Search This Blog

Thursday, April 20, 2017

Merit-based admissions

People who do well in JEE Advanced will mostly tell you that merit is a one-dimensional beast, never mind that it consists of performance in three heterogeneous disciplines, mapped into a single number by the simple addition of three numbers without checking if these numbers are comparable or represent something which can be simply added. It is further believed that admission to higher education is a reward for performing well in an exam under the conditions of extreme pressures. And hence the only justified admission policy is to admit students in the decreasing order of merit or performance in that exam (modulo any laws passed by the parliament, but of course, not before they have pointed out that merit has been violated by those laws).

Should a university consider admission as a reward for performance. Or should it be something else.

Well, it has to be something else. In fact, no university will couch its admission policy in terms of rewarding past performance. University goals are varied: Who is likely to do well in our programs? or Who is likely to do well even beyond our program and be successful and enhance our reputation?

If university goals are somewhat in line with what I have mentioned above, then it is obvious that the admission through a single test is not ideal. No one can perhaps argue that performing in a particular test can both be an indicator of future performance in Computer Science, as well as an indicator of future performance in Economics. No one can argue that India has the best admission process since this is the only major country in the world that does not take language abilities into account for admission. No one can argue that the ability to answer MCQs is a strong indicator of the ability to write long notes necessary in any studies.

The reason for following a single exam is that it provides an objective, transparent mechanism while simultaneously measuring one variant of irrelevant merit. In other words, the most important outcome of JEE based admission has been that we have kept out pressure from the rich and influential as well as won most court cases. And by no means, this is a small achievement. In India, one may genuinely consider keeping pressure out and winning court cases as more important property of admission process than getting students aligned with the university goals. However, over a period of time, people doing well in JEE have started believing that the prime purpose of JEE is to evaluate merit, and that there can only be one definition of merit, which is the JEE rank.

Besides the obvious inadequacy of a single number predicting success of all programs in all disciplines in all institutes, there are other problems as well.

If a university asks the question who is likely to do well in our programs, and let us assume (despite absence of any data) that performance in PCM has a strong positive correlation with performance in Mechanical Engineering. What if we were compare the performance of people who have 80% in JEE after intense coaching, whose parents were rich and provided a good study environment whether at home or in a hostel, with the performance of people who have 75% in JEE without any coaching, and who come from modest backgrounds, and are the first generation learners. When both these groups compete, who is likely to do better. Most universities would have data to show that latter group performs better, and hence they would actually create admission policies to offer more admissions to the latter group.

When we point this out, an immediate reaction is, but how do you evaluate deprivation, and what has been the impact of that deprivation, and how much extra "credit" should be awarded to compensate for that deprivation so that we can compare the merit of the two. Now, in this question, there is an inherent belief that JEE score is a strong indicator of future performance of every sort. Gold standard of transparency and honesty (?) has been converted to Gold standard of merit. Do you ever question why people with same total in JEE score are ranked differently based on marks in some subject or the other. Does that imply different merit. Isn't it true that luck plays a part. If you had a headache that day, your merit may be way down. That any exam performance is only an indicator within a significant band is conveniently forgotten. But when it comes to any affirmative action, we want exact data, why 5% and not 4%. Why not ask data about JEE also.

Let us do another thought experiment. Again, let us assume that PCM performance is an indicator of future performance in all programs. Consider two possible admission decisions. One, take the top 100. Two, take the top 90. Study whether there is sufficient diversity in the class. If certain backgrounds are missing, let us fill the last 10 seats through them. Now, there is a belief (and educationists may even have data, I am not one) that diversity is good for education. That diverse inputs will cause more innovative projects. That having diversity in your peer group in college will prepare you for a much more diverse workplace that you are likely to face. So both performance within the university and success beyond the university is likely to be enhanced if the class is more diverse. Because of these reasons, universities encourage student exchange programs besides having diversified student admissions to begin with.

