Search This Blog

Friday, December 7, 2018

Academic Advising

The college life is very different from school life. In schools, most of the decisions are taken for you by teachers and/or parents. In any case, there are few academic decisions to be taken. Except at the beginning of 11th class, most of the courses are fixed for you. And that makes transitioning to a college harder.

Academic advising involves a trained staff or a faculty member guiding the student on the choice of courses (in case of electives), what courses can be done in any given semester (given the issue of pre-requisites, and graduation requirements of the program), etc. This becomes even more important if the student has failed a few courses and may not be satisfying the pre-requisites of many advanced courses. Besides, universities sometimes offer choice of second major, minor, dual-degree and other flexibility in the programs. Students are not just confused about the rules around them, but also whether they are consistent with the educational goals of the student. There could also be issues regarding whether to take a break and do a semester long internship, and finally what should they be doing after graduation. Should they go for higher studies or a job or something else. Which place to study. What kind of job, and so on.

As should be clear from above, the academic advising becomes important when a student has options but if left alone may not exercise those options in the best self interest. For a long time, Indian institutions did not give many options to students. You got admitted to a specific program - so you don't need help in choosing your major. You had a fixed curriculum, with fixed courses to be done every semester. So no choice there. May be you had an elective where only 2-3 courses will be offered and one had to take one course. There was no flexibility of second major, minor, etc., so no decisions to be taken. In such a system, only those students needed advise who had failed some courses and needed to restructure their programs. Academic advising in India, therefore, has traditionally been limited to handling academically deficient students.

Over the last two decades, things have changed quite a bit. With UGC and AICTE prescribing choice based credit systems for all academic programs, students have many decisions to take every semester. But the systems for providing this help have not been developed. In good colleges (like IITs), not only there are a lot of electives and they can do courses in which ever order they want, but there are options like minor, second major, dual degrees, etc. But the students are left to seek advice from their seniors alone. They can advise based on their own experiences and hence cannot really help another student with all options.

The colleges have given up on this extremely important responsibility by stating that the students are adults and ought to know what is best for them. This is quite irresponsible since even adults do not always know what is best for them, and even worse, may not even have all the information to make an informed decision. This becomes even worse in India where the websites are often not updated, and getting the information itself can become a project.

Consider an example. A student had 14 courses to do to complete the graduation requirement. He has been a somewhat academically weak student and has never performed well even with 5 courses in a semester. He planned to do 6 courses in each of the two semesters and two summer courses and thus hoped to graduate in a year. Of course, if he fails even one course in this plan, he will have to stay back for a semester necessarily. If someone had advised him properly and told him that the chance of succeeding in this plan are very slim and hence he should plan for 3 semesters and a summer to complete these. In the alternate plan, he would have done only 4 courses in each of the three semesters ad two courses in summer.  With lower academic load, he was more likely to learn better, with hopefully better grades, and likely better future options. But he registered for 6 courses and failed. So the transcript has more failures. The pass courses have poor grades and his confidence is shattered and he will have to spend that extra semester any way.

So many final year students (even those who have no backlogs) regret their decisions of doing something or not doing something. They wish someone had told them about the options early on. But with advising being so closely linked to poor academic performance, they never did approach anyone for advice.

I recall that when I was Chairman of Senate UnderGraduate Committee (SUGC) at IIT Kanpur, I had organized an advising session for 2nd semester students where faculty members from different departments told them about the excitement and opportunities in their respective disciplines. At the end of the advising session, so many students chose to change their program and unlike the normal behavior where most students apply for change to "more popular" programs, that year a lot of them applied to change to "less popular" programs since they realized that this is more in line with their interest and popularity of a program does not really matter for their career.

Most good universities (and even no so good ones) abroad invest a lot in academic advising. Happier graduates who attain their goals are more likely to be good brand ambassadors and good donors later in life. Early detection of possible academic problems can result in interventions before the problem becomes too serious and the student's program is terminated. I hope we too can invest some more resources into this extremely important service in our universities.

Thursday, September 6, 2018

JEE Mains 2019

JEE Mains for the next academic year admissions have many new things. First of all, it will be conducted by National Testing Agency (NTA) and not CBSE. NTA is a new organization created by Government of India exclusively to conduct large public tests. This, by itself, shouldn't matter to applicants. But NTA is bringing in a few crucial changes.

There are two important changes. One, you can give JEE mains twice in a year (January, and April) and as per NTA website, it will be treated as single attempt. JEE Mains can continue to be attempted for three consecutive years - the year of giving 12th class exam and the next two years. The difference is that in each year, you can attempt it twice and the better of the two scores will be counted. The second change is that the exam is online. No option for a paper version.

In each month (January and April), the exam will be conducted on three weekends. On each weekend, it will be held in four sessions. Of course, students in each session will get a different question paper. So the most important issue here is that of normalization. What if some paper is easy and some other paper is tougher. This is an important issue since CBSE did not do normalization across multiple papers and people were left with the feeling of unfairness. But NTA has announced a process of normalization.

They will look at your score relative to score of other students in your session alone and assign you a percentile score. This is the score that would determine your rank. If you give JEE both in January and April, the higher of the two percentile scores will be counted as your performance.

Let us examine if percentile is a good way to normalize. The basic assumption here is that if the distribution of academic preparedness among students is same in two exams, then the relative rank of an individual student is likely to be similar. So even though marks obtained by students could be different and would depend on whether the questions were easier or tougher, but the ranks won't be very different.

This assumption is a reasonable one when it comes to considering different sessions within the same month. It is so because NTA would randomly divide the applicants into all sessions. Random allocation of sessions to large number of applicants is the best way to ensure that academic preparedness is similarly distributed in all sessions. However, can we assume that academic preparedness in January will be same as academic preparedness in April. I don't know, but it seems unlikely. And that is a question mark on the normalization scheme.

The other question mark is that while it distributes students randomly, an individual student can still get a vastly different rank depending on which session the student has been assigned to. For example, if I am extremely good at Mathematics and relatively poor in Chemistry then I will have a very different rank in a test with tough Mathematics and easy Chemistry versus a test with easy Mathematics and tough Chemistry. So, one will still have to make sure that different question papers are not vastly different in their difficulty levels. The normalization procedure they have chosen will be good for taking care of some variations in the questions but not a drastically different type of paper. And, whether one can normalize across two months is very questionable.

I hope they could have allowed students to choose the weekend at least, if not the specific session. If they had too many applicants for some weekend, they could have stopped taking applications for that weekend, and if there were too few applications for a particular weekend, they could have reduced the number of sessions in that weekend from 4 to 3 or even 2. This would have been a huge relief to students. I am sure that those taking 12th class board exams would prefer earlier dates so that they have more time to concentrate on their board exams.

The next question that I am getting asked often is should one take JEE Mains twice.

Well, the only negative I can think of in taking JEE twice is that if you were to perform poorly in January, you may be so disappointed that you spoil your board exams. If you are confident that you will not be deterred by one poor performance, go ahead and take it twice. If you are not adequately prepared for JEE Mains in January, 2019, treat this as a practice test in real testing environment. (You can't get more real than this!) On the other hand, if you are reasonably well prepared, my prediction is that it will be easier to get a high percentile score in January than in April because a large number of students will be taking the January test as a practice test. The repeaters must take JEE Mains in January, since they must already be well prepared and they can take advantage of many students taking it as a practice test.

Should you take the April test even if you get a good score in January. Well, if you get a 99 percentile kind of score, which will mean that you would be in about top 10,000 ranks, may be you can avoid April test. (I am assuming that about 10 lakh students will take the exam either in Jan or in April or both. Hence 99 percentile would be around 10,000. But exact numbers may vary.) But unless you have a very high score (which guarantees eligibility for JEE Advanced, and which guarantees admission to one of the top non-IIT choices if you don't perform well in JEE Advanced), there is no harm in trying for an improvement.

At the end, please note that I am writing this based on what I have read on NTA website. You should visit their site often, just in case there are changes in the rules and processes, or just in case, I have misread or misunderstood them.

Best wishes to all potential students of IIT Kanpur.

Wednesday, September 5, 2018

Remembering my teachers

Today is Teachers' Day, a day we celebrate in memory of Sarvepalli Radhakrishnan, our second President, and a teacher himself, who was born on this day, 130 years ago. I thought I will recall my education and some of my teachers on this day. I have been incredibly fortunate to have been educated and influenced by a large number of great teachers. I would guess that this is a statement that anyone who has seen any success in life would make.

The first teacher to remember is my own dad. He was a school teacher, officially designated as a teacher of political science, but every principal asked him to teach Mathematics. Our house was full of books on Political Science and Mathematics (and very few on history, the other Master's degree he had) and I would read a lot of them, and asking him questions about topics which I would perhaps study in school a few years later (and in case of political science, never). I had read many SC decisions at an age when my classmates perhaps did not even know what is Supreme Court. And I had won pretty much any award in mathematics in school. He was a great teacher not just because he could teach the two subjects so extremely well, but because of all the life lessons he would impart to me and my siblings.

