How to Increase Consumption of online courses in India

In my previous blog, I had talked about institutions moving instruction to online. It was a hybrid model where most of the lectures were online, but students spent some semesters on campus, not for lectures, but to do labs, projects, extra-curricular activities, make friends, and may be a few classroom sessions too. If institutions could recruit a star faculty at higher salary to teach larger classes online compared with two good/average faculty members for in-class teaching of small classes, they could save money and improve quality of education.

But a natural question is how can our poor quality educational institutions with no flexibility to recruit star faculty, improve their delivery of education through online. And the answer is quite obvious, they can outsource their teaching. There are several providers of online courses who would also certify that the student has learn enough to satisfy the course outcomes. These include SWAYAM (of Government of India), Coursera, edX, and niche players like Harappa.

In fact, augmenting one's own offerings of in-class instruction with online courses from other platforms seems like a better thing to do than to have everything (whether online or offline) done by in-house faculty. But will this really improve quality of education as assumed above.

We need to make some assumption here. What I have been thinking has been best articulated in a recent tweet by @learning_pt

(7/10 offline) > (9/10 online) > (3/10 offline)
last category most applicable to Indian colleges

What they are saying is that a good teacher (rated 7 on a scale from 1 to 10) in a class is better than an excellent teacher (rated 9 on a scale from 1 to 10) online. But an excellent teacher online is better than a poor teacher (rated 3 on a scale from 1 to 10) in a class. (Would love to read any research in this. Currently, it is an assumption which seems right based on one's own experience.)

There is no denying that a large number of our educational institutions have a terrible shortage of faculty. As a result, many institutions recruit temporary faculty every semester at very low cost, which often results in very poor quality teaching. If temporary faculty is not recruited then permanent faculty is so overloaded that they can't do justice to the instruction. In some institutions, the permanent faculty is itself of poor quality.

A simple solution would be for the college to tell its students that out of 5 (say) courses they are supposed to study in a semester, only 4 will be taught in class rooms, and they should register one course online that the college will specify (say, on SWAYAM). And the college will pay for the cost of doing that online course, since it saves the cost of hiring a faculty, which is likely to be higher.

In this solution, students get better learning, college saves money, infrastructure requirements go down, may be faculty teaching load also reduces. So every single stakeholder is better off in this model. Considering this, UGC came out with its guidelines in 2016 in which it encouraged universities and colleges to offer up to 20 percent of the curriculum in online mode.

However, adoption of online has been painfully slow till now. Why is it that a solution which is providing benefit to all stakeholders is still not being adopted.

There are several issues, which essentially show that there is lack of autonomy and such changes will not happen through UGC diktats.

The most important issue is that most of our higher education system still runs in the archaic affiliated college model where a university controls everything about a college. So a college cannot decide the syllabus and whatever syllabus the university has decided will not match 100 percent with what is being taught in the online course. So unless the university agrees that the specific online course is equivalent to the course that university had designed, the college cannot ask its students to do that course. And the university does not agree.

Why does university not agree. Several reasons. One, it does not have time to figure out whether an online course is equivalent (called "mapping" in UGC parlance). Improving quality of college education is not a priority item for most universities.

Two, there is a uniquely Indian penchant for fairness. This implies that either ALL students in ALL colleges agree to do the university defined curriculum, or NO student in NO college should do the university defined curriculum and instead do the online course. Anything else will be deemed unfair. So a decision cannot be taken at the level of a college. It has to be university-wide and forced upon ALL colleges affiliated with the university. So, if one college lacks faculty in one area and wants that particular course to be taken online, but another college has faculty in that area and wants to teach the university prescribed syllabus, there is a problem. And the simple solution is to not take any decision, which means that everyone does university prescribed syllabus, even if faculty is not available in a college. (And this issue of fairness is not just in poor colleges and universities. It is there even in IITs. You can't have the same course being run independently by two instructors. So no innovation. Everyone agrees to common minimum standards.)

Three, most universities in India do not understand the concept of doing a course on Pass/Fail basis. So they will attempt to integrate the performance of student in that online course in their own transcript and not know how to interpret the grade given there as marks of the university. Should an "A-" grade by online course provider means 80 percent marks of the university or 85 percent or something else. An easy solution would have been to consider this course as being passed but not contribute to your CGPA or marks.

Four, there is a fear of the unknown, rather, a fear of the technology. Is this really working. Can their evaluation be really trusted. (No one will say that their own evaluation can hardly be trusted.)