Let us for the sake of argument assume that there is data to show that diversity helps. What should a university do? Should you admit 100 students whose average performance will be 60, or should you admit 90 students whose average performance will be 70 (helped by diversity) and 10 students whose performance would perhaps be only 50. There is certainly a plausible argument for not admitting strictly through merit (if at all merit can be determined by a single number). But ask someone who was 91st in our imaginary list. Will he agree. If you assume that admission is a reward for performance in a particular test, you will not like what the university is doing.

Universities may have other goals. For example, should we admit those who are the smartest and will benefit very little from university education, or should we admit those who are likely to gain a lot more from university education. So on a scale from 1 to 10, should we admit students at 9 who can be helped to reach 9.5, or should we admit students at 7 and take them to 8.5. In other words, should "merit" be the only criteria for admission or should we also look at what will benefit the society most. This question is particularly important when the society is funding the cost of that education. Again, I know what the "merit" crowd will say, but I will only suggest that most universities around the world will admit both kinds of students in some proportion. (We also are mandated by law to do affirmative action, which means admitting students at lower "merit" but they are likely to benefit the most from our education.)

A lot of universities have a stated goal of helping the society that nurtures them. One way to do this is to prepare entrepreneurs who will create jobs in that society. Depending on the geographical location (and particularly so, if the location does not attract a lot of investments), universities may prefer to have some weight to the nearness criteria, since students belonging to that society are more likely to stay there beyond graduation. A lot of universities will, therefore, admit more students from within the state even when that magic number representing merit is lower for these students.

In summary, the insistence on admitting based on a single number is misplaced. As long as the single number is considered a proxy for transparent/honest process, the insistence has a logic, but when this number starts getting treated as a proxy for merit, there is a problem. On the other hand, it is necessary for public funded universities to not only admit the best, but also give an appearance of admitting the best. As long as public universities like IITs have a transparent/honest process, they should have flexibility in both defining the merit (as combination of multiple parameters, perhaps), as well as bypassing merit to achieve the multi-objectives of  the university.

Sunday, April 16, 2017

Improving the gender ratio in IITs

Recently, there have been media reports on a decision taken by Joint Admissions Board (JAB) of IITs where by IITs would create some extra seats for women so as to improve the gender ratio in the under-graduate programs. Two of the reports are here and here. Earlier, JAB had asked a committee headed by Director of IIT Mandi, Prof. Timothy Gonsalves, to look into the ways of improving gender balance in IITs. This decision is apparently one of the recommendations of the committee.

Though the details are sketchy, it seems that there is a goal of having at least 20% women in the under-graduate class in stages. For 2018 admissions, the goal has been set as 14% which will increase by 1% every year to reach 20% in 7 years. In recent years, number of girls admitted to IITs number around 9%. To achieve a 14% ratio in 2018, they will have to increase the number of seats by 6%, and all these 6% will be filled exclusively by women candidates.

A little over a year ago, I had suggested in this blog that we must do some research into why women are not getting selected in larger numbers despite their performing extremely well in Physics, Chemistry and Mathematics in class 12th and also they perform well after they get selected in engineering. My own contention is that there is an inherent bias in the society which restricts the coaching options for women. The fraction of women in Kota coaching classes is very small, for example. Even in larger cities, the fraction of women in JEE coaching is less than the fraction of women in science sections in schools. Sometimes, it could be lack of willingness to pay high amounts for a girl child. Or it could be the concern for their safety.

I am not sure what data the committee looked at, but apparently they did find that part of the reason for lower number of women is societal bias, and hence that bias needed to be compensated by some mechanism. And what other mechanism do we know of in this country but reservation.

Is reservation the best mechanism to achieve gender balance?

The obvious shortcoming of reservation or a quota system is that its benefits are not directed towards the disadvantaged class but a larger class. The additional seats may not all go to those women whose parents have refused to send them to outstation coaching, or even to a more expensive coaching within the city. At least some of the seats will be taken up by women who actually go for expensive or outstation coaching.

But note that this is the problem of reservation based systems in general. Aren't OBC reservations benefiting students whose both parents are well educated and can hardly be called "Educationally Backward." But in India, we always argue that if we don't use simple proxies for disadvantages faced by people, the whole system will be gamed by rich and influential. In case of simple proxies (like caste for socially and educationally backwardness), some non-deserving people may sneak in, but it helps those who need such help. Something similar is likely to happen with women reservation as well. A few non-deserving women will get admission, but overall, it will help compensate the societal bias to some extent.

