I agree with the argument that if online edu. requires more conscientiousness, we should expect a lower success rate. I understand that gwern doesn't advocate

I agree with the argument that if online edu. requires more conscientiousness, we should expect a lower success rate. I understand that gwern doesn't advocate against online education, but it's worth noting: even if the conversion rate through online edu is lower than expensive, sit-in-a-classroom edu, we can expect more people to become more educated, provided the economic or status rewards from online edu aren't much less than the expensive variety - because more people will want to try. To increase the online edu success rate, look for mechanisms that make online learning more rewarding - perhaps some of the Zynga playbook - e.g. providing immediate public status gratification, or involving friends. To specifically address akrasia (of course, having lower cost means more people will begin what they never really intended to do), allow more flexibility in deadlines (if the course has a fixed schedule, allow skipping assignments until the next run), offer a type of completion with cheap physical "classrooms" for proctored tests/assignments (would increase the value of the signal - you really performed the task yourself), and optionally including time for watching lectures with others.