Vacant Seats in IITs

Every year, the issue of vacant seats in IITs is raised by potential students, their parents, and politicians. It is argued that seats in premium institutes require a lot of investments and they have a huge demand. And hence something that has taken a lot of public money and has a lot of public interest, can not be allowed to go waste.

For a long time, the argument of IITs used to be that we announce a larger number of seats than what we could really afford to admit. So if the public investment was for 100 seats, we are announcing 110 seats and hence if 10 seats remained vacant, that is really not a waste. This explanation is no longer acceptable to public or to politicians, and all IITs are under tremendous pressure to fill that last seat. And hence we have gone from a single round of seat allocation to multiple rounds of seat allocation. Now, increasing the number of rounds will only have an incremental effect on the vacancies. There is no guarantee that having more rounds will fill all seats. On the other hand, the public and politicians do understand this, and therefore, very small number of vacancies (say 1-2 percent) do not make headline news. It is only when some IITs have 5+ percent vacancies that the problem is highlighted.

Having a larger number of rounds, unfortunately, only forces the problem on to lesser endowed institutions. When you try to fill that last seat in an IIT, the student comes from an NIT or BITS or some other good institution. And by offering him/her the admission in IIT, you are only creating that vacancy in another institute. Yes, one can argue that a seat in IIT is more valuable than a seat in an NIT or BITS, but only marginally so, and if keep doing more and more rounds, we are essentially forcing the rest of the country to either start their semester after IITs do, or to do admissions after the semester has started. This is certainly not fair. We must come up with a solution where the admissions stabilise several days in advance of the beginning of the semester for everyone.

To solve the problem, understanding the reasons for the problem is very important. And if we look at the admission process in India, we find the following issues stand out:

• The entire admission process for all colleges and universities have to happen in a few weeks in June and July. This restricts the number of rounds, doing any interesting way of admitting students, and students jumping from one institute to another every few days. We do this because for us it is a non-negotiable red line that college admissions must be based on 12th class capabilities (whether 12th class performance, or performance in another test which is based on 12th class syllabus). No where in the world would you see that the admissions to college are based on 12th class. Typically, the admission process starts several months in advance, almost one year in advance in some cases.
• For us, the number of seats is sacrosanct. We can admit so many students and no more. The public at large does not understand even basic statistics (most international surveys show Indians as amongst the poorest in skills typically taught at schools). All over the world, the universities offer admissions based on historic data on acceptance rates. So you offer more admissions than what your target student number is, and with some people not accepting your admission offer, you will hopefully be close to the number you wanted. It will not be exact, but that is acceptable.
• It is also sacrosanct for us that a student be able to apply for admission in thousands of colleges simultaneously. Most parts of the world would charge an application fee for every program, not to mention for every university. A student should be able to have a reasonable guess based on his/her performance where they can get admission, and they should then apply to a few places only. Of course, in India, the problem is that the admissions is based on a lottery (euphemistically called JEE), and the result of that lottery comes too late for the student to do all the research about admissions. If all the parameters of admission are known to the student at least a few months in advance then only the student can figure out where to apply.
• The government order which states that a university can only retain Rs. 1000 if the student withdraws even at the last minute has created a huge mess in the country when it comes to admissions. Most of the top students come from middle class or higher background (since it is so important to have access to coaching, quality books, a decent environment at home to study, etc.). The parents of these students can easily afford to block a lakh of rupees (or two) by accepting admissions in multiple places, knowing that they can get almost the entire money back even at the last minute. This has resulted in a large number of vacancies that an institute comes to realize only on the first day of classes, and that is pretty late for admissions.

Given these facts, how do we ensure that IITs can have admissions of more students (closer to their so-called sanctioned strength) and are under less social and political pressure to do things which can only hurt academics (like admitting students after the classes have started).

It is obvious to me that starting the admission process early (say, after 11th class) is not going to happen any time soon. And government is unlikely to permit institutes to have a graded policy on refund (Rs. 1000 if you withdraw 10 days before classes start, Rs. 5000 if you withdraw 5 days before the classes withdraw and Rs. 10,000 is you withdraw one day before the classes start). So the play is really in terms of extra admissions.

Ideally, one should be able to look at the statistics of the last five years, figure out what percent of students have not joined, and add that number to the number of seats, clearly articulating what the sanctioned strength is and what is the extra seats for a particular batch to handle attrition. The problem in this scheme is that the numbers are really statistically irrelevant if we consider each program and each category independently. If we see Computer Science, General, at IITK, may be we will see just one vacancy in the last 5 years. Should that translate into an extra seat next year? And can we really articulate this extra seat in CS in the way that public will understand and won't put pressure on us to increase that seat on a permanent basis.

It is also clear that if any system is seem to disregard reservation in slightest form, it would most probably be struck down by courts as illegal. So vacancies have to be looked at category-wise.

So here is my proposal:

Consider the number of seats in each category where the student who has been offered admission has not joined. Next year, that many extra admissions in that category will be made by the institute. The hope is that the number of students who don't join roughly remain the same every year. (To take care of situation where suddenly in one particular year, too many students have not joined, which may increase the batchsize next year substantially, one can do some fine tuning to this. For example, we could look at the lower number of vacancies in the past 2 years.) Note that we are not taking into account the attrition within these extra students. So we are already being conservative. So the chances of a batch being significantly larger than the sanctioned strength is very unlikely in this proposal.

The important question is: Which program do they get admitted to? And the answer is that the program will not be specified at the time of admission. Now, this may sound terrible that one is getting admission without knowing the program, but in reality, it is not as bad. First of all, it is extremely common in good universities for the students to get admission without deciding their major. Second, in IITs, a whole lot of people take admission just because of the brand name and not because they are excited about the discipline that they will study. Third, everyone knows what are the least popular disciplines in each Institute. Those are the disciplines one is almost sure to get. So one does have a fairly good idea of what is the "worst" case scenario.

If we assume that the number of extra admissions would be equal to the number of students who would not join the institute, then assigning the program is simple. We could either do an internal sliding after the last date of joining, based on their counseling choices. This would really be trivial. Alternately, and in my opinion, a better method would be to wait till the end of the first year, and decide their programs based on the branch change rules of individual IITs.

What if the number of extra admissions is more than the number of students who did not join the institute. So we have a few extra students in this batch. Note that this is likely to be very small number. We could just increase the number of seats in each program proportionately (actually, it will come to an increase of just 1 seat in a few disciplines in most IITs, which is not a big deal).

Of course, in the long term, finding ways to offer early admissions, and having a penalty for late decision on withdrawal would make life simple for most admission seekers.