Quora Feeds
Active Member
Tommy MacWilliam
Three big changes: grading in MOOCs will be improved, MOOCs will cost something (and stop being so massive), and the production value of MOOCs will increase.
First big change: grading in MOOCs will be improved.
Right now, the biggest problem MOOCs face is that of grading. Many of the early successful MOOCs have been technical courses, and that's no coincidence. For many introductory CS, math, science, and engineering courses, creating computerized, automated grading systems is pretty simple. For example, we can automatically generate math and science problems (potentially with randomized parameters) that can be solved by a computer, and then we can simply compare the computer's correct answer to the student's answer. Similarly, automated grading systems can analyze the output of a student's code for a CS problem and compare it to the output of code that we know is correct (i.e., written by the course staff), which means we can tell students if their code is correct or incorrect with pretty high confidence. It's also pretty easy to create a computerized system that can support things like "the student's answer can be within 5% of the computer's answer and still be considered correct" or "this answer is correct but this answer is also correct as well".
Doing the same thing for less technical courses, like English or history courses, is significantly more difficult. Right now, many humanities MOOCs resort to asking simplified questions, like multiple-choice or true/false questions, since computers can easily auto-grade those. What we can't easily do right now is automatically assess the quality of an essay. Turns out that's just a really hard problem for computers to solve with today's technology, but it's one that many people are working on solving. Since feedback on long-form answers is a pretty significant part of your learning experience in non-technical courses, I'd expect solving this problem to be a top priority at many of the organizations behind today's MOOCs. Even in technical courses, it's not easy to tell a student "your code is well-written, but its design could be improved here" or "while your answer to this math problem is right, you could improve your approach by using this method instead", so improvements here will be valuable across the spectrum of online courses. Some MOOCs have tried using peer-grading to solve this problem, but that results a in pretty big variance in quality, since the feedback you get is totally dependent on how much effort your grader puts in. In this model, you might feedback from someone who is putting significantly less effort into the course as you, which isn't too helpful. Once we can provide richer feedback on more complex questions (which is probably a long way out), the MOOC experience will be significantly different for students.
Second big change: MOOCs will cost something (and stop being so massive). There's something cool and exciting to be taking a course along with 150,000 other students for free. What's not cool and exciting is when you post a question on the discussion forum and it gets immediately buried by the thousands of other questions being asked at the same time. Many students require one-on-one or small-group interactions to fully understand concepts, which is why tutoring programs exist and many colleges strive for small class sizes. Without lots money to pay staff, courses simply can't provide that kind of support structure, so in today's MOOCs, students aren't going to get that kind of intimacy. Don't get me wrong, the volunteers you'll find in a MOOC forum are incredibly helpful and diligent, but many students would (and will) pay money to get a video chat session with a trained staff member. Separately, producing a MOOC takes a huge amount of resources and cash, so in the ideal world, MOOCs are financially self-sustaining.
Here's an example of what I mean:
This is the pricing structure for CS50x, a course on edX. The course is still offered for free, but for a small fee, you get access to staff-curated discussions, office hours, and more. The $350 here is way lower than the thousands of dollars it would take to enroll in course offered through a distance education program, so students are much more likely to pay the fee if they really want to learn. Naturally, the community of paying students will be smaller than the community of free students, so in a sense, MOOCs will start to feel smaller. This freemium model is a win-win for both students and courses, so I'd expect more courses to adopt this in the future.
Personalization can also be increased through the MOOC platform itself. For example, if you're taking a math course and the platform detects your performance on a particular topic is lower than your usual performance, additional resources or practice can be automatically suggested. Similarly, a MOOC platform can use your performance in a course to recommend courses you might be interested in taking next. So, making MOOCs feel smaller can come both through additional human resources and continual technological improvements.
Third big change: the production value of MOOCs will increase. You know what's really nice about teaching an introductory calculus class? Introductory calculus isn't going to change much year to year. Same with Shakespeare; it's not like he's been putting out a lot of new material lately. So, material prepared for the 2013 offering of Shakespeare 101 does not need to be completely redone for the 2014 offering of Shakespeare 101. Rather than starting from scratch, more and more MOOCs will have an existing corpus of material they can build on, which means instructors can spend time refining and improving materials instead of scrambling to get them completed the first time.
Along the same line, running a MOOC once gives instructors a ton of data points. In a single run, instructors learn what worked for students, what topics were particularly difficult to understand, and how processes could be streamlined. Naturally, the second (and third and fourth) time around, instructors can take everything they learned into account to produce an even better course. You simply don't get that kind of data with 20 students, and modern MOOC platforms have made the collection and analysis of this data even easier.
Finally, as universities invest more and more into MOOCs, quality of videos will dramatically improve. Many early MOOCs were filmed in a professor's kitchen using a webcam and a PowerPoint deck using Comic Sans. Now, some universities have recording studios, Hollywood-quality cameras, and a dedicated production staff to create much higher-quality videos. As more and more universities launch MOOCs, the landscape will naturally become more competitive, which will drive quality up across the board. Here's one such recording studio at Harvard:
Fundamentally, the future of MOOCs will scale the educational experience better. There's lots to be done from both a technological and pedagogical perspective, but at the end of the day, MOOCs will likely introduce gradual changes that iteratively improve online education.
