Delink Teaching and Learning

This is a topic on which I have been thinking of writing for some time. And then I read a few days ago an article by Bibek Debroy in Indian Express. The title is "Like advanced nations, India must delink classroom teaching from student learning."

There is far too much focus on teaching in Indian institutions, whether schools or colleges. In this context, I recall a curriculum workshop that we had organized at IIT Kanpur almost two decades ago. A well known Professor from MIT was also an invited participant. During the tea break, I asked him one question. How would MIT or any other US university would evaluate an application of a student from a college from where they have not admitted students in the past. If the student has a 9.5/10.0 CGPA and claims to be a topper, how would they compare him/her with candidates from other good institutions.

He replied that he would look at two things. First, how many courses do they do in a typical semester. Second, whether those courses are teaching skills and technologies or are they doing basic stuff (hopefully followed by projects, but pedagogy is difficult to ascertain from the transcript).

He asked me to find out the number of courses (in terms of equivalent of standard 40-lecture a semester courses) in IITs, NITs, other good institutes, and then in the affiliated colleges (where the Technical University dictates the curriculum). And sure enough, we found at that time that IITs were teaching 40-42 courses, NITs and other good institutions were generally teaching 45-50 courses, and most technical universities were requiring more than 50 courses (there are semesters when you do 7 courses).

The implication was that if you are forcing the students to learn 7 courses in a semester, there is no way an average student will learn anything. To really learn a course well, you must do some self study, revise the lectures, do home assignments, carry out some projects, etc. All this must mean that you are investing about 10 hours a week in the course. Would an average student spend 70 hours a week in studying. No way. Today, in most of the colleges, students complain if they have to do anything more than 40 hours a week. (During the admission season, one of the common question asked was how many hours would one have to study in that college. Or why one should join IIIT-Delhi which insists that you spend 40-50 hours a week on academics, when in other colleges, you don't even need to attend lectures.) So, if they have more than 5 courses, their focus will immediately be on how to get good marks and not on learning.

When we have such a serious shortage of faculty resource, and when higher teaching results in lower quality of learning, isn't it obvious that a university should teach less - this will improve learning as well as reduce costs.

The curriculum is decided on the basis that if we teach many things, the student is likely to retain some. Since a large number of our faculty in such poor colleges has come from similar programs, they are very reluctant to experiment with teaching less and hoping that student will learn most of it. And one gets absolutely hilarious arguments - our students are not like IIT students, they are not as smart. Well, if your students are not as smart, and we at IITs believe that students are only capable of learning 5 courses a semester, you should probably be teaching your students only 4 courses a semester. If they are not as smart, how come they will be able to learn 7 courses which even the so-called smarter kids at IITs can not do.

Besides the quantum of teaching, which Mr. Debroy focuses on in his lecture, there is another important issue. Even if we are teaching 4-5 courses in a semester, do we focus on learning. Unfortunately, not. A lot of even good faculty (good in the sense of their strong knowledge of the subject being taught) would focus only on having a great lecture - the transparencies should cover the material well, the diction should be clear, should be well prepared to answer all questions, and so on. Is this good enough for learning? Unfortunately not. It is now well understood that even adults can concentrate on a task for only about 10 minutes. That is why you find many MOOC lectures of that duration, The Khan Academy videos are of similar duration, etc. It is important that in a classroom environment, you reboot the class every 10 minutes. May be ask a question, ask them to think or whatever. But you talk to faculty members and they will tell you that they can't afford to "waste" time in the lecture. The syllabus is so vast and everything has to be covered. As long as the focus is on covering the syllabus and not uncovering the syllabus, the quality of learning will remain poor.

Also, if the focus is on learning, immediately, an instructor would consider having assignments and projects. But, if the focus is on teaching, why bother about projects which do not enhance the status of an instructor as a great teacher. And you can see in most places where you have 7 courses a semester, there are no labs, projects, assignments, etc. They really don't care about learning.

Besides teaching, the only other focus is on job. Let students get some job somewhere somehow. That is a parameter the colleges will use to attract the next batch. That these students will have no career without learning is completely ignored by every stakeholder - teachers, students and parents. And hence the focus on skills and technologies that one may get asked about in the job interview. The curriculum is decided on the basis of what industry HR folks tell colleges. And since every HR guy gives a different technology name, the colleges just take a union of these and teach them (and of course, teach them without projects).

Our regulatory bodies also encourage teaching since learning is more difficult to measure by them. It is easy to measure the inputs - teacher-student ratio, teaching load, average class size, and so on. Of course, there is some realization in accreditation bodies and they are now asking for evidence that learning is taking place.

Is there any center on teaching and learning at the higher education level (not school level education). Hardly, any. While any good university abroad would have such a center. We all know how to teach, and we don't care whether they learn. So why do we need to invest in teaching and learning centers.