The next teacher to whom I owe a lot was the principal of a small private school, Childrens' Home School (which no longer exists) near our home. She was truly generous, and allowed me to study without having to pay any tuition. School teachers did not earn much in those days and there was no way my father could afford to send us to a private school. (Teachers' salaries have become high only after 6th pay commission.) The primary school system in the government sector was run by Municipality while the secondary and senior secondary schools were run by Delhi government. The municipal schools were terrible. (I studied in one for a year, and if I had to study there for 5 years, I don't know if my education would have been as good.) So an offer to study for free in a private school was really something that I can never forget and has been a reason why I have been contribution a part of my salary for education ever since I started earning.

In my secondary school, Ramjas School, a school run by Ramjas Foundation and supported by Delhi government (and hence no fees), there were lots of good teachers. It was considered one of the best schools of that era. But strangely, the two teachers I remember the most are the sports teacher and the one who taught me gardening (we had a subject named, Socially Useful Productive Work, SUPW). The first one, Mr. Mallik, was managing the Bharat Scouts and Guides program at the school, and this school had the largest number of presidential scouts every year. If 8 scouts were chosen to represent Delhi, 6 or more would be from this school. He also managed the hockey team of which I was a member. He would make us work really hard, often 2-3 hours after the school times in harsh weather. Mr. Gupta, the gardening teacher, was not just teaching gardening but love for nature. He would also often discuss current events. He was a hockey fan, and would often discuss the decline of Indian hockey after the world cup win in 1975. So he wasn't bothered about the syllabus, and we loved him like anything. He was also a teacher whom we trusted blindly. If anyone did something silly, one could lie to all other teachers and principal, but if he asked you, it would be impossible to lie.

For my senior secondary (11th and 12th), I went to SBM school. We had shifted home and hence I had to switch schools. The teachers were completely focused on 12th class results. But when I met the teachers individually and explained to them that I wanted to prepare for JEE as well and therefore, I wouldn't be able to spend several hours a day answering 100s of questions every day, they not only agreed to give me less number of homework problems every day, but also ensured that the school bought a few books that would be useful for JEE preparation, since I may not have been able to afford them otherwise. I fondly remember Mr. Gupta, our Maths teacher, who was truly concerned about our future all the time, and was willing to spend any amount of time after the school hours, if we still had questions unanswered in the class. The vice principal was a terrible person though. (So even a community as great as that of teachers have a few black sheeps.) I remember we were playing the Hockey final at the intra-district level, and I worked hard to score an equalizer almost in the last minute. This guy was very unhappy that 11th and 12th class students are participating in all sports and extra-curricular activities. He went to the match referee and told him that the school concedes defeat and also when the district team is selected to participate at the state level, he should not choose anyone from our school. Our sports teacher was none other than Mr. Mallik Jr., son of our sports teacher in my previous school. He was like his father, always taking our side. He protested strongly, but Vice Principal had his way. After this incident, the VP was always mad at me. On more occasions than one, he would ask me to come to him, and then he will tell everyone that I am a proof that Goddess Saraswati and Goddess Laxmi can't live together. (That I had intelligence but no money.) He would make fun of my cotton clothes, or an inexpensive school bag or something else.

IIT Kanpur was fantastic. I must have enjoyed more than 90% of my classes, and the rest weren't bad either. Those were the days when teaching and learning was considered as the primary reason for a college to exist. It is not that faculty didn't do research then, but clearly the focus was on students' learning. With every teacher being so great, it is difficult to name a few, but I will try. The one I loved the most was Prof. Gautam Barua (now at IIIT Guwahati). The CSE department was a small one and they took all decisions by involving students. I was the representative of my batch in the Department Under-graduate Committee (DUGC). Every semester the faculty would meet to decide the next semester's courses and I would suggest if Prof. Barua could teach a course to our batch. And he would gladly agree. He not only taught Operating System to us (which was his research area), but even Data Structures, which was typically taught by theory persons. He also taught for the first time a course called "Distributed Systems" which was a precursor to the course on Computer Networks. I remember getting so excited by the course on Distributed Systems that I would spend long hours in library (the only semester in which I went to library) and read article after article in the journal, "Computer Networks and ISDN System." I also did my BTech Project with him and thus more than 1/6th of the credits done in all 4 years are with him. I would later apply for graduate studies and explicitly state in my Statement of Purpose that I wanted to study networks, when everyone would tell me that Theory was the hot area and easier to get scholarship in. And I specifically wanted to go to University of Maryland since I had read so many papers of Prof. Satish Tripathi. Other Computer Science professors who have helped me become what I am included Prof. Rajeev Sangal (now at IIIT Hyderabad), and Prof. Somenath Biswas (now at IIT Goa).

Among the non-CSE faculty, Prof. R N Biswas was the most fun who taught us the common compulsory course on Digital electronics. Other great teachers included Lilavati Krishnan (Psychology), and Mohini Mullick (Philosophy). I was really excited about Maths, and have taken almost all electives in Maths. My transcript would show about the same number of Maths courses as Computer Science, perhaps more in Maths. Prof. S K Gupta, Prof. Borwankar, Prof. S P Mohanty, Prof. B L Bhatia, Prof. R S L Srivastava and many more. As I said earlier, with so many great teachers, it is difficult to list everyone, but IIT Kanpur was home to the best teaching at that time.

At University of Maryland, I was fortunate to have Prof. Ashok Agrawala as my PhD supervisor, and as a life coach. We have discussed almost anything and everything under the sun, and have received valuable advise way beyond my thesis topic. One of the things I remember was his insistence that I be able to communicate my PhD work to a layman. Only after I came up with how I would explain my work to someone who hadn't known anything about computers or networks, did he allow me to submit the thesis. And that has helped a lot. Prof. Satish Tripathi (now at SUNY, Buffalo) was around whenever I felt low. On two occasions when I almost quit my PhD, he would leave everything aside and spend a couple of hours to explain why I should not give up. I also took a course of Prof. Pankaj Jalote (later, a colleague at IIT Kanpur, now at IIIT Delhi) and even interacted with him on research issues (and have joint papers). Prof. Uday Shankar was my MS supervisor and really taught me how to write research papers. At UMCP, in those days, the systems research group had four Indian professors, but students from about 10-12 different countries were there. And these students taught me history and culture of their respective countries.

I end this by appealing to all my readers to consider making a small gift to your favorite education provider, not just on the occasion of Teachers' Day but periodically. Teachers do a great job, but often the students need additional support. I just did before writing this.

Thursday, August 16, 2018

IIT Gandhinagar Decennial

The second generation IITs have completed 10 years (except two of them, which started a year later). Recently, IIT Gandhinagar organized a function to mark the occasion. They invited me to be present and publicly thanked me for whatever little I may have done in the last decade. This was a good time to recall many great initiatives that have been taken up by IITGN.

When Prof. Sudhir Jain became Director, IIT Gandhinagar in 2009, he asked me to be one of the many friends that IIT Gandhinagar would come to have, and spend just one day a month at that IIT. Given the logistics, costs, and overhead of travel, I promised to visit alternate months and spend on an average 2 days in each visit. And sure enough, this was my 59th visit to Ahmedabad in just a little over 9 years.

When I look at IIT Gandhinagar, it is obvious what a great leadership can do to an institution. And a great leader is not just one who can keep coming up with great ideas, but has an eye for leaders around him/her who will come up with equally great ideas themselves. Poor leaders, on the other hand, can only promote mediocrity or worse.

Though a lot of innovative thinking has gone into the policies and processes at IIT Gandhinagar, some of them have simply followed common sense. Indeed, the best example of the difference is that they treat an 18-year old as an 18-year old. It is common sense, and yet most institutions would think of under-graduate students (even 21-year olds) as juveniles. The Institute respects its students as adults, trusts them to do the right things always, and students don't let the Institute down. Treating them as adults mean not involving parents, unless the situation is serious. Adulthood means that they have a greater say in running their own affairs (hostels, festivals, discipline, mess) and have a significant say in running of the Institute. In return, the students ensure there are no headaches for the Director from their side There is huge affection for the Institute, and that shows up in statistics like the fraction of alumni giving a gift to the Institute in a year - perhaps the highest of all Indian institutions.

