This is an audio conversation from the CIO Talk Network podcast. The timestamps in the chapters and transcript below refer to the audio recording.
About this conversation
A CIO Talk Network conversation with Charlie McMahon (Tulane University), David Kuntz (Knewton Inc.), hosted by Sanjog Aul.
What the conversation established
The conversation established that big data and adaptive learning can personalize education that was long delivered the same way to everyone. David Kuntz of Knewton Inc. said 30% of high school students and 46% of US college students do not graduate, and argued that moving content online lets his company collect and analyze what each student reads, answers, and struggles with to recommend what they should do next. Charlie McMahon of Tulane University said education has not changed much since Plato, pointed to flipped classrooms and MOOCs (one artificial intelligence course drew 160,000 registrants against an expected 8,000), and said the near-term gain is mining data to improve student recruitment and retention, since students often leave over homesickness or money. Both agreed the bigger prize is giving teachers insight into the 130 or 140 middle students they rarely get to know, while guarding against data privacy risks and a widening gap between elite and online education.
Host Sanjog Aul talks with Charlie McMahon, Vice President of IT and Chief Technology Officer at Tulane University, and David Kuntz, VP Research and Adaptive Learning at Knewton Inc., about reforming education with big data. McMahon and Kuntz describe how analytics and adaptive learning can personalize instruction for millions of students, from flipped classrooms and MOOCs to Knewton’s recommendations that tell a student what to work on next. Kuntz points to dropout rates of 30% in high school and 46% in US college as evidence that something is not working, while McMahon frames the practical payoff at Tulane as better student recruitment and retention. Both weigh the risks, including data privacy and a sharper divide between elite and online education.
Contributors
- Charlie McMahon, Vice President of IT and Chief Technology Officer, Tulane University
- David Kuntz, VP Research and Adaptive Learning, Knewton Inc.
Host: Sanjog Aul, Founder and Host, CIO Talk Network
On this page: Key takeaways · Chapters · Explore more · Transcript
Key takeaways
- David Kuntz of Knewton Inc. said the education system is failing too many students, citing that 30% of high school students and 46% of US college students do not graduate. He argued that moving content online for the first time makes it possible to collect and analyze each student’s work and personalize instruction for millions of learners.
- Charlie McMahon of Tulane University said education has changed little since Plato and is now transforming quickly through new media. He pointed to Tulane pulling cable TV from residence halls so students consume content over their IP connections, and to flipped classrooms and MOOCs that deliver instruction at low marginal cost.
- David Kuntz of Knewton Inc. said the company’s aim is to maximize the probability that a student reaches their goals by recommending what to work on at each moment. He compared the idea loosely to a Netflix or Amazon recommendation, paired with analytics that give teachers and parents insight into how a student learns.
- Charlie McMahon of Tulane University said massively online courses can recreate a good teacher’s reach through social structures among students. He described an early artificial intelligence MOOC he joined that expected about 8,000 students but drew 160,000 registrants who formed cohorts and helped each other learn.
- Charlie McMahon of Tulane University said the near-term payoff of big data at an institution like Tulane is better recruitment and retention. He noted students often fail to return between years because they are homesick or lack money, and suggested mining social media to recruit small cohorts of students who fit together.
- David Kuntz of Knewton Inc. said data should serve both students and teachers, not replace the relationship between them. He said a typical high school teacher may have 150 to 180 students and know only the ten best and ten worst well, so analytics can flag an at-risk student and explain why.
Chapters
- 0:00 Introduction: reforming education with big data
- 2:30 Evidence that education needs reform
- 4:33 Dropout rates and the shift to digital
- 6:54 From Plato to MOOCs at Tulane
- 10:33 How Knewton personalizes learning from data
- 17:47 Corporations training workers on demand
- 21:00 Privacy and mining personal data
- 23:53 Helping teachers reach the aha moment
- 33:09 Measuring baselines and outcomes
- 36:55 Recruitment and retention at Tulane
- 47:58 Risk of a sharper education divide
- 50:14 Prescription plus insight for students
Explore more
- Challenges of Big Data
- Big Data on the Big Screen
- Making Sense of Big Data
- Big Data: Big Fuss or Big Biz?
