IIIT Delhi: An Innovative Admission Process

This appears to be a month when I will write less about my ideas, and write more about the innovation in higher education that some institutes are carrying out. Today, I want to write about the novel and innovative admission process that IIIT Delhi has adopted for its under-graduate program. (Actually, one can learn a lot from their PG admission strategies also, but I will focus on UG admission today.)

First, the mandatory disclosure. I am a member of their Academic Senate. However, I have attended my first meeting only recently, and share no credit for this innovation.

First a bit about Indrapastha Institute of Information Technology (IIIT) Delhi. It is an Institute solely focused on IT education and research, set up in 2008 by Delhi Government, and is a university empowered to give its own degrees. Within three years of its existence, it has built an awesome reputation. Excellent faculty. A visionary Director.

The process for under-graduate admission starts with the question: What type of students do we want. A question that is rarely asked by universities and colleges in India. As a first cut solution, one will argue that we want students who know their science well, and have an aptitude for engineering or more specifically IT. Aptitude! They already are different from everybody else. But they go much beyond this initial answer.

Should we admit students who have reasonable aptitude and are excellent in science. Or should we admit students who are reasonably good in science, and have excellent aptitude. Most institutes would vote for the former (that is, if they at all think about it this way). It is then perhaps assumed that 60% (or whatever) marks in 12th class is sufficient indication of reasonable aptitude. So they will have a test based on Physics, Chemistry and Maths, and rank applicants based on score in the test. But IIIT Delhi is different. They think aptitude is lot more important. So they will consider 80% marks in 12th as sufficient academic preparation, and take an entrance test which tests aptitude. The ranking is based exclusively on this aptitude test. (As an aside, this is the method for selecting students from within Delhi. They do admit some non-Delhi students which is done through AIEEE. They do not ask them to come to Delhi to give the aptitude test.)

The story does not end here. The next question they ask is whether it is really important for someone studying IT to know a lot of Physics, Maths, and Chemistry at the 12th class level. And their answer is that just like engineering colleges are starting to offer basic biology related courses in their curriculum, the requirement of even Physics and Chemistry knowledge is to a minimal level, and if needed, that knowledge can be imparted in the Institute. So, the eligibility is just Mathematics with any other 4 subjects in 12th class. So, they are creating an opportunity for students who have done Maths, Commerce, Economics, etc., in the 12th class, to get into a BTech (IT) Program. It is interesting and I certainly admire the courage of IIIT to conduct this experiment. (I am told that not many people from non-science streams applied this year. Perhaps the message did not go across. Hopefully, more students will take advantage of this from next year onwards.)

The minimum eligibility is 80% marks in Maths, and 80% marks overall in 5 subjects in 12th class. They say that the cut-off is on the higher side since they do not want a very large number of students to apply for a small number of seats. And you fill up their form after the 12th class result is out. They are not in the business of creating a huge profit through their admission process. Their cut-off is high enough that students who only did coaching by ignoring their 12th class completely are less likely to be eligible. In their own little way, they are encouraging school education in Delhi, without burdening the kids with the pressure to get close to 100% as some of the colleges in Delhi University do.

And finally, they seem to have done their home work well in terms of how to judge aptitude for IT. So the exam consists of five areas: Comprehension, Data Interpretation, Mathematical Ability, Analytical Ability, and Decision Making.

The entire process is online. One fills in form, takes the print out of the admit card, and just reaches the examination center. From the last date of filling up the form to the announcement of results, it takes only 18 days. (There is no typo. It is not 18 weeks.)

The admission process and many other innovations being carried out at IIIT confirm my belief that leadership is extremely important for the success of an institute, even more so for a new institute.

JEE 2011 Admits Errors in Question Papers

JEE has announced the answer key on its website.

It took JEE only 5 weeks to announce the answer key. And it must have hurt the organizers of the world's toughest exam to admit that they had SEVEN questions which were either ambiguous, or had a printing mistake, or were simply wrong. And these questions are worth 36 marks. That is huge when 4+ lakh students are vying for less than 10,000 seats.

