments to the code to allow unsuspecting users not only to identify themselves as humans but also help digitize books.

This approach of bringing together two seemingly unrelated problems is what ultimately gave birth to Duolingo. The business model is for the language lessons to forever remain free to the users, but to start charging companies for the user-generated translations.

How does it work?

The site currently offers tracks in German, French, Portuguese, Italian and Spanish for English speakers, as well as English for Spanish, Italian and Portuguese speakers.

For each language, Duolingo displays a long skill tree. Each step in the tree contains several lessons. Every lesson needs to be mastered for the step to be mastered, which then unlocks the next steps in the tree.

Does it actually work?

Broadly speaking, yes! Duolingo makes learning languages fun and engaging. Here’s how:

  • The lessons and quizzes are geared towards practice rather than theory and find exactly the right balance between challenge and fun.
  • The whole experience is gamified (losing lives when answering incorrectly, gaining points for correct translations, being awarded increasingly unattainable medals).
  • The social features keep users engaged (leaderboard, forums).
  • Advanced learners have the flexibility to unlock specific portions of the skill tree to start directly at the lessons relevant to them.
  • An iPhone app has gathered rave reviews and an Android version is in the works.
  • The Duolingo team is very reactive and users are encouraged to report bugs (both technical and language based).

Overall, the interface is incredibly well thought-through and the feeling is one of great quality. It’s easy to forget when using Duolingo that you haven’t had to install an expensive piece of software on your computer.
For instance, the software is typo-resistant so you don’t lose precious points because of fat fingers.

Is Duolingo perfect then?

It's getting close but here's what's in the way:

  • Even though the constant stream of improvements is impressive, Duolingo sometimes feels a bit ‘beta’ and users (especially as they advance through the skill tree) need to be willing to contribute to the site by raising issues with the support team and suggesting amendments to quizz solutions.
  • Decay was recently introduced in the skill tree. Now steps are not ‘mastered’ for life but need to be regularly refreshed. Whilst this is a much more realistic approach, some bugs have been preventing users from actually practicing their most ‘decayed’ skills.
  • Finally and perhaps more importantly for the long term survival of the company, the Immersion tab which contains articles to be translated is actually the least engaging part of the site. The article selection is very narrow and the overall experience feels much less polished than the skill tree. If Duolingo is to continue to thrive, let alone make its VC backers rich, it will need to sort this out.

To conclude, Duolingo not only is a fantastic tool to pick up languages from scratch but is also very valuable to more advanced learners. The site keeps introducing new tools, new languages and improving the overall experience at an impressive pace.
With its amazing user interface and incredible business model, Duolingo is pushing the boundaries of what is possible in the realm of free and open online education.

Write a comment below or share this post!

Wednesday, 17 April 2013

Do online courses have an issue with dropouts?

Bill Gates, Steve Jobs and Mark Zuckerberg dropped out of university. Sergey Brin is officially on leave from his Ph.D. studies at Stanford.

There are quite a few less glamorous examples of college dropouts however.

In 2011, the Harvard Graduate School of Education published a study showing that only 56 percent of US college students completed four-year degrees within six years.

What is the situation in the world of the MOOCs?

In the seminal AI class launched in September 2011, less than 20% of the students enrolled earned a certificate. Although course’s dropout rates may vary and data is hard to come by, a figure of 90% can be considered conservative. Duke University’s first MOOC, A Quantitative Approach to Bioelectricity saw 346 students take the final exam and 313 pass it, when 12,735 had registered.

What can explain the MOOCs’ dropout rates?

Registering for a MOOC is by definition free and non binding. Most students register without considering the hard work often necessary to obtain a certificate. A lot of students might even be registering to be kept informed about courses with unannounced start dates.

Importantly, none of the students leaving a MOOC they have registered for, without a certificate of completion will be left saddled with student loans.

It is perhaps no surprise that many more students register for courses than expect to follow them to completion. According to Daphne Koller of Coursera 70% of the fee-paying Signature Track students graduated from their courses. This clearly shows that making a financial commitment increases retention.

Does any of this matter?

On the issue of student retention, free online courses shouldn’t be held to the same standards as traditional universities.
Student retention is nonetheless an issue for MOOCs to be widely perceived as a force to be reckoned with in the future of education.

Write a comment below or share this post!

Monday, 15 April 2013

What are certificates of completion worth?

Most Massive Open Online Courses offer certificates of completion to students. In the seminal Stanford AI Class, participants received a PDF certificate signed by the professors, Peter Norvig and Sebastian Thrun, with the grade point average and percentile.