Another simple idea is to have liberal rules. When I was a student at IITK, we were really free birds. The rules did not matter as anything we wanted to do, we could request and it would be considered on merit and not in light of the rules. And if this request was for something worthwhile, it would be approved. But things did not stay the same 3 decades later. With student-faculty ratio declining sharply, it wasn't possible to consider each request on its merit. A lot of times, we started hearing, "rules are meant to be followed" and we found out that rules did not allow many things we took for granted as students. IIT Gandhinagar had always said that they wanted to learn from errors older IITs had made. And they decided that right from the very beginning, the flexibility will be inbuilt into the rules, and not depend on someone taking a positive decision on your application on a case-to-case basis. And the flexibility shows in branch change rules, having flexibility in curriculum like minors, honours, double majors, and many many more situations.

My favorite visits to IIT Gandhinagar have been to attend their annual Academic Advisory Council meetings and their annual Leadership Conclaves. The two are held on consecutive days. For AAC, they invite academicians who have lots of ideas and are from any part of the globe. In LC meetings, they invite people from industry, government, alumni, and academia. Typically, they will discuss academic issues in institution building in AAC and administrative/governance issues in institution building in LC. New institutes have to be open to external viewpoints since they may not have sufficient internal capacity. But even with this caveat, IIT Gandhinagar has been exceptionally outward looking. One of the advice that I heard during these meetings was that while any educational institute will take a long time to be truly world class, one need not wait to start thinking and functioning like one. So when you have to create a policy for anything, think about how a world class university would do this. This way of thinking and doing things has ensured that IIT Gandhinagar's march towards excellence has been faster than what it would have been otherwise.

Even though a new institute struggles with several problems - faculty recruitment, building of infrastructure, and so on, IIT Gandhinagar has always proclaimed that as a public funded institute, it must be open to not just its students and faculty but many more people. So they allow students of other colleges to spend a semester doing courses at IIT Gandhinagar. They have a large program of summer interns on their campus. They even have special programs for foreign students in winter. They run several training programs for teachers of other colleges. They allow non-students to set up companies in their incubation center (and now building a technology park). And all this seems to help them a lot. They attract more applications for Masters and PhD programs in a single department than the MTech/PhD applications in IIT Kanpur in all departments put together.

IIT Gandhinagar has several agreements with foreign universities and companies for summer internships, semester exchanges, research collaborations, etc. They want their graduates to have a global exposure. A very large fraction of their students would have spent some time abroad during their program. However, not all agreements are fully funded. This means that under normal circumstances, some of those programs are not accessible to students coming from financially challenged backgrounds. But not at IIT Gandhinagar. They have programs that will provide support to such students so that they too have equal access to any and all programs that IIT Gandhinagar is involved in. Much of this support comes from gifts from friends. They raise more philanthropic funds in a year than some of the old IITs.

One of the most innovative programs that they have is their Foundation Program. Every under-graduate student admitted to IIT Gandhinagar goes through a 5-week program in which their are large number of talks, workshops, discussion sessions, projects and outdoor activities. Students develop a broad set of life skills including creativity (music, drama, painting, art), sports, leadership, empathy, societal concerns, and inculcate values and ethics. This is their most famous export, and now many IITs and even other institutes have started doing similar programs.

Another interesting program is Explorer Fellowship in which a student is given a small budget to travel across India for 6 weeks. The budget of less than Rs. 1000 per day is to be used for travel (only non-AC travel allowed), stay, food, and any other personal expense. They must visit a state in South, North and North-East of India. The students not only understand the diversity of the country, but also understand the circumstances under which our people live. Another summer fellowship scheme is "Gram Fellowship" in which the student is expected to live in a rural setting.

No wonder that they care so much for everyone working for them. The facilities they have asked the contractors to create for construction workers, for example, are absolutely amazing. Each outsourced employee gets some benefits which are unheard of at other places. For example, they have created a Children's Education Fund through which some educational expenses of children of contract employees are taken care of.

There are far too many innovative ideas that IIT Gandhinagar has implemented than what I can list here. But suffice to say that my association with them for the last 9+ years have been extremely fulfilling. There are lots of faculty members with whom I have interacted with and all of them have given me new ideas and perspectives.



Friday, August 3, 2018

MTech program needs change

There was a time when MTech students were the major research manpower in our technical institutions. The number of PhDs were very small, and faculty needed some hands to work in their labs. We also needed some people to help us as Teaching Assistants to support our under-graduate teaching. So the focus of MTech program was not that we are preparing them for industry. That had to happen for us to attract anyone to the MTech program, but the program was not optimized for those outcomes. Since we needed MTech students for our research and teaching support, we would offer good students not only fee waivers, but also a decent stipend.

On the input side, there were always those few students in NITs and other good institutes who had enjoyed their under-graduate education, were more curious to learn more, did not want to commit for a PhD program (and hence getting scholarship from abroad was difficult) and if IITs offered a program which did not require depending financially on parents, why not spend 2 years in a good environment (and perhaps get a tag of an IITian as well). It was a pleasure to work with these students.

But slowly, situation is changing. The number of PhD students has gone up quite a bit in the last 10-15 years and is still climbing. So the dependence on MTech students for doing our research has declined. The top students in NITs and other good institutions are going directly for corporate jobs in larger numbers. In fact, we hardly see any applicant from NITs and IIITs these days. Also, the starting of dual-degree programs has meant that our own BTech students (who are better trained than most of the students we are now recruiting in our MTech program from affiliated colleges) are continuing as MTech students. We are also becoming more open to the idea of some of the senior Under-graduate students to help us as Teaching Assistants in our courses. (Some IITs allow even 2nd year students to be TAs.)

The motivation of the students joining the MTech program is mostly to get better trained than what a typical private affiliated college has been able to do, and improve the quality of job that they can get. From a job in the IT services industry, they want to move up the value chain and join a product company (and I am told similar motivation in other disciplines as well).

So if the motivation of students have changed, and the requirement of faculty has changed, shouldn't that get reflected in the way we handle our MTech program.

What would prepare our MTech students better for industry. Currently, we have roughly 1 year of course work and 1 year of thesis. (In IITK, it is 0.75 years of course work and 1.25 years of thesis, since we had the maximum shortage of PhD students historically. So a greater focus on MTech research.) To prepare these students better for industry, we need to improve their core knowledge, and expose them to several new ideas, and finally, get them to integrate/synthesize all they have learnt by doing a substantial project. So we need to increase course work, and reduce the thesis to a project.

This will also free up faculty time, since guiding an unwilling student for 3 semester and then eventually writing their thesis is not a worthwhile investment of time when you have PhD students on one hand, and dual-degree students on the other.

Of course, if there is someone really interested in research and is academically well prepared for it, there should be an option to do that.

More importantly, if we are no longer depending on MTech students for research support or teaching support, we don't need to have a fee waiver and a stipend for everyone in this program. The financial support should be contingent upon someone either being selected (after a careful selection process) as a teaching assistant, or someone doing quality research work. In fact, research support may come through the sponsored research route rather than using the Institute funds. This will also allow us to expand our MTech program substantially and be more useful to industry as a result. This will also push those students who are interested in research to think of joining our PhD program.

In summary, an MTech program may have about one semester of refresher courses, one semester of courses which is exposing them to advanced topics in multiple areas (breadth), and one semester of courses which provide them depth in the area they want to work more, and finally, one semester project work. Of course, one size never fits all, and the basic template could be tweaked if someone comes better prepared, is interested in research. We could allow project work to be replaced by appropriate internship experience in industry. Create several such flexible options around the basic template. Financial support only if one gets selected as teaching assistant or a faculty member is willing to recruit you on his/her sponsored research project. There may be a few need-based scholarships.

Thursday, August 2, 2018

Proposal for Lateral Admission in IITs

Several years ago, I had written a blog article suggesting that IITs offer lateral admission. Whenever I have discussed this with people, a few questions have come about - how would you compare students from different colleges, would you have another form of entrance exam for entry into 2nd year.

My response has been that if we are convinced that lateral admission is good for IITs and good for students, then we will find a way out. So before asking for implementation details, are you convinced that it is a good idea. The obvious win for IITs is that the single exam misses out on a lot of meritorious students, and we get a second chance to identify them and recruit them. And this second chance, if designed carefully, will involve not just an exam on one day, but looking at their one year effort, and hence a better predictor of future success than a single exam can ever be. The obvious win for the students is that they get a second chance to get into the institutions of their choice. And, of course, if some student is not convinced that this is good, they don't have to apply.

Good thing about lateral admission is that we don't need all IITs to agree on this. A single IIT can choose to do it. There is some agreement that we will all do silly things together when it comes to admission to first year of under-graduate programs. But there is no such agreement for lateral admissions, and there have been lateral admissions in 60s, 70s, and 80s.

The only top tier institution that admits students laterally (that I know of) is IIIT Hyderabad. They admit students only to dual-degree (BTech+MS), from those who have completed 2 years of engineering in some other college, and these students have to spend 4 years in this program. The idea is that if these students were to complete their bachelors in their current college, they will take 2 more years and then if they were to do a Masters program, that will be another 2 years. So anyone who is looking for doing both Bachelors and Masters is better off doing a much stronger Bachelors and an integrated Masters. Of course, someone only interested in a UG degree or hoping to do a Masters from abroad or from top IITs/IISc in India will not be too keen on this program. So they have ensured that they are not inundated with thousands of applications, and they can choose those students who are well prepared and are ready for a research degree.

