TeachLab with Justin Reich

David Joyner

Episode Summary

This week on TeachLab, host Justin Reich is joined by Dr. David Joyner, Executive Director of Online Education at the Georgia Tech College of Computing. Together they discuss the challenges and advantages of online learning, the hard shift to remote learning under COVID, and David’s new book The Distributed Classroom. “The distributed classroom is really about asking the question, ‘Can we take one classroom experience and distribute it across students who can commit to different levels of attendance, different levels of in-person attendance, different levels of synchronous attendance, things like that, such that you get to have as much of the experience as possible within your individual constraints?’" - David Joyner

Episode Notes

This week on TeachLab, host Justin Reich is joined by Dr. David Joyner, Executive Director of Online Education at the Georgia Tech College of Computing. Together they discuss the challenges and advantages of online learning, the hard shift to remote learning under COVID, and David’s new book The Distributed Classroom.

“The distributed classroom is really about asking the question, ‘Can we take one classroom experience and distribute it across students who can commit to different levels of attendance, different levels of in-person attendance, different levels of synchronous attendance, things like that, such that you get to have as much of the experience as possible within your individual constraints?’" - David Joyner

In this episode we’ll talk about:

 

Resources and Links

Check out David Joyner’s book The Distributed Classroom

Check out David Joyner’s paper Components of Assessments and Grading At Scale

Learn more about Dr. David Joyner

Check out Justin Reich’s book Failure to Disrupt: Why Technology Alone Can't Transform Education
 

Transcript

https://teachlabpodcast.simplecast.com/episodes/david-joyner/transcript

 

Produced by Aimee Corrigan and Garrett Beazley. Recorded and mixed by Garrett Beazley

 

Follow TeachLab:

Facebook

Twitter

YouTube

Episode Transcription

Justin Reich:                 From the studios of the Teaching Systems Lab at MIT, this is TeachLab, a podcast about the art and craft of teaching. I'm Justin Reich. This week, I'm joined by David Joyner. David is an educator, author, and the Executive Director of Online Education and the Online Master of Science and Computer Science at Georgia Tech's College of Computing. His research focuses on online education and learning at scale.

                                    David is the co-author of the recent book The Distributed Classroom, the latest in a series on learning in large scale environments, published by the MIT Press, series editor, yours truly. The Distributed Classroom proposes a vision of the future of education, in which the classroom experience is distributed across time and space without ... I think I messed up a word. The Distributed Classroom proposes a vision of the future of education, in which the classroom experience is distributed across space and time without compromising learning. And we're excited to talk with David about his book today. David, welcome to TeachLab.

David Joyner:               Thank you. It's great to be here.

Justin Reich:                 So you're the Executive Director of Online Education at the Georgia Tech College of Computing, probably not a position that they had there 20 years ago. What is being the executive director of online education in a program like that mean?

David Joyner:               It's a position they didn't have two years ago, and they created it right before COVID hit, right before everything became online education, which was fantastic timing or terrible timing, depending on whether or not you're the one in the role, which I am. So, yeah. So most of what I do is I run our Online Master of Science and Computer Science program. We started in 2014. We're up to 12,000 students as of this semester, just finished graduating our, I think, ninth graduating class, which took us to 6,500 graduates. So we describe it as a MOOC-based master's degree, which is we try and take the lessons from massive open online courses and apply them to the for credit space. So we deliver a master's degree, but online and at scale in a way that lets us let anyone in who meets our minimum qualifications, no matter where in the world they are, everything like that.

                                    In addition to that, I teach our online undergraduate class. It's a CS1 class, presupposes no prior computer science knowledge, which we developed to find out if we can extend the lessons we learned in the OMS CS program to the undergrad level. That also exists as a massive open online course. It's had about 5,000 students at Georgia Tech complete it, as well as about 15,000 in the MOOC version complete it.

                                    What's novel about that one is that it's also the course we use to deliver our dual enrollment section of CS1. So we also have high school students from across the state who enroll in that class, alongside current students at Georgia Tech. And it's a fascinating experience because we accommodate non-traditional students. I've had a semester where our youngest student was 15 or 16 and our oldest student was in their 70s. And they interact on the forum, they help each other out, and never know that that big chasm exists. So it's been fascinating.

                                    So most of what I do is run those programs, teach those classes. Most of what I do is answer email, but that's because of the nature of what we do nowadays. So ...

Justin Reich:                 Email and meetings.

David Joyner:               Yep.

Justin Reich:                 That's the week of an academic. So you have described these kinds of learning environments as distributed classrooms. What are the key features of distributed classrooms?

David Joyner:               So when we talk about what we think of as the distributed classroom, we're looking at where we come from, which is that we have our online MSCS program, which exists at one far end of a spectrum, and then you have traditional learning environments, which exist at the other end. Traditionally, you have all your students in a room together at the same time and the same place, and you can talk to them all live and you can interact with them all live. And then what we've done over time is said, "If you can't commit to being in Atlanta at 9:00 AM, Monday, Wednesday, Friday for 17 weeks at a time, then you go to this online version of this class, which is online, asynchronous. You don't have to be here at any given time. You can interact whenever you're able to join."