- Making Big Data Smaller
- Big Data’s Movie Magic
Transcript
Introduction: reforming education with big data 0:00
Sanjog Aul [00:00:26]:
Today’s topic is Reforming Education With Big Data and our guests for today’s show are Charlie McMahon, who is the VP of IT, and CTO with Tulane University. Good morning, Charlie. How are you?
Sanjog Aul [00:01:26]:
Now we also have David Kuntz, who is the VP of Research and Adaptive Learning from Knewton. Good morning, David. How are you?
Sanjog Aul [00:01:38]:
Now, so with the school year just starting, we wanted to take a look at the topic today, which is close to my heart and like many people, I’m a parent and I want the best for my kids. I want them to be able to learn at their own pace, be challenged, and have the absolute best and yet sometimes that’s difficult and the other kids in my child’s classes are given the same methods and textbooks as thousands of others. So how can we start thinking about technology to help our kids meet their full potential? So to that end, Charlie, the first question I’d like to pose here that what do you think is the evidence to show that education really is a need for reform? There have been, of course, you know, people who are pro off of this,
Sanjog Aul [00:02:24]:
they are supporting this particular agenda, or some other people say everything is good.
Evidence that education needs reform 2:30
Charlie McMahon [00:02:30]:
Well, this is an interesting and, of course, complex question to deal with. It’s, I think, undisputed that higher education in the United States is the model that all of the rest of the world aspires to. We do better at educating at the college levels than any other nation in the world. Now, having said that, that doesn’t mean everything is rosy. There are reports that you see in the news over and over again of people being lured into very expensive educational processes that have questionable outcomes, that leave people without knowledge or skills or better opportunities for employment, or better opportunities for living a good full life that contributes to society. So I think if I look at where we have evidence for reform, for the need for reform, it really centers around those institutions at the margin that don’t provide good value to their students and we see examples of that in the press almost weekly these days.
Sanjog Aul [00:03:54]:
So, David, when you come from your vantage point and of course your organization is attempting to do what was otherwise almost seen as very difficult or even impossible in terms of getting a fundamental insight that how people are really getting educated, what is it that they really need and how do you build a fit for purpose, highly intelligent education system? So what is your pitch to the rest of the world? Why we need a reform? Why should we change the way we are getting our kids educated and the way we went ahead and took our education?
Dropout rates and the shift to digital 4:33
David Kuntz [00:04:33]:
Right and now, that’s a great question. If you look at the educational system today, I mean, one of the things that you see is that, you know, dropout rates are high, you know, 30% of high school students don’t graduate, 46% of US college students don’t graduate and yet we live in a world in which increasingly having both a broad and deep understanding of the world in which we live is critical to being a productive member of society, of getting a job, keeping that job and advancing in that job and so clearly something is not working right. We’re also at a really interesting point in history right now where education is moving from print, from the Gutenberg innovation of hundreds of years ago to the digital innovation of today. All of the content is moving online and that affords opportunities for us that have never existed before to provide highly personalized education to millions and millions of students. That’s just a new fact about the world
David Kuntz [00:06:11]:
and given that we have that opportunity and given that we have these failures in the educational chain, it’s an opportunity that I think we need to seize.
Sanjog Aul [00:06:25]:
Now, Charlie, when we look at the way the world is changing, or perhaps it’s moving at a fast pace, we want new skills, we want to prepare people for far more complex areas in which they want to be trained in. Do you think all of those areas we understand well today? And is our education system morphing towards that direction where we are more progressive and more at par with the pace at which the world wants us to be?
From Plato to MOOCs at Tulane 6:54
Charlie McMahon [00:06:54]:
Well, I do believe we are transforming ourselves at a fairly rapid pace. I mean, if you think about how education has been delivered since Plato, we haven’t changed an awful lot, you know, over these last few thousand years. Now we have different media that are available to reach large numbers of people very efficiently, very effectively. You couple that also with a society of learners who are predisposed to consume information in that manner, and I’ll give you kind of a crude example here at Tulane University, we are pulling cable TV out of our residence halls and are going to let students consume entertainment content completely over their IP connections, over the Internet, so to speak. We see that propensity of our new students interacting with information in the palm of their hands, also spilling over into the way they like to get educational content. So the things that you’ve seen recently, flipping the classroom, having short instructional videos that students can access prior to coming to class for a lecture has proven to be very effective. You think of things like MOOCs, where we’re able to deliver large amounts of information to students at a very low marginal cost as a new method of imparting information. Now, we don’t have that right yet.