But what about smaller universities or autonomous colleges. Why aren't they adopting online courses. Well, a little bit of above. But there is another big problem. We have been pushing online (SWAYAM mainly) as a way to improve quality. And I too am guilty of this as this blog itself started with the assumption of quality, that an excellent online teacher is better than a poor in-class teacher, and hence online. It hurts egos to admit that an online course will be better than what some of them can teach.

Let us consider Tier 1 institutions. Almost every faculty there would be in the good to excellent range. So their own teaching is likely to be better than any online course by even an excellent teacher. So they refuse to consume online courses. So one IIT would not allow its students to register for a SWAYAM course taught by a professor of another IIT. Notice that if online was pushed as a way to provide flexibility to students - allow them to study electives that their own faculty do not offer, allow them flexibility of learning at any time (and if a lot of students do one course online, scheduling of remaining courses will become so much simpler and benefit everyone else too), allow faculty to offer esoteric electives for which there may not be enough interest in one institute but together with students from elsewhere, the course may become viable, and so on. As long as the focus is only on quality, Tier 1 institutions would resist allowing these courses to be credited by their students.

Let us consider Tier 2 institutions. Many of them do recruit poor quality faculty on temporary basis every semester, and could benefit immensely by using online courses. To some extent, the reasons given above hold, particularly, lack of exposure to technology and hence a lack of trust of these courses. Also, these temp faculty provide a very useful service to permanent faculty. Further, these institutions are desirous of getting good NIRF ranks and good score in accreditation. Our accreditation agencies and NIRF ranking does not recognize online teaching in the sense they will still seek student-faculty ratio and not look at the average class size of courses taught in campus. So, if they have to recruit faculty anyway to maintain the student-faculty ratio, might as well ask them to teach instead of doing nothing for that salary. Part of the reason is also that if Tier 1 institutions do not consider these courses as good enough for them, why should we consider them good enough for us. We are only marginally behind them in quality, if at all, will be the argument made.

So, each tier has its own reasons to not consume online courses. How do we then encourage online education in India. One possibility is that the experience gained during Covid-19 would have broken all mental barriers and convinced a lot of stake holders of the utility of online courses. If that happens, great. But I suspect that we might go back to our old habits, mainly because everyone has moved online without much preparation, and most of us have done our own courses online with poor infrastructure support and not really consumed other courses. Hence the experience and feedback may not be that great to sustain this beyond Covid-19 situation.

What do we need to do?

We need role models. If a few reputed institutions can start consuming online courses (perhaps in addition to creating online courses), and their students and faculty can communicate to the rest of the world that they are happier with the quality, choice, flexibility, etc., it would encourage a lot more institutions to try.

GIAN program, in which India invited faculty members from around the world to teach compressed short term courses, had desired that those courses be delivered live to multiple locations and everyone offers those courses for credit in their respective institutions. I don't know how many GIAN courses were offered for credit even in host institutions (where it was in live classrooms), not to mention in other institutions where it was beamed live. Now, GIAN has made it mandatory that the course be recorded in a MOOC kind of fashion, which is a positive step. But not enough. We need to ensure that these courses will actually be consumed by some institutions for credit. In my opinion, a course should be approved only when the host institution, along with at least a few more institutes agree to make it available for credit to their students. Otherwise, it is difficult to justify the huge cost of GIAN courses.

We need more autonomy to our institutions. They should feel empowered to take decisions. For example, they do not know if they are allowed only SWAYAM courses, or they can use any course by any institution.

We could incentivize institutions. May be some additional budget if an institution agrees to consume online courses. So, not only they save money, but they also get additional money.

And in all our communication, we focus not just on financial saving, or quality, but the flexibility it provides to our students.

What is interesting is that our student community has long been a fan of online education. If you look at the statistics of global players like Coursera, Indian students consume a large number of courses on their website, including paid courses. My own students keep telling me how much they learn from online content compared with in-class lectures. So our primary customers will be very happy if we can allow them some online courses.

Online Higher Education in India

I have been excited about online education from a long time. I recall my nervous beginning of giving a lecture on Computer Architecture course almost 35 years ago with students sitting in dozens of location and watching me on TV, and asking questions through telephone. When I returned to India, and joined IIT Kanpur, I tried something silly about 25 years ago. I sent out an email to lots of companies, offering to teach their employees Computer Networks for free. There were many registrations. And the model was that I will take a regular class in IITK. I would record my lecture (audio only), would transcribe it, include some pictures, etc., and send that material by email. My external "students" liked it, but it was too much work and I stopped it after a few weeks.

