Tuesday, May 01, 2007

Tabula Rasa on Top Engineering Institutions and their mission

This post of mine from yesterday ended thus:

Let me get back to Bill's question here: Do our top engineering institutions think of women's underrepresentation as a problem?

Picking up on it, Tabula Rasa offers one of the best comments I have received so far, allowing us to take the discussion forward. I think it's worth quoting in full. Here's the first part:

This is an excellent question but it's getting passed by in the discussion. First off, I detect a strong normative flavor to some of the comments -- that (a) these TEIs *ought* to think of women's representation as a problem, and, moreover, that (b) they *ought* to do something about it, most likely by (c) catching the culprit, the overly difficult jee, which needs to be (d) simplified with the introduction of questions that are however harder than at board level.

Working backwards. (d) An astute comment pointed out that there isn't very much conceptual space between a question that is harder than the board level, and one that is at the current jee level. (c) I and others have stated elsewhere that we don't believe one can remedy deep-rooted systemic biases / ills with "tweaks" to an examination that a tiny percentage of the population is interested in.

I agree that there is certainly a strong normative streak in the discussion (my end of it, anyways). There is a lot else in this part that I agree with, so what I say next is not really about what TR said, but about the impression he seems to carry of this discussion. In particular, his paraphrase of the normative line of thinking makes it seem that the case I built up against the JEE (and in turn, the Top Engineering Institutions (TEIs)'s ) is weak. I want to correct that impression; I believe my case is fairly strong; here's why:

Formally, TEIs' students have a pre-requisite: a pass in Class XII with certain minimum marks. In principle, they could just use some objective criteria for picking the top students from high schools, and be done with it. If they chose this method (let's not worry now about normalizing across boards, rank order, who gets computer science at Kanpur, etc), they would have a large number of women.

A long time ago, TEIs decided to introduce an extra layer between high schools and them: let's call this layer the Exam. One of the results is that the student body in TEIs is overwhelmingly male.

Now consider these two choices: direct admission based on high school marks, and admission through the Exam. While both are perfectly fair 'selection machines', we know that they produce significantly different results. In particular, we know that the latter introduces a skew in the student body, a skew that keeps a lot of bright, diligent, top-of-the-class women out of TEIs. This leads me to conclude that the Exam route is unfair to women, since it seems to (prima facie, if you will) introduce a bias. And yes, this bias exists over and above the bias imposed by the society.

Well, that's a three-paragraph summary of my case, and I have not even mentioned the 'evil' of coaching centres! I am willing to change my view if some detailed studies show that the Exam route is (a) necessary, and (b) fair to everyone; but as of now, my case against the Exam, built on publicly available information, remains strong.

* * *

Moving on, TR says:

That leaves us the two questions of whether the TEIs ought to think of under-representation as a problem, and whether they ought to do something about it. My opinion (again, *opinion*) is that it's not really their job to worry about such things, let alone try to go about fixing them with the limited toolbox at their disposal. What's the point of a Top Engineering Institution, after all -- to produce Engineers and Engineering; not to Help Remedy Society's Deficiencies. That's what government is for. One might argue that the reservations effort is/was a step in that direction -- I personally don't think it was the right step, but I think *that's* the type of question to be asking because it is harder. Focusing attention on the TEIs in this case is a red herring, plus it makes us feel good since it absolves *us* of any responsibility for actually *doing* anything except looking socially concerned.

I wouldn't give the TEIs such a fantastically free passage as TR appears to. As our elite institutions, TEIs should certainly not shy away from self-introspection to see if their procedures have an implicit bias, and if so, take credible steps to address them. If they shirk this duty, their effect on the society is seriously corrosive indeed; they end up legitimizing (encouraging?) bigotry, expressed through statements such as "women can't solve problems", "women are good only for Muggu exams", or just "women are not good enough!" (We have seen some of these in several comments).

Also, as TEIs employing Big Brains, they have a responsibility to get things right -- not just academic things, but also social things. They should be sensitive to the social currents (and they should strive to stay ahead of them) because their very existence and sustenance depend on their socially conferred legitimacy. In other words, if they don't do the right thing (and their legitimacy comes to be questioned), somebody else -- read, government -- will force them to. [BTW, this applies to private institutions as well, except that government will have to use softer options (such as withdrawing research funding) to force its will down their unwilling throats.]