And this is not all. Apparently, there are more questions, even in Chemistry, where the "official" answer is being considered unsatisfactory by the students. But IIT will not take cognizance of these complaints. The answer key is final.

And one wonders why did IITs take 5 weeks to announce the marking scheme. If they had done so in 5 hours, they could have taken inputs from all, and then finalized the scheme. But now, they have to announce the result on 25th May, and they can't change the key at this late stage, even if some errors are genuine. The doubt about JEE 2011 will always remain.

Of course, given that they weren't interested in listening to the students to begin with, they did the next best thing, in terms of giving marks to all, or giving marks for multiple correct answers.

But doing the best under the circumstances does not mean that it won't hurt anyone. For the ambiguous questions, if someone found the answer, what is now deemed correct, but did not write the answer, the candidate does not get any benefit. Of course, the hope is that there are very few such students, and some unfairness to them is more acceptable than to redo the exam and delay the entire admission process by two months. (And I agree that one needs to be practical about these things.)

The Right to Information Act has forced JEE to give out some information in the last 5-6 years, and in each of these years, there have been some serious doubts about the conduct of the exam. It makes me wonder. Was I lucky to have passed JEE a few decades ago. May be they were equally inept at conducting the exam at that time too, and I was declared selected by mistake. I must thank my stars that there was no RTI then. Otherwise, someone else may have studied in IIT in my place, and that someone would have been writing this blog, instead of me.

I think IITs should seriously consider giving up JEE. They just can't handle it. A professional organization should be entrusted with conduct of such an important exam.

AIEEE Paper Leak

Yesterday, there was chaos across the country. The AIEEE paper was found in the possession of a person in Lucknow (and later in coaching Mandi of Kanpur, and other places as well). Apparently the paper was being sold for huge sums, some media reports mentioning figures as high as Rs. 6 lakhs. Why on earth someone will pay lakhs of rupees for a paper, when the probability that such a leak will be found out is very high. But that is not what I want to discuss.

I am amazed by the handling of the incident by CBSE. It was quite remarkable that CBSE had thought of this possibility and had a full back up in terms of duplicate exam material which was already stored in each city where there was a center for the exam, and this material could be distributed to most of the centers within a couple of hours.

(Contrast this with JEE paper leak in 1997, which, co-incidentally also happened in Lucknow. We had to postpone admission by 5 weeks, causing havoc to admissions in all other engineering colleges in the country, since most students who get admission to IITs, will leave their other college, causing vacancies there. There was no plan B.)

Unfortunately, there was AFMC exam in the afternoon, and thousands of students were scheduled to give both. Worse, some schools were centers for both AIEEE and AFMC exam, and therefore, could not re-conduct AIEEE after 3 hours. CBSE lost no time in giving relief to all affected students. They announced that those who had to give AFMC or those whose center cannot conduct a late AIEEE exam will be able to give AIEEE on 8th May. Taking such quick decisions must be applauded. Of course, the problem did not end here. Some students complained that they had both AFMC on 1st afternoon, and another exam on 8th. CBSE again came to their rescue and said that they will be able to give exam on 10th May.

What this means is that CBSE was ready with not just one extra set of examination material, but is also ready with multiple sets of question papers. That is excellent planning. So despite a major problem on hand, CBSE will stick to its timeline for announcing results.

When 11 lakh students are told that they have to stop writing an exam, and give a different exam a couple of hours later, this is bound to cause problems. Chaos is bound to happen. But the fact that more than 10 lakh students could give the exam the same day is something that speaks volumes about how AIEEE is planned (as opposed to JEE where we take double the time to announce the result for less than half the number of candidates, with no transparency).