Charlie McMahon [00:08:56]:
I was talking to provost at Stanford who’s been one of the leaders in developing MOOC technology, and he acknowledges we don’t have it right yet, but we’re going to get it right. That type of technology is going to be transformative, but we’ve got more work to do to get the technology right so that it delivers these customized educational experiences for our students, both formal and casual learners.
Sanjog Aul [00:09:29]:
Very interesting response. Charlie and David, I’m going to come back to you based on what Charlie just said. If we have the new type of plumbing, which is the way we disseminated information to people, that is transforming and we are going to make full use of it. Now come to also the people who are consuming that information, they understand the device and whichever device they are on, they would want to consume, but what they want to consume or how they want to learn. People are looking for personalization in any and every aspect of their lives, including education. So sitting from a student standpoint, yes, they’re going to be very happy if, instead of having to go to this class or maybe even in the class, they can learn from their iPads, but they also want some personalization in that area as well so that it doesn’t look like it was a broadcast from a professor and every student got exactly the same message, whether or not they were truly interested in it. What do you think? Am I going in the right direction in terms of the trends that you see?
How Knewton personalizes learning from data 10:33
David Kuntz [00:10:33]:
Well, so, yeah, first, I totally agree with Charlie. The technology is providing access in ways that it never has before and that’s a really important step. What Knewton is trying to do is bring highly personalized, high quality education to each individual student and when you, as the student, you know, are online interacting with a MOOC or with an online learning management system, everything that you do, the homework assignments that you do, the tests that you take, the activities that you engage in, and the material that you read, all of that generates data that is collectible and analyzable and we use that data to try and understand how a student learns, what they know and what they don’t know, and what kinds of approaches to learning work best for them. And the reason for doing that is because what we’re trying to do is maximize the probability that a student will achieve their educational goals and it might be that right now, getting this piece of content, getting this little bit of instruction in this way, in a graphical form or using this kind of analogy, or in a little chunk as opposed to a big chunk, is what’s going to move you forward most.
David Kuntz [00:12:19]:
Whereas 20 minutes from now or tomorrow, it might be a different approach and so at each moment we’re trying to figure that out, and that helps. Then I think, I mean, whether a student is on their iPad or in the class working with a teacher or professor, and that teacher or professor has a dashboard that gives them information about how that student learns, all of that ultimately provides better educational outcomes.
Charlie McMahon [00:12:49]:
Let me echo what David has said here. There is, I think, a trend for letting students decide what they want to study but where I really believe we are headed as our next step is wrapping some intelligence around mining the data around the best way to present a body of information to the learner. You know, we are still in an environment where someone is going to say, here is the content we want you to learn and at the end of that process, we’re going to do some kind of assessment to certify that you have learned that. You know, right now that’s primarily driven by a university who puts together a curriculum and says, if you’re going to be proficient in calculus, here are the things you have to know. I can see in a very near term where that might be a corporation that comes and says, we’re hiring a cohort of workers for a class of jobs that need to be filled. And we’re going to define the curriculum that needs to be learned and present it in a MOOC like environment and the ability to sense and individualize not so much the content, but the way that content is delivered to our consumers, to our students is going to be a really important emerging trend in education.
Sanjog Aul [00:14:26]:
So Charlie, you’re saying that almost every type of education that we are getting today, which is collegiate or elementary, eventually it’s going to become almost like a vocational education where it will be fit for purpose and that too will be driven by a corporation.
Sanjog Aul [00:17:21]:
So, Charlie, you did mention a very interesting phenomena that you see that is likely to happen where professionals will be created on demand by a corporation who have a need that they want to meet. Now, that becomes almost like a vocational training of sorts, which is on demand but is that going to become fabric of how our people will get educated and join the society and workforce?