In demanding that our TEIs keep their end of this social bargain, we are not asking them to somehow step in and "Remedy Social Deficiencies." We are just asking them to first remedy their own little deficiencies. That's all.

Here's the final part of TR's comment:

Getting back to the TEIs -- are they completely off the hook? IIRC one of the comments from the original article that sparked this whole discussion off was an IITian professor bemoaning the decline in "sparkling eyes" in their class. As someone who teaches for a living, I strongly believe that the last person to be blamed for this is the student. TEIs are viewed very differently today from what they were twenty years ago -- how many engineering graduates today want to spend their lives as engineers? TEIs have become nothing more than a stepping-stone to other, "greater" things. This reflects in part on where the good jobs are, and in part on how society's values are changing. But someone who can teach to save their life should be able to generate sparks in the eyes of half their class, easy. If classfuls of India's best and brightest are sitting around getting bored at some of India's most exclusive undergraduate institutions, it signals to me that there's a problem that's not with the way they got there so much as with what's happened afterwards.

I am in total agreement with TR on this one. On that positive note, let me stop here.


  1. Anonymous said...

    TEIs can respond to social situations and stay ahead of the curve. My favourite example remains MIT. By consistent, conscious efforts , they have diversified the profile of students. At present 50% of the undergraduate classes are women.

    This happened not due to some blind desire for equity but because data showed that the performance of women was under predicted by SATs . They needed to be compensated by 25 points.This enabled MIT to continue with its academic mission: to PRODUCE great engineers and scientists, not serve as adjudicators of abstract generic merit.

    One article making a case for institutional changes coming from a great female engineer is
    this . It covers a fair amount of ground.

    This necessarily means introducing subjectivity in the admissions process and frankly , I dont see why not?The copout of undue influence speaks little of the integrity of the professors, in which I have complete faith. Secondly we do have a homegrown example in the IIMs which do the same. It would definitely require courage, and I think we can demand it from them.

    As to sparkling eyes, please. The 17 year olds of yesterday were babe in the woods, with their world view much smaller simply because India was a much smaller, restricted place back then. No internet,no infosys, no MNCs, no Mckinseys . The IITs frankly have no idea of responding to a world which has rapidly changed around them. The educational pedagogy is 40 years out of date, MIT has been running a UROP program for now approximately 40 years, IITB started one recently. We have not been able to even copy something as cool as IAP or the MIT tech competition. The standards of rigidity are blatantly Neanderthal, Feynman was busy changing his major 2 times back in 1939, nobody can achieve that even today in India.I mention these because these are not activities which need a lot more money, but they require a different way of thinking and a commitment to the student as a young adult.

  2. Anonymous said...

    I agree with the comment by 'as'. IITs should admit more women not for ensuring social equity but for their own betterment. It somehow looks that the JEE is equivalent to playing games on the computer where women seem to have a raw deal. Is that what IITs want to achieve?

  3. Anonymous said...

    Abi, now that I have gone through all your articles, I can summarize them in these two points. Please correct me if I am wrong.

    a. The present screening system (JEE), by its very nature or because of social trends, is unfairly discriminatory towards women.

    b. The current batch of students are not (as) bright (as the previous batches). Hence JEE is not able to identify the best talents.

    I cannot connect with the second point. If you are choosing the top 4000 students out of nearly 200000 appearing for the exam, any method you use (except lottery, ofcourse) will give you above average students. It might be true that another set of 4000 are better than the chosen ones, but to say that the original set are not intelligent enough to be in IIT is frivolous. If instead of 6 we had 100 IITs, many more students with average intelligence will get in. Would the professors still lament the lack of quality students in that case? Every professor wants bright students, and it is natural. But is it not a kind of intellectual elitism to say that only the brightest should be able to avail quality education? The responsibility lies in the IITs to turn the above-ordinary students into extraordinary men/women. Indeed the success of IITs lie in a large part on not the quality of students that get in, but by the things they learn in IIT.

