Search This Blog

Sunday, October 11, 2009

Poor gender ratio in IITs and other top institutes

I thank my good friend Salil Dave to have forwarded the following link on admission stats at MIT. Interestingly, MIT selects ~20% of the girls who apply and only ~8% of the boys - since more than twice the number of males apply than females, the net effect is that the incoming class has about the same number of boys and girls. Assuming that the applicant abilities are not a function of gender, this is a deliberate admissions policy to have a 'balanced' class.

I thought it was pretty obvious that neither the abilities of women are any less than men when it comes to science in 12th class, nor are their performance in engineering disciplines any less when it comes to engineering institutes. In fact, one can go over the results of 12th class board exams in any state in India, and one will find that amongst the top 100 or top 1000, the girls actually perform marginally better than boys. And, when one looks at performance of girls in engineering colleges, it has been reported that there is no difference.

Given all this, the obvious fallout is that if an institute is keen to admit the best possible students, it should try to have a class which is gender balanced. If your admission process does not result in a gender-balanced class, there is something wrong with the process, and it needs to be relooked into.

Less than 10 percent of the students admitted to undergraduate programs at IITs are girls. In view of above, this is extremely serious. There may be a large number of reasons for this serious imbalance. Parents may not want to spend money on coaching of their daughters, when they have no problems to do it for their sons. Parents may not want to send their daughters to far away places like Kota for coaching. (And there is no denying that most people who get selected in JEE today come in after a serious amount of coaching.) May be females can't take the stress of a single exam deciding lifetime of choices. May be there are other reasons as well.

It is important that IITs try to understand those reasons and then try to come up with an admission process which ensures that despite all such reasons, they still get the best possible students, which they are not getting right now.

It is interesting to note that such a problem exists only in the top 50 odd institutions (amongst engineering ones). The reason is not difficult to see. Due to coaching, the relative ranking of boys go up by a few 10s of thousands ranks. If we just look at 12th class performance, then, as we said earlier, in top 1000, there are 500+ girls and 500- boys (because 12th class exams are generally without excessive coaching). But in top 1000 of JEE or AIEEE, there will be only 5 percent girls. This percentage of girls keep increasing as ranks lower and at 50000+ ranks, one starts seeing about 50 percent girls. So all institutions who take the students from top 10,000 ranks would have the worst gender ratio. Institutions who are in the middle and take admission from ranks 10,000 to 50,000 primarily would have a better gender ratio, and institutes who admit students beyond 50,000 rank, would have no gender imbalance.

Any good university in the world will study the profile of students joining it, and if it notices that a certain class of good students are missing, it would study whether it is because they are not attracted to the university or it is because the university's admission process is in some way excluding them. In either case, some steps will be taken to remedy the problem.

The top institutions in India too should relook at their admission strategy and find ways to admit more girl students not because of some equal access issues, but because it will improve the average performance of the class.

Unfortunately, it is easier said than done. Particularly looking at IITs, the JEE has always been a holy cow. You can't criticize it. To argue that JEE has not been admitting the best students all along will cause major ripples amongst the alumni, because it is directly questioning whether they were the best. And no faculty member in IITs, in general, would want to admit that they teach anyone but the best. So there is no incentive to change, and as long as National Commission for Women is silent on this, so much the better.

So, changes in JEE are almost always pushed from the politicians and bureaucrats (and then we complain about lack of autonomy, interference, etc. if you don't fix your house, someone else will). And politicians have so far focused on caste based access. May be they haven't realized the gender-based inequity. But one day they will. I only hope that that day comes sooner than later, because a mechanism to improve gender equity will only improve the quality of students in IITs and other top institutions.

Now, the obvious question is what can be done. There are multiple ways to handle this problem. One is, of course, to make major changes in JEE so that coaching has less possibility of enhancing success in JEE. That is a very long-term issue, and I have written about it in my earlier blogs as well, and I really don't think a major well thought of change to JEE is possible.

The other possibilities are based on assumption that JEE will remain by and large what it is. One possibility is to give extra credits (marks) to women to compensate for the difficulties they in general would have faced in JEE preparation, and through this means increase the number of girls in the merit list. This is a standard way of providing easier access to disadvantaged groups in several countries. Even in India, there are instances of attracting sports-persons to a university by giving him/her credits (add something to 12th class marks before preparing the admission merit list). A prominent central university was giving preferential treatment to people from disadvantaged backgrounds in this fashion before they were forced to apply reservations.

And, of course, the option of reservation is always there, something that we all in India understand extremely well. What we can do is to look at the current ratio (say 10 percent), and reserve a slightly higher number of seats for women (say 15 percent). Let us do this for a couple of years, and study the performance of men versus women in these batches. If the performance of women is similar to men, then we increase the reservation to 20 percent, and so on.

Thursday, August 27, 2009

IIT JEE should encourage school education

Last week, there was a meeting of Joint Admission Board (JAB) of IITs, which includes Directors and JEE Chairpersons of all IITs. It was held at IIT Madras. A newspaper report mentioned that the JAB decided to setup a committee to decide the cutoff marks in the 12th class examination for JEE-2011. The committee will look into the issue of normalization of marks of various boards, amongst other things.

I hope the report is correct. It is high time that IITs consider school performance in a serious way for admission to IITs. All across the country, the emergence of admission tests for admission to higher education has ensured that students do not take school education seriously. When the top half of the class does not take school seriously, it discourages teachers, and learning of even the bottom half is seriously hampered. Also, we find that it has resulted in lack of many skills which should normally be developed at the school level, including the most important one of communication and language skills.

