IITs, JEE And The Girls Quota: The Entrance Exam Needs A Redesign

  • The real fix for minimising biases in the system is an academic redesign of the JEE for admission to IITs.
  • The Learning PointSaturday, April 29, 2017 5:02 pm IST
    IITs are India’s premier engineering institutions.
    IITs are India’s premier engineering institutions.

    The "Idea of India" has been leading to a national obsession with finding disparate groups, disadvantaged in some way – and trying to solve problems specifically for those groups, letting the mess spill over into other areas. The latest in this series is the introduction of a quota for girls at the Indian Institutes of Technology. I do think that given no other action, some effort to give representation to 50 per cent of the population is better than no effort at all. But I also think that this is a terrible fix because it flies in the face of gender equality if you think about it.

    Sounds contradictory, I know.

    The real fix which I have been warning about for ages is a necessary but minor academic redesign of the Joint Entrance Examination (JEE).

    Had this been the exam for a private university like Ashoka, it'd be entirely their prerogative, but since this is a publicly funded institute, certain expectations need to be met.

    Let's get some things straight first.

    Coaching by itself is not a problem. In fact, it is a boon.

    It is an example of how an unregulated economic partition can thrive and create capacity when our highly regulated school system has collapsed.

    If anything, access to coaching has made education available at scale, and as a commodity, which might sound repulsive, but it is most certainly better than almost no education at all.

    In the absence of a respectable schooling system, coaching made education accessible at scale. This was of special value to people who couldn't go to good schools, simply because there aren't enough of them.

    Kota and Hyderabad opened up the gates of respectable higher education to those who could not attend schools like Delhi Public School or DAV or top ICSE schools.

    In 2013 when the JEE started factoring in board examination scores, there was a very naive expectation that coaching chains would start teaching English and fifth subjects and register as schools.

    Nothing of the sort happened. Why? Because, surprise, surprise, the school regime is one where government interference is heavy, and you cannot even make a well-deserved profit after all that hard work and hassle. Most of the problems attributed to coaching are a consequence of problematic test design.

    Selection biases at work

    An exam will be biased towards those who put in extra effort towards it, and there'll also be a bias, where people interested in a particular exam have a greater likelihood of taking coaching for it.

    The ratio of girls to benchmark against

    The question is not "there are 50 per cent girls in the population" so why isn't that reflected in the IIT-JEE selection list.

    The real benchmark is that the percentage of girls topping and scoring over 95 per cent in Physics, Chemistry and Math (PCM combination) in the CBSE and ISC examinations is commendable.

    A quick look at the list of CBSE toppers for 2015 suggests that the fraction of high-scoring girls from the science stream is easily over a third, perhaps more. This is the ratio we must benchmark against and ask ourselves why this isn't reflected in the JEE merit list.

    The reason I cite the example of CBSE and ISC is that their syllabus is quite in line with the IIT-JEE syllabus, though the questions are far easier. Frankly, the CBSE/ISC level of Physics, Chemistry, Math is far closer to the standards required by an engineer than that of the JEE.

    That said, the JEE does something critical which board exams seldom accomplish: it forces you to think, and it does a good job of identifying science and math olympiad winners.

    The board exams tend to be uninspiring, have predictable recall-based questions and will be far easier to game if used as the selection criteria.

    Also, we need to remember that girls are not the only group biased against. There is natural or unnatural bias depending on the board, gender, region and income group.

    Admission processes are always biased, but we need to try eliminating any stark, unnatural ones. As things stand, the proportion of CBSE/ISC kids qualifying is over ten times that of most state boards; the proportion of students getting in from Andhra Pradesh is far more than that of Tamil Nadu or Kerala. Whoever prepares the most for an exam, or goes through better quality schooling, has a greater chance of getting in.

    All we should do is eliminate any unnatural friction, which in the case of the JEE is primarily the secrecy around the exam.

