Friday, September 28, 2012

Coursera Placement Services

(Post by Mark Lewis, Associate Professor of Computer Science at Trinity University)

I haven't finished the second week of course material for my Coursera course, but something just came into my inbox that I felt warranted an immediate post. Below is an e-mail from Coursera about placement services. It makes sense. This is one of the possible options for services like Coursera and Udacity to make money.
Hello Mark Lewis,

Thanks for enrolling in Functional Programming Principles in Scala! At Coursera, we want to help you achieve your goals, whether those goals include finding a job or just learning something new. On that note, we're piloting out a free job placement service to connect you with great professional opportunities. If you would like to participate, click here ( to complete the Placement Services settings page and opt-in to the service. After you opt-in, we will share your resume and other information you provide, with a small number of carefully selected partner companies and allow them to introduce themselves if there’s a match. 

In order to give you more control, we will not share any of your contact information with employers beyond what you choose to provide in your resume, personal website or social-network profiles. Finally, you will always have control over whether or not your information is shared with employers and you can use the opt-in field on the Placement Services settings page to opt-in or out at any time. In this early stage of the pilot, most job opportunities will be US based and your course performance can only be shared for a few selected courses. However, we will offer a broader number of opportunities and include more courses as we expand the service.

We're excited to connect you with great companies and new opportunities as we develop this service! If you'd like to participate, opt-in here! (

Andrew Ng, Daphne Koller and the Coursera Team
I feel that this should bother traditional schools on two fronts. The lesser front is that this could be something that truly sustains MOOCs to make them viable for the long term. Of course, being viable in the long term doesn't matter if traditional colleges retain a significant edge in value that makes people want to attend them. The more significant aspect of this, in my opinion, is that if MOOCs prove to be useful to both employers and to students in making the connection between them and finding jobs for people, it will turn them into a serious threat to traditional colleges.

I know that many of my colleagues at small, liberal arts schools like to tout the benefits of being a well-rounded, well-educated individual. Many look down on the idea that college is about getting jobs. However, they have to consider the cost of college and how much monetary value people place on being well-rounded and well-educated. I believe that there are few people in the US who can afford to value those qualities much above $10-20k in and of themselves. Given that college typically costs 10x that much, the reality is that people pay for college because they see it an investment in their futures. That investment pays off when they get better, higher paying jobs. If you take away the job benefits of college, only a small, elite subset of the population would continue to pay for college. That subset is far smaller than the number of students that need to be attending colleges for the current number of institutes of higher education to survive.

So the reality is that if most of the faculty in the US want to actually keep their jobs, they need to also appreciate the fact that they are imparting skills, qualities, and abilities in their students that do help them to get better jobs. That is a big part of what students and parents are paying for.

If the MOOCs can demonstrate the ability to hook students up with employers in a broad and general way, I believe that is when they become true competition for traditional higher education. At that point it doesn't matter if they are lacking many of the aspects that we value in traditional higher education. The MOOCs don't have to be a copy of the current model, they just have to provide similar benefits because when you get down to it, that is what people are really paying for. If they can provide similar benefits while costing between nothing and a tiny fraction of the cost of college, they have a remarkable edge.

I would really like to hear if people doing other MOOCs with Coursera or Udacity have gotten similar letters. I acknowledge that I might be getting this mainly because I am signed up for a MOOC in Computer Science. It might be that other fields aren't getting this type of treatment yet. If you have taken a MOOC, please comment and let me know.

I will close with a somewhat tongue-in-cheek third threat that this poses, the possibility of luring away faculty from normal colleges. Of course, professors become professors for a reason. We love what we do. Our job has its own special rewards that are simply impossible to find in other fields. However, as my first post indicated, I think that one of the great strengths of MOOCs, as they currently stand, is in the area of continuing education. I feel like faculty should be some of the most likely people to utilize that service. How many job offers can a professor be given before at some point he/she decides to try out doing something different?

Wednesday, September 26, 2012

Google Course-Builder

I watched a Google Hangout On Air today sponsored by the Google Course-builder team.  The event was led by project head Peter Norvig and featured an interview with Udacity founder Sebastian Thrun.  It was the second Hangout on this topic since the project was announced less than two weeks ago.  Last week's hangout featured a conversation with Vi Hart of Khan Academy and Math Doodling fame.
Screenshot of the Google Course-Builder team, plus Thrun.
Google Course-Builder team, plus Thrun. 

Those interested can watch the entire recording below.  I will simply share a few bits that I found interesting.

Norvig stated that Google is committed to the Coursebuilder project and that they are responding quickly to user concerns.  I know some probably worry about the level of investment required to implement a course on this new platform, given Google's history of abandoning projects.  Thrun asked whether Coursebuilder might one day become something like YouTube, where course creators can upload their materials, and Norvig said yes.