I think something more interesting may happen here. Once the parents know that getting admission to IITs is somewhat easier for women, they may actually be more willing to get them coached. Currently, one of the reasons for not investing in their coaching is that the chances of success are so low, and the expected return on investment is consequently low. They are willing to invest in sons' coaching even with that lower expected return, but will invest in daughters' coaching if the expected return is higher. As a result, a greater percent of women may succeed in the admissions process on their own. My gut feeling is that the percent of women in the normal process will keep increasing and they will need only 5-6 % supernumerary seats even as the goal improves from 14% to 16% and all the way to 20%. And because of this hope, I am positive about the reservation.

Are there other methods that they could have used to increase women admission?

Absolutely. JNU has had the scheme of deprivation points in their admission process. Under this scheme, they would add a few points to the other pieces of evaluation based on some criteria of background of the candidates. One of the criteria is gender, and a small benefit accrues to female candidates.

IITs could do something similar. They could increase the marks of every women candidate by some small number in a way that in the top 10,000 ranks, there are exactly 1400 women. And now women have ranks based on this new marks. One could a priori decide what is the maximum number of marks to be added, and if to ensure that there will be 1400 women out of 10,000 ranks require a higher number of marks to be added, then we will still stick to the maximum marks. (And, of course, we would know how many of them were in top 10,000 before these extra marks, and how many have been added, and create that many more seats to satisfy the current policy of not reducing the number of seats for categories not part of new reservation.)

Of course, it is easy to expand this mechanism to implement all sorts of reservations, and we will have data on exactly how much difference there is between various categories. So we could increase the marks of all SC students in a way that there are exactly 1500 of them in the top 10,000 (subject to the maximum number of marks, note that even now there is a limit on how much lower we will go in the merit list).

This mechanism is very useful when you have very small reservations, for example, in case of Physically challenged students. A PH-ST student is competing for a 0.2% quota (3% of 7.5%), which means that in a large number of programs, there will be 0 reserved seats in any given year. But if you add enough marks to their score that they represent 3% in 10,000, they will be able to seek admission to any seat that they deserve at their performance level.

This mechanism will be useful if we want to compensate for any other bias or discrimination or deprivation that candidates have faced.

Of course, what we are arguing now is that a 15% reservation over 10,000 seats means that the reservation is overall and not in each program. This would mean that they may get slightly less than 15% in some programs and slightly more than 15% in some programs. We only need to make sure that the distribution of marks are such that it won't lead to very high or very low presence in the popular programs (which I suspect will not happen). I don't know how courts will look at it, but it is worth trying.

Can we improve gender balance without any affirmative action?

That would be the least controversial and best method, in general. But that would require a lot of research, and we normally want to solve the problem without doing research. For example, if the hypothesis that lower women representation is due to societal bias and consequent lack of investment in their coaching turns out to have some merit, then perhaps we need to have the entrance exam (or at least some components of it) which are not impacted by such high pressure coaching. Small amount of coaching would be enough. One way to do that is to have speed tests, I am told, instead of very difficult to remember tricks. On top of that, something that government has already asked IITs to do, we can get study material prepared by IITs. And, of course, we are also seeing development of apps where by a candidate can practice for speed tests and get feedback sitting at home, all at a very low cost.

Of course, if the research shows some other reasons behind gender imbalance, we will need to tackle that properly.

Will this lead to more demands of diversifying student population?

Tamil Nadu has a little over 5% population of India, but it does not send 5% students to IITs. Muslims have about 15% population in India, but the fraction of Muslim students in IITs is much smaller. Wouldn't there be demands for increasing their representation.

Of course, there will be. But note one thing. The system proposed is saying that if 50% population does not have even 20% representation then there is something wrong somewhere, and we need to do something about it. So, the goal is not to ensure representation aligned to population fraction. Also, this is the population which seems to be doing much better in pretty much every exam in the country, except JEE advanced. If there are other groups which meet these criteria, that too can be studied.