                                    And that's great for accommodating a much broader set of learners. We have lots of people who can't commit to attending during work hours because they're trying to transition jobs or something like that. We have lots of people who can't commit to normal hours because childcare and things like that. So we're able to open it up to people who can't commit to the level of in-person attendance and regular attendance that you need to to do an in-person program.

                                    But what's interesting about all that is that there's so much in the middle between those two extremes. On one extreme, you have in a room together, same time, same place, and the other extreme, you have on your own whenever you want, whenever you want, wherever you are. And there's so much we can do in between those two extremes that we see with our students forming things on their own.

                                    There's so many places where you'll have students who'll say, "I can commit to attending synchronously. I can be online 9:00 AM, Monday, Wednesday, Friday. I just can't be in Atlanta," or you'll have students who say, "I can commit to a time. It just has to be a time after hours." And what that means is that there's this resource available, there's this structure available for different students that they can commit to that lets them preserve some of that in-person or some of that live interaction that's really valuable without going all the way to the extreme of a completely asynchronous experience, just because they can't commit to being there in person.

                                    The distributed classroom is really about asking the question, "Can we take one classroom experience and distribute it across students who can commit to different levels of attendance, different levels of in-person attendance, different levels of synchronous attendance, things like that, such that you get to have as much of the experience as possible within your individual constraints?"

                                    And what's nice about the modern technology that we have is that more than ever we're able, I think, to build an experience that actually distributes it across multiple points in that spectrum, instead of designing separate classes that are, "This is the online section. This is the in-person section. This is the synchronous section." So that's what we mean when we talk about a distributed classroom.

Justin Reich:                 So the book primarily uses the online master science program in computer science, but I'd love to hear you talk about, so we could differentiate a little bit, what does this distributed classroom look like in CS1? What are some of these concrete examples of middle spaces between fully in-person and fully online that you can create for people who are just putting their foot in the door of computer science?

David Joyner:               Yeah, absolutely. So the CS1 actually is the best example of that, because it's accommodating a slightly more traditional audience. So what we have right with CS1 and in terms of its vanilla setup is it's an online experience, it's asynchronous. And quite honestly, that initially was done both for scale and for the fact that I think CS1 is just a really nice topic to teach online. It's such a practice oriented field. It's really about getting in the weeds on practice problems, actually doing some coding. It's not the traditional model of attend class for 50 minutes, and then two days later try the homework. You can't learn CS that way. So we built it that way [crosstalk].

Justin Reich:                 If the thing that you're trying to learn is how to sit in front of a computer and type stuff, it's a reasonable way to teach people, to have them sit in front of a computer and learn to type stuff.

David Joyner:               Yeah, absolutely. I love the model of it. We've all been in those classrooms where the teacher's teaching something and we're a little confused, but they're not going to stop because they can't just stop and not everybody in the class is confused, or maybe some of us are bored. And it's really nice to have this environment where we can say, "You're going to learn for five minutes and then put it into practice for 15 minutes or 20 minutes or 30 minutes or two minutes, depending on what you individually need, because you're not sitting in a classroom with a bunch of other people, all who have to go on the same pace."

                                    So we designed it with that kind of thing in mind. But what it's allowed us to do is create this kind of experience where, when a student is able to commit to attending something in person, for example, we have an optional recitation that they can come to, they get some supplemental work, they get to talk to other people, they get to work through problems with classmates, work through problems with teaching assistants. But if they can't attend that, then they're not locked out of the rest of the experience just because they can't be there in person and on campus.

                                    The place where that's been really interesting, though, has been in the past couple years, as we've been forced to actually go more online with that class. We found some things that we were able to do that really embody what I think this distributed classroom view was about. The first one was last fall, fall 2020. There were a whole bunch of students who were going to come to Georgia Tech to start their academic career from China. And because the world was on fire, they weren't able to come.

                                    And so instead of saying, "You just have to put off your enrollment for a year, our new campus at GT Shenzhen agreed to host them for that first semester. And they were able to take some classes on the ground there. But they were actually able to take this class together with our students who were actually in Atlanta, as well as our students who would've been Atlanta if, again, they could have come to campus, all in one big class.

                                    And what was really interesting there was that they were able to have their own recitation in Shenzhen, China because of the climate, the nature of the virus restrictions there at the time. They were actually able to have their own group of students meeting in person every week with their own teaching assistant, getting their own individual support while they were officially enrolled in a class that was being delivered in Atlanta all the way across the world, because the nature of it allowed us to distribute a portion of the experience to wherever you happen to be. And if you are in a place where you have a cohort of classmates to talk with live, you're able to do that.