We may have to think of our own goals for lateral admission. But the IIITH model is a good one. Many IITs have backtracked on admission to dual-degree programs through JEE because it has been felt that a 12th class student should not be forced to choose a Master's program in an area when s/he is not aware of anything in that area. IIT Kanpur is one of those IITs. And we are noticing that few students voluntarily take up dual-degree programs. As a result, we have a shortage of good quality MTech students, and a shortage of good Teaching Assistants to support our UG teaching. But offering dual-degree after 2nd year of college is fine since the student has been exposed to the discipline, and has a good idea of whether s/he may be interested in Masters program.

We could, therefore, seek admission from students who are about to complete 2 years in their respective colleges. Ideally, we could ask for all the details of projects, internships (if any), other activities in college, and short list on that basis, followed by an interview. But since many of my readers would not trust IITs to be objective, or would fear that this may lead to corruption or at least court cases, may I suggest a test as well. If we could negotiate with GATE that they can allow 2nd year students to give GATE, and also specify a part of the curriculum which is meant for them, it would solve the problem of an objective test. Since GATE is now an online test, such complications can be easily handled. A second year student will be shown only half the paper which is of topics which are pre-announced, and will get a score different from what senior students get. And we could shortlist students based on this score. We could even have some topics/questions which are meant for only these students. In an exam which is taken by 12 lakh students, some additional students will not be a major challenge. (Actually, this could be GATE as well with no changes at all. But that might give some advantage to students from colleges which have more core courses early on.)

So we shortlist students on the basis of this version of GATE. And then we interview and do whatever else we normally do in our MTech admissions, including looking at their two years' record. Once we admit the students, we could go through their transcript and decide what courses in IIT can be waived for them. If we have a doubt about the quality of the course, we could have a small test to check the academic preparation. After 2 years of studies and a good GATE score and good performance in interview, at least one year worth of credits will be waived, and hence this is advertised as a 4-year program. But for some students, many more credits may be waived, and they may even complete their dual degree earlier than 4 years (even in three years, for that matter). We could even consider those who are in 3-year programs in universities (and hence have done less engineering courses).

We could also have lateral admission after 1st year. Here, we could offer a 3.5 year BTech or a 4.5 year dual-degree. To reduce the number of applications to a level where an IIT can handle them without having to conduct a national wide admission test, we could go back to JEE - only those students are eligible who had a JEE rank of less than 50,000 (or whatever will give us reasonable number of applications). And then from those small number of applications, look at all aspects of student life, performance in 1st year, and have an interview. If we have enough seats for this round of admissions (and in several IITs), we could even encourage most students to not repeat JEE but join a good college, since the probability of success after JEE may not be much higher than probability of success in this round, particularly for those who missed coming to IIT by only a few marks.

While we are at it, let me also write about the scheme that Kakodakar Committee had suggested, which isn't lateral admission, but something similar. It suggested that IITs could admit students who have completed 3 years of UG. These students would do one year of UG in IITs, but will promise to stay on for 2 more years doing MTech (and hopefully, PhD). To enable this, IITs would enter into an agreement with colleges (it suggested only NITs, but could be expanded to any university) where the colleges agree to transfer credits done at IITs and more importantly, agree to not give UG degree till a year later, because otherwise, the student may just do a one year of UG at IIT, go back to his/her college, get the UG degree and never come back for MTech. (I wouldn't mind that at all, but that is what was proposed.)

So we could admit students in not just 1st year of UG, but in all other years of UG as well.

Tuesday, July 31, 2018

Encouraging students to excel: Preference in Admissions

Yesterday, media had a wonderful news that five students who represented India in the International Physics Olympiad have all received Gold Medals. This is the first time that Indian team has come back with five Gold medals. The previous best performance was 4 Golds and 1 Silver.

And then the analysis started. It was pointed out by my friend, Prashant Bhattacharji, that India has been performing well in Physics Olympiad for a long time, and surprisingly to me, our performance in Maths Olympiad is not so stellar, Chemistry Olympiad even lesser. He then pointed out that Physics syllabus for Olympiad is very similar to JEE syllabus, while Maths syllabus for Olympiad has some differences with JEE syllabus, etc. The news papers also reported that while China too has 5 Gold Medals, their overall ranks are higher in most years because our team does not perform well in the experimental physics, something that is not tested in JEE. Almost every year, team members of Physics Olumpiad do extremely well in JEE (this year, two of them are in top 10 ranks, and other two have good ranks, while one student is now in 12th class), while the correlation between JEE performance and Maths Olympiad performance is lower. As an aside, three of the team members have joined IIT Bombay, while one is joining MIT in USA. The fifth one, I hope wants to join an IIT, though, I am sure would have an offer from MIT as well.

How do I interpret all this.

For most parents in the middle class, the biggest point of stress today is admission to a quality educational institution (both at school level and then at the college level). If a student is even marginally interested in science, s/he is under pressure to prepare for competitive exams like JEE or NEET. So one should study all science subjects equally, and study all those topics well which are part of the syllabus for these exams. You aren't allowed to spend more time on the subject that you really love and want to excel in. Also, you aren't expected to "waste" your time in labs. Is this the way to encourage excellence? Of course, not.

But things aren't going to change as long as there is serious shortage of good colleges. Parents will ask their wards to focus primarily on JEE/NEET type of goals. While the biggest problem, obviously, is the lack of sufficient quality educational institutions, the second biggest problem is that all these quality institutions have only a single admission process based on a single test. If we want that students should excel in whatever they are interested in, we will have to create an incentive scheme which is acceptable to parents.

One such scheme could be that students who represent India will be offered admission to top places related to the subjects that they have excelled in. Students who don't represent India but were part of the training camp (which is typically 20-30 students) could be given bonus marks (as IIIT Delhi does) and hence strong preference in admission. In fact, the national level Olympiads could come up with a list of top 100 in that subject who would get that preference in admissions. Of course, the exact mechanism is not important, and various institutions can come up with their own mechanisms.

Note that the goal of admission process is to select meritorious students who are likely to perform well in the program. There is no doubt that these students will do that. Also, this can be a vehicle for early admissions to students, something that is seriously lacking in India and causes much too unnecessary stress. So a student could be offered admission after 11th (because s/he got into Team India at that age) subject to reasonable performance in 12th class.

As an aside, it is being reported that all five of them studied with coaching classes. So may be we can give some credit to coaching classes, which are trying to help those who are suffering from poor schooling in India.

To end, specific suggestions in this article are not important. The point is that our top institutions should be open to the idea of admitting students who have excelled in different forums.

Monday, July 30, 2018

IoE at State level

As I have been saying for more than a year, my hope from IoE process was that if 10 private institutions are given complete autonomy, they will improve in their quality of education and research and that would create a constituency for greater autonomy for everyone. And my reason for disappointment with the process was that only 2 private institutions were selected (Greenfield does not count since there is no comparison between pre-autonomy and post-autonomy era in their case), and even from these two, one's autonomy has been withheld. I really wish they had given autonomy to the best 10 institutions without bothering about whether they would be in top 500 in 10 years or not.

A question that I have been asked is whether graded autonomy wouldn't serve the purpose. The answer is NO. IoE autonomy is greater than the highest level of graded autonomy. Most importantly, IoE autonomy is taking them away from state government control by converting these private universities into deemed to be universities (if they are not already deemed to be universities). And many states have controls which are sometimes even worse than what UGC/AICTE impose. For example, in some states, tuition control is so rigid that there is no way you can follow all UGC/AICTE guidelines on faculty:student ratio and the give the salaries to faculty members that are recommended by UGC.

Over the last couple of decades, I have been hoping that the competitions between states to attract the best educational institutions would ensure that they come up with policies giving more autonomy to their institutions. I was hoping, for example, that other states will notice how Rajasthan has so many private players setting up good quality institutions and will realize that it is because Rajasthan has very light regulatory framework for private universities. But this hasn't happened. On the contrary, in some states the regulatory framework is becoming more rigid, reducing autonomy for their private universities. Unfortunately, quality is not a focus of our education policy, only cost is. Worthless degrees at low cost are considered better than good education at higher cost, particularly in an election year, and the changes done in an election year cannot be undone after the elections.

So given that worthless degrees at low cost will remain the focus of higher education policy of most states, is there any way we can have some quality institutions in such state (other than central government ones and the deemed universities). Recently in a meeting, I heard this solution. (So this is not my idea, but I loved it, and hence sharing it here.) Have an IoE scheme at the state level. Just convince the state governments that they can have 100s of colleges where all their voters can get admission, but just give autonomy to 5 private universities in the state based on whatever criteria of quality you decide.