It is a difficult decision. But one that I think could lead to attracting better talent by IITs. I am hoping that the "quota" part will remain very small and will eventually go away, and that IITs will implement other ways to attract talent, including changes to JEE.

Added on April 22, 2017:

Prof. Timothy Gonsalves, the Chairman of the committee on improving gender imbalance in IITs has made a posting on his FaceBook wall giving a summary of what went behind the report. I strongly encourage everyone to read that.
 An excerpt from the same:
Would admitting girls with slightly lower ranks compromise on quality at IIT?
 A study in IIT-Delhi looked at the final CGPA of male vs. female students. Over a period of 13 years (2003-2015), females outperformed males consistently by an average of 1 grade point, despite having lower JEE ranks! This amazing finding gels with our experience as teachers in other IITs also. It is an indication that this cohort of young women is extraordinarily talented and highly trained despite the disadvantages of growing up as girls in India.
I am told that the difference in grade between girls and boys of similar JEE ranks at IIT Delhi is a whopping 1.5. It proves beyond a shadow of doubt that giving some push to women whether through quota or bonus marks, or whatever, will actually admit better students to IITs.

Tuesday, April 11, 2017

International Student Exchange Programs

In the last 2 years that I have been in Delhi, I have attended many a meetings with folks from different countries trying to figure out how the institutions in India can engage with institutions in their respective countries. There are two models which seem to working well.

One is that of twinning wherein a student in India joins an Indian institution, does some course work for two years, and then seek admission in a partner university abroad, which recognizes the credits completed here. The student spends two more years at the foreign institute, and get a degree from there. This is a win-win situation for everyone involved (commercially, at least). The student wants a degree from abroad, and gets it at a cheaper price than spending all 4 years abroad. The parents are happy that they didn't have to send their ward abroad when s/he was too young. The foreign university is happy that they are getting at least two years' tuition from a foreign student. The local university is happy because they can actually charge a bit more for the two years than what they may charge for their 4-year program.

The other is that of research collaboration. Two researchers meet somewhere, may be in a conference, and they decide to collaborate. Much of the interaction can happen over Internet, and a few visits can be supported by their respective projects. On top of that, there are government to government schemes under which they can apply for projects jointly, and while getting big moneys in international projects is difficult, collaborations can certainly happen.

But what about things beyond this. In terms of teaching programs, can we have student and faculty exchange programs where our students can spend a semester or two in the foreign location, and their students can spend a semester or two on our campuses. The same could be done with faculty. Can our PhD students work in their labs and their PhD students work in our labs. This is where one does not find any solution. For our students to go to North America or Europe is very expensive - travel, lodging and boarding, as well as tuition. For their students to come to India, well they don't think of it as an option. If they were coming to our campus, it would be easy to argue for tuition waivers. They don't pay tuition here, and our students don't pay tuition there. But we don't see a 2-way exchange.

So in every such meetings, there will be complaints about lack of two-way exchanges.

In a recent meeting, I asked the representatives of various universities whether they have an MoU with a university from Bangladesh or Sri Lanka or Mayanmar. The answer was on expected lines, none of them had an MoU with any university in these countries. The reason was supposedly obvious. Their universities were not great (though I can tell you that most universities represented in that room would not be better than good universities in these countries). Now, your student does not want to go to a university which is roughly similar to that of your quality, and you are wondering why an American or a European student does not want to spend a semester at your campus.

I recall that long time ago, when we were discussing relationships with top universities in Senate of IIT Kanpur, one professor had said that only after IIT Kanpur has had a good working relationship with HBTI, MNNIT, and other decent colleges of UP, would it realize how to have a relationship with MITs of the world.

A gentleman from US asked a question, "What is the goal for student exchange?" A pin drop silence. Frankly, the goal is only to brag about it, or enable our students to go to US/Europe with fee waivers. If the goal was what normally universities say, greater cultural diversification in the class, then we could achieve that by having more students from Africa, Central Asia, South East Asia, etc. Our focus on these regions is fairly limited right now.