Corporations training workers on demand 17:47
Charlie McMahon [00:17:47]:
Well, so I don’t know that it will become a dominant model, but certainly I can envision, you know, a few things that might trigger that to become a reality in the near future. Imagine an economy that is booming and corporations are finding it hard to get new graduates with the skills they want to fill the jobs that they have. So we’re in an efficient market and in a capitalistic society, when we have a demand, we’re going to find a solution to supply that need and I can see a corporation stepping forward much in the way you might see in Japan, creating cohorts that substitute for some of the things we think of traditionally as benefits of getting a college education.
Sanjog Aul [00:18:45]:
David, from the area that you come from, which is helping the organizations understand what does the person as an individual student want to learn and what are their strengths and weaknesses and accordingly, tailor make education that is to not have them become mechanics, but rather engineers. That is not just they know how to do certain things, but also why, so they can improvise as the world changes around them. What is it that you are after? What problem are you trying to solve as an organization yourself?
David Kuntz [00:19:18]:
Right, yeah, the challenge. So Computer Aided Instruction has been around for a long time. You know, 30 plus years. People have been working on ways to take advantage of the power that a computer brings to help the educational process. The big challenge there is how to do that at very large scale, how to do that for millions and millions and millions of students in essentially real time so the access problem is one that we’ve talked about already but if the access problem is solved and students are engaged with activities online so that data collection is possible, there’s a technical challenge around how you manage to process all of that information quickly enough and appropriately enough but the bigger challenge is that humans are complex and we need to model their behavior
David Kuntz [00:20:23]:
not just their behavior, but their learning patterns. Recollection and forgetting, you know, affective states like frustration and boredom, all of those kinds of factors play a role in a student’s ability or propensity to learn and so part of our challenge is to model those things so that we can then make good predictions and recommendations to students and teachers and parents about what those students should be doing next.
Privacy and mining personal data 21:00
Charlie McMahon [00:21:00]:
Yeah. So David, you really hit on what I think is a very important point. We have lots of data that exists about ourselves and our preferences and our habits and there’s actually a big discussion around privacy and control of that data. Now, Microsoft did something fairly interesting a year or so ago. They had a couple of researchers sit down and ask this question. What will you be doing 285 days from now at 2:00 in the afternoon? And you think, boy, that’s a question that can’t be solved but it turns out that if we allow information about our habits or the routine things of our lives, you can pretty accurately predict what you’re going to be doing, where you’re going to be 285 days from now at 2:00 in the afternoon.
Charlie McMahon [00:22:03]:
You know, so to the extent that we are willing to let our data be available for analysis, we can learn all kinds of interesting things that make our lives better. You know, some crude examples, if you have ever ordered a book on Amazon, you know, after you’ve ordered one or two, Amazon starts making suggestions that say if you like this one, maybe you’ll like that one or if you’ve ever ordered a pizza from Domino’s, you know, they will remember that last time you got a thin crust pepperoni with breadsticks and a 2 liter Pepsi. You know, by letting that information be available, letting people mine that information, we can do things for those individuals that make their lives easier and better and actually add value. Much the same thing that you’re talking about, David, in understanding how people learn and analyzing deeply the information about the process they go through to provide value in the learning environment.
Sanjog Aul [00:23:21]:
Yeah, David, what do you think is more important? Enabling the understanding of a professor or an academic institution to understand each individual and then they are in a better position to lead those individuals? Or is it better to enable that individual who is the subject or the person who’s actually going to be educated or who will be benefiting from all this analysis and accordingly, whatever tactics you’re going to use to get them impart education, who are you focusing on?
Helping teachers reach the aha moment 23:53
David Kuntz [00:23:53]:
Well, it’s not an either or, it’s a both. I mean, think about some of the best teachers that you’ve had when you were growing up, right. One of the things that made them great teachers was that they were engaging, certainly. But another aspect of it was that when you were struggling with something, they were able to come up with an explanation, an approach, an analogy, something that you were able to latch onto that made it click for you when you said, oh, oh, yeah, now I get it, right. What we’re trying to do is increase the frequency and probability of that aha moment and that comes from two different directions, both by providing the student as much as possible with the kinds of instruction and the kinds of information that are most suited for them, but also by providing their teachers with insight into that student that ordinarily they couldn’t get.