                                    We've now extended that to our campus in Lorraine as well. And so in my class this semester, I've got students both in Atlanta and in GT Lorraine, Lorraine, France, where our European campus is. And they're able to interact in the same forum, talk with same TAs, get the same support, no matter where they happen to be. And then if they're in a place where they have several classmates, we can have dedicated on ground support for them.

                                    It's a model that I think has a lot of potential in other places as well, because we talk about a lot of classes that you really need a lab experience or you need to be able to work on some physical prototype or something like that very often are the things that are very hard to teach remotely. But if you get to a point where you're able to say, "All right, we don't need everyone to be here in Atlanta. We just need you to have five people local to you, and we can justify creating that fiscal infrastructure for you to do what you need to do there," you can start to create these little pods of people all around the world who can get that social learning, get that classmate interaction, and get that authentic experience, even though they're not able to be in Atlanta.

Justin Reich:                 Now, are there high schools that host cohorts like this? Are most of your high school enrollees for this class doing things on their own, or in addition to folks in France and Shenzhen, do you also have people in different in places in Georgia doing this together?

David Joyner:               Yeah. No, that's a great point. So we do have high school students who are dual enrollment students in the CS ... If I say CS 1301, that's our official number for it, but it's a CS1 class. So we have high school students around the state of Georgia who are part of the dual enrollment program who enroll in that class as well. We have 80 of them this semester alone.

                                    And it's a really nice experience there, too, because the dual enrollment program in Georgia and I'm sure in other states as well has been around for ages. But it typically has such a high barrier to participation that you don't see nearly as much adoption as you could, because you have to require students who are high school students around the area to come to campus twice a week, which that's a big deal. It's a big commute. It means they can't be in their other classes. It's not the usual model if I just walk across the hall and I'm at my next class. It's a big commitment.

                                    Because it's online, they're able to join the class when they wouldn't have been able to make those kind of commitments otherwise. And so we see students from a much broader range of counties across Georgia than we would see if we require the traditional kind of participation. Generally speaking, we don't see many schools send us multiple students the same semester to the course. So there's not much happening at the individual schools right now, but that's something I'd really like to see us get toward, because as this picks up and expands, there's no reason why Norcross High School about 30 minutes north of Atlanta couldn't host an entire cohort of dual enrollment students all in the class together, who could have their own teaching assistant dedicated to them, because we can justify that if there's 20 students at the same school all enrolled at the same time. I'd love to see it get to that point.

                                    What we have seen is a couple of charter schools around the country have also started to offer their own sections of the course. And what they do is they use our material, but they hire their own teaching assistant and they handle their own grading and they handle everything like that on their own. But they're able to hire a teaching assistant who doesn't need to be on ground with their students because they're meeting online, and they're able to take advantage of our instructional content and all of our assessments, which are all graded automatically. And so they can create a CS1 course without hiring an in-person CS teacher that leverages some proven material, some high quality material, and it also includes some human support without having to go out and find an actual CS teacher who can be in their classroom five days a week, because that's really hard to find.

                                    So that's the other thing I'm really excited about with that. It lets us distribute to places that can't have access to CS1 education otherwise.

Justin Reich:                 In the earliest days of massive open online courses, one of the conversations that came up frequently was is it better to understand a MOOC as a course or as a textbook?

David Joyner:               Mm-hmm (affirmative).

Justin Reich:                 Is it the whole enchilada, or is it really the instructional materials? And it sounds like you're trying to answer both like, "Well, if you need us to provide the certification, if you need us to provide the registraring, if you is to provide the instruction, the grading, we can do all that. But also, if you just want our instructional materials, then we've done two things. We've built everything in a way that you can just pick that up and take it. But also, there's a sort of ethos behind it, which says, "We are here to serve as many learners as we possibly can." Instead of bragging about how many people we turn away, about how many people get rejected from the Georgia Tech College of Computing each year in the admission cycle, we're going to start bragging how many students around the state we can get Georgia Tech college credit in every given year." Does that sound like a fair analysis, or is there other ways that you would think about it?

David Joyner:               Oh, yeah. So I'm pretty sure you are actually in my office right now because I'm working on a paper, and that's exactly the section I wrote today, was that 30 years ago, 40 years ago, there was an ethos in some colleges about ... It was the old adage, "Look to your left. Look to your right. One of the three of you will not graduate from this institution or won't be here next year." And at the time, that was considered a point of pride, that we're so rigorous that we fail out one-third of our students.

                                    And fortunately, things changed. And now, it's at the point where having a high retention rate is valued and it's part of what makes a college reputable. And that goes into all the rankings and things like that. But in return, what colleges do is they get more and more selective. So they're able to say, "It's good that 97% of our students come in, because we were really selective about who we wanted to have in the first place."

                                    But that locks so many people out who were qualified for the experience and would've succeeded given the chance. And that's been part of our mission with this, is to make it so you should never be proud to reject someone who could have succeeded. You should try to do whatever you can to accept everyone who has the potential to succeed. And that's the foundation of that OMSCS program, is that we let in anybody who meets our minimum bar. And as a result, our acceptance rate or something like 75, 80%. It's sky high.