I believe that the right thing to do by our universities is to attract exchange students from comparable or less developed countries than us. This will help us in multiple ways. It is more likely to succeed than to try wooing students from US/Europe. Thus the first goal of cultural diversity in the class (and a side goal of doing better in ranking) will be achieved. This will make us learn what are the challenges that foreign students face, including but not limited to getting a visa, police registeration, finding accommodation in nearby localities, and so on. It would be easier for our students to spend a semester in these places and gain an exposure of different cultural setting. Once we have all this knowledge and experience, we will be able to come up with better ideas to expand the scope of exchanges to richer countries. Right now, there are many attractions that we can market to students of richer countries, but we really don't know how to leverage them, how to prepare a program suited for different classes of students.

And just like we are offering a twinning program to our students, we could get into agreements with universities in less developed regions that they will send us their students after two years of training and we will train them for 2 more years and give our degrees. Why should we always be importer of education service. We should try to become an exporter of education service.

At a national level, bringing students from such countries also projects our soft power, create goodwill, create ambassadors for life.

Sunday, April 9, 2017

NIRF 2017

So, the second round of rankings are out. And every single problem that I could have imagined with a ranking system, I find it in these rankings. I had written two blog articles when the NIRF framework was announced in September 2015. They are here and here.

Let me summarize my most important objection to NIRF. While a lot of stake holders use ranking for taking important decisions, and therefore, it is important for the universities and colleges to participate in ranking games, there are several inherent problems with ranking games. The biggest problem is that a single linear rank gives a hugely distorted view of the academic landscape. And even to get that, the data is simple not available in reliable fashion. So a lot of fudging takes place.

Having lots of rankings in private sector is still ok, since most stakeholders will take each of them with a pinch of salt, and hopefully do a bit more homework about the universities they are interested in. While stakeholders may not understand the exact nature of problems with rankings, they are skeptical of any private ranking whose aim typically is to make money out of this information. Also, having multiple rankings mean that to some extent stakeholders realize that either rankings are inconsistent or there are indeed multiple ways to look at universities. But when government creates a ranking, stake holders may use it as the primary information and that would be disastrous.

The second major problem with a government ranking is that it leads to inconsistency in public policy. On one hand, having more out of state students is considered positive for educational outcomes, and on the other hand, we could have 50% reservation for in-state students in the institutes managed by the same government. Opposite for women participation. It is said that having a 50-50 population is good for educational outcomes, but nothing will be done to attract greater number of women students to IITs.

And, third major problem with government ranking is that the government is further reducing its already low credibility. There is no way one can verify this huge amount of data. So there will be huge errors in data leading to some strange ranking. Who should be held responsible if a potential student trusts these rankings, get admission, and then find the truth. What if this happens at a government institute. Will some heads roll, or will we just say, "you are stupid to trust government, and don't deserve anything better."

Last year, when NIRF ranking came out, there were lots of questions, and the answers were rarely provided, and really, the only answer was, "We did not have enough time. So some things may have been overlooked. We will do a better job next year." I personally was very upset at the lack of transparency. We were told that the data consistency would be maintained by way of complete transparency. We will all know what all has been submitted by various institutions and that would keep a check on each institution from lying. When I approached some of the folks in NIRF, I was told that due to very short period within which they had to come up with the ranking, they had taken some ad hoc decisions which may be difficult to explain to public, and hence only that year, the transparency is not there, and the 2017 ranking would be completely transparent.

Well, I don't see the submissions of various institutions on NIRF website this year too. I am told that this year, it was compulsory for all participating institutes to keep a copy of their NIRF submissions on their own websites, and if any institute does not keep that information, they will be removed from the ranking. I just checked several universities, and there is at least one in the top 10 who does not have NIRF data on their website (at least I search on the main page, used their search and used google search with "NIRF " as the quey. One other university has only partial information on its website and some crucial pieces of information missing. Others too may remove it in due course. To be transparent, the easiest thing would have been to put up all the submissions on NIRF website (may be they could have asked the institute to submit sensitive data such as placement details separately, and the rest could have been published by them). I don't know why they couldn't do this.

