The online Inquirer article dated December 4, 2018, presents an overview of the policy, enacted in 2018, of creating supernumerary seats for women at the undergraduate level at the IITs, in order to ensure that their enrollment reaches a target of at least 20% of the undergraduate class by 2020. We are grateful to the BSP for writing on this important subject and seeking to communicate the thinking behind the policy, via interviews with 4 people who were closely involved with its conceptualisation and implementation. We feel that the article, on the whole, does a good job of conveying the background to the policy, and the reasons why it was felt necessary. However, we found certain aspects of the way the details of the policy were presented and interpreted to be somewhat misleading, and in some cases at odds with what those interviewed were actually seeking to say. Gender diversity is essential to every institution, not just to create a healthier environment for both genders, but also to aid in the process of learning, as a diverse community brings out a multitude of perspectives and opinions. Given the current state of female representation in IITs, we believe that the provision of supernumerary seats for women is an essential step towards creating a better community inside the campus and eventually creating a balanced society. In this response, we present a slightly different perspective which seeks to address some of the implicit criticisms of the policy contained in the original BSP piece, and clarify some of the misconceptions we feel might have arisen therein.
Should we think of this as ‘reservation’?
While the scheme is certainly an affirmative action scheme which is designed to increase opportunity for women by compensating for social disadvantage, the way it operates is in certain key respects different from what in India is normally understood by ‘reservation’. Hence we feel that the use of the word ‘reservation’ so prominently can convey a misleading impression. (Though this is not to say that we think the existing SC/ST/OBC reservation schemes are undesirable: those are rooted in a separate social background, and later we briefly discuss their importance as well.) In several of the interviews the points of difference have been enunciated, and these have been well brought out in the small print at the end, but the first and dominant impression conveyed by the title and layout of the piece still ends up being that this is no different from reservation. Under this scheme, supernumerary seats have been created for female candidates in addition to the existing seats and girls are not eligible for the gender-neutral seats until and unless the female-only seats are all filled up first. This is the key difference between the supernumerary scheme and the existing reservation schemes. Hence, the gender-neutral seats remain reserved for men unless the number of girls outcompeting boys purely in terms of JEE ranks exceeds the number of female-only seats, in which case the whole scheme becomes redundant and the outcome is exactly equivalent to what would have happened without this scheme.
It’s funny that our society is so male-dominated, not just in terms of numbers, but also thoughts that when one aspect of the policy seems to be benefiting males, it has been referred to as the bright side of the story. The fact that the gender-neutral seats effectively remain reserved for males for at least the next few years (or in other words, until the entire scheme becomes redundant, as mentioned) is indeed an advantage for males. An effort being undertaken for that section of the society which has remained in the dark for so long is intensely debated, but the moment it seems to be benefiting males, it is considered to be brightening the society.
The huge graphic which shows a girl standing atop the word reservation is so deceptive at first glance as the assumption behind the graphic appears to be that the current status of men and women is equal, and then reservation is lifting women higher and giving them an unfair advantage. Given the facts presented by the committee members and the well-known percentage of women in IITs or any technical institute for that matter, a graphic based on such an assumption is unwarranted. All the efforts of the interviewees to explain why there is no such unfair advantage being given at all are effectively undermined by this very prominent graphic being placed right at the centre of the piece, unfortunately conveying precisely the opposite message to what they wished to convey!
On merit and meritocracy
The notion of merit — which seems to be the most important driving factor behind the scepticism of the idea of supernumerary seats — is by itself so poorly defined. No single exam can be the sole test for merit, and this is illustrated by the performance of girls at IIT which is incoherent with their performance in the JEE, as emphasised by all the interviewees (in other words, the JEE significantly underestimates the potential of women to perform well once admitted to IIT; see also the below figure). Schemes like supernumerary seats should be seen as attempts to make the exam closer to an ideal selection process which takes more aspects into account rather than simply the outcome of a 6-hour exam. Many social factors affect exam outcomes, as has been pointed out in the interviews. And our goal shouldn’t be just to admit students good at a particular kind of exam problem; we believe it’s equally important to have a diversity of ideas, and a healthy and welcoming environment where students from all backgrounds can flourish. So one exam can’t be sacrosanct. We need to consider multiple factors, as universities globally do. We seriously need to re-think what we mean by merit, before we insist on always putting it above all else. This scheme is a step in that direction. Of course, once the number of girls increases and the society starts offering a more conducive environment to girls preparing for engineering, the scheme will have fulfilled its role and will no longer be needed.