In early 2000s, I recorded all the Computer Networks lectures twice, once in a studio environment, and once in a classroom environment, and these were used by a lot of people to learn networking. But till now, I have not offered a real online course, and I will remedy that shortcoming soon. However, my learning has moved online. While I continue to learn new things through books and other printed material, or by attending lectures in person, more of my learning takes place through stuff available online, from Khan Academy, Webinars, Coursera, and so on. And I am convinced that increasingly, everyone will get more of their learning online.

But can we offer a full degree program online. Frankly, even I was skeptical till last month. But Covid changed that. And yet, I am not suggesting a fully online degree, but a hybrid model in which one spends some time on campus and some times learns from home.

What are the typical concerns with online learning?

1. Is learning from lecture as effective when the lecturer cannot have a personal interaction.
2. Can learning be as effective when there is no peer learning, no interaction in hostels, for example.
3. What about labs.
4. What about non-curricular aspects of learning, having fun on campus, teamwork, making friends, creating networks.

Let us first address the issue of effectiveness of lectures. If I give a lecture in a physical classroom to a bunch of students, and I give a lecture using video conferencing to the same set of students, what would be better. Quite likely, the lecture in the classroom. But that is not the right way to look at it.

People have looked at effectiveness of small versus large classes, and initially they asked the question: If I were to give a lecture to 50 students versus if I were to give a lecture to 100 students, which class would have learnt better. And the answer was the class of 50 students. But then the question was changed. If an average teacher were to give a lecture to 50 students versus if a rock star teacher were to give a lecture to 100 students, which class would have learnt better. And this time the answer was the class of 100 students.

Similarly, the question in the context of online learning ought to compare a physical classroom lecture by an average teacher versus an online lecture by a rock star faculty. The answer is likely to be an online lecture by a rock star faculty.

What a college (at least a private one) can do is to recruit the best teachers at twice the salary of "average" teachers, and ask them to handle a class size which is four times of what the "average" teachers were handling. If you look at load on a teacher in a typical decent college, there is a faculty to student ratio of 1:20, and each student is doing roughly 5 courses a semester. So on an average, a faculty is teaching about 100 students in a semester, typically in two courses of 50 students each. A rock star faculty can easily teach an online course with 400 students. That is pretty trivial, in fact. So everyone benefits. Students get better lectures. Institutions save money, not only in terms of faculty salary, but also in terms of physical infrastructure, and can share some of those savings with students and reduce their tuition. Faculty salaries go up substantially. So the job attracts best people.

The second point about peer learning needs to be handled. The online pedagogy will have to be different than classroom pedagogy. In classroom teaching, students are invariably meeting outside the class, in the corridors, in hostels, in library, in cafe, and they will ask questions, clarifications, from each other, thereby enhancing learning. Someone teaching online will have to encourage this explicitly by having discussion boards, may be some points for greater interaction there, by asking collaborative assignments, and so on. And the model that I am looking at for now, there would be some opportunities to get together on campus as well. More on that in points 3 and 4.

The third point is about labs. In the simple model that I am proposing, I am assuming that students will be coming to campus occasionally to do labs. If the college is primarily a local college, we could ask different batches to come to campus on a different day, like 1st year on Monday, 2nd year on Tuesday, etc., or consecutive weeks, like 1st year in 1st week of the month, 2nd year in 2nd week of the month, etc. In case of institutions with students from all over the country, we could ask them to come for one month at a time. Like 1st year in 1st month, 2nd year in 2nd month, etc. So we would complete all labs easily, giving enough time to students to explore, do projects, etc., and yet require only 1/4th of the infrastructure on campus, at least hostels and classrooms.

The fourth point is about extra-curricular, campus experience, etc. I would modify my approach a bit and insist that first year students stay on campus full time even if most of the courses may still be online.

So here is the overall scheme:

First year students live on campus full time. The lectures are online, but there could be some interactions in person. The labs are held regularly. They get to know their peers, make friends, etc. Once they know each other, they would continue to interact in future semesters when they are spending most of the time working from home.

Depending on how much you value on-campus experience, you could make second year off campus or on campus or give an option to students. If they are off campus, they need to come to campus for some time during the semester, as mentioned above.

For third and fourth year, one of the issues would be that there are many electives, and hence class sizes are small, and given that our strategy is to hire rock star teachers at double the salary, it would not make much financial sense to ask those faculty members to teach small classes. They could teach some classes but we could collaborate with other institutions and offer cross registration so that each class becomes a reasonable size. This could be done through organizations like Coursera, and not necessary through mutual agreements. Our faculty could offer a course which can have registration from other places, or we could offer our students an option to register for a course offered by some other university. Again, these batches would spend some time on campus, scheduled in a way that 2nd, 3rd and 4th year students don't come at the same time. Also, if anyone wants to spend the entire semester on campus, we may allow it with an extra cost.