                                    And we get in conversations with people who say, "Well, it can't be a reputable program if it lets 80% of people in." Why not? If those are the people who have the potential to succeed in the program, why wouldn't you let them in? And I'd rather let 10 people in who will ultimately not succeed than reject one person who would've, as long as we're not charging them an arm and a leg, which is the other defining feature of that master's program, is that the tuition is extremely affordable. It's about $500 per class. So we're able to give people the benefit of the doubt, give people a chance, because we also know we're not saddling them with five figures of debt just to enroll in one semester and then fail out.

Justin Reich:                 So in your book The Distributed Classroom, one of the things that you described as a defining feature of the distributed classroom is minimum necessary compromise. What's minimum necessary compromise, and how would you map it to the CS1 class that you've been describing for us?

David Joyner:               Yeah. No, absolutely. So I think of minimum necessary compromise as there exists this traditional experience where you're in the class with your classmates with the teacher at the same time and same place, regular schedule, and there's a high cost associated with that. And that's the structure that locks so many people out. It locks them out either because they can't afford it, because they aren't competitive with the students who are going to get into the program, most likely because of what opportunities they've had earlier in life, or because they can't commit to that regularity because of other obligations they have, taking care of family, everything like that.

                                    And so it says, "What is the minimum part of that traditional experience you would have to give up to be able to participate in the class?" Is it that all you have to do is give up coming to Atlanta Monday, Wednesday, Friday, 9:00 AM? You can do everything else. As long as you don't have to come to Atlanta. It can be going somewhere local in your city. If you live in New York, you might be able to commit to going to a classroom in New York. You can't move yourself to Atlanta for access to that. Or it could be you can commit to doing something synchronously at a given time every day, but it just can't be during work hours or can't be during hours when you have your kids at home. It has to be after bedtime or something like that.

                                    And so it just asks the question, "What's the minimum amount you have to give up to participate in that educational experience? If you can't be there at the same time at the same time on a regular schedule, what do you need to allow you to? And then in turn, what can we say about what you learned in the process anyway?" And that's where the online master's program serves as a good example of it, is that it gives the exact same credit, it gives the exact same diploma, everything about it's the same. They're transferable. You can move credit from one campus to another if choose to move to Atlanta later because we're able to say, "Despite the fact that you're compromising, despite the fact you're giving up attending in person, you're still doing the same assessments, you're still learning from the same faculty, you're still accomplishing the same learning outcomes. And so we're comfortable giving you the same credit."

                                    MOOCs actually give a good example of this as well, because MOOCs don't give academic credit. So in that case, it's a case of you're not able to commit to the tuition prices of coming to campus. You're not able to commit to attending synchronously. You might not have the prerequisites. You might not have the prior qualifications. If it's a class that's offered as a graduate level class for credit and you don't have a bachelor's degree, you can't take it for credit there. But you're still able to commit to learning the content, and so you get some other kind of certificate out of the equation.

                                    So it's just what parts of the experience can you maintain and what parts can you commit to? And what can we say about your learning in exchange? And it's no longer the case that you have to be here for us to say anything about what you learned.

Justin Reich:                 What kind of expertise has Georgia Tech have to assemble in order to be able to generate these programs? What's been the hardest thing for you and your colleagues to have to learn over the last few years to make these work?

David Joyner:               Oh, gosh.

Justin Reich:                 Or were you sitting on top of all the expertise that you needed and just ended up deploying it in different ways?

David Joyner:               No, I think it's been a really interesting experience, because there's been a lot we had to learn, but a lot of what we had to learn was just how close we were in the first place. And one of the things actually in the book that is what comes up in conversation a lot is that we underestimate how distributed we already are. As we've slowly shifted over the past 20 or 30 years towards relying on learning management systems and grading things and sending grades back via those systems, answering questions on forums, having chat rooms and things like that for our classes, we've gravitated towards a place where we're actually very distributed in the first place.

                                    And so one of the things I try and help our faculty understand who are about to undertake developing a new class is that the majority of what you're going to do in teaching this class is not going to change when you go online. You can do the same projects. You can do the same assessments. Most things you can do pretty similarly, because you are already so heavily relying on technology to distribute the experience in the first place. We're not at a place anymore where you have to be in class to turn in your homework at the front of the room at the end of class. We're at a place where we're submitting it online anyway. We're grading it. We're sending back grades online. So much of what we do is already well suited to this. This is the last step. It's not the first step. It's just saying, "That one thing that I'm still doing live and in person, now can I move that piece to a distributed kind of area?"

                                    And over the past couple years, we've seen everyone's had to do that. So we've seen it's been under the worst conditions. So I think a lot of the negative experiences people have had have been because everyone had to do it on two weeks notice in the middle of the biggest crisis most of us have ever faced. But what we've been able to do, what we've been able to find is that we were already pretty close to there anyway. If the pandemic had hit 10 years ago, I'm pretty sure schools would've had to shut down for months instead of going to virtual learning for a while, because we weren't at a place where we had the technological infrastructure to just have this last piece fall in place.