    Now coming to the first point. The major purpose of introducing JEE was, as you pointed out, somebody felt that the board exams are not an accurate metric for measuring intelligence. The same arguments you are making against JEE might have been made against XII exams at some point of time. There are some problems with our school leaving examinations too. The XII exam results (at least for CBSE) are skewed against people living in rural areas and heavily depends upon the kind of schooling one gets. Replacing JEE in entirety might not be the best solution. What we can have is a kind of deprivity index (which can be a combination of sex, region, school, sports and extra-curricular brilliance, or anything) which can be applied to results obtained in any pattern of examination that is decided by the IITs.

    Lastly, I must point out that CGPA is not the be-all-end-all of IITs. The CGPA might correlate heavily with XIIth board marks, but the IITians success in professional life (engineering, management, social work, teaching) is not directly proportional to the grades. What made them successful is the things they had learnt in their 4 or 5 years in IITs, and CGPA might not be reflecting it accurately.

  4. Anonymous said...

    Are such socialist experiments fun stuff to do with other people\'s money on other people\'s lives?

    Do we do them with our own money, on our own lives?

  5. Tabula Rasa said...

    ha! as i typed that comment this morning, i was wondering whether i should post about this whole jee thing myself. looks like i've been bangalored -- in more ways than one :-D

    two brief points.
    1. in response to let's not worry now about normalizing across boards, rank order, who gets computer science at Kanpur, etc
    well yes, we could choose to not worry about this *now*, but surely we'll have to worry about it at some stage, which is when we're going to come up against the point that i, chitta, sudeep, and others have brought up, that ultimately the objective of the jee is to rank and not rate, and *all* ranking exercises are discriminatory in some sense. yes, the chosen ranking exercise results in a certain segment of the population opting out. the question then is, how can we design a ranking exercise that leads to less opting out? and here we again enter into opinion territory.

    2. in response to the "fantastically free passage" but, again it's a difference of just opinion. i believe the iits are possibly failing their mandate in making engineering an exciting career option. however, i believe this is true across the board. if engineering was deemed a desirable profession for all, less people would shy away from trying for it.

    abi, i'd also like to apologize if i'm coming across as intransigent. i'd like to blame this on the nature of the forum we're using -- i look into your blog a couple of times a day these days, often in a rush, and hence have to pick up on only a single point or two to discuss. also, writing makes us seem to take more extreme stands than we may actually espouse.

    also, i agree with you that some of the comments in the thread have been quite regrettable. they certainly don't represent my views. (but i can't resist adding -- is such bigotry the fault of the iits or a social ill in general? ;-)

  6. Vivek Kumar said...

    @Abi: I am not sure about your "total agreement" with Tabula Rasa about that last paragraph.

    This whole debate gathered momentum only after your quotation of an article where a Prof said that the "gruesome" JEE is the main reason for the loss of "sparkling eyes" etc.

    You quoted that article in a highly approving tone. You called JEE "dreaded and dreadfully brutal". This criticism of JEE by some IIT faculty members was called by you as "a welcome development".

    How does all that translate to a "total agreement" with TR's conclusion that ".. there's a problem that's not with the way they got there so much as with what's happened afterwards"??

    Apologies in advance if I am much mistaken. I would welcome some clarification.

  7. Rahul Siddharthan said...

    Vivek -- I think there is a problem both in how they got there and in how they fare afterwards, but the problem is the same: the IIT approach prefers single-minded focus on textbook material and problem-solving to a breadth of interests.

    The JEE doesn't just discriminate against women: it discriminates against anyone who wants to "have a life" from the age of 15 to 17. And, from what I've been reading and have heard from friends, life inside IIT is much the same.

    Electrical engineering may be a fascinating subject, but in order to be fascinated, you must be capable of being fascinated by many different things, and realise that class quizzes and exams are not the ultimate goal in life; and in order to fascinate you, the professor must have a broad range of interests with which s/he can draw connections, and also realise that class quizzes and exams are not the ultimate goal in life.

  8. Vivek Kumar said...


    "And, from what I've been reading and have heard from friends, life inside IIT is much the same."

    I don't know how you got this impression, but life inside IIT is certainly not all about quizzes and assignments.

    You can "have a life" (I would want to know what activities you include in that, BTW) and still do well at quizzes/exams/assignments to have a CPI/CGPA that would be respected by recruiters and by foreign universities alike.