Just imagine if 20 states declare 5 private universities each as autonomous and these are also able to get autonomy from UGC under the graded autonomy scheme, we would really have 100 universities who can then try to improve their quality of education and research way beyond the expectations of our regulators. If we don't worry about top 500 but only about the quality, there is a greater chance of our universities being in top 500.

Sunday, July 29, 2018

Aadhaar Challenge: A political stunt?

Yesterday, Chairman of Telecom Regulatory Authority of India (TRAI), Mr. Ram Sevak Sharma, threw a challenge to the world. He gave out his 12 digit Aadhaar number (which I don't want to publish here) and tweeted:
"Now I give this challenge to you: Show me one concrete example where you can do any harm to me!"

The Internet world quickly found out the phone number linked to his Aadhaar, and slowly more information started coming in, his previous and current addresses, date of birth, his frequent flier number, his email addresses, PAN number, bank account details, voter-ID card number and all associated details, alternate phone numbers, the model of his phone, his pictures and that of his family, current location, secret questions to his email addresses, and so on.

Much of this information was in public domain anyway. He kept making two claims: One, all this information could have been found (and perhaps has been found) without knowing Aadhaar. Two, putting all this information in public domain results in no harm to him.

I have refrained from writing on Aadhaar, since a large number of pro and anti-Aadhaar people are divided on political lines, and hence most of the debate is ill informed. Also, the debate does not have to be "Aadhaar compulsory for everything" versus "Kill Aadhaar." But that is how it plays out in public arena.

Coming back to his challenge, I wonder if he has an agenda. Otherwise, a smart and wise man that he is (after all, he is a fellow alum of IIT Kanpur :-) and a Distinguished Alumnus Awardee), he would know that such challenges do not prove anything. If indeed someone is able to get information that can potentially harm him, he will keep arguing that this information was not found using Aadhaar (and most hackers would not reveal their methods). But on the other hand, if no one is able to get any important information in a short period of time, that is not at all an argument in favor of Aadhaar's security. May be it takes more time. So, either way, it proves nothing.

Also, he is big enough man to fight legal battles and has a large network to undo most of the harm, if something does happen. So it is very little risk to him personally. But throwing such a challenge is not in national interest. Since there is a possibility, however small, that some harm may happen, that some people may actually succeed in hacking. That a person at such a responsible position is throwing such a challenge is just so sad. The only reason I can think of is that he is playing to the gallery and his supporters and supporters of Aadhaar will be very happy with him.

He has been arguing that just knowing bank account number will not harm him. What he has not yet said is whether the bank account numbers should be in public domain. If indeed his bank account number has been found using Aadhaar number, isn't that a failure of Aadhaar. Of course, the supporters would argue that the bank account number may have been found not from UIDAI site but from some other source. But the issue is different. Even if we assume that five feet thick and thirteen feet high wall is enough to secure data on the servers inside those walls, shouldn't this be the responsibility of UIDAI to secure the entire Aadhaar eco-system. Shouldn't every Aadhaar center be secure. May be not 5' by 13' wall, but 1' by 7' wall :-) In fact, I would go a step further. How businesses and government departments keep Aadhaar and use them should also be controlled by UIDAI. If they have no control over such use, they shouldn't insist on compulsory sharing of Aadhaar. Of course, today's discussion is not even touching upon the issue of government potentially having access to every interaction that happens between me and UIDAI.

I know most of the information that people have found out about Mr. Sharma can also be found about me, but unlike him, this reality gives me stress. Of course, many will argue that no honest person needs to worry, almost suggesting that if I am stressed about it, I must be dishonest. But the way our government systems and courts work, undoing any damage is extremely slow and expensive, and I don't want to go that route.

Added on 29th July:
-------------------------
Another much more detailed and well articulated article on why this challenge is irresponsible.
Issues with TRAI Chairman RS Sharma publishing his Aadhaar Number, challenging hackers to harm him by Nikhil Pahwa

Wednesday, July 25, 2018

Should you drop a year and take JEE again

Yesterday, Amit Paranjape shared a link to a story of students taking a year off after 12th class to do interesting things. That led me to think about a large number of students who take a year off after 12th class to exclusively prepare for competitive exams like JEE. Is this the right thing to do.

The argument in favor of a one year drop are many. If you had a certain level of performance while preparing for 12th class and JEE together (just taking this as an example, I guess it will be true for any other competitive exam as well), your performance will certainly be better if you don't have to go to school, you don't have to take all those school tests and board exams, and you don't have to study "useless" subjects like language. Apparently, the data also supports the theory that the chance of a repeater are pretty good. If we look at yearly Advanced JEE reports, one finds that among those who are admitted to IITs, about half are those who passed 12th class that year, and half passed 12th class in the previous year. If we look at media (including social media), one finds many stories of people who succeeded in improving their performance in their second attempt, but hardly any stories of people who did worse.

So, shouldn't this be an obvious thing to do. With life expectancy of 80 years and rising, what is one year, if you can have a much "better" career afterwards.

But there is a problem. When we say that 50% of an IIT class is of repeaters, it does not tell us how many of them had succeeded last year as well. How many of them have improved their ranks substantially. If you are willing to invest one year of time, and resources (for coaching, for example) and lost wages for a future one year, it better be for a significant improvement in ranks and not a minor improvement in ranks.

Let us consider a student who has received a rank of up to 8000 in Advanced JEE this year. (I am using all examples of unreserved class, but same arguments hold for reserved classes as well.) This student could have received an offer of admission to Earth Science program at IIT Kanpur, and if he barely studied enough to pass all courses in the first year, he could shift to any program that closed at 3600 or later (through the program change policy of IITK after one year). If he worked hard in IITK, he could get a change to a program which is closing at 2800 or higher this year. So an improvement of 5000 ranks next year is completely useless since what he will get next year, he could have got that after joining IITK this year itself. Getting a better CPI to be able to get the top three programs - CSE, EE and Maths - is hard, but so is to get a rank within top 1000 next year. You can't depend on change of program after first year, but you can perhaps depend on it more than you can depend on getting a top rank in JEE Advanced (and being in top 1 lakh in JEE Mains as well).

Does it make sense to drop a year if you had a rank giving you an unpopular NIT seat and you hope that next year you will at least get a popular NIT seat. I still wouldn't recommend dropping a year. There are several institutions outside the JOSAA system which are as good but have significantly lower cutoffs for admission because they do their own admissions, and in some cases are far more expensive. I am talking about places like LNMIIT, DAIICT, etc. These are as good in terms of education as most NITs, but are not preferred because many students and parents go for brand and not quality, and also they will require an addition lakh or more per year. But remember, one year drop has costs too and as I said above, not just cost of coaching, but also cost of lost wages. So, go for a private place which is as good as an NIT.

So the only situation in which dropping one year may make sense is where you are sure of not a small improvement but a very substantial improvement. You had a JEE rank of 75,000 this year, and you are expecting a rank of 15,000 next year. Or you had a JEE advanced rank of 20,000 this year, and are expecting a rank of 5,000 or better next year. Small improvements are not worth this investment.

But what is the probability of major improvement. Very small!

While the argument stated at the beginning (that we will have exclusive focus this year) is attractive, one also has to see if one will continue to have motivation to study hard for one long year. Would you not feel lonely since most of your school time friends have gone to colleges. How would you answer your neighbors, relatives and all sundry uncles as to which college you are studying in. Would you not be stressed by the thought that you might perform worse than this year. In the coaching class, everyone is a new face, no one whom you had known for years in school. These conditions can break the motivation level of most people. And once your motivation is gone, so is your chance of significant improvement.

So when should one drop a year.

If you can identify a reason (other than giving 12th class and JEE together) that caused your JEE performance to be worse than what you believe to be your capability and that reason is essentially a low probability event, then it would make sense for you to drop a year. For example, if you fell ill on the JEE day or in days leading to JEE affecting your last minute preparation badly (otherwise, you were doing well in your coaching exams). You had a family crisis (like a major accident or worse, a death) which obviously would have affected your preparation and performance. Or any other big reason like these.

Of course, situation of every student is unique. Whether they can maintain motivation for entire year will depend on their will power. Whether something seriously bothered them and affected their performance will have to be decided by them. Whether it is possible to go for a much more expensive option for education will depend on their financial background. Worst case scenario will be different for each student. For example, someone who has performance extremely well in 12th class and can get admission to a good college based on that even next year is taking a smaller risk and hence is not likely to be affected by stress through the year. So take everything said in this article with a pinch of salt and see what applied to you and what doesn't. But, in general, if there is even a slight doubt, don't drop a year.