    The problematic exam design and a system of secrecy

    There is a limited number of seats, and there is a massive tussle for said seats. You can't stop people from studying 24x7 or taking additional coaching.

    But there are specific issues with the IIT-JEE which I will address.

    Exams are not bad

    As of now, India is a poor country and Indians cannot afford the rounded assessment which American and British universities focus on.

    The upside-down thing though is, currently the JEE comes with a huge price tag, which means that the process caters to the relatively rich. Students from CBSE and CISCE have nearly ten times the success rate of state boards.

    Possible action items and issues to be addressed

    (a) Complete lack of transparency about the extent and scope of the syllabus

    Every coaching centre or publication tries to thrive on this information gap by selling overly complex problem sets.

    After you take the exam, you realise that you could have done with just one book, but by that time the damage is done. Lots of folks end up joining coaching centres even if they had the potential to just study on their own. This part is unacceptable because a publicly funded educational institute cannot implicitly force students to pay for a private channel.

    There is nothing wrong with the coaching itself but the unnecessary requirement created for it, much of the time. Who is to say that someone who went to an expensive private school wasn't unfairly advantaged in the process after all.

    The solution to this is simple: IITs should recommend text books.

    This does not mean that questions should be repeated verbatim, but that the principles and facts required to handle the exam are bounded.

    And that kids don't get taken for a wild goose chase.

    (b) Arbitrary gates to publicly funded corridors

    Only 32 per cent of the students study pure sciences at the Class 11-12 level, and it is unclear what percentage of them opt for PCM.

    In general, only Mathematics is required for many of the branches.

    As things stand, the IIT-JEE process is a microcosm of problems which will develop once the new National Testing Service kicks in.

    The exam is a Physics, Chemistry, Mathematics exam and the result is interpreted by using a simple summation of the marks in those three papers.

    Biotechnology and Biochemical Engineering have limited relevance of Physics and very little use of Mathematics. The exam doesn't have bio at all. This, of course, can be fixed by just having a Biochemistry section in the Chemistry paper – it already exists in Class 12 syllabus.

    Programmes in Computer Science and Electrical Engineering don't need Grade 11-12 Chemistry.

    Programmes in Industrial Engineering and Economics don't need Grade 11-12 Physics or Chemistry.

    Architecture does not require either Physics or Chemistry. See what I mean?

    Arts, commerce and medical stream students could easily handle a variety of courses which are made inaccessible to them.

    The solution will be to interpret the PCM scores differently, and each department can assign a weight to the subject(s) it considers most important.

    The current "one nation, one merit list" principle does not make much sense as the desirable attributes for Chemical Engineering could be vastly different from those required for Electrical Engineering.

    A basic language and functional English test are perhaps required now. Too many complaints of folks unable to write a single sentence.

    This itself is an area where girls might have an edge over boys, but the reason for the inclusion of English or a language section shouldn't be for the purpose of gender-balancing but academic reasons alone. All major college entrance tests and standardised examinations, such as the SAT and the GaoKao, do have a language section.

    (c) What is wrong in Kota and Hyderabad is not the coaching but the schools

    Students register in dummy schools, skip essential classes in language, practical work and the fifth subject.

    The JEE was meant to be something in addition to the Class 12 exam, but this bypass is unfair because students in other states have to handle their regular schooling (as they should).

    The fix for this does not lie with the IITs, though the inclusion of functional language skills in the JEE will help. While it is natural that an entrance exam can only test on a least common denominator regarding subjects, the school system does have a role in imparting education in the areas of language, practical work and exposure to specialised electives like Biotechnology, Computer Science or Psychology.

    As it is, many of our schools offer very few options in subjects. Some don't have a Biology option for the Science stream. Schools in developed countries routinely offer fifteen or more subjects even at the Grade 10 equivalent.

    Take a look at the subjects offered by this school in England, which is hardly an outlier or an exceptional school. This is a bit of a digression, but this