His answer to virtually all of the audience questions was basically "Of course you can do that,you just have to build it yourself (e.g. charging fees for courses, integrating popcorn.js, etc.)  Norvig indicated that, unlike Udacity, Coursera and EdX, Google wasn't interested in providing courses, only in providing the platform (Google's Powersearching course, for which the platform was designed notwithstanding, I suppose).

Norvig says that in teaching his own Artificial Intelligence class as a Massive Open Online Course on Udacity, he learned that

  1. Community mattered.  Students learned from each other.
  2. Motivation was more important than information
  3. Learning is done by the student, not the teacher.  
All three of these lessons put the instructor and the lecture material as secondary to the learner and, in particular, groups of learners.  community was a recurring theme, with Norvig bringing out a colleague to speak to how they generated community in the Powersearching class by emailing participants reminders, linking directly to the forum from the course activities so that people could share strategies and questions, and using the Google+ Hangouts tool to form study groups.

The conversation with Thrun was fun to watch, though they both seemed somewhat nervous or uncomfortable.  The conversation highlighted the experimental nature of both Udacity and Coursebuilder, the fact that they were reacting to what had come before (e.g. MIT's OpenCourseware) and the need to come up with pedagogies for online and lifelong learning.

They discussed the trade-offs between synchronous and asynchronous courses, where synchronous courses generated excitement and community and asynchronous courses were ultimately more convenient.  Thrun indicated that he envisioned a future where these two could be reconciled somehow through self-pacing.

They discussed the question of learner motivation, hitting on gamification but also on ultimate employability of students completing these courses (Norvig's AI course in particular).

Tuesday, September 25, 2012

First Week of Class: Functional Programming in Scala on Coursera

(Post by Mark Lewis, Associate Professor of Computer Science at Trinity University)

I just finished the first weeks lectures and assignments for Functional Programming Principles in Scala so it seemed appropriate to write a post on my first impressions. We were told in e-mails that this course has over 32,000 students enrolled in it. (Update: another source showed 40,000 as the number for the first week.) The fact that it is being taught by Martin Odersky, the creator of Scala, is inevitably a big factor in getting many people to register. Normally I don't know if a course on functional programming would garner so much interest, but there is no doubt that functional is seeing a general resurgence and the fact that Martin is the person behind the language is a huge factor. Nearly every Scala community I belong to has had notices about this course.

The quality of the teachers is probably one of the greatest strengths of MOOCs. The nature of these courses means that they fully enable the "superstar effect" to come into play in education. One teacher can now serve every person in the world who has an interest in learning about a particular topic. Granted, they do so through a one-way distribution mechanism of prerecorded videos and scripted automatic assessment. Then again, that doesn't sound too different from what actually takes place in large lecture halls. The internet just happens to be the world's largest lecture hall and anyone across the world can make it to lecture, regardless of actual physical location.

In this regard, Martin Odersky does not disappoint. The lectures are clear and insightful. The examples are definitely well thought out. I found that I appreciated and enjoyed what he had to say despite having significant experience in this area myself. Given my background, I did find the option to watch at faster than normal speeds to be quite helpful. 1.5x seemed about right for me. Going faster than that made the audio hard to understand. This is, of course, one of the great benefits of the video format. You can "attend lecture" when it suits you and you have control over things like speaking speed as well as the ability to rewind and watch segments again if needed.

I assume that the general format of the videos is something that is associated with Coursera. Early on there were scenes of Martin sitting in his office, but most of the time it was either slides or screen captures. The interface for the slides, with Martin's hand and a pen that could point and write were things I particularly appreciated. It reminded me of being back in school years ago when teachers would write on overhead projectors. This had a more personal feeling than a straight screen capture. IMO.

The occasional breaks for quizzing and to let students think about questions were very useful from a pedagogical standpoint. I did run into one problem though where I used a feature of Scala that hadn't been discussed yet and had the automated system tell me that my answer was incorrect because of it. This is the type of thing a human could easily spot and comment on. Being told that an answer was wrong when I knew was correct was a bit frustrating to me and I could see that really throwing off a novice student.

After the lecture videos, there was a little assignment with three problems. They were very well done and posed some challenge for the experienced student because you had to solve them with a limited set of constructs. The setup for working on the assignments and submitting them was easy to follow and the feedback was nearly instantaneous. They also provided test suites so that you could check your code against some tests before you submitted it.

One the whole, I was really impressed with the experience. It is only the first week, and there is a lot more to see, but this convinced me that for people like myself, Coursera and other MOOCs can provide a remarkably powerful resource for continuing education.