Monday, July 23, 2018

Tyranny of the last year's closing ranks

Finally, the IIT admissions are over. For the entire month of June, we had thousands of parents and students asking a million questions about various institutes and programs, how good they are, how good is the placement, what is the scope of these programs, and so on. After a massive amount of research, most parents have come to the conclusion that parents in the previous year were very intelligent, they had done all the due diligence, and therefore, it is best to create an option list in the same order as the last year's closing ranks. Of course, there are exceptions, and hence you would see that the order in which various programs close this year is slightly different compared with the order in which programs closed last year.

If we look at IIT Kanpur (and for illustration purposes, let us just look at unreserved seats and their closing ranks), last year (2017) its programs closed in the following order:

Computer Science (CSE), Electrical (EE), Maths and Computing (MTH), Mechanical (ME), Chemical (CHE), Economics (ECO), Aerospace (AE), Civil (CE), Physics (PHY), Material Science (MSE), Bio Sciences and Bio Engg. (BSBE), Chemistry (CHM) and Earth Science (ES).

The 2018 ordering of the programs in terms of their closing ranks is: CSE, EE, MTH, ME, ECO, CHE, AE, PHY, CE, MSE, BSBE, CHM, and ES.

If you notice Economics and Chemical Engineering have exchanged places and so have Physics and Civil Engineering. These are minor changes as the closing ranks for both Economics and Chemical Engineering were very similar both last year and this year. It is just a matter of a couple of students preferring one over the other. The same is true for the closing ranks of Physics and Civil Engineering.

What is rather interesting is that while those who got a rank in JEE 2018 were thinking of the order in which they should write their options, those who got a rank in JEE 2017 were also doing exactly the same. IIT Kanpur has one of the most liberal policy on change of under-graduate program, which is based on the first year's performance. At the end of first year, students apply for change and can give options in the order of preference.

The order in which 2017 rankers gave options for various programs at IITK in 2018 is: CSE, EE, ECO, MTH, ME, CHE, AE, CE, PHY, MSE, BSBE, ES, and CHM.

There are minor changes in the two list: Civil and Physics have exchanged places, and also Chemistry and Earth Science have exchanged places. But the big story is about Economics. Not a single Economics student has applied for change to another program and only students with a CPI of 9.0 (out of 10) have been able to get Economics. If we look closely, 2018 students have considered Economics, Chemical and Aerospace very similarly, with closing ranks of 2718, 3171, and 3174. But the 2017 students' preference for these three programs is vastly different with transfer closing at a CPI of 9.0 for Economics, 8.4 for Chemical and only 7.7 for Aerospace (with Mechanical in-between at 8.8).

Also, while the 2018 closing ranks for Physics, Material Science, BSBE, and Earth Science are vastly different at 3586, 4549, 5537 and 8104 respectively, they are almost same for 2017 batch. Indeed, any student from any of these four programs could have taken a transfer (and can still take) to any other program within this group.

Also note that students who have finished their first year in IIT are the most active responders on social media regarding counseling information. So what they are thinking in terms of which program is better, they are saying it loud and clear to everyone else. So what is happening. Why is there substantial difference in the two batches.

My hunch is that 2017 batch students are more independent of their parents, while 2018 batch students are very much dependent on their parents for the decision. And hence, when they realize that the perception of placement is very different from reality of placement, they can take action based on this new information. Also, some of them become confident of doing what they like.

The parents are more likely to be affected by the tyranny of the closing ranks.You may write whatever you wish in your blogs, FaceBook, quora answers, websites, but all that will convert the thinking of only a small number of people. As I had mentioned in another blog a couple of months ago, many institutes don't do any "marketing" because they realize that in the short term, it won't work. But, of course, they are wrong. A small impact every year will result in major change over a decade. And that is why it is important to keep giving the right information to all potential students and their parents every year.

Note on Program Change at IITK: This year we have allowed change of program to 87 students in a batch of 821 students. Also, we allow program changes even after 3rd and 4th semester. I am quite sure that at least 25-30 students will get their programs changed in the next two semesters. This means that almost 15 percent of the batch would have their programs changed. That is a very large number.

Since there is a request to make public more details of program change at IITK, here is the article on this.

Monday, July 16, 2018

Mountain out of molehill: Is 2 = 2.0

After a few days of suspense, the Madras High Court has stayed the single judge order asking IITs to reorder their ranks. (That the order really wouldn't have changed anyone's rank is not important.) And after a few days' gap, the counseling has restarted.

Let me recap the issue. The following was the instructions on the exam paper.

  • For each question, enter the correct numerical value (in decimal notation, truncated/rounded off to the second decimal place; e.g. 6.25, 7.00, -0.33, -.30, 30.27, -127.30) using the mouse and the on-screen virtual numeric keypad in the place designated to enter the answer.

Now, there were two interpretations of this line.

One, this instruction is meant only for those answers where the answer can be truncated/rounded off, and more specifically, was not meant for integers.

Two, presence of "7.00" and absence of any integer in the possible solutions implies that even integers had to be written up to two decimal places.

In my opinion, the first one is the right interpretation. However, the second one is a "reasonable" interpretation and since IITs are more interested in finding out whether you know the answer or not, we may award full marks for those who used second interpretation.

Second thing, I said was that instructions only suggest what is the minimum you need to do to get full marks and doing more should not be penalized. So even if writing two zeroes was suggested and was the correct interpretation, writing an integer as an integer is certainly a better answer and hence should be given full marks.

The opposite view point that was presented to the court and was written by lots of people as comments on various social media, including this blog was that the second interpretation is the correct interpretation and the first interpretation is so wrong that those who followed that should be penalized. And, following instructions is of supreme value in exams like JEE.

Frankly, I am bothered by the opposite view point. I am unable to look at it as a difference of opinion. There is something else at play here. What they are saying is that lack of an integer in the list and presence of 7.00 implies that integers had to be written in a particular way. Do they have any idea of what is implication, what is proof, etc. And then, saying that it is not enough to show that you know the answer, but following instructions is an integral part of an exam. Do they really believe that IITs are looking for students who can follow the orders and not think independently. If yes, IITs need to do a lot more in terms of improving its image.

In an exam, if I ask them to write names of any 4 prime ministers of India, and someone writes five names (all correct), should I penalize this guy for not following the instructions. Obviously, additional information, if wrong, can lead to penalty, but should there be penalty if the additional information is correct.

And as, someone commented, should IIT penalize someone who comes only one hour before the exam, when the instructions were to reach three hours before the exam. (I may be off in terms of exact time.) Obviously not, since following instructions are useful only to the extent of peaceful conduct of the exam and not beyond that. Similarly, following instructions in the exam are useful only to the extent of displaying your ability to solve problems in the exam and not beyond that.

To me, this alternate opinion is borne out of inability to take responsibility for one's actions. Almost every student tells his/her parents/friends that they deserved better in any exam than what the result indicates. I talk to so many JEE passed students every year for counseling, and the conversation always start with, "I was expecting a better rank, but...." And they are always looking for excuses to convince themselves and others that they are indeed better than what that rank indicates. (And they very well may be, the rank indicates nothing about them as a human being.) And as soon as they find out any excuse that anyone has used, they are very happy and start using that excuse themselves. Sometimes, even small coaching places, will tell their students that they deserved better rank, and it is all because of someone else (IITs mostly) that their ranks have suffered.

Because of this reason, one can easily predict that the kind of problem we had this year, will actually happen every year. Such issues have arisen every single year in the last few years, and will continue to happen in future too irrespective of how good the language is of all instructions and all questions.

IITs are normally very reluctant to change since they think that following what was not questioned in previous years would ensure less court cases this year. But I think they should learn from JEE 2018 where there were major changes (women reservation - I am really surprised there is no court case on this, fully computerized exam - there was a case, but dismissed early, numerical questions with answers not between 0 and 9). Irrespective of how many changes you bring about, all those who want to blame IIT JEE for their performance will latch on to one thing and the other changes will go unquestioned, and they will latch on to that one thing, even when there is no change.

The only way to reduce court cases is to reduce the value of the exam. Give students option. If not doing too well in one exam leaves good options open for students, they will feel more confident of owning up their performance and not look for excuses.

Till that happens, court cases will happen irrespective of what IITs do. So they shouldn't be afraid of changes. Changes do not increase the number of court cases.




NEET 2018: The solution is normalization

If there is one exam worse than JEE, it is NEET. And it is compulsory for all medical admissions. Earlier, we have seen different question papers in different languages, but no attempt to see if the two papers have similar difficult level, or normalize the marks if they aren't. That is not a surprise to those who have followed CBSE's conduct of JEE Mains in the past. There is no normalization in the multiple papers they offer (online and offline).