David Kuntz [00:25:06]:
If you think about a high school teacher, a typical high school teacher, they may have 150, 180 students that they see every day and, you know, they may know well the five or ten best students and the five or ten worst students and the other 130, 140 students in the middle they just don’t have the opportunity to get to know as well. By being able to provide them with the kind of insight that they could only get ordinarily through extended interaction with those students, which they simply don’t have time for, they can contribute to that aha moment in ways that perhaps they couldn’t before.
Sanjog Aul [00:25:53]:
So, Charlie, if you were to look at the way the education system is, are there any systemic problems which this approach is likely to solve for us?
Charlie McMahon [00:26:11]:
Well, I think that when you talk about education at scale, certainly there are issues that these technologies are going to help solve. David put his finger really on the essence of a good teacher and a good learning experience, where the teacher has an opportunity to observe the struggles that a student is having and find ways to reach that student to help them break through an understanding of the concept. The key in these massively online environments is how to replicate that for hundreds of thousands of people simultaneously. And in my view, the way that’s going to happen is by leveraging social media types of interactions that many of our new matriculating students are already familiar with. I’ll give you an example. One of the very early successful MOOCs was a course on artificial intelligence and it happened to be something I was interested in, so I signed up for that course. It was amazing.
Charlie McMahon [00:27:30]:
The social structures that evolved almost instantly with this course. I remember signing up and the expectation was we were going to have about 8,000 students. There were 160,000 that registered. We had cohorts within that group that formed natural affinities for each other who kind of helped fulfill the role that the teacher typically has in David’s earlier explanation. So the ability, for example, for students to help each other solve problems, answer questions, enhance understanding has got to be enabled by social media and it turns out that we’re pretty efficient about that, that pretty quickly students understand when someone, you know, when Charlie says, here’s what I think, that Charlie’s opinion has a certain value based on the collective judgment of the group. You know, either I’m usually right or I’m usually wrong, but that gets understood pretty quickly and by the way, that’s a proxy for that
Charlie McMahon [00:28:50]:
some employers are starting to look at to say how good an employee is this person likely to be if they’ve been in a MOOC, in this social environment where they’ve worked autonomously, where they have been able to leverage groups and lead projects. All of a sudden now you’ve got information to be mined as a prospective employer that is value add on top of simply can they pass the test or not?
Sanjog Aul [00:29:22]:
So if you were to look at David, where we are with respect to creating the tactics so that we will be better in terms of imparting education and students will be better in terms of receiving the education and make full use of it, how are we measuring the baseline and how are we measuring the incremental improvement that may have happened? Based on some insight that you gathered and based on that, some new strategies were deployed to incrementally improve the way the exchange happened and education happened and the learning happened.
Sanjog Aul [00:32:29]:
So David, when we are talking about anything, in order for us to make investments and we want to see results, there has to be a baseline and there has to be a measurable improvement or degradation for that matter, or some sort of a measure that will tell us whether we are going in the right direction or not. So what is the starting point any educational institution can have as a baseline and what should they be measuring as they go about finding out more about the students and the teaching methods and effectiveness and then make changes and then later on find whether all that money and all that effort really paid the dividends or not?
Measuring baselines and outcomes 33:09
David Kuntz [00:33:09]:
Right. No, great question. I mean, and there are both internal and external measures that are worth looking at. You know, the big kind of high level things are things that we mentioned already, things like dropout rates, you know, those are long term indicators, but of course, absolutely critical ones. If we start to see dropout rates fall, if we start to see passing rates increase, those are good indicators that the educational process is improving. There’s a danger there in that you could easily make dropout rates go down and passing rates go up by circumventing the system, right, by simply promoting students and there have been some school districts that have been engaged in that activity in years past
David Kuntz [00:34:16]:
But at the bottom line, you know, we’re trying to make sure that students are able to know and do the things they need to be able to know and do in order to be productive members of society. So another kind of thing that we look at, if you think about the kinds of indicators that we have right now, we have report cards when students are young, we have standardized test scores, college admission tests, you know, high school exit exams, and we have, you know, postgraduate assessments of various sorts. Those are very coarse indicators that happen very sporadically. Being able to collect and analyze the data that students generate over the course of their educational career enables the creation of essentially this learning profile that at any given moment provides direct insight into what a student knows and doesn’t know, what they’ve mastered and what they have, and that I think, more than anything else, will be able to tell us whether education is improving and whether educational outcomes are improving.