Fortunately, I am currently at an institute where the focus on learning is tremendous. We have for each course, a well defined course outcomes. What is it that a student be able to do at the end of the semester. At the beginning of the semester, a faculty member discusses his/her plan with a couple of colleagues and explain how the instructor would ensure that learning outcomes will be met. What kind of assignments and projects will be there. What kind of in-class interventions are being thought of. How extra help will be provided to students who may be slow learners in that particular course. The entire discussion takes no more than half an hour, but that 30 minutes investment really provides a lot of useful feedback to the instructor. We then take a feedback from the students of the course 3 weeks into the semester. The usual end-of-semester feedback may not be very useful since the instructor may change next time. But an early feedback which the instructor can use to make changes to the course/pedagogy really makes a big difference. And the feedback is quite simple: Write what is going right with the course, and write what can be improved. No more questions. Anonymous. And finally, in the end-of-semester feedback, we write down the learning outcomes, and for each learning outcome, ask the student to comment on whether s/he feels confident of having learned that. So we have a decent idea of whether in any course offered in the Institute some expected learning outcomes were not met. Every semester, in one of the faculty meeting, we ask those faculty members who have got great reviews to say in few minutes what innovations they did in their course so that others can learn from their experiments.

If you notice, all that have stated above does not take too much time away from research. Most of the things take very little time. The total time spent on teaching without thinking of learning outcomes is really not much different than the total time spent on teaching when you are always conscious of learning outcomes.

Great teachers somehow intuitively know what is right for learning. But the challenge is to have average teachers and make sure that their students learn. And for this, the focus will have to shift from teaching to learning.

Wishing each one of you a very happy and safe Diwali. Darkness is nothing but ignorance, and the way to remove darkness is not by lighting a lamp, but by ensuring that our youth learns.

Double Major in UG Programs

Recently, Dr. Pushkar (BITS, Goa) forwarded me this article in Washington Post.
"Meet the parents who wouldn't let their children study literature." An interesting observation in the article is that the number of students majoring in two subjects in US universities has increased tremendously in recent years. About 40 percent of students go for double major in their under-graduate programs in US. And the reason for this is not that the world is becoming more complex, requires multiple skills to solve problems, need more inter-disciplinary approach, etc. The reason is that parents (like in India) are impressing upon their wards to study STEM fields in larger numbers, and the students are not as interested. Their love and passion is in something else. So as a compromise between the wishes of their parents and their own interests, the students study double major - one major of parental choice, another of their own choice.

So, in this blog article, I am just thinking aloud of the possibilities in India.

In India, the parental pressure is huge. (Refer to an earlier blog article on this topic: Parents and JEE Counseling.) In most cases, they decide everything that their son or daughter should do - which exams to give, which college to join, which discipline to major in, etc. And there does not seem to be any hope of that changing in near or even distant future. Will double major option work in India. Will students prefer to study another subject.

There are several differences between the US system and the Indian system, of course. I believe that in the Indian system, the parental pressure starts much too early, and a fair number of students have been brain washed completely to believe that STEM (Science, Technology, Engineering and Mathematics) are the only useful disciplines to study. So I am not sure there would be many interested in second major (except those who got admission in an unpopular discipline and can only study a popular discipline through the option of second major). Though some of the new universities like Ashoka are doing a remarkable job in making liberal education popular, and this resistance to anything other than STEM might change over the next few years.

The other major difference between the two systems is that in US, when a student joins a STEM major, a non-STEM major of high quality is available on the same campus, and one can take a few courses out of interest, and then decide to do even more courses to complete the second major, all on the same campus. In India, we have all but banished non-STEM disciplines from institutes offering degrees in STEM fields. There may be one or two non-STEM disciplines which may have sufficient offering to complete the second major.

Another big difference is that our graduation requirements for a degree are much higher than any US university. Our universities and institutes tend to believe that if you throw a large number of poor quality courses at the student, s/he will remember at least something out of them. With that kind of graduation requirement, the double major would invariably mean a 5-year program. Also with rigid structures like what course to do in which semester, planning a non-standard path becomes very difficult.

But, despite all these reasons, I believe that if we can start expanding our offerings, there will be demand for any high quality program. There will be enough 18-year olds even in India who may not be able to say no to their parents, but would be excited about studying what they like to study. And while the 18-year old may study two majors because of compulsion and interest, what one is likely to find at the end of those four years is that there is a significant demand for that combination in the job market. There is serious shortage of every expertise in India, and more so of people who know something about two or more disciplines.

At IIIT-Delhi, we are seriously considering having such programs available to our future students. As a small institute, focused on IT, we may not be able to offer several options, but we certainly want to offer a program which will allow students to get IT background as well as background in something else. The something else would have to be defined carefully. We are considering "Social Sciences" as that something else. We are also thinking of "Design" as that something else. We might in future consider "Finance/Economics" as yet another option. All such programs will be designed in a way that one can complete them in four years, and there is enough CS/IT content which is typically core discipline content in a good quality BTech program, and there is enough content from the second discipline which is typically core discipline content in a good quality BA/BSc program. We rolled out one such program this year (but both disciplines are STEM disciplines - CS and Applied Maths).

Flexi Fare on Indian Railways

I love dynamic fares and I wrote about their need on Indian Railways long time ago (2011) in this blog article. I had talked about several ways in which fares could be made flexible.

I was excited when the Railways introduced dynamic fares in a Rajdhani like train between Delhi and Mumbai, and I wrote this blog article in December, 2013.