Also, transparency is not just about making submitted data public. It is also how NIRF has interpreted it. How that information has been converted to marks. Yes, they do give score in the five categories. This is appreciated, but they should give out information in each of the part of major factors. And also tell us various parameters that they have used in ranking.

In the small bit of research that I did, reading several submitted reports, this is what I find:

One of the top 5 business school is claiming that in the past 3 years, 100% of their graduates have got a job through campus placement. Frankly, it is unbelievable. Not a single student opting out of placement to start a company, for example. I see this happening in all IITs, NITs, etc., but not at the top business school of the country. Not even one graduate deciding that may be PhD is a good idea. Not even one graduate going back to the company where s/he came from 2 years ago, and not participating in the campus placement. (I thought even Government of India sponsors a few candidates for MBA at such places, and they are expected to join back, and not seek campus placement.) Also, every faculty member, except one has an experience of 5 years or more. Have not not hired any faculty member in the last 5 years, or everyone hired in the last 5 years has left them, or they only hire faculty with 5 years of experience. All this could be true. But a bit unbelievable. And at least as far as the student placement is concerned, if the data is correct, then it is quite sad.

In one university, the submitted data for consultancy amount is X, and the summary sheet for that university put out by NIRF shows more than 10X. I can understand 10X. Someone typing the information could have made a mistake in placing decimal. But having an arbit number which is unbelievably large is rather strange.

In one university, I saw the number of faculty members in NIRF data sheet to be unexpectedly large. I went to the university website. The NIRF submitted data for this is not on their website. Then I went to the website of each department and counted the faculty members shown there. The number is less than half.

The information on papers has been outsourced by them. I am told that in one university, the numbers shown in the datasheet is very less compared to what this university is claiming. The university folks tell me that NIRF sent them an email several months ago telling them the number of papers that they have found. The university wrote back giving its numbers, which was much higher. Now, if both of them are looking at the same database, the only reason for this difference could be that the search queries are different. In particular, faculty members use different names of the university (like IIT Kanpur, I.I.T.K., Indian Institute of Technology Kanpur and so on), and it is possible that they have searched for some names and not other names. It would have been absolutely trivial for them to share their search query, or ask for our search query. But no communication from them, and at the end when the rankings are out, they have simply used their own data. To me, this is height of callousness and incompetence. And we are going to use such data to take important decisions on funding and autonomy.

I see in the data sheets for engineering colleges that they asked for median salary and from management schools, average salary. Why this difference? I thought median salary captured the performance of placement much better, and must be for everyone. But anyway, I checked the numbers. While many institutes have given numbers which are believable as median salaries, there are many where the numbers are simply not believable. My guess is that they have given average salaries rather than median salaries. (Averages are invariably very high compared to median for most colleges.) In fact, about two institutes, I am sure this has happened. I guess in some places this may have happened inadvertently, but in some places it may be a deliberate error.

Then there is this data about capital expenses (not counting building construction). Will this number vary from year to year very drastically. What if it is shown as a small fraction of last year's expense. Shouldn't it raise a suspicion that perhaps last year, they had some construction going on and this year, there is none. Shouldn't they seek some verification of data at least in cases of suspicion.

And note that I am only talking about universities ranked in the top few in some category or the other. I understand that it is not easy to verify data for all colleges and universities, but they can always have a 2-stage or 3-stage process. That is, get data from all colleges which is not verified. The only "stick" is that all this data will be posted on the NIRF website, and if any questions are raised, they will be investigated and if a serious error is found, the information will be given out to press and they will be barred from ranking for some time. With this, you finalize top 120 or so univs in each category, and for these univs, some level of checking can be done. May be some proofs can be asked for. May be someone can check for information on the univ website, and so on. If any glaring errors are found, they are out of ranking. This way, there will be better trust in top 100. And finally, those whom you are going to declare in top 20-25, there should be yet another level of data verification. May be an agency can be hired to actually visit the university and verify everything. So the top 20 or so ranks would be based on very high quality data.

(Of course, the problem of a linear ranking not reflective of all diverse strengths of universities will always remain.)