See Questions On Quora
Continue reading...
Three big changes: grading in MOOCs will be improved, MOOCs will cost something (and stop being so massive), and the production value of MOOCs will increase.
First big change: grading in MOOCs will be improved.
Right now, the biggest problem MOOCs face is that of grading. Many of the early successful MOOCs have been technical courses, and that's no coincidence. For many introductory CS, math, science, and engineering courses, creating computerized, automated grading systems is pretty simple. For example, we can automatically generate math and science problems (potentially with randomized parameters) that can be solved by a computer, and then we can simply compare the computer's correct answer to the student's answer. Similarly, automated grading systems can analyze the output of a student's code for a CS problem and compare it to the output of code that we know is correct (i.e., written by the course staff), which means we can tell students if their code is correct or incorrect with pretty high confidence. It's also pretty easy to create a computerized system that can support things like "the student's answer can be within 5% of the computer's answer and still be considered correct" or "this answer is correct but this answer is also correct as well".
Doing the same thing for less technical courses, like English or history courses, is significantly more difficult. Right now, many humanities MOOCs resort to asking simplified questions, like multiple-choice or true/false questions, since computers can easily auto-grade those. What we can't easily do right now is automatically assess the quality of an essay. Turns out that's just a really hard problem for computers to solve with today's technology, but it's one that many people are working on solving. Since feedback on long-form answers is a pretty significant part of your learning experience in non-technical courses, I'd expect solving this problem to be a top priority at many of the organizations behind today's MOOCs. Even in technical courses, it's not easy to tell a student "your code is well-written, but its design could be improved here" or "while your answer to this math problem is right, you could improve your approach by using this method instead", so improvements here will be valuable across the spectrum of online courses. Some MOOCs have tried using peer-grading to solve this problem, but that results a in pretty big variance in quality, since the feedback you get is totally dependent on how much effort your grader puts in. In this model, you might feedback from someone who is putting significantly less effort into the course as you, which isn't too helpful. Once we can provide richer feedback on more complex questions (which is probably a long way out), the MOOC experience will be significantly different for students.
Second big change: MOOCs will cost something (and stop being so massive). There's something cool and exciting to be taking a course along with 150,000 other students for free. What's not cool and exciting is when you post a question on the discussion forum and it gets immediately buried by the thousands of other questions being asked at the same time. Many students require one-on-one or small-group interactions to fully understand concepts, which is why tutoring programs exist and many colleges strive for small class sizes. Without lots money to pay staff, courses simply can't provide that kind of support structure, so in today's MOOCs, students aren't going to get that kind of intimacy. Don't get me wrong, the volunteers you'll find in a MOOC forum are incredibly helpful and diligent, but many students would (and will) pay money to get a video chat session with a trained staff member. Separately, producing a MOOC takes a huge amount of resources and cash, so in the ideal world, MOOCs are financially self-sustaining.
Here's an example of what I mean:
This is the pricing structure for CS50x, a course on edX. The course is still offered for free, but for a small fee, you get access to staff-curated discussions, office hours, and more. The $350 here is way lower than the thousands of dollars it would take to enroll in course offered through a distance education program, so students are much more likely to pay the fee if they really want to learn. Naturally, the community of paying students will be smaller than the community of free students, so in a sense, MOOCs will start to feel smaller. This freemium model is a win-win for both students and courses, so I'd expect more courses to adopt this in the future.
Personalization can also be increased through the MOOC platform itself. For example, if you're taking a math course and the platform detects your performance on a particular topic is lower than your usual performance, additional resources or practice can be automatically suggested. Similarly, a MOOC platform can use your performance in a course to recommend courses you might be interested in taking next. So, making MOOCs feel smaller can come both through additional human resources and continual technological improvements.
Third big change: the production value of MOOCs will increase. You know what's really nice about teaching an introductory calculus class? Introductory calculus isn't going to change much year to year. Same with Shakespeare; it's not like he's been putting out a lot of new material lately. So, material prepared for the 2013 offering of Shakespeare 101 does not need to be completely redone for the 2014 offering of Shakespeare 101. Rather than starting from scratch, more and more MOOCs will have an existing corpus of material they can build on, which means instructors can spend time refining and improving materials instead of scrambling to get them completed the first time.
Along the same line, running a MOOC once gives instructors a ton of data points. In a single run, instructors learn what worked for students, what topics were particularly difficult to understand, and how processes could be streamlined. Naturally, the second (and third and fourth) time around, instructors can take everything they learned into account to produce an even better course. You simply don't get that kind of data with 20 students, and modern MOOC platforms have made the collection and analysis of this data even easier.
Finally, as universities invest more and more into MOOCs, quality of videos will dramatically improve. Many early MOOCs were filmed in a professor's kitchen using a webcam and a PowerPoint deck using Comic Sans. Now, some universities have recording studios, Hollywood-quality cameras, and a dedicated production staff to create much higher-quality videos. As more and more universities launch MOOCs, the landscape will naturally become more competitive, which will drive quality up across the board. Here's one such recording studio at Harvard:
Fundamentally, the future of MOOCs will scale the educational experience better. There's lots to be done from both a technological and pedagogical perspective, but at the end of the day, MOOCs will likely introduce gradual changes that iteratively improve online education.
See Questions On Quora
Continue reading...