When they introduced dynamic fares on a system-wide scale in the name of "Premium Tatkal" I wrote another blog article praising the move, even though I said that this was a poor implementation of dynamic fares.

My hope was that Railways will learn about market, about demand and supply, and will learn to price their product better through this experiment.

They have taken the next step and introduced something called "Flexi Fares" on Rajdhanis, Shatabdis, and Durontos, the premium trains of the Indian Railways. The fares will increase with bookings, 10% after 10% seats are filled, upto a maximum of 50%.

And this time, I am beginning to believe that Railways is not really thinking about the market, about demand and supply. They are only thinking of higher revenues. Thinking of higher revenues is a legitimate business strategy. However, when you think of higher revenues without thinking of what the market can bear, it may not lead to desired goals.

I understand that Dynamic fares or Flexi fares are the only politically correct way to raise fares, and with a background of no fare hike for almost a decade, several large fare hikes in quick succession are needed to bring railways back to some kind of financial health and Mr. Prabhu is doing that. What has happened is that the fares on these trains have just gone up by 50% with a few tickets being sold at less than that.

My problem with the hike is just that. That it is not introduction of dynamic fare. The fares are completely fixed and do not depend on demand. Yes, there are a few price points. Ideally, Railways should do what airlines do - keep following their own bookings for every flight, keep looking at the competition, and decide fares based on their estimate of the demand and supply around that particular flight. If the 10% seats get filled up the day reservations open 120 days in advance, that is a very different demand than if the 10% seats get filled up over the next 60 days. Second, they also need to look at their competition - status of other trains close to Rajdhanis/Shatabdis/Durontos - are people booking them instead, or are they too not having heavy bookings, as well as status of flights and at least good quality bus services.

A truly dynamic fare will be more acceptable to public than what IR has done. It would really hurt to see a half empty coach when I have paid the "surge" pricing. Dynamic fares would have meant that in the lean season, the increase is smaller but in the holiday season coming up, the fare would go up higher.

In terms of fare increase, I think this was needed, and this was overdue. These are premium trains, and no poor person travels by these trains. There is no reason for subsidizing higher classes in premium trains.

A lot of comments on social media say that these trains would become unviable. They wouldn't be able to find enough AC-2T passengers on Delhi-Mumbai or Delhi-Bangalore Rajdhanis. Yes, there won't be many end-to-end passengers on most Rajdhanis on most days. Already, at least in the current lean season, air fares are cheaper than AC-2T fares on Rajdhanis. But this is where the planning by Railways become crucial. My favorite example for these discussions is Lucknow-New Delhi Shatabdi. The Lucknow-Delhi air fare is cheaper than Executive class fare, and the fare on Chair Car after 50% hike would be right there too. So in the lean season, I would expect the passengers to take flights. But come busy season, people will be back to Shatabdis since the air fares will be much higher. However, if Railways can earmark more quota from the intermediate stations, like Kanpur, from where there are no flights, they can really double the fare and still find passengers. But will Railways give more seats to Kanpur and other intermediate stations. Similarly, I would expect much fewer passengers on Delhi-Howrah Rajdhani traveling end-to-end. But I would expect more passengers to board from Kanpur, Allahabad, etc. if the quota from these places go up.

At least till the regional air connectivity scheme takes off, giving more seats to inter-mediate stations would give a lot of dividend to the Railways.

My fear is that increasing the upper class fares substantially without linking them to demand and supply and without a plan to fill those seats through quotas of intermediate stations, would cause the AC coaches to remain vacant, and then someone is going to say, there is not enough patronage of such trains, and let us stop this train, or remove upper class of service. So we could have reduced capacity not because there was reduced demand, but because that demand was not managed properly.

By the way, if the Railways finances were more transparent, and someone could show that the cost of transporting an individual in AC-2T over long distance like Delhi-Bangalore is inherently more than a flight seat over the same distance, I would have no problems if long distance upper class seats are reduced substantially. If economics of transport options is such that for a certain route trains can't compete with flights, we shouldn't mourn the death of passenger trains in those sectors. May be we can have more freight trains instead on those sectors. However, my fear is that in the absence of any transparent way of apportioning costs, such decisions could be arbitrary.

One of the criticisms that I consider valid is that Mr. Prabhu is trying to avoid taking tough decisions. Is he willing to increase fares or pass costs in Mumbai suburban. Why not increase fares of 2S and unreserved tickets, which are heavily subsidized. The fare hike in Rajdhani/Shatabdi/Duronto is just testing the waters. If there is enough noise, after a certain period, we will see partial roll back. If there is not enough noise, the experiment would be extended to other express trains as well. However, from all what I am hearing, even if this experiment is extended, it will be only for AC classes in mail/express/superfast trains, or at best SL class in a few trains. Without tackling the biggest sources of subsidy, you aren't going to take Railways out of that proverbial ICU.

We don't need flexi fares or premium tatkal fares. We need market linked fares. And we need market linked fares not just for AC classes, but also for SL class. And, of course, we need to recover a greater part of the cost of suburban and passenger trains as well.