This year, a whopping 49 questions in the Tamil version of NEET had wrong translation. Well, at least they translated so that everyone gets the same set of questions. Better than completely different papers of the past. But 49 wrong questions out of 180 total questions. Obviously students would complain. CBSE refused to do anything about it. The matter goes to high court. CBSE argues that students should have read the English version since they had already issued the instructions that in case of any problem in translation, English version would be considered authentic. But they couldn't answer a simple question. How would the students know which questions are wrongly translated, and hence they should look at English version. And if it is to be assumed that everyone understands English question paper, why give options of local language to begin with, or was it the case that students from vernacular background are required to read both papers and spend extra time (without any compensation, of course).

The court was aghast at such arguments. CBSE should have suggested how one could fairly treat the 24,000 students who had taken the test in Tamil language. And sure enough, the court ordered that all 24,000 students should get full marks in all 49 questions.

Obviously, it is unfair to others. It is no one's argument that all these 24000 students would have solved correctly all 49 questions, if there was no error in the translation. And by awarding them full marks, they have been given vastly higher marks than what they would have scored in an error free exam. And thus, they have been placed higher than those students who took exam in English.

How do we know how many marks these 24000 students would have scored in an error free exam. We really can't say about an individual. And then there are second order effects like wasting time on a question which is now ambiguous, say. But statistics does allow an approximate answer. We can normalize marks based on 131 questions that everyone has been able to attempt.

The problem is that CBSE has never learnt how to normalize marks. They haven't done it in JEE before. They do a completely horrendous "normalization" of 12th class marks, which is nothing but grade inflation. So they don't know how to provide at least a statistically fair solution to these 24,000 students.

Is there a way to study correlation between 131 questions and 49 questions. Can we do this analysis for different categories of students - all India, students from Tamilnadu, students from TN Board, etc., and figure out which class is most similar to the group of 24,000 students who chose Tamil as the medium of question paper. Is there strong correlation of some questions out of those 131 with each of these 49 questions. How difficult is each question for that cohort. Based on these types of questions, one can start looking at what would have been the likely score of the student if there was no error in translation and if they were like an average student of the cohort that we are comparing them with.

But, I suspect I know what will happen eventually. CBSE will look at the marks of 131 questions for each of those 24000 students. We will then see how many marks did other students getting identical marks in 131 questions got in 49 questions. We will assign that many marks in 49 questions to those students. This is a very simplified version of normalization but something is better than nothing and courts might agree with this, and frankly, CBSE will only care what will go through the courts.

Any solution based on statistics will cause heartburns. But it is the best you can do, short of conducting the exam again.

Sunday, July 15, 2018

Institute of Eminence Report

In the last one week, a lot has been said about the Institute of Eminence process and report. There are far too many questions about the way this policy has been handled. And my take on this is, that it may have been strategic. Make so many errors, violate your own policy so many times, that different people will comment on different issues and there would not be one single issue on which all the people are united in criticism.

The report that has been made public on UGC website says nothing about how specifically these institutes were decided. There are a lot of general things that they say (but do not seem to follow themselves). When I was reading the report, I was repeatedly asking myself, if this is a competence issue, or is it that they were not willing to spend enough time on such an important assignment (after all, other than the chairperson, members are extremely busy people, but then they should have refused to be a member rather than sign such a poor report).

The report clearly says that the primary goal of the committee was to look for universities which would be in top 500 of any popular world ranking within 10 years. One would have assumed that those universities which are already in top 500 would be automatically selected, unless this committee is convinced that those universities have seen their zenith, and are now on the declining path, and in the next 10 years, they will be out of top 500 list. And if committee is convinced of this, it would actually give reason for the same. So what is the basis of keeping IIT Kanpur out is not clear.

The committee talks about private institutions not doing enough research and hence the possibility of them getting into top 500 is less. And then, it ignores the private university which arguably has the best research output (Amritha University) and because of this research output they already have comparable or better ranking in various lists compared with the two private universities that have indeed been selected. There was certainly a possibility of having the 4th private university in the list.

In the report, they mention that some of the newer universities are so small (less than 3000 students) that they couldn't really become world class, but at the same time recommend a university with 0 students to be in the list. Also, when the goal is to have a university in top 500 in 10 years, how come a university which is claiming to be in top 500 in 13 years has been selected. Jio University seems to have suggested that it will start in 2021 and will be in top 500 by 2031. Based on their only higher education venture (DAIICT) which is nowhere in the world rankings even after 20 years, it seems difficult to believe that they will be in top 500 within 10 years of starting, but in any case, the goal was 10 years from the selection, and they themselves are saying that they will not be in top 500 within 10 years from selection. For a good analysis of this decision, read the following report by Prof. Sandeep Shukla.
Money cannot buy excellence in education, but Jio's 'Eminence' tag is worth crores.

This is not to say that Mr. Ambani cannot be the promoter of a great university. I hope Jio will be one. But as of today, there is no reason to believe that they will be in top 500 within 10 years. Indeed they themselves are saying that they won't be. So the tag is clearly misplaced.

(People are too sensitive. When I say DAIICT is not in world ranking, I am saying only that much. I am not saying that it is a bad institute. It is a good institute, but not in top 500 ranks in the world.)

The committee says that many universities applied only to get autonomy from UGC and they can be dealt with within the graded autonomy regulations. But why should that be an issue. My motivation may be to gain autonomy, but you please look at me from the perspective whether I can get into top 500 ranks in the next 10 years or not.

The committee has recommended 8 public universities. When they had space for 2 more, they have still decided to ignore those universities which are already in top 500 list, the primary goal as stated by the committee. One would have expected an explanation in the report (and not now to the media, even that is not forthcoming on this issue though).

On what basis has government chosen 3 out of 8. There appears to be some feedback from the committee, since the Chairman of the committee has said that IIT Madras was not selected because Chennai has bad weather because of which international students and faculty don't come to Chennai and hence their international ranking is poor.. Gimme a break. Weather in Chennai much different from Mumbai? And other IITs have a large number of international students and faculty members?

IIT Kanpur is a curious case. Not only it is not in the list of 8 universities, it is not even in the list of specialized universities focusing on science and technology. A university which has been consistently in the top 5 of the country in all rankings over the last so many years being ignored without giving any reason as to why the committee believes we will not be in top 500 in the next 10 years. Do they know something that we don't know. Was our presentation so poor. Was our proposal so poor. (I was hoping to work for an Institute of Eminence. I am disappointed though I guess I have the option to work for Jio :-)

Combine all this with what all has happened in the last one year, and the optics is really poor. Reopening of the application submission even though 100 universities had already applied. Changing the shortlisting after the committee had decided to shortlist only 40 institutes and asked them to give presentations.

What is certain is that the four members of the committee have not displayed any eminence in signing this report.

Other articles on IoE saga:

UGC's laziness has led to needless 'Institutions of Eminence' controversies by Prof. Pushkar
Imminent Eminence: Ambani's Egg by Prof. Mukul Kesavan
Institutes of Eminence status given without field visits, rankings by Anubhuti Vishnoi
Jio Institute: Why the Modi government is misguided in giving the eminence tag to a select few by Arihant Pawaria
The 'Institutes of Eminence' falls woefully short of what India needs  by Maheshwer Peri
After UGC, Expert committee too disappointed in selecting Institutions of Eminance by Prof. Pushkar

Tuesday, July 10, 2018

Higher Education Commission of India

The government is proposing to set up a new regulatory authority that will replace UGC through an act of parliament. It will be called Higher Education Commission of India or HECI.

At the outset, this proposed bill appears to be a product of complete confusion in the Ministry and their right hand not knowing what the left hand is doing. But after a moment's reflection, it would seem that the right hand indeed knows what the left hand is doing and is not particularly happy about it and may even want to cutoff the left hand.

Before you propose a solution, it is usually a good idea to define a problem and then search for a solution. Of course, sometimes a solution is proposed first and the problem is searched appropriate for that solution. So what is the problem. UGC, of course. You dimwit. You claim to write about higher education, and you don't even know that UGC is the main reason for most of the problems of higher education. I would politely argue that I know UGC is part of the problem, but do you know what aspect of UGC's functioning is the problem.

As per the media reports, the government seems to believe that UGC was not able to function properly because it spent enormous amount of time in deciding grants. If that function is removed from there, they will focus on academic matters, and Indian education system will soon be the best in the world. But wouldn't it be a drain on the time of IAS officers. How will they cope up with this huge extra workload. And if you are going to put extra manpower in MHRD, why not put that extra manpower in UGC. And a voice from the clouds reminded me that I am supposed to be a professor of Computer Science, and hence should have heard of not just cloud computing, but also AI, ML and the government's favorite these days, the blockchain. And somehow the combination of all these technologies will appear in the form of a software, that would disburse the money to all the universities and colleges (I am sure through the Aadhaar verified accounts), and everyone will be happy ever after. In fact, it will improve autonomy since funding decisions will be taken through Artificial Intelligence (real intelligence being in limited supply) and we, of course, all know that AI cannot have any biases and softwares have no way to be tweaked or controlled. But I have a query. If we can develop such a software, why not keep the server in UGC building, why should it be in MHRD building.