Notice that I have not discussed the parameters per se, only the poor quality of data that they have against those parameters.

The parameters are also a problem. They are strongly in favor of bigger institutes. I wonder why. I can see that a bigger class to some extent leads to peer-to-peer learning. But beyond a point it does not help. And then what helps is having a variety of disciplines, a variety of courses available on the campus. But just having larger number of students does not guarantee that variety. There are many other serious problems with the parameters, but may be that is for another blog. But I wonder if they did some research to establish correlation between those factors and better teaching/learning or better research. I also wonder if they did any sensitivity analysis with their parameters. (What if I change the parameters slightly, does it result in major changes in the top order.)

I did not write last year after NIRF ranking were published since I wanted to look at what they will do after they have had enough time to do things right. But unfortunately, they have not utilized the time effectively.

Of course, the good thing is that most students/parents who have approached me for admission related queries, do not seem to bother about NIRF ranking. I hope it stays that way.

Added on 9th April:

More interesting stuff.

Jawaharlal Nehru Centre for Advanced Scientific Research which has no teaching program and is considered a tiny research center is ranked the third best teaching place in the country, overall. Such is the stupidity of this ranking that a place without a teaching program has been ranked 3rd best place in India in teaching.

Homi Bhabha National Institute shows its annual budget as 473 crores, and has 973 faculty members. Basically, the entire budget of DAE institutions. Are they really educational institutions. That they have been declared as a deemed university to encourage their scientists to get PhD in-house now means that they can declare the entire budget as university budget, and all scientists as faculty. And NIRF 2017 admits that. Further, with 973 faculty members, they have 252 publications listed in Web of Science, easily the worst ratio of all research places in India. But guess what. They are in the overall list because of a very strange rule. The number of faculty members will be deemed to be 10% of the number of students, irrespective of the actual number of faculty. So in case of HBNI, it will be assumed that these 252 papers have been written not by 973 faculty members, but 310 faculty members. Absolutely crazy stuff.

This article in wired points out that there are drastic changes in 2016 ranks and 2017 ranks. Can quality of universities change this drastically from year to year. It specifically mentions a university having jumped the rank from #83 to #12. It means that either the ranking last year was arbitrary or ranking this year is arbitrary. How do we know that it is not arbitrary this year.

If you look at the overall ranking in teaching learning, you would find that most of our Agriculture universities or Veterinary colleges have a far superior teaching programs than IITs. May be we should handover IITs to Ministry of Agriculture. They seem to be running far better academic institutions.

A very sensible advice from prof. Ashish Nanda, Director, IIM Ahmedabad, "Rate, don’t rank, academic institutions" in Hindu Business Line.

How do you explain a huge difference in the rating by NAAC and ranking by NIRF.

Added on 11th April, 2017:

A blog by Dhruv Baldawa, "Why we should not be ranking our educational institutions"

Added on 23rd April, 2017:

IIT BHU raises objections on NIRF ranking 2017, says list based on ‘incomplete data. Here is the news report.

Tuesday, April 4, 2017

Should I take admission in Computer Science?

After my last blog article,where I had said that the quality of CS education is terrible in the country, I have received a few career counseling requests where the question is: "Should I study ECE or other engineering disciplines instead of CSE?" These are students (or their parents) who will be seeking admission to under-graduate programs in engineering this year.

A few quick comments: The problem of poor quality education is widespread and not restricted to Computer Science. I just know a bit more about CS and hence can say things more confidently about CS. But I don't believe that other disciplines are doing much better. Second, take admission in a discipline which you want to pursue as a career, and not worry about recent placement data. There will always be jobs for well educated persons. Third, if you have no personal interest in any discipline, seek admission in a college where the quality of education (in a holistic sense) is better. Again, ignore placement claims.

But I know from my past experience that the previous paragraph is completely unsatisfactory. Many parents want an unambiguous answer - would there be more job opportunities for CSE graduates or ECE graduates 4 years from now. They are not going to be happy with the answer that there will be more job opportunities for smarter people who have learn both engineering skills (including problem solving, analysis, synthesis, etc.), soft skills and life skills (communication, leadership, etc.) technical skills (programming, use of several tools), and have a broad based education to get the ability to pick up whatever is needed to be learnt on the job. This is independent of the discipline that you study.