We could now make another change. I am assuming scheduled, synchronous lectures for the first year. May be second year, too. But from 3rd year onwards, the courses are not through synchronous lectures, but through recorded lectures, the standard MOOCs. The idea is that after 2nd year, they are more responsible and don't need a strict timing. Also, by the end of 2nd year, they would have learnt some basic skills that they can even think of taking up a job and doing the rest of under-graduate degree in part-time mode. This will really help students from financially weaker backgrounds.

Now, one may feel that the quality issue has not been addressed adequately (though I believe that all this with rock star faculty can deliver even a better quality). To the extent there apprehensions are justified, we should also look at the benefits - significantly lower cost to the student, requires much less physical infrastructure on campus, thereby allowing campuses in the middle of big cities to expand their intake, and ability to learn any time from any where. The last point enables students to take up jobs and move to part-time education, a big plus for those from disadvantaged backgrounds. On the other hand, it also enables to finish degree sooner than four years, for someone who can spend more time on his/her education. So, even if there is some marginal decrease in quality (that I cannot perceive), the advantages are tremendous.

There are challenges, of course. Initially, rock star faculty may not want to shift to a college offering online education even at a higher salary because it is not clear whether this model will be a success in the market. UGC regulations may come in the way for now, though they are constantly updating their rules. The bigger problem will be accreditation and ranking. AICTE requires a teaching focused college to have a faculty student ratio of 1:20. They need to change that and find other proxies for quality teaching. Similarly, NIRF and other rankings consider faculty-student ratio as a strong proxy for quality of teaching, and the entire model here has been conceived based on hiring fewer top class faculty who can handle larger online classes.

Engineering Admissions 2020: IITs should use JEE Mains

In a normal year, this would have been the time when the 12th class board exams would have been over, JEE Mains would be over, and some students would be preparing for JEE Advanced for admission to IITs, which would have been held around 20th May. But this is anything but a normal year. The country had to be locked down. The exams had to be postponed, and we haven't even finished conducting 12th class exams, not to mention admission tests.

In the most optimistic scenario, JEE Mains is proposed to be conducted in the first week of June. To the author, this does not seem feasible. One can conduct a national level exam only when there are not many hot spots, and some limited travel to nearby cities is allowed. JEE Mains is taken by almost 10 lakh students, and you will find everyone being accompanied by 1-2 parents/elders. So we are talking about movement of 30 lakh people (including exam center persons, invigilators, etc.). I doubt if we would be in a position to allow that kind of movement anytime soon.

So, the next semester is unlikely to start in the last week of July. There will be delays. The question is how soon after the situation has improved can we start our colleges. Conducting JEE Advanced in the same format as every year will add five weeks extra to the delay. Since there is a joint counseling of a large number of engineering institutions, the admission to other institutes cannot happen before IITs are ready to do theirs. Even if we were to abandon the joint admission strategy, there is really no point in other institutions doing their admission early, since IITs are the most sought after engineering institutions in the country, and students will leave other institutions when they get admission to IITs, wasting a large number of seats in the next tier institutions.

There are two issues to be considered: Are the advantage of JEE Advanced to IITs so significant that they would go to the extent of imposing a 5-week delay on all other engineering colleges. And the second issue is whether this would be fair to students, and what all we need to do to ensure the maximal fairness in a bad situation.

First, let us take the issue of fairness to students. I am sure some of them (particularly those who would not do well in JEE Mains) would claim that they would have done better in JEE Advanced. They may even challenge the change of admission process in a court on the grounds that they have been working with a specific plan for two years or more, and it is unfair to change goalpost at a late stage. And indeed, some of that feeling would be genuine. Also, some people who took JEE Mains in January and got a good enough score to succeed for JEE Advanced may not have registered for JEE Mains in April. They may now want to take JEE Mains.

So we need to do multiple things. First, all students who had taken JEE Mains in January but did not register for April version should be now allowed to register for the same. To ensure that each one of them would have reasonable time to focus their preparation to the JEE Mains now, we should immediately announce that it won't be conducted in at least 6 weeks (or whatever duration) from the date of announcing change in admission policy. And finally, to mitigate the feeling that JEE advanced is a better predictor of talent, IITs should agree to have a much more liberal branch change policy for this batch. This would mean that even if you think you were good enough for a better rank in JEE Advanced, prove that by getting a better CGPA in the first year in the IIT of your choice.