    In my definition of "have a life", I am including things like music, dance, sports, treks, social service, debating, politics, quizzing, photography etc. So unless having a life implies only one activity - dating - I think IITians have as much life as students from any other campus.

  9. Vivek Kumar said...

    @rahul: I should added in the last comment itself, I totally agree with you on:

    "the professor must have a broad range of interests with which s/he can draw connections"

    I wrote a post alluding to this a while ago.

  10. Abi said...

    AS: Thanks for those wonderful links! From afar, SAT appears to us (to me, at least) to be a temporary ideal to work towards. But, the two links you provided, along with those I got elsewhere (see my next post) have led me to refine my views about such exams. The stuff about a systematic bias (however small they may be) tells us that we should be very careful in interpreting the results that come out of such tests.

    Without getting into the details of your critique about the internal functioning of the IITs, let me just say that all our institutions should become more student-centric, allowing a lot of flexibility in choice of courses (and later, careers too).

    Anoop: Point (a) you have made is broadly my view; the only amendment that I would make is to say the discrimination due to JEE is in addition to the ones women face in society. The best evidence I have is that at the end of high school, girls do at least as well as -- if not better than -- boys (in terms of the number of girls and boys above any cut-off, including, say 95 %). Our entrance exams -- including JEE -- seem to say to them "doing well in school is immaterial!"

    I have not asserted your Point (b). That assertion was made by a bunch of IIT faculty members, who were probably frustrated by the lack of interest and motivation in their classes. I believe they have taken the easy route to take the line that the students are not as good as those of an earlier era.

    My view is that just as the IIT students are among the best that India has to offer the IIT professors are also among the best are also among the best that India has (let me emphasize that 'among' is the keyword in this sentence). If they want to snipe at each other, let them. I will stand back and watch.

    Now coming to the first point. The major purpose of introducing JEE was, as you pointed out, somebody felt that the board exams are not an accurate metric for measuring intelligence.

    No. I have not said that, and nobody in his/her right mind would make this claim. If board exams are useless, what's the point of using them as a pre-requisite?

    Entrance exams solve a different problem: since there are many different boards, and since their standards of curriculum, difficulty level of exams, grading policies vary considerably, a "common" exam appears to be a good way to select students. My point is that in doing so, and in selecting a male-dominant set of students, the entrance exams introduce a bias.

    In a comment (on a different post), Confused pointed out that perhaps it's the board exams that are biased. (Compared to JEE, yeah, they do have a more pro-women bias! ) If someone wants to make that claim, that's fine by me. But, I don't know of anyone from the IITs (or those who suggested any of the entrance exams in the first place) who accused board exams of bias.

    Coming back to whether JEE results are a more accurate in measuring intelligence (required for doing well in the IITs), the internal study -- which showed that board exam results are a better predictor of academic performance at IITs -- points to a strong possibility that this claim is hollow.

    Which brings me to the last point you raised: "I must point out that CGPA is not the be-all-end-all of IITs." I agree.

    But I have to wonder what you mean when you said, "What made them successful is the things they had learnt in their 4 or 5 years in IITs, and CGPA might not be reflecting it accurately." So, what is it that an IITan 'learns' in IITs that his/her "CGPA might not reflect accurately?"

  11. Rahul Siddharthan said...

    Vivek -- yes I've known some very multi-dimensional IITians, and some very one-dimensional ones, too. I suppose it all depends on what you're like to begin with.

    Abi - the question of how well you fare in life afterwards was dealt with (in the US context) in an article I linked to recently. If you haven't read it, do -- it's really interesting.

    Here's a relevant extract:

    “As a hypothetical example, take the University of Pennsylvania and Penn State, which are two schools a lot of students choose between,” Krueger said. “One is Ivy, one is a state school. Penn is much more highly selective. If you compare the students who go to those two schools, the ones who go to Penn have higher incomes. But let’s look at those who got into both types of schools, some of whom chose Penn and some of whom chose Penn State. Within that set it doesn’t seem to matter whether you go to the more selective school. Now, you would think that the more ambitious student is the one who would choose to go to Penn, and the ones choosing to go to Penn State might be a little less confident in their abilities or have a little lower family income, and both of those factors would point to people doing worse later on. But they don’t.”

