- I’m interested in online education technologies as both an instructor and a researcher.
- I’m a member of the Technical Advisory Committee to EdX, helping to guide the evolution of the open source EdX platform.
- I'm the Faculty Advisor to the UC Berkeley MOOCLab.
Online education—IDEAS (Innovation in Data-driven Education-as-a-Service)
What makes research in online education interesting? For me it’s the “three legged stool” of rich because of scale, rich despite scale, and delivery as a service.
- Rich because of scale: with large cohorts of students (online or in the classroom), you can not only gather and analyze much richer datasets, but also take advantage of student work products in new ways. AutoStyle (work in progress) attempts to use the submissions of other students in the same cohort to incrementally guide a given student toward the stylistically-optimal way to express a program.
- Rich despite scale: At scale, what mechanisms increase students’ sense of engagement and of being part of a cohort despite the fact they may never meet their peers in person, and how can we help instructors better understand what’s happening in such large cohorts?
- Delivered as a service: With online learning technology taking the form of Web services accessed by standard clients such as browsers, the potential exists for reducing research results to practice much more quickly and “baking in” best practices directly into the software.
Some recent papers:
- Structuring Interactions for Large-Scale Synchronous Peer Learning [CSCW 2015, to appear] (by D Coetzee, Seongtaek Lim, Armando Fox, Björn Hartmann, and Marti A. Hearst). Our MOOCchat system lets us form groups of online learners on-demand and use text chat to let them discuss answers to questions. In a Mechanical Turk-based experiment, we found that discussing challenging problems leads to better outcomes than working individually, and incentivizing learners to help one another yields still better results. We are now trialing this in MOOCs.
- Should your MOOC forum use a reputation system? [CSCW 2014] (with Derrick Coetzee and Profs. Björn Hartmann and Marti Hearst) Previous Berkeley research investigated why the reputation system used by the programmer discussion forum StackOverflow is so effective. In a controlled A/B experiment, we investigated whether similar systems can improve sense of community and learning outcomes in UC Berkeley’s software engineering MOOC, CS 169.1x. While we found no effect on learning outcomes, we did find positive effects on forum engagement.
- Chatrooms in MOOCs: all talk and no action [Learning@Scale 2014] While MOOC students say they enjoy having chatrooms incorporated into MOOCs, the chatrooms do not appear to increase student engagement or result in other positive community effects.
- Monitoring MOOCs: Which Information Sources Do Instructors Value? [Learning@Scale 2014] With such a large number of students it is now suddenly much easier to see trends and diagnose the significance of student problems. We interviewed over 90 instructors to learn about what sources of information they value when trying to understand student behavior/engagement in their MOOCs. Among other things, we find that quantitative indicators aren't enough, and instructors are very interested in understanding what's happening in the forums. More info.
Online education—outreach, policy, etc.
There’s a lot of media misinformation and hype regarding online education, so I’m helping Berkeley do its part to inform both our own faculty & students and the wider world about the potential opportunities and uses (along with pitfalls and misuses) of this exciting new combination of technologies. A few recent writings on this:
- MOOCs in Higher Education: Current State and Perspectives. In March, a group of academic leaders in online education from the USA and Europe gathered in Dagstuhl, Germany, to assess the current state and desired future directions for both MOOCs in academia. Eight provocative positions are put forward, in hopes of aiding policy-makers, academics, administrators, and learners regarding the potential future of MOOCs in higher education. The recommendations span a variety of topics including financial considerations, pedagogical quality, and the social fabric. The final report ("Manifesto") was prepared by Pierre Dillenbourg (EPFL), Armando Fox (Berkeley), Claude Kirchner (INRIA), John Mitchell (Stanford), and Martin Wirsing (Ludwig-Maximiliens-Universität Munich).
- From MOOCs to SPOCs: How MOOCs Can Strengthen Academia. An op-ed in Communications of the ACM, in part a response to editor-in-chief Moshe Vardi’s piece on Will MOOCs Destroy Academia?
- What We’ve Learned from Teaching MOOCs, a piece written by Dave and me for the EdX teaching blog, about lessons for instructors new to the medium.
- What’s the proper role of MOOCs in higher education? I wrote this for the Berkeley “faculty experts weigh in” blog to discuss an under-recognized use case for MOOC technology: SPOCs, or Small Private Online Courses.
I teach about every other semester. (Most recent: Fall 2013; next: Spring 2015) Learn real software engineering the way it's done by the pros, while building a cool SaaS app as you go.
John Canny and I ran a CS 294 graduate project course in Fall 2013 about autograding and other innovative online education technologies.
I used to think I was a good instructor, and I had the HKN ratings and other awards to prove it. But as I absorb more of the learning sciences literature I'm beginning to see my methods as hopelessly ad-hoc. [intlink id="437" type="page"]I'm trying to get better about that[/intlink].
[intlink id="410" type="page"]Archival proposal[/intlink] for an undergraduate course on Internet Services I wanted to do while at Stanford. UC Berkeley CS169 is essentially a much better version of the course I proposed back then. :-)