And now an interesting outcome of this. Since the students would have given several different exams (the question paper on 1st, 8th, and 10th, and the online version would be different), there would be some way to compare the performances of the students across those exams. And if the exam can be held on 1st, 8th, and 10th May,
why can't it be held on 1st December, 1st January and 1st May (just giving three random dates to make a point that it is possible to organize the exam on multiple dates through out the year - and may even allow students to sit for 2 of those exams - something that I have been arguing for in many of my posts).

Added on 3rd May: CBSE has announced that the exam will be held on 11th May for all those who could not give the exam on 1st May. So, it will not be on 8th and 10th, but only on 11th.

Let me add that all the information above is from media, and if you are an affected student, please contact AIEEE/CBSE for authentic information. Do not hold me responsible, if any of this information turns out to be incorrect.

Stress due to multiple entrance exams: A Proposed Solution

Yesterday, I had written on this blog why it is difficult have a common entrance exam for all engineering colleges. While discussing the blog with a few colleagues, I realized that the problem or the source of stress is not preparing for multiple exams, but the logistics of taking multiple exams.

A student typically sits for 5-6 exams in the months of April/May. Having 1 or 2 less exams is not going to remove or even reduce preparation effort. If the exams have very different syllabi, then either of them cannot be removed anyway. The two universities want different skills or knowledge to be tested presumably for some genuine reason. For example, the engineering admission test would have a different syllabus from the medical admission test. On the other hand, if two exams have similar syllabi, then preparation for the second exam is a small incremental effort, and removing that exam will not reduce the effort much.

The major problem for the student is not the number of exams, but the associated logistics. Sometimes, two exams are on the same day, or they could be on consecutive days, but the centers are in two different cities. Scheduling of exams is inflexible, because they depend on availability of schools in different cities. Most Schools are only available on Sundays, while some schools are also available on Saturdays. So large exams like JEE and AIEEE are held on Sundays, and exams with smaller number of candidates are held on Saturdays. And as I argued in my previous post, all these exams have to be held within a 5-6 week period in April/May, so the number of distinct dates available is much smaller than the number of exams to be held, and some clashes cannot be avoided.

So the primary problem is that the number of exams to be conducted is more than the number of suitable days available to conduct those exams. This results in scheduling conflicts. This results in having to find ways to reach from one city to the other in time, and find appropriate accommodation and transport options, that too, just before an important exam.

The obvious solution that all administrators have been suggesting is to reduce the number of exams. That will take away scheduling conflicts, and will give some peace to the student before every exam and would be easier on him/her to work out the logistics for each exam. But as I explained yesterday, this is extremely difficult to achieve (and perhaps not even desirable, since it gives more options to the students).

There is another solution to the problem. That is to increase the number of days over which these exams can be held. There are two ways it can be done. If we want to hold these exams only in April and May, then find ways to hold them on working days. Alternately, consider the option of holding them prior to April, perhaps through out the year.

To hold the exams on weekdays, we could have dedicated infrastructure only to hold exams. Suppose we were to create about 2-3 lakh seats just for examinations, spread over the 50 largest cities of the country. Given the number of exams of all sorts that we have (not just the admission tests, but exams for jobs in all PSUs, exams by UPSC and State PSCs), such a facility will be in use on most days of the year, and can be a profitable investment. And when there is an exam like AIEEE, which has more than 2-3 lakh candidates, only these large exams need to be held on Sundays, since they will need schools in addition to this dedicated infrastructure.

Alternately, the exams, particularly the big ones, need not be held only in April and May. They could be held on several days, including dates prior to the board exams. In such a scenario, the students can chose the dates for such exams in a way that there is no scheduling conflict with other exams they are planning to take. Of course, as I argued in my previous blog, holding an exam on multiple days invariably would mean a computer based standardized test. And once we have a large question bank, and the process of such tests has been streamlined, there is no reason why these tests can not be held through out the year.

I believe that increasing the number of days on which admission tests can be held is a much more viable option than to reduce the number of tests to one in each discipline (engineering, medicine, law, and so on).