    It would be really interesting to see that sort of study done here... but one complication is that you don't just "get into IIT", you "get into computer science at IIT Kanpur" -- whereas at Penn admission is the important thing, you can easily switch your major. As somebody pointed out, this may be why JEE AIR correlates poorly with CGPA: a civil engineering topper at IIT Kharagpur would probably have a much lower AIR than a mid-level CS student at IIT Kanpur.

  12. Abi said...

    TR: Thanks for visiting to comment on your own post (made from Bangalore ;-).

    I'm still not sure why Chitta, Sudeep and you are so hung up on this exercise about ranking. I do understand your concern that board exams produce a large number of students with 90+ percent in each subject. Let me try an idea by you:

    Think of an exam with questions that are pre-tested on a sample population, and have differentially weighted. As a matter of abundant caution (otherwise, this exam will slide back to the old, UNimproved JEE!), we might ensure that none of the questions is so tough that it is answered correctly by less than 25 percent of the students!

    If this exam has 1000 marks, is it really difficult to design it in such a way that, for example, 90 percent in the board exams gets mapped to 900 marks? If you have three such exams (mathematics, physics and chemistry, in that order, so that, for example, (x, y, z) gets a better rank than (y,x,z) for x>y>z) you have (nearly) a million unique slots! If you want more unique slots, change the mapping to equate 90 percent in the board to 800 marks in the entrance exam. Therefore, this method of testing should be quite easy for a group of statisticians to design.

    This method of testing does require that questions are standardized. To my knowledge, none of our entrance exams is. Going back to what I said in the post, this is another thing that our IITs (for JEE) and CBSE (for their own board exams and for the AIEEE) have not bothered to 'get right'. I have ranted about the lack of standardization and the noise in the results that come out them here).

    Vivek: Ah, I see now where the problem is. Looks like the extract from the Outlook article led to a misunderstanding (for which I'm partly responsible) that I actually agreed with *everything* the IIT professors said. Thanks for asking for that clarification!

    Here's a little background: We went through an interesting discussion sometime ago about whether one should agree with (or, approve of) the stuff one cites on one's blog. My answer, by the way, is "no"; the fact that something is interesting is good enough for me to cite it on my blog.

    Did I find the Outlook piece interesting? Absolutely. [When IIT professors badmouth their students in public, yes, it's certainly blog-worthy!]

    Did I have to agree with (or approve of) everything in that article -- and particularly the stuff about 'sparkling eyes' and all that? Hell, no. As an outsider, I know too little about IIT students to be able to judge them one way or the other. And, here's another fact: I actually know some IIT professors who just adore their students!

    I view the bash-fest between IIT professors and their students from the outside -- as a spectator. I have no interest in taking sides on this fight, though I did express some (admittedly guarded) views on this matter in a comment on your blog.

    You see, Vivek, among the three points that you mentioned in your post, it's only the second -- gender disparity -- that really matters to me. [I have a strong view on the third one too -- coaching centers -- but only so far as it's related to the second).] It's not just about the bias against women. It's also the bias against students from poor families and rural areas. When this discussion started, I saw coaching centers as the primary mechanism through which this bias operates. Now, with all the reading that I have been doing, I am leaning towards the view that the exam design itself may introduce a bias and may need a serious re-look.

    So, coming back to the Outlook article, what did I really approve of? I approved of the fact that IIT professors were willing to go public with their bashing their entrance exam [their bashing of their students was just an interesting sidelight!] That is certainly a welcome sign. I agree agree with their view that "this 'gruesome' examination ... 'reduces young people to automatons' who '[rely] on pattern recognition.'
    I used that particular (controversial) extract to get back to a pet rant of mine about entrance exams in general and about the "dreaded and dreadfully brutal" JEE in particular.

    Thanks for giving me this opportunity to set the record straight.

  13. Abi said...

    Okay, while I was busy typing my last comment, looks like Rahul and Vivek have been exchanging some interesting views about IITans. Let me sit back and enjoy the fun!

  14. Tabula Rasa said...

    Sorry, I couldn't follow your example (of the proposed new exam). Could you please explain it again?

  15. Vivek Kumar said...

    @Abi: echoing TR.. could you make a new post with your proposed JEE?

  16. Pratik Ray said...

    What about an exam which has 50% MCQ (like IIT JEE screening) and 50% long answer type (like board exams)?