Of course, continuing education isn't really the purpose of this blog. This blog is about using these techniques in the context of normal college education and specifically how it could be useful for, or threatening to small, liberal arts institutions. I want to address this as four questions.

  1. Could the Coursera course be a drop-in for a course at a Primarily Undergraduate Institution (PUI)?
  2. If not, why not and what is missing?
  3. Could this MOOC be used to facilitate a normal PUI classroom?
  4. If not, why not? If so, what benefits would it provide?

Based on the first week of the course, here are my answers to these questions. I will try to revisit these in later posts to give some continuity to my observations.

  1. No. What I saw from the first week would not be a drop-in substitute for one of my courses or hopefully any other course on this topic at a PUI. While I can give that answer definitively, I also think that it would be reasonably easy to largely close that gap.
  2. The main problem is that there was too little of the course. There was only about an hour of video lecture and this is for a seven week class. The second week has slightly less than an hour of lectures. Compare that to ~40 hours of class meeting time for my normal courses. Granted, administrivia is all pushed outside of the lectures, but even given that, I feel they would need 1:30-1:45/week and a few more weeks to be a proper substitute for a 3-credit PUI course.

    Also missing were a lot more exercises and evaluation questions. The platform provides the ability to add these things, and they were used. However, I would want to see some exercises and questions that come at the end of each lecture as well as more inside of it. The assignment questions were appropriate, but other exercises that build up knowledge one step at a time would be required since you can't ask the instructor questions.
  3. Yes. In fact, after watching the first week's lectures I almost wanted to force the students in my "Principles of Functional Programming" course to sign up and watch. I didn't for two main reasons. The first is that we are currently working in Scheme, not Scala. While the thinking is similar, the students struggle enough with the very simple syntax of Scheme. Even though 70% of my students know Scala, I still don't feel it would be appropriate to try to mix things up like that in their brain. In addition, the first week of the MOOC was during the fifth week of my class. That causes a bit of a synchronization problem.
  4. I have already been playing with blended learning techniques and recording my own videos for students to watch outside of class. The system for the MOOC takes this an extra step up. I feel like if it weren't for the little problem where our semesters are not synchronized, I could ask students to basically take the MOOC and then add my own exercises and discussion on top of it to bring it up to the time level of a 3-credit course and also to fill in gaps when students don't grasp something and the resources associated with the MOOC aren't sufficient for them to fill in the gaps.

In general, I have a very positive view of this MOOC, but I have serious questions about whether it can function well for traditional students who are beginners and who lack direction or motivation. I can see some ways in which the MOOC could be altered to help close those gaps, but there are other things that I still can't see happening without a little bit of human interaction with someone who does understand the material. Perhaps more insight will come to me in future weeks. Even with those limitations, I feel that I could build a class around a MOOC where my role was more of a facilitator and it would make my workload for a single course much lighter, freeing me up to teach more courses or to do other things for the University.

Sunday, September 23, 2012

Blog Introduction

The purpose of this blog is to document activities and share ideas as part of a critical look at MOOCs and their relationship to liberal arts education. There is a lot of buzz about how education is one of a number of areas that is currently ripe for disruption by technological innovation. Developments like OpenCourseWare have made educational materials freely available for many years now. However, the ability to have free video distribution through sites like YouTube has opened other doors. Khan Academy aims at giving lessons on many different subjects and quite a few include quizzes and other forms of automatic assessment., some of which are gamified to further motivate students. The arrival of MOOCs in the fall of 2011 introduced another twist with fairly complete models of instruction available to large numbers of people.

In the year since Stanford offered two MOOCs from their CS department, the field has grown by leaps and bounds with many of the top national Universities jumping into the fray. At the time I am writing this, there are nearly 250 MOOC offerings split between Coursera, Udacity, and edX. These offerings include quite a few courses in the Humanities and social sciences as well as the STEM fields. Each MOOC will enroll 10,000+ students from around the world and many will give certificates of completion to those who stick with them and do the required work. While it might not be exactly clear what the future holds for these MOOCs, one thing is clear, change is coming fast.

To help make sense of this change, Trinity University has applied for funding from the ACS to initiate an informed dialog on MOOCs and the liberal arts. To deal with the fact that things are moving so quickly, Trinity has put up funds to get things going before we hear back from the ACS about the proposal. To make sure the dialog is informed, a number of faculty from Trinity and several other ACS institutions have agreed to complete a MOOC and record their experiences and thoughts. This blog will serve as part of the information distribution aspects of this work. readers should also feel free to make comments and contribute to the dialog. In spring 2013 the discussion will move beyond this blog as Trinity hosts video conferences between ACS campuses to help get all faculty thinking about the challenges and opportunities provided by these new modes of education.