Sanjog Aul [00:35:52]:
Charlie, if you were to invite someone like Knewton as an organization who provides this insight or claims to provide this insight, and they were to work on pay for performance versus a flat fee upfront, and this is just a makeup scenario, what is it that you would set as expectations that if I am going to be able to get this quantifiable results, only then I owe you X otherwise not and what challenges would Knewton have in accepting those because if you put something as high as, or a goal like, okay, if my dropouts reduce or if my grades incrementally increase across the board, but that’s not a direct measure, that’s not actually giving you any idea about the cause and effect, because there are multiple variables which would come into play besides putting in a Big Data solution so coming in context with the Big Data solution, if applied in the education systems, what performance indicators, what KPIs would you be looking for?
Recruitment and retention at Tulane 36:55
Charlie McMahon [00:36:55]:
I’m going to narrow the scope of your question a little bit and just talk about an institution like Tulane, an, I’ll say, elite, you know, PhD granting institution, and so this is going to ignore, I think, very real societal issues that we might have with dealing with, you know, dropout rates, as an example. No, I have to say something about this because it’s an important issue. You know, when we focus on improving dropout rates, we really are ignoring some of the societal issues that we have in folks with lower incomes where it may not be possible for an individual to go to school exclusively full time for 4 years or 5 years and get a degree. You know, we may be looking at individuals that have to work a semester and take a semester off, take a semester off to work to save up enough money to pay for the next semester but I think it’s a little bit unfair to certain segments of society to say this is, you know, one of the outcomes that’s really desirable.
Charlie McMahon [00:38:23]:
Now, having said that, with my caveat, if I were to invite Dave into Tulane and say, here are a couple of issues that we deal with that we want you to help us with and help us by mining the data that you can access to help us focus on two things, one on recruitment, and the second on retention. By recruitment, you know, we need to attract enough of the right students to fill our classrooms. Just like any other business, university is better when it operates at full capacity, it’s the most efficient from a cost standpoint, the most appropriate way. We, all universities, struggle with these two issues. Can we attract the right kind of students? And once we attract them, can we keep them? And this goes, you know, certainly to the things that David has been talking about.
Charlie McMahon [00:39:26]:
You know, how do we understand the progression students are making day to day, year to year, in moving towards completion of a degree? Some of that is certainly knowable by mining information about their performance in the classroom. Some of the things that are important to us really deal with social issues and so I’ll say, for example, at Tulane, when students fail to come back, say between their sophomore and junior years or between their freshman and sophomore years, there are two things that typically we see topping the list. One is that they’re homesick and by that I don’t mean they want to go home to mommy and daddy, but, you know, maybe most of their friends went to a different school and they want to be with their friends and the second comes down to money, you know, they just don’t have enough money to continue their education and have to stop for a while.
Charlie McMahon [00:40:34]:
So if David were able to come in and provide data that helped us attract the right students, and once we’d attracted them, helped us keep the students in school, that would be a big win.
David Kuntz [00:40:49]:
I can’t help with the money, and I don’t know that we can help very much with the homesickness.
Charlie McMahon [00:40:58]:
Yeah, but actually, I think you can, David. You know, I think that by mining social media, you can understand that when we’re recruiting a very desirable student from a location, that maybe that’s the only student from this school or city, that by mining the social media, you may be able to say, Charlie, you need to go attract these three other students also, because they’re part of the social clique and they fit your profile of a student and now we have created a little mini cohort that ensures that all three of these students may have a more successful experience at a place like Tulane.
Sanjog Aul [00:41:47]:
Hey, you got a million dollar idea from Charlie on this show David, so what would you do further to expand on what Charlie just mentioned in terms of how he would have liked to utilize your organization to help meet that need? And how are you going to compare this with many other use cases where you may have had conversations with other leaders within educational institution and what are they looking for?