Then we hear of he second problem of UGC. It does not have sufficient teeth. By the way, how many teeth does it have? My friends in universities are mortally scared of UGC. It can send completely arbitrary diktats to universities and ask them to comply even when it does not have sufficient teeth. Do you really want to give them more statutory powers. I mean, Indian universities probably are more controlled than in any other country, and you want to increase that control.

And I always believed that the government policy was to give more autonomy. I hear about this Institutes of Eminence scheme, and the graded autonomy scheme and Autonomous colleges and so on. The whole idea seemed to be to slowly increase the autonomy in the higher education sector, but with one master stroke, you are going to undo all of that and put all sorts of controls back on the universities. (Cutting off left hand!)

HECI can not just issue guidelines on anything under the sun, but also demand compliance and take the strongest possible action, including removing the right to grant degrees, if it is not satisfied with the compliance.

It can even ride roughshod over state legislatures. Anytime a legislature approves setting up of a new university, HECI will intervene and say, not so soon. First you need authorization from HECI. That is federalism - Indian style.

Whenever anyone has criticized UGC and wanted its demise, the hope was that if somehow we can just remove lots of UGC diktats, our universities will start improving. But this new act believes in exactly the opposite. You must control much more than now for the universities to improve.

So you have an interesting state of the affairs with Prime Minister's office repeatedly reminding MHRD that we need to give more autonomy to universities, and MHRD repeatedly trying their best to not only continue license permit raj but expand its control.

I am extremely suspicious of anyone who says, "here is the new rule, and this rule applies to everyone except me." If this rule is good for others, why is it not good for you. The proposed law wants to keep Institutes of National Importance outside the regulatory framework of HECI. And most of the central government universities (IITs, NITs, IIITs, even the PPP model IIITs, IISERs, SPAs) fall under the INI nomenclature. Which means that HECI is meant to control primarily state universities (including private ones) and deemed universities. I think a common regulator for all will at least keep the regulator abreast of best practices and there is a hope that the regulator may suggest those best practices to others. There is also hope that if HECI tries giving a stupid diktat to IITs, there will be enough hue and cry that the diktat will be forced to be removed, thereby saving our universities. The framers of the law have figured this out and decided that anyone whom the media is likely to support should be out of regulation. I suggest that they add a line saying that besides INIs, MHRD reserves the right to keep any other university out of HECI's regulatory control. This way any future university who has as many friends in media as IITs can be quietly kept out of regulation.

Then there are drafting issues which make the Act confusing. The members are required to be scholars and more specifically persons of eminence and standing in the field of academics and research, etc. And then it says that Secretary of Higher Education, Secretary of Skill Development and Secretary of DST shall be members. The first two are usually IAS officers and not persons who have a standing in the field of academics and research. And then, one doyen of industry who too is unlikely to be fitting the bill. So, may be they don't need to specify the qualifications of the members.

All members to be appointed by the Government. No role of states. I am not sure how we could bring in non-Government players, but there should be some way. May be the chairman and vice chair can be appointed at the higher level, as in many other commissions, by Minister, leader of opposition, and one more.

There is a confusion about conflict of issues and resulting two year cooling off period for members. Two current professors will be members. If they retire/resign or cease to be professors during their term as a member, then they stop being member of HECI from that day. On the other hand, the moment their term ends, they must cease to be a professor as well for a period of two years. So during membership of the Commission, they can be professors, and that is not conflict of interest, but after their term is over, they cannot be a professor since that will be conflict of interest. I think there should be an exception for professors, since otherwise, you are essentially saying that only those on the verge of retirement can be members. You do need to allow younger blood to be members.

Members are appointed for five years. But suppose a member resigns without completing his/her five year term. The new member appointed will be for a three year term. So, if the previous members stays for 5 years, the next member gets a five year term, but if the previous member stays for 4 years and 364 days, the next member gets a 3 year term. This appears to be a drafting error.

Bottomline: The bill wants an extremely tight control over higher education.

Monday, July 9, 2018

Institutes of Eminence

So, finally, the first chapter of this saga has been written. But we are promised that there are more chapters to be written still. For we wanted to select 20 institutions of eminence and in the first round, only 6 have been selected, leaving 14 slots to be filled in future.

Congratulations to IISc Bangalore, IIT Bombay, IIT Delhi, BITS Pilani, Manipal, and Jio University (yet to be set up, but with Reliance behind it, one hopes it would really be transformational).

I have been a supporter of IoE scheme, but the result is rather disappointing. I supported this for the following reason. I assumed that in the Government sector, most of the institutions selected would be those who already have a huge amount of autonomy, and what they really get is some small change to try and improve further. In fact, in multiple IITs, there were discussions whether it is worth applying for, since the fear was that the autonomy would be less under IoE scheme than what we currently enjoy. To me, the real game changer was the selection of 10 private universities.

I have been a strong votary of giving huge autonomy to every university in the country. But most people remain skeptical of it. I felt that if 10 universities get as much autonomy as was being promised, it would surely lead to all of them becoming much better than what they are already and that would happen in a relatively short period of time, say, 5 years. If academicians, administrators, politicians, all of them see that 10 universities have improved so much in such a short period of time, and the only change is autonomy, they will all think of removing all shackles of other universities, and this will lead to the golden period of India.

But now, with only 3 private universities being given that autonomy, and one of them being a new university (which means that we cannot compare pre-autonomy and post-autonomy period of this university), the impact of this experiment in five years would be so much less. Less number of people will be convinced about giving autonomy to every university based on this experiment, and hence the announcement has been a huge disappointment.

Of course, the minister has said that this is only the first list. So there is hope of more lists. But I guess it would be difficult to do this process again very soon, and then there are elections. So the next list is at least a year away.

If media can be trusted, apparently there is no dearth of good government institutions, but they could only find three private institutions. So, if they can select a 4th private institution, they would immediately announce 4th government one as well.

I am disappointed that Ashoka University is not in the list.

I am also disappointed by some statements in media attributed to the Chairman of the committee, Mr. Gopalaswami. He said, "Where we saw an institute had not improved its accreditation and ranking scores in three cycles, it does not inspire trust that it will be able to meet this goal" [of breaking into top 500 ranks in 10 years].

Accreditation happens in 3 to 5 years, and hardly anyone has gone for three cycles. So I would assume he is talking about ranking and in particular, NIRF, since private sector rankings cannot be trusted. Now, NIRF is still a work in progress. In three years, there have been significant changes in its methodology and in the level of participation. To use such a ranking for denying someone Institute of Eminence status is exactly what a babu is trained to do. Babus cannot do subjective quality evaluation. They need objective numbers to support them. I expected better from this committee. Of course, I understand that this is not the only thing they have looked at. But even pointing this out means that this was definitely an important enough criteria. That this was not the only criteria is proven by the fact that Amritha which has improved rank every year and has had an excellent rank in all three years is not in the list. They were 14th in 2016, 9th in 2017, and 8th in 2018 in the universities ranking.

But overall, a small step forward. Hope this small step leads to bigger steps forward in due course.

Added on 9th July:
-------------------------------------
After posting this, the maximum comments (not here, but on other social media sites) have asked if Manipal is better than IIT Kanpur. Well, the straight answer is that in the disciplines IIT Kanpur operates (Engineering and Science, mainly), IITK is better. But there are many disciplines that Manipal has (like Medicine) in which obviously there can be no comparison. If the reports saying that the committee had actually decided on a list of 8 government and 3 private universities, but government took a policy decision to have an equal number of government and private, are true, then I believe that government would do us a great favor if it admits to this. That way all government institutions can claim that they are those 5 unannounced ones. I am convinced IITK would be in that list of 8, and hence even by this metric, not worse than Manipal. Of course, I must add that Manipal has been continually working to improve its teaching and research. They have improved their NIRF rank every year and since the goal was to find private institutions who will be in top 500 in 10 years, they were always a serious contender.

Second most common comment is about Jio University. Should a greenfield university be named as an Institute of Eminence. Well, it is too late to ask. The call for proposals clearly mentioned that greenfield universities are welcome and that they will be evaluated on the basis of plans that they have for the university (and the financial strength and commitment of promoters to carry out those plans). If people did not protest then, they shouldn't protest now. The best comment I received in this regard is from Dr. Bijoy Panigrahi (IIIT Naya Raipur) who suggested that we should have had three categories - government, private and greenfield. I fully agree (and new IITs could be considered in the 3rd category), but it is too late to suggest. I wish this suggestion was given in 2016. So as of now, let us hope that they will start soon, and have great plans and a great team to materialize those plans.