Monday, February 13, 2012

Prof. V.G. Idichandy on JEE

The Fifth Estate, the "official magazine" of IIT-M, has a wide ranging interview of Prof. V.G. Idichandy [Part 1, Part 2]. First a brief intro:

In a career spanning more than 40 years at IIT Madras, Prof. Chandy has held numerous positions in the campus administration – Dean of Students, Deputy Director, Chairman – JEE and many more. He has played an instrumental role in setting up many of the institution’s current systems and is widely known for his camaraderie with students of the institute. He retired from the role of Deputy Director in October, 2011."

The second part has a section devoted to his views on JEE:

On the JEE System

During his association with JEE, he made a general observation that a diligent student with good marks in the 10th and 12th did well after coming in to the IITs but that there was no correlation between performance in JEE and performance at IIT. [See this Indian Express report from 2005 on Prof. Idichandy's study. To the best of my knowledge, it's not (yet) in the public domain.]

It’s a 12th std boy who’s writing the exam, what’s the point in asking him to solve M.Sc. level questions? This way we were purposefully driving them towards coaching institutes.

As the way forward, a suggestion was made that every board should select their top students; that is to say, those who scored more than mean plus two times the standard deviation, and ask them to write the exam.

We also suggested collapsing the two-part examination into a single objective paper that tested your analytical skills and comprehension.

The latter proposal was accepted, but the former was not, because the minister of the time remarked that the cutoff being based on statistical quantities was not something most people could understand. Instead, it was decided to insist on a first class. So only 60% and above were able to write the test. Prof. Idichandy was still not convinced.

We weren’t very happy about that, because the purpose wasn’t served.We had also decided to curtail the number of JEE attempts to two. There were people who stayed in coaching institutes for years together, till they succeeded; at that time there were people who were admitted to an undergraduate programme here even at the age of 28. Such was the attraction towards getting into IITs!

Views on Branches being allotted based on AIR JEE Rank

Another proposal was that branches should not be allocated on the basis of JEE rank. As he put it in his own words:

I’ve been fighting for this for a long time, I have not succeeded so far.

His explanation for the same was such: once you get into IIT after several years of rigorous exercise, you’d most likely be spent and exhausted; you’d want to relax and enjoy your first year. As a result, 30% of students fail in Physics, Chemistry and Math – these are all students who have cleared JEE. So the idea was to not allot a branch on entry, but to give them time to prove themselves or earn a branch.

Admission into a particular branch need not only depend on an All India Rank.

Instead, he suggests, one could study basic engineering, understand what each branch has to offer, and then make a well thought out decision – maybe at the end of second or third semester; until then, they would all have common courses. This way they are also away from the pressures of parents as well as those of society, a society that believes only people in Computer Science and Engineering would survive. Aptitude for a branch is never taken into account at all.

The senates of every IIT had deliberated on this before, and it was almost decided upon a few years back, but then two IITs objected, and there was no consensus. This year, once again, it has come up as an agenda item, and will be discussed again.

He backed up the idea of a branch of study called General Engineering, which wouldn’t specialise in a particular discipline.

Why should every student who get into IIT go through the rigours of high funda mathematics which are required for design and analysis? Many students would be happy to choose such a stream – why should you do a specialised degree in Aeronautical Engineering if you’re intending to do an MBA afterwards?

A student undergoing this degree could choose to specialise in a suitable branch at masters level.

Unified entrance examination:

The pool of students writing AIEEE and JEE are almost the same; at least for the first five lakhs. He further questions the need for two separate entrance exams. A more important change that has come about is the adequate weightage given to 12th standard performance. The fact that a 90% in different boards may mean different things is being compensated for by calculating percentile instead of using the absolute marking – a lot of analysis is being done in this regard.

2013 students have been given sufficient notice regarding 12th standard marks weightage”, he concludes.


  1. Dheeraj Sanghi said...

    The problem in India is that we want to solve all problems at one go, and don't want to solve problems incrementally. So, it is either everybody gets a branch at the time of JEE Counselling, or nobody gets a branch at the time of JEE Counselling. There are lots of intermediate steps. How about filling up 95% of seats in specific branches at the time of JEE Counselling, and 5% getting no confirmed branches, even though seats are there, and then at the end of 1st year, we have a liberal branch change policy, and everybody gets a branch at that time. I am not saying this is perfect, but just an example that it is possible to think of making incremental progress. When you try to change the world in one go, the resistance will be too much to handle.