The Inquirer article suggests that the observed CGPA difference between boys and girls might be confounded by different category distributions for the two genders prior to the supernumerary scheme (specifically, there being proportionally fewer girls admitted in the SC/ST/OBC categories, relative to the General category). However, in doing the analysis we did seek to control for such possible confounds. If girls’ higher average CGPA were accounted for just by the category distribution, then the same would also be largely captured by girls having higher JEE ranks overall, as there is a pretty strong relation between rank and category. But the higher CGPA result for girls remains significant even if one controls for JEE rank. In other words, if a boy and a girl with similar JEE rank are compared, then the girl will tend to have higher CGPA after coming to IIT. So this should also largely take care of any category confound. And it’s worth noting that there is also an effect that runs in the opposite direction: branches which tend to have higher CGPA (and higher JEE ranks) tend to have fewer girls. So on the whole, the result of girls performing better remains pretty robust even when one seeks to control for factors like JEE rank.
Box plots of gender-wise CGPA distribution for the 2012-entry batch at IIT Delhi (a similar trend is seen across 13 years of data, 2003–15). The thick central lines indicate median CGPA, and the blue boxes indicate the 25th- to 75th-percentile range. It should be noted that these distributions are despite the fact that the highest JEE rankers are disproportionately boys. So if we compare a boy and a girl with similar JEE ranks, the difference is expected to be even greater than that observed here. (Figure courtesy Prof. Subhashis Banerjee.)
The article also presents a student survey which shows that 89% of respondents believe there is ‘no coherence’ between the CBSE exams and the JEE (presumably meaning that performance in one is not predictive of performance in the other). The article uses this survey finding to suggest that the data cited by the interviewees on girls performing relatively much better in the CBSE exams is irrelevant. This completely misunderstands the reason for citing that data! In fact, the survey finding only serves to strengthen our point, which is that the JEE is in effect biased against women (for a variety of reasons as has been indicated in the original interviews). So naturally, we don’t expect coherence between the results in the two sets of exams. But the point is that, unless we somehow believe that the JEE is absolutely perfect, and the CBSE exams absolutely meaningless, we must accept the natural conclusion that non-JEE assessments favour a higher estimation of the performance of girls in maths and science. So to be fair to girls, we should seek to compensate for the way in which the JEE specifically underestimates their aptitude, as this would in fact be a more accurate assessment of their ‘merit’; and this is exactly what the supernumerary scheme seeks to do.
Will the scheme benefit only ‘privileged’ girls?
he article states that “[I]t’s an open and quite well-known fact that the caste-based reservations with similar promises have benefitted the already privileged among these communities”. To begin with, this claim is just false, under any meaningful definition of the word ‘privilege’. Whom is this ‘well known’ to? We are not aware of even one scientific study or article which demonstrates such a thing. On the contrary, there is plenty of social scientific literature showing that in our society, caste itself is the strongest marker of privilege, and this persists to the present day. Whether you equate privilege with wealth, or social status/mobility, or any other meaningful measure, in our society it correlates extremely strongly with caste. To cite just one example, a team of social scientists from IIM Ahmedabad, Duke University, and the University of Manchester found that caste has a large disadvantaging effect even amongst people purportedly from the same ‘class’ or economic status. For instance, the son of a domestic worker is 3 times more likely to become a scientist if he is ‘upper’-caste than if he is SC. (Iversen et al., EPW 2017.) We can provide further examples of such published research to anyone interested.
So the idea that a large number of SC/ST/OBC students who have been enabled to study at IIT via reservation, are somehow ‘privileged’, is completely incorrect. In a fair society, students from these communities (which collectively are estimated to make up 70–80% of our society) would have been admitted to IITs roughly in proportion to their population. In our current society, even with reservation, their proportions are lower than in the overall population. Without reservation, they would hardly be represented at all, which just goes to show how underprivileged they still are, how many social barriers they continue to face. Until the time that such barriers are fully dismantled, reservation is an essential component of a fair admissions policy.