Justin Reich:                 One of the things that you mentioned was grading and assessments, and we should probably, for folks who haven't read The Distributed Classroom yet or don't know about the Online Masters of Science and Computer Science, you describe the course as MOOC based. But I think when most people think of MOOCs, they think, "Oh, so it's all auto-graded problems. It's all just the computer will check my computer code, or in many, many MOOCs, it's just all multiple choice or something like that."

                                    But this is one place where your program diverges quite substantially, because to host these thousands and thousands of students every year, by far the largest master's program in computer science in the country, you're probably graduating something like 10 or 15% of all students who are in a master's degree in computer science here. You all hire an army of teaching assistants every year. You've got something like 300 of them or something like that.

                                    It's connected to this idea of minimum necessary compromise where you thought, "What is it that we want people to be able to do? Can we compromise on human grading? No. We really can't. We're [crosstalk] going to have to find a way to keep human grading because that's not a compromise we're willing to make."

David Joyner:               Yeah. No, exactly. That exactly right. I'm glad you asked that, too, because I once had a conversation with somebody where we talked about the program for an hour. And at the end of the conversation, he said, "Well, it's great to hear you think you've had some success. I just don't think you can give a master's degree based on a bunch of multiple choice tests." I realized, "You've had this misconception about what we do this entire conversation that never came up, because for us, it's so fundamental that that isn't the way we work, that we don't even often think of mentioning it."

                                    So you're right. This semester, we employ 450 teaching assistants for my classes alone, including my undergrad class. I have over a hundred. So I have more teaching assistants this semester than we had students the first semester I taught or co-taught a class in the program.

Justin Reich:                 If you had had 50 more teaching assistants, it would've been a small business required to have a vaccine mandate-

David Joyner:               Exactly.

Justin Reich:                 ... under the original proposal of the Biden administration.

David Joyner:               Yeah.

Justin Reich:                 You're basically running a business, which is larger than a small business.

David Joyner:               Yeah, exactly. And that's the fascinating thing. So my PhD student Bobbie Eicher has done some research just to try and find out, because one of the things I think that makes our program great, if I do say so myself, is that our faculty have all the control over their own classes that they want. It's not this kind of model that I've seen from some other places where it's a very heavily, sanitized is almost the right word, where it's really delivered by a team of instructional designers instead of the faculty. There's lot of merits in that approach of having experts really take care of the day-to-day experience. But I think what makes ours [crosstalk].

Justin Reich:                 That's a lot of what you would find at for-profit online schools, the largest online schools, where basically there's a team of instructional designers that build a course package, that build this kind of online textbook, and then you hire, as a different set of people, instructors who facilitate the learning experience. And probably in the worst of those circumstance, you basically have people who don't know the curriculum that well just being there to vet people's assignments and shuffle them through.

                                    In the best possible version of this, you could imagine that you're just giving different educators different talents and strengths the ability to specialize. If you don't have to do any curriculum to design because someone's given you a really terrific curriculum, then you can just focus on the teaching and learning. But you were describing that that's not what you're doing there. Your faculty design the course, facilitate the course, but they facilitate the course with an army of up to a hundred teaching assistants.

David Joyner:               Yeah, exactly. I think it's the case that at this level, if you're doing a graduate degree in computer science, the reason why you do that at a place like Georgia Tech is because you want to take classes from the people doing their research out the world. The people are right there on the cutting edge, and you can bring in that expertise into the classroom, as opposed to the more well established classes where you can grab the textbook off the shelf and know that this field, it's like calculus. This field is well enough established that we can start to optimize the curriculum, as opposed to just concentrating on making cutting edge content available.

                                    But, yeah. So we have 450 teaching assistants this semester, and they are absolutely incredible. One of our early findings was that we didn't think that online students would be interested in being teaching assistants, because we know they're mostly mid-career professionals, they have families, we don't pay competitively with what they make in their day job obviously, and we don't even have what we have for on campus students, which is a tuition waiver, which is really the the nice reason to be a teaching assistant is that you get free tuition. Our online students don't get that. And even if they did the tuition is so much lower that it's not worth nearly as much.

                                    So we didn't think online students would be willing to be teaching assistants. We found out we were completely wrong. They love it. They flock to it. We turn away far more people than we can hire each semester. And they're better suited for the role than other students are, because they actually know what it's like to be an online student. So they've been at home in their basement office working on their own. They know what it's like to ask a question on a forum and have hours go by and not get an answer. They know how isolating that feels. And they bring that expertise to their role as a teaching assistant.

                                    So as a result, they end up being just fantastic at it. And they also bring in that professional experience. So one of my TAs, he's a teaching assistant for my educational technology class. And he, as a professional background, has worked at the executive level for several major textbook publishers. So he brings in an expertise in educational technology that goes far beyond what you would usually get from a course staffed by the great students from the previous semester, who are great people, but they don't have 20 and 30 year careers in the field.