  2. GT Brown said...

    "but that there was no correlation between performance in JEE and performance at IIT."

    Did they account for restriction of range? This is a common problem with such analysis (eg. predictive validity of SAT/GRE on undergraduate/graduate grades).

  3. Abi said...

    @Dheeraj: I agree with you -- IITs could have tried all kinds of little experiments including the one you mentioned -- admitting a subset of people into the 'general engineering' programme, and allowing them a year or two to choose their specialization.

    There are other such admission-related experiments: direct entry to a subset through Olympiads, and board exams. This would have allowed them to tweak their admission policies (based on experience with students coming in through multiple routes) to get the kind of student body they need / deserve.

    It's rather sad that they stuck single-mindedly to the JEE ...

    @GT Brown: JEE is a 'pure' entrance exam -- whose primary purpose is as a 'filter'. IITs don't bother with the kind of justification about its predictive validity.

    Unfortunately, all we know about Prof. Idichandy's study is through snippets littered here and there. The study itself is not available online.

  4. Vikram said...

    @GT, I believe you are talking about correlating score ranges (say 680-720) to performance instead of the specific scores themselves. I think that seen in that sense most of the students who qualify for entry to the IITs would be in the 760-800 range, maybe even 780-800. The admission rate is only around 1% so all the kids who get are in the 99th percentile.

    So in some sense the JEE-GPA correlation study seen in an American context would mean that there is no correlation between the very top end of the SAT scorers and their GPAs in college. That might well be true, but the thing is American universities admit students based on a combination of GPA and SAT scores, and the combination of the two does predict college GPA quite well:

  5. rahul said...

    I have one interesting observation..Why there are so many ex-students of IIT computer science dept who are professors at top notch universities? I never found many Indian metallurgists or physicists in US universities..Why average cgpa of CS and ECE students is much higher than that of Metallurgy or Mining students?

    From my experience at IIT, I would say top 1000 rankers are always good. There is no difference between rest 3000.

    My 2nd point is why IITs are always giving emphasis on quality of incoming students..Can't they add value to their education? In US, people don't even do calculus in high school. (except some good schools)Why teach so many courses in 1 semester. In US, there r 3/4 courses per semester. Here we have 5 theory and 2 lab courses. Assignments are not regularly given..If given, profs never check them properly. So why should one expect them to solve if people copy them from good students. Personally I think it is the hardworking students in the departments like Metallurgy or Mining who are the biggest losers at IIT. Because they will be mocked by their batchmates if they take their studies seriously, and finally their batchmates will earn more after doing MBA and laugh at them as they will be doing core job or higher studies. And the curriculum of these branches at IITs is outdated. So even if some students r going for PHD, they r not able to do well like their american counterparts because here they r teaching powder metallurgy, and in US, people r doing organic electronics and high level nano stuffs.

    Whatever entrance exam is taken, the situation will not change if venture capitalists are considered GODs in the campus and students do one night study( not in 1st year but after interacting with their seniors).

  6. rahul said...

    The situation is same in all engineering colleges as people do engineering to get a fat pay package not to make devices or solve the energy crisis. My friends who did well in board exams and could not get into IITs are also selling soaps. The best method is to make 85% cutoff in board exam. But the admission should be based on entrance exams as it hones problem solving skills

  7. Digbijoy Nath said...

    Rahul, I agree to your point that most (smart) students study engineering to make money by pursuing an MS/MBA later, and NOT to solve energy problem or make new devices. In fact, most people's (whaetever profession) idea is to increase bank balance for a secured future (basic human tendency)..

    And you are again right - Material Science/Metallurgy departments in India are yet to catch up with the extremely sophisticated levels of materials research as in USA ... here, Materials Science department works in sync with ECE and Physics department to study novel materials like double perovskites, organic materials, flexible electronics, spintronic devices, etc. besides classical/physical metallurgy (steel, iron, etc.)