David Kuntz [00:42:16]:
On the admission side? Again, the measures that we have right now, I mean that are widely in use right now, are fairly coarse and so, I mean, the admissions officers at Tulane or at other universities are quite skilled at identifying characteristics that will make a good mix, that will make a good community at their university. Adding to the information that they use to process that information about how students learn and work together, I think might also be really valuable. One of the things that we do at present is provide a mechanism, for example, in our math readiness product that lets instructors create study groups, blends of students who, along a variety of different dimensions, students who are struggling with the same thing, for example, or students who are working on the same kinds of material because in an adaptive learning environment, all your students may be all over the place in terms of what they’re working on. But most importantly, we also enable instructors to put together groups of students who are complementary in certain ways, whose approaches to learning are complementary or whose knowledge base is complementary so we can identify students who are strong in one area and weak in another, and pair them with groups of students who are weak in that first area and strong in another, so that they can teach and learn from each other and that on the retention side, at least in the universities where that’s been implemented, this is something that both students and instructors alike seem to be very happy with.
Sanjog Aul [00:44:18]:
How about looking at the loss of human element, if you went to Big Data, ish, if you will, in terms of your ability to identify what that human being wants or does not want. So the relationship between professor and anyone who is in academia and the student is a very sacred one. When we start bringing too much of computer based intelligence on how I should assess that individual as a student, would there be a risk that we would lose that relationship? And we will try to use them as subjects versus individuals. What could be the negative aspect of introducing way too much computing and or Big Data based or big data centric insight to manage your students and the relationships?
Sanjog Aul [00:47:40]:
So in short Charlie, would introduction of Big Data and trying to get analytics and insights govern everything that we do with education? Would that sever the relationship between a professor and a student and make it more a scientist and a subject?
Risk of a sharper education divide 47:58
Charlie McMahon [00:47:58]:
Well, so we do have risks with this, with this technology. One of the real risks that concerns me is that while using MOOC like technology can reach millions of more people than can currently access the university education, we are creating an opportunity for more elite, you know, a sharper divide between people that can get an elite education versus, say, a MOOC or an online general kind of education. I think there is some risk there, not so much the scientist subject, but a continuation of the haves and have nots, just offering more opportunities to people that currently have none, which I think overall is going to be a positive. I don’t think this is a negative. I’m also going to say that in answering this question, when I think about this question, I think about this from the perspective of a guy that’s been in technology for a long time. If I were to think about this from the perspective of a new matriculating student, a high school graduate that’s going on for higher education, I might look and say, you know, I’ve had great, wonderful social interactions with my peers for a decade, with people that I’ve never met, and I may not think as a new student that not having that one to one interaction with my favorite faculty member is necessarily that big of a detriment.
Sanjog Aul [00:49:48]:
So coming to you, David, if you were to look at all the different data that you’re able to help collect, do you also have a recipe for success on where all this could be effectively utilized? Or would you take pride in just showing them the insight and let the academia and their own management figure out what they use? Is there at all a prescriptive angle to this? Or are you just going to provide the insight and that’s where it stops?
Prescription plus insight for students 50:14
David Kuntz [00:50:14]:
No, again, it’s both. So I mean, the core part, or a core part of what Knewton does is to provide a recommendation. I don’t really want to say analogous to a Netflix or an Amazon recommendation, but you could think of it kind of that way in that what it says to the student is here’s what you should be working on right now and so that’s a core part of what we do, but the other part is analytics that provide that insight back to parents and teachers and instructors. It’s the combination of those two, the direct prescription to the student and the additional insight to those around the student who help that student learn, that I think provides the greatest overall benefit. I mean, just as a simple example, just knowing that a student is at risk, right? Being able to tell a faculty member, this student is, you know, in danger, right? Is important, right? And being able to tell them, here’s why that student is at risk, here are the things that you might think about to help mediate that risk, I think is even better and actually affords, going back to what you and Charlie were just talking about, a greater opportunity for student teacher interaction than they might have had before because now the instructor knows something about that student that enables them to have a kind of direct and useful conversation with them where before they might have just been a face in a sea of faces.
Sanjog Aul [00:51:58]:
Yes, there are some things which we still have to get a good answer for, but at least it looks like it is a movement in the right direction.