                                    And what's even more remarkable about that is he's been a TA for the class for seven years now. On campus, you're a TA for maybe two or three semesters, and then you graduate and you move on. Online, you never have to move on. We can hire alumni as teaching assistants. And now, that's actually almost the majority of our teaching assistants, are alumni of the program who just want to stick around and continue helping out.

                                    And this is my roundabout way of saying, so as a result, the vast majority of work students do in program is human graded. My PhD student Bobbie did some research to try and find out what is the fraction of work that's project based, what's homework based, what's exam based, what's human evaluated, what's AI evaluated, what's peer evaluated, because we know some programs and some other even degree programs use a lot of peer evaluation.

                                    And what she found was that 75% of the work that students do for their degree in our program is evaluated in whole or part by people, so not by an auto grader, not by a multiple choice thing. It's actually evaluated and graded by human graders. And where we do use automatic evaluation, it's very often running code against a real simulation kind of thing. So it's authentic to code. It's not auto graded for the sake of convenience. It's auto graded for the sake of that rapid feedback cycle kind of thing.

                                    And so the human is still very heavily in the loop in the program. And it's what makes the program, I think, so good and so much fun to work on, because you really are working with a lot of really great people who bring this great diversity of perspectives and experiences to the classroom environment in a way you couldn't do if you weren't online in that scale.

Justin Reich:                 What are some of the either quality control or community building? How do you help your hundred TAs for CS1 be really good, be in sync with each other, those kinds of things?

David Joyner:               Yeah. So I should say for CS1, I actually have about 15 TAs. The hundreds [crosstalk] of TAs are for the graduate program. So I'd say that that's the other thing that's interesting, is that every class kind of has its own identity, its own community in a way that I think is part of what makes it a really, really nice program, because you're not getting this very sanitized kind of experience where every single class, everything is going to operate the exact same way, regardless of the affordances of the content. The professors and the teaching assistants are able to tailor things to whatever works best for their own content, for what they want to accomplish in the class.

                                    And so I teach four classes at the graduate level, in addition to the undergrad class, and they all are actually pretty different. So one of them is heavily project based, where students propose their own project. One of them is heavily code based, where students are actually coding little mini projects to run against simulations. One of them is heavily essay based, where students write a whole lot during the semester. Actually, all my classes involve a lot of writing. So that's just my bias as to what you should do at the graduate level. But they're all different enough.

                                    And I think the way that those classes end up facilitating their community really does come back to those teaching assistants, that they have the leeway in this environment to have a lot of impact on the class experience, that they're there on the forums, they're there in the chat, they're there wherever the class is delivering its experience, constantly managing and facilitating this interaction with students. And we're able to find people who are just fantastic at that.

                                    So I think being the TA who is responsible for managing the forum community is a very specific kind of role. And there are individuals who are just fantastic about it. They're just such a positive presence that when they post, you can just feel the excitement coming from that person. And it really comes back to the fact that we're able to bring in this diversity of people. We're able to bring in people who are ... You're a really cheerful, uplifting person. You're going to be in front of students. You're going to be active on the form answering questions, because everyone's going to feel good having interacted with you. You, on the other hand, are a very rule based, policy based person. You're fantastic for running our plagiarism checks and making sure that students aren't copying essays from repositories and things like that, because you're very detail oriented and you're able to chase those cases down. You, some other person, are great at getting a pile of papers and giving great feedback on them on time, so students get that rapid kind of feedback.

                                    In a traditional class, one person has to do all of those roles. When I just started teaching, I had to do plagiarism checks myself. I did a lot of grading myself. I was active on the forum myself. And what I found was that I'm good at answering questions. I'm good at facilitating discussion on a forum. I apparently have that kind of excitement that comes through. I'm really not good at those other roles. I'm not detail oriented enough to follow through on some of the plagiarism checks and things like that. But we have this size that lets us find the exact right person for each individual task, instead of forcing professors to wear every single possible hat to deliver a class in the first place.

Justin Reich:                 So David, I know you, and I think the book describes this as optimistic about this as a model, which deserves more consideration. And in this conversation, we've hit on a bunch of these things. You can just reach more people. Even if we were to stipulate, "Hey, not everybody likes online learning. Not everybody is well suited to this or that," and I think there's good research to believe all of these things, but there are all kinds of high school kids around Georgia who can't make it to Atlanta, who can't arrange their schedule, who would really benefit from this. And so let's find those folks. Let's find working professionals who want to do this and that you're able to ...

                                    I think one of the things I appreciate about The Distributed Classroom is that the two ways to think about large scale online learning, one of them is to say something like, "The ideal learning experience is a one-on-one tutorial, and everything is worse than that. And we just have to figure out how we can approximate it as closely as possible with computers or other things like that. But we'll never actually get there."