The same kind of suggestion is then made for the women’s supernumerary seats, that they will end up benefiting ‘privileged’ women and giving them an ‘undue advantage’. This is again deeply problematic and again goes against the whole point of what all 4 of the interviewees were trying to say. Before talking about ‘undue advantage’, one has to acknowledge that all women in our society, just by virtue of their gender, face an undue disadvantage! There is so much evidence for this, which is what all of us have talked about in the interviews, and we have further elaborated on above. When a scheme is created to compensate for such social disadvantage, to refer to it as giving some people an ‘undue advantage’ is simply not accurate. And it must be noted here that the percentage of supernumerary seats is consistent across all categories. So, within the group of women admitted under this scheme, women from the SC/ST/OBC communities will be represented in the same proportions as mandated for the overall student population. Whereas prior to this scheme, the numbers of women admitted were somewhat disproportionately skewed towards the General category. So, contrary to the suggestion, the scheme will especially ensure greater representation for underprivileged women.
Is the scheme based on flawed assumptions? Would alternative schemes have been better?
The article talks about a very interesting way of solving the issue of underrepresentation of women in engineering, which is educating and encouraging female students to pursue engineering in middle school itself and hence motivating females in all sections of the society to take up engineering. This indeed would be an ideal alternative if at all possible, but what surprises us here is the fact that we expect students and their parents from all sections of society to understand something which we, members of one of the elite institutions of India, are ourselves not ready to accept yet. In fact, the whole point of encouraging gender diversity at IITs is to set examples for younger girls to pursue engineering.
The article declares the assumption of a gradual increase in the number of girls (relative to the 2017 baseline) to be a dangerous one. However, this assumption is not actually part of the policy’s design. In fact, the policy makes a conservative assumption that without supernumerary seats, the number of girls admitted would remain constant at 2017 levels for the next few years. And hence it preserves the number of seats available to boys at 2017 levels, whereas, without the policy, gradually improved performance by girls would naturally have led to a decline in this number! It is also not true that 2017 is some kind of major outlier (with 9.3% of IIT admits being girls); the proportion of girls admitted to the IITs has been roughly constant at 8–9% for a long time. So using 2017 as a baseline is as justified as any other baseline.
Thus, unless and until the number of girls gaining a seat purely on the basis of the JEE does go up substantially, the male candidates are not at any disadvantage as their seats remain unoccupied by girls. More importantly, the very purpose of the scheme is to provide more role models and more confidence to parents to let their girl child pursue engineering. These have been identified as major hurdles for equal representation and the policy aims to counteract these; hence the anticipated increase in number of girls is core to the policy’s success and should be seen as the expected outcome of the policy, rather than an underlying assumption.
Regarding the issue of infrastructure and the space crunch which has been raised several times, there is no denying that this needs to be addressed, and construction of two new hostels and several other buildings on campus is actively underway. But this was a pre-existing problem, and it’s worth noting that the total number of additional students admitted under this policy in 2018 (which the article states to be a ‘large number’) is in fact just about 0.5% of our total student strength. So the implications of this policy for the infrastructure problem are really very small, and this was a problem that needed addressing independent of whether we adopted this policy or not.
To conclude, we feel that some of the conclusions and overall impressions conveyed by the Inquirer article regarding the supernumerary seats policy were not in sync with the actual facts and data, and we have sought to rectify these. No policy is 100% perfect, but we believe, for the reasons discussed, that the policy at hand is a significantly progressive one, with the benefits far outweighing any possible disadvantages. And while fairness and equitable opportunity for women are the primary drivers, we strongly believe that the scheme will be beneficial for all of us, and in particular for men too. A better gender balance on campus will naturally foster more frequent and healthier cross-gender interaction. And the IITs are mandated to use technology to solve social problems and improve people’s lives, and 50% of society is women. To understand and address society’s problems, we need proper representation from every section; each brings its specific experiences and perspectives. A diversity of views and ideas strengthens us all, academically and intellectually.
(We would like to thank other members of the IGES core team for their thoughts and inputs.)
Article by: Sumeet Agarwal, Aditi Jha