(Notice that prior to admission, it is assumed that placement is a function of college and discipline. But after people do get a job, their opinions change and then the placement is because they are smart and the college has done nothing for them.)

First the bad news. The number of jobs created in the Indian IT sector has been less in 2016-17 compared to 2015-16. And it is expected that in 2017-18, the number of jobs will be even lower. If you consider the fraction of graduates getting a job across all universities/colleges, Information Technology has been a poor performer for a very long time. Just about 2.5 lakh jobs for about 10 lakh graduates. But CSE/IT have been popular programs because people don't look at the unemployment data, and only employment data, and that too top packages and not average or median packages. And the employment numbers were very high. Unfortunately for them, a lot of routine jobs are going away and these numbers are likely to keep coming down.

As per this report in Business Standard, Infosys recruited only 5500 employees in the first 9 months of the year as opposed to 17500 in the corresponding period of last year. Also, most IT services companies are re-skilling their employees as AI and Machine learning removes many routine jobs. Other reports talking about reduced IT hiring this year are here, here, and here. Note that most of this reduction has happened before Trump became President of US and therefore, does not take into account the possible restrictions on visas, outsourcing, etc.

There was a 20% decline across the sector in campus hiring this year. It may be noted that 20% decline is not uniform. Some of the poorer colleges will see 100% decline (companies won't visit them at all), while the top tier colleges may not see any decline at all.

It is expected that with more automation, and with developments in AI and Machine Learning, more and more routine jobs will be done by software or robots, and even where a human must be in the loop, the component that human will actually carry out would reduce. And hence, one should expect a more serious decline in IT services jobs over the next 4 years.

And hence, if the primary reason to join a program is immediate campus placement (and not learning, or interest), then one should check if that campus is dependent on IT services sector jobs for placing its students. If the majority of the students are joining IT services, there could be issues with placements 4 years hence. (And just to add, there are perhaps less than 40 CS/IT departments in the whole country where a majority of graduates are getting core CS jobs. IIIT-Delhi is one of them.)

Is Computer Science as a discipline on the decline. Absolutely not. In fact, it is one of the most exciting phases that we have seen yet. The kind of problems that we are able to solve today with cloud computing, high performance computing, big data, AI, deep learning, mobile computing, ubiquitous wireless Internet connectivity, and so on, we could only imagine just a decade ago. Computer Science has truly changed the way we live our lives and will keep changing it further as the future unfolds. This is really the time to study Computer Science and contribute to the creation of a wonderful future for all of us and the next generation. Also, Computer Science professionals are interacting with professionals from so many disciplines that irrespective of what your interests lie in, you could still study Compute Science and study a discipline of your passion and solve problems at the interaction of the two disciplines. (Keeping this in mind, IIIT-Delhi started the program on CS and Applied Maths last year, and is starting two new programs this year - CS and Design, and IT and Social Science.)

A typical question I get is this. By doing BTech in ECE, one can still take GATE in CS and do an MTech in CS, but by doing BTech in CSE, one is stuck in CS discipline. My answer is that in general one should not depend on change of disciplines in future. If you are so motivated that you can study CS material for GATE in your 3rd or 4th year, then you can right away take CS, study hard, and you can have a good career in CS. The problem of CS jobs is only because of poorly trained and poorly motivated students. On the other hand, there will be more opportunities for people with broader skillsets in future. (And that is why we are hoping that our new programs at IIIT-Delhi will be very successful.) 

So, if you want to study Computer Science, go right ahead and choose CS. The only exceptions to this general advice is: if you are only able to get admission in a college which has a poor quality teaching program. And once you join the program, make sure that you work hard to take as much advantage of your four years in college.

If you are not interested in anything, but want to do a CS program only because of placements. Check if the college graduates are too dependent on IT services companies for placement. If yes, be warned that IT services job are likely to become hugely competitive in 4 years.