                                    The other way to think about large scale learning is that there may be things that we can do that are just better. What if your teaching comes from not one person who's good at some things, but not others, but from a whole team of people who are all really good at the thing that they're focusing on? There could be some really good outcomes for that.

                                    So March 2020, along comes a pandemic. Courses all around the world have to shut down. And certainly, in those early days, I thought to myself, "Wow. This could really be a moment in which massive open online courses or the way we've described these things as online, these could really break through in credit bearing higher education, because the alternative, if someone somewhere is at a state university and they all got sent home in March of 2020, maybe they had to stay home in fall of 2020, either they have to borrow something like the Online Master's in Science and Computer Science and use that as the online textbook and have some kind of local facilitation or some professor who's never taught online before has to go sit in front of their home office webcam and teach on Zoom with two kids who are not in school crawling under their feet.

David Joyner:               Yeah.

Justin Reich:                 It couldn't possibly be the case that having thousands of college faculty across the country basically by themselves invent courses on Zoom is better than using all of the existing resources that all the large scale learning folks have built over the last decade. But that is exactly what happened.

David Joyner:               Yep.

Justin Reich:                 Exactly what happened as far as I can tell in almost every credit bearing organization of higher ... Apparently, zillions of people signed up for MOOCs, but I think they are mostly people doing it for leisure time or because their work requirements changed or they had more time at home or something like that. But the professor who's teaching microeconomics who had never taught online before just hopped in front of his webcam and taught some crummy version of online Zoom in a box microeconomics instead of taking one of the many classes is that edX or Coursera has built for hundreds of thousands of dollars and using those. Why wasn't the last couple of years the big moment for those kinds of approaches for the distributed classroom? Is it a demand side problem? Is it because students really just want their own teacher? Is it some kind of supply side problem, like universities don't know how to take advantage of these resources? What do you see as the obstacles there?

David Joyner:               Yeah, I think there's a couple things, because one part of it, I think, there's an analogy to software engineering actually, which I'm a computer science person. So everything's an analogy [crosstalk].

Justin Reich:                 That's fair, given the topic.

David Joyner:               Yep. But is that very often when you're talking to junior engineers and software developers, there's always a bias towards wanting to build something yourself as opposed to finding a library that does it. And it's I think something that we very often have to overcome, is actually looking at the resources out there, because it's that kind of thing where I know the effort that's going to be required to teach on Zoom. So it's familiar, it's comfortable. Even if you never taught on Zoom, you can do an analogy of, "Okay. Instead of teaching in front of class and putting my slides on the projector, I'll share my screen," and you can pretty quickly figure out what it's going to be like. And it's known, it's familiar. You know what kind of time is going to be involved, whereas going out and seeking other resources, you don't really know how long that's going to take. And it doesn't actually end up taking that long, but you don't know how long it's going to take.

                                    And we very often, I think, choose the thing that's familiar as opposed to the thing that's a little bit less familiar that could have ended up being better. I think we always see this with tasks we streamline as well, that we do something the hard way several times, because we think it's going to take a while to learn a way to do it easier. And then we actually one day sit down and do it and realize that we've cost ourselves so much time over the years by not doing it what would've been the easier way, just because the easier way involves some learning.

                                    I think that's part of it, I think, is that it was the default option. It was known. People could figure out what work was involved. I think there's also a concern that if I'm teaching macroeconomics, for example, if I don't teach the material myself, if I go out and find videos from the internet that are much better produced than what I could produce, but they're not mine, then there's this fear that it's not really my class anymore. And students might not respect me as the authority. The school might not respect the idea that this is my class, and there's this possessiveness and insecurity about if I'm not the one presenting it, then it's not mine.

                                    And I think that's not an unreasonable thing to feel. I think one of the models that we saw, at least at Georgia Tech, I'm sure other schools found that too, but I think it's my favorite model of some of this approach, is when faculty, they'd take videos and they'd take course material from Coursera and edX and YouTube and Khan Academy and all the different places that are doing so much great work to make material publicly available, and they'd use it, but they wouldn't just plop it in and say, "All right. This week, you're going to watch lesson five from this Coursera course on macroeconomics."

                                    Instead, they would say, "Hey, everybody. Welcome to this week of this course. This week, here's what I want you to do. I want you to watch this content from this place. And here's what I want you to pick up on this, because this is going to connect to what we're going to do next week." And I call it the hosted model, just because it reminds me of a talk show host, of introducing different segments and things like that, where they're still the guide to the course, they're still the arbiter, they're still the person who is ushering you through this course. They're just not the person who has to be the sole font of knowledge for everything. But they still play a very significant role.

                                    I think it is the kind of thing that, if you're first getting into the environment, you don't know how much work you're going to have to put in to make that kind of experience happen, but once you've done it, you find it's actually not nearly as difficult as you would anticipate. And it's also quite maintainable. So it's the kind of thing where you're able to then the following semester follow up and reuse a lot more than you would've reused otherwise, because the videos haven't gone anywhere.