From the perspective of IITs, there are two advantages of continuing with JEE Advanced. One is the fear of court cases. And second is a feeling of autonomy. One could also add a feeling of JEE advanced being a better exam in detecting talent, but I will dismiss that since IITs have never even cared to think what kind of students they would want to admit, they have never made their data public, or done research themselves in terms of any correlation between the performance in JEE and performance in IITs.

Considering how Supreme Court has been largely supportive of anything reasonably done by the Government, I think the courts would definitely consider Covid-19 as a very special circumstances in which some tough decisions needed to be taken in the interest of saving 5 weeks of several lakh students across the country. As far as the feeling of autonomy is concerned, I think that JEE Advanced is an important symbol of IIT autonomy, and it must be preserved. The fear would be that if this year IITs agree to admission through JEE Mains, they will be pressurized to accept students from JEE Mains every year. So, the government would have to give a very categorical assurance that they will not put pressure on IITs to disband JEE Advanced next year.

With this, it should be possible to admit students in IITs through JEE Mains this year.

While we are at it, let me also suggest how JEE Mains could be conducted as early as possible. As we discussed in the beginning, conducting JEE Mains means allowing 30 lakh people to travel over a period of one week. Now, if a lot of these people would have to travel a couple of hundred KMs, then one can conduct such an exam only after there is normalcy in pretty much all parts of the country. Managing permissions for lakhs of people to do inter-city travel if there are still restrictions would be a nightmare and not worth hurrying up the conduct of the exam. But if NTA can increase the number of test centers so that there is a center in most districts. This would mean that most candidates would have to do only an intra-district travel on the day of the exam. Such an exam can be conducted once we have virus spread under control in large parts of the country.

We need to take a quick decision, since it impacts conduct of JEE Mains also. If we don't decide within the next few days, then we would have by default decided to delay admissions in all engineering colleges by an additional 5 weeks.

Note: The main issue that I want to address is delay of 5 additional weeks. I would be happy if IITs can think of other ways to address this delay.

Be Ready for an Online Semester in August

The lockdown has been extended by another 19 days, as expected. Whenever we do start relaxing the restrictions, it will have to be gradual. My own guess is that the two sectors that will be opened at the very end would be education and entertainment. Will we be ready to start classes in July/August. Seems very unlikely, at least not residential campuses with students from all over the country. The assumption here is that even when we flatten the curve, we will need to maintain vigil in the form of social distancing for a long time. And maintaining social distance in crowded lecture halls, hostels, dining halls, and all other facilities will be a challenge. Also, there are likely to be significant restrictions on organizing large events, student festivals, contests, etc.

Assuming that social distancing requirements will not go away in a couple of months, or even 4-5 months, we really need to think of how to manage the next semester in our educational institutions. We really have three options in India. One is to postpone the next semester by two months, assuming that social distancing norms will get relaxed in an additional two months. But notice, one is taking a risk. What if, they don't. Will we postpone the semester by another month, or two months. Second option is to stagger the semester. Bring in old students in July end, as usual. And bring in new students two months later, when social distancing norms are likely to be more relaxed. The assumption here is that if we have only 75% students (considering a typical engineering under-graduate institution, for others, the number would be different), we could possibly maintain harsher social distancing norms, and only when the norms ease out, do we admit the new students. This has two assumptions, namely, one can maintain social distancing norms with 75% students, and that social distancing norms will sufficiently relax within an additional two months. Again, a significant risk.

So, we should look at the third option. Go online.

In the current semester, a large number of educational institutions have been forced to move instruction online. Faculty had to be trained quickly. Internet access had to be strengthened, investments had to be made in tools for managing online education. Now that all this has happened across the country, we should use the intervening period to look at the difficulties faced in moving online and try to fix them so that we can provide a better learning experience to our students. Many of the problems faced due to lockdown will not be there next semester. If a student does not have Internet access at home, next semester, the student can be asked to go to a nearby place, a friend's home, for example, to gain that access. We can even have local colleges allowing other students to sit there at non-peak hours and use their Internet bandwidth. So all the learning of this semester and all the investments of this semester will be leveraged for one more semester. We can start the semester on time.

And we can have a combination - some batches online and some batches on campus - only so many on campus that we can ensure compliance with social distancing norms. This is really what was also expected when MOOCs started becoming popular that one day, the university experience will not be a continuous four years on campus, but learn online for some semesters, perhaps even part-time, and learn on campus for some other semester to enjoy that wonderful experience.

In the long run, it is clear that Covid-19 would change the education sector completely. Online learning would become a very important part of any educational program. Next semester, we may choose online education for some batches because circumstances are forcing us to. But, in general, we should start getting ready to have an online strategy.