There have also been some studies done which showed that the performance of students in IITs has a stronger correlation with the 12th class marks, and a weaker correlation with the JEE rank. If that be the case, it is clear that insisting on higher 12th class marks for admission to IITs (and indeed to all higher education institutions) will result in better student intake than the current system. Considering the quality of examination and in particular weaknesses in proper invigilation during the exam, it is too risky for IITs and other higher education institutions to only rely on 12th class marks for admission, and at least in the medium term, exams like JEE and AIEEE are here to stay.

To have a very high cutoff in 12th class marks is not easy to implement due to differences in various boards across the country who conduct these exams. And indeed, one of the tasks that this committee has been assigned is to find a way to normalize the 12th class marks across all boards.

One point that JAB seems to have missed out is that there are two distinct goals for looking at 12th class marks. One goal is to only select those who do extremely well in 12th class. If you follow this goal, then one would suggest a cutoff for CBSE as 85% or 90% and then one will have to do normalization for other boards etc., because there may be boards in which getting so many marks may not be possible.

While this goal is laudable and must be pursued in the medium term, there can be another goal in the short term. This goal can be to ensure that the students preparing for JEE also take 12th class seriously. This goal can be met with a much lower cutoff for 12th class marks. A cutoff, which disqualifies about 1 percent of the selected candidates. If you look at the merit list of 8000 students of this year, find out the 12th class marks for all of them, and decide the cutoff which will be missed by about 80 students, then it is likely to be slightly less than 70%. (We, at LNMIIT, looked at the 12th class marks of all admitted students this year - a small sample indeed compared to IIT system as a whole - and found out that the lowest 1 percent were below 68%. Since we take students after IITs (AIEEE rank upto 18000, and EML rank upto 9000), if the same exercise is done by IITs, the chances are that the 1 percentile will be at 69% or so.

Assuming these numbers to be true, JEE could keep the 12th class cutoff at 70%, and disqualify about 1 percent or may be slightly higher number of students. If you disqualify about 100 students on this basis, it will shock everyone and people will start taking 12th class more seriously.

Another thing we noticed was that almost all candidates who were below 70% marks in 12th class were from CBSE system in our case. Again, if out of 100 students disqualified, most were from CBSE and only a handful from, say, UP Board, students from UP board will not go to court saying that they are being discriminated against and that they should get the benefit of normalization.(Like no one has challenged the 60 percent limit so far.) And even if they challenge, courts are not likely to be sympathetic to them if only 1-2 percent students are being disqualified.

Since JEE has often been accused of destroying the school education system, this simple step of increasing the cutoff from 60 to 70 percentile will cause a lot of serious students to be back in class, and JEE would have done a great service to the school education system in the country.

I would, therefore, like to suggest that while a committee can try to find a medium term solution to select only those for IITs who do extremely well in 12th class, but in the immediate future (JEE 2010), one may consider increasing the cutoff to 70% (or at least 65%) with appropriate concessions to the reserved category students.

Monday, August 24, 2009

How to Choose an Engineering College

Summer is the time for a lot of stress amongst school leaving students and their parents. They are terribly confused regarding the college they should join for higher studies. They don't know enough about various options, and that adds to the stress. Media tries to help by publishing rankings of colleges. However, there are several problems with those rankings. Most rankings are only for the top few institutes. Also, rankings are based mostly on perceptions, and on unverifiable data supplied by the Institutions themselves. Many times, they use parameters unrelated to quality (like size). Further, they rank the entire institution, and not a specific program.

Students (and parents) typically look at last year's closing ranks, which means that they miss out on new and exciting places. Or they look at placement data narrowly (even if one student got a Rs. 10 lakh job, the institute must be good). And another bias is that government colleges are better than private colleges. Some NITs are excellent, and a large number of private institutions offer poor quality education. But there are many private institutions with much better faculty profile than NITs.

So, what should students do. The solution is simple. They should do their own research and come up with their own rankings.

In my opinion, the quality of education is primarily driven by quality of faculty. So, if there is only one thing that you want to check in a college website, it is the quality of faculty. You should look for not just the number of faculty members, but also their qualifications, whether they are PhD or not, where did they study, are they active researchers and publishing.

The second most important criteria is the academic freedom that the Institution has. In IITs, NITs, IIITs, and deemed universities, there is complete academic freedom to change curriculum, provide flexibility to students, have their own exams, etc. To faculty, it is very demoralizing that someone else will test what they have taught to their students, and generally speaking (with some exceptions), quality of teaching is better in institutions who have their own exams.

Of course, having freedom does not mean that it will always be used effectively. So, one has to look for more information. Third most important criteria is the curriculum. Do they change it frequently enough. Judging the curriculum will be difficult for a student of, so seek help from a person in that discipline. For example, in Computer Science, if they teach you multiple programming languages, they are more interested in spoon feeding than teaching concepts. In general, and this is counter-intuitive, better departments will have fewer courses in the curriculum. Weaker departments would like to spoon feed you a lot of information in the hope that you can say something when it comes to placement interview. Also, more flexible system would have more electives, and you would be studying topics of your interest. Of course, make sure that those electives are actually offered, and are not just on paper.

Other parameters are less important in my opinion, but do matter. These include infrastructure. How many labs, how many PCs, what is their library holdings, how much is the Internet bandwidth, sports facilities, etc. Another parameter is what happens to the students after they graduate. This includes campus placement, but more importantly, how many of them go for higher education. Remember, the most important skill that college education is supposed to impart is to learn how to learn. If a college is successfully imparting that skill, then it must result in substantial number of students to think about higher studies. And finally, brand equity does matter. Have they got any ranking in any survey. What is the closing rank of last year. Have they been accredited.

Now give weights to these parameters, and start surfing the web to collect relevant information. This is hard work, but remember, you are about to take the most critical decision of your life. Don't depend on others advice alone. Have your own research to back up your decision.