                                    But it's not what we very often have in MOOC world as well, which is once it's produced, maintaining anything is really difficult, because you have to ramp up the studio space and find your video editor and everything like that again. You're still doing something pretty lightweight and organic with it. So you can then say, "This semester, actually, I think I'd like to add this lesson or remove that lesson or add some more content here" in a much easier kind of fashion, whilst the taking advantage of these great resources that are already available online.

Justin Reich:                 That's great. I think those are great supply side arguments, why faculty didn't go this direction. I'm really interested in the demand side piece, which is that if the argument that the polished online materials are really better than what people can hack together, it seems to me there would've been in various places student uprisings, "Hey, why do I have to take this lousy class that this guy is just doing his lectures online and we're sitting on Zoom, when I know from my other life experiences that there are these really good online classes that we could be taking advantage of?"

                                    And I have not heard yet of that demand side surge of interest, which also makes me think that there's a bias from students, too, that people actually want to be taught by a person that they have some affinity with, some connection to. If I signed up to go to Penn State or SUNY Buffalo or ASU, I actually want a professor from Penn State or SUNY Buffalo or ASU, because they're part of my team and they're part of my tribe and part of my community and things like that.

                                    But I think some of what you describe in The Distributed Classroom, in the long run potentially addresses some of that in the sense that a big part of what you are building are these communities. The courses that you're creating are not libraries of videos with projects that people check, but a living community of people that are helping each other and supporting each other.

                                    So if you could put yourself in front of the Secretary of Education of the state of Georgia and say, "I think some of these ideas really have some legs for K12," what would you say to whoever is the commissioner of K through 12 education in Georgia? What do you think middle schools, high schools can learn from the experiences that you all have had at Georgia Tech?

David Joyner:               Yeah. I think that the biggest thing that ... So I should say high school does present some very different challenges, and middle school obviously as well, in the sense that one of the key defining factors behind what we've done in the MSCS program is this idea of, for this class, there's only a couple dozen people in the entire world who can teach this material. So naturally, we have to scale up how many people that person can teach, because we're not going to find other people. At the high school level, you have sometimes a different equation with things like calculus and English and history, where the supply of teachers is a much bigger pipeline of people coming in to teach that.

                                    The question then becomes, What can we do with these kind of technologies and experiences to improve the outcomes of students?" And right now, I think teaching is one of the hardest jobs in the world, because teachers are forced to play so many different functional roles. I can't think of any other role in the world where the same person has to be the person who knows all the content, the person who can give good individual feedback, the person who can monitor progress, the person who can also handle students' mental struggles, the stress that goes along with things, but also be the disciplinarian, to play all those kind of roles all at the same time.

                                    It's the reason why, when we talk about things like intelligent tutoring systems, we say, "Are we going to replace teachers with AI?" No, because if we can replace teachers with AI, we've run out of things to teach, because that's the hardest thing for an AI to do is to teach.

Justin Reich:                 Yeah. That notion of the job of a K12 teacher having been impossible for a while and becoming increasingly impossible, I think, really resonates when we think about all that we're asking teachers to do, to be advisors, coaches, subject efforts, graders, curriculum designers. It's just a list of demands, which mere mortals can't do.

David Joyner:               Yeah.

Justin Reich:                 And there are good reasons to want schools to more complex and sophisticated in some ways. The world is more complex. There's more knowledge that people need to be successful. But we're not going to be able to have our schooling systems produce better educated, more well rounded, more specialized students just by asking the teachers who work in them to be able to do more and more and more.

David Joyner:               Yeah.

Justin Reich:                 Well, David, it's been terrific having you here on TeachLab. Really appreciate the work that you're doing to share this model with the world. And I think it's exciting to see how it'll continue to grow and change and how we can learn from it in the future.

David Joyner:               Yeah, absolutely. It's great to be here. I'm always excited to share some of what we do, just because whenever we have these kind of conversations, I realize things that we're doing that are novel that don't feel novel to us because we do them every day. But then we have a conversation and find out that idea that we take for granted is something that the world can use. So I love having these conversations.

Justin Reich:                 That's great. David Joyner, he's the Executive Director of Online Education and the Online Master of Science and Computer Science at Georgia Tech's College of Computing. And he's the author of the recent book The Distributed Classroom from the Learning in Large Scale Environments series, published by the MIT press.

                                    I'm Justin Reich. Thanks for listening to TeachLab. I hope you enjoyed our conversation. Be sure to subscribe to TeachLab so you don't miss any episodes from our new season. In today's show notes, you'll find a link to David's book The Distributed Classroom and his website, along with a link to my book Failure to Disrupt: Why Technology Alone Can't Transform Education if you'd like to dig deeper into what we talked about today.

                                    This episode of TeachLab was produced by Aimee Corrigan and Garrett Beazley, recorded and sound mixed by Garrett Beazley. Stay safe until next time.