While it is always nice to have something to show for one’s hard studying, what are these certificates of completion really worth?

Here are the main issues to keep in mind when thinking about their value today and going forward:

  • Brand:

“Contingent on academic performance, you will get a Statement of Accomplishment stating that you completed this course. However, no certificate will be given from Wharton / Penn and successful completion of this course does not make you a Wharton / Penn alumnus.”

There is a limit therefore to the value of brick and mortar brands on MOOC certificates.

  • Academic standards:

Coursera’s model is based on partner universities using its platform. The courses on offer range from 4 weeks quizz based courses targeted to a wide audience, to 12 weeks courses with much heavier workload.
Platforms with carefully curated and mostly in-house content such as Udacity offer less variability (course difficulty is also clearly identified as Beginner, Intermediate or Advanced).

  • Distinctions:

As with traditional brick and mortar universities, Udacity offers different level of certification depending on the academic achievements of the students (from a Certificate of Accomplishment to a Certificate of Accomplishment with Distinction to a Certificate of Accomplishment with highest distinction).

Since its courses are asynchronous, Udacity’s certificates of completion can be earned by taking a series of automated tests at any time. Course staff sometime have a hard job policing the forums to avoid any leaks between students and other temptations to cheat.

  • ‘Verified certificates’:
To address the risk of cheating and/or identity fraud, MOOC platforms are increasingly prepared (for a fee) to verify students’ identity and provide proctored exams.
From $30 to $100 per course, Coursera’s signature track now offers students in a small selection of courses the opportunity to obtain a verified certificate of completion (the student’s identity is verified via photo ID and unique typing pattern).
Udacity has partnered with Pearson VUE since 2012 to offer proctored exams through Pearson’s extensive international network of test centers.

  • Transferability of credits:

Coursersa has worked with the American Council on Education’s College Credit Recommendation Service to approve five of its online courses for credit.
Udacity is offering some courses in partnership with San Jose State University. For a $150  enrollment fee, students can earn credits transferable within the California State University system.

  • Recognition by potential employers:

Back in 2011, Sebastian Thrun offered to put the very best students from the AI class in touch with some of Silicon Valley’s most exciting companies. Now, both coursera and Udacity plan to generate significant revenue via career placement.
The fact that Google, Nvidia and Autodesk (amongst others) have co-created courses with Udacity also goes to show employer’s pragmatic approach to identifying and hiring talent.

  • Extra-academic criteria:

Perhaps not surprisingly, even in the online world, companies wish to look beyond just good grades. Indeed, companies recruiting via Udacity’s Career Placement Program have expressed interest not only in the students’ academic track records but also in their contribution to the course forums.

In conclusion, we will probably see a world with traditional degrees (at traditional prices) for residential students, fee-based ID-verified online courses and free online courses.
In any case, whilst there still is a lot of uncertainty surrounding the value of MOOC’s certificates, they are as central to the future success of MOOC platforms as they are to the motivation of students.

Write a comment below or share this post!

Friday, 12 April 2013

A review of Khan Academy

Khan Academy is not MOOC platform. With several millions of monthly users, it certainly nails the first 3 letters of the acronym. It doesn’t however provide courses as such.

It is nevertheless worth having a detailed look at Khan Academy. If the birth of the MOOCs can be traced back to the seminal Stanford AI Class of late 2011, the professors behind the AI Class admitted to having been heavily influenced by Salman Khan’s TED talk earlier that year.

A hedge fund analyst with multiple degrees from MIT and Harvard, Salman Khan tutored his cousins in maths and quickly started doing so via youtube videos for a better efficiency.

Despite the low production-value of the videos, his cousins told him they actually enjoyed being able to rewind and repeat at will. It was soon obvious that they were not the only ones watching either...

As of April 2013, and after receiving grants and award from the likes of Google and Microsoft, Khan Academy is one of the World’s leading non-profit organisation in the education field and its Youtube channel totals a quarter of a Billion viewers and 900,000 subscribers.

Khan Academy initially focused on primary and secondary education but started to cover other topics as well (GMAT practice, computer science and art history and topical economic issues have been gradually added).

Whilst the initial ‘markers on a blackboard’ video format has remained largely unchanged, detailed statistics are made available for students (also available for teachers in a flipped classroom setting) to monitor progress.

Khan Academy is deservedly seen as a great tool for children and educators in primary and secondary education.
Despite the lack of structured courses, with Salman Khan’s outstanding tutorials on Quantitative Easing or the Chinese currency and US debt and other topical issues, Khan Academy has a lot to offer to other demographics as well.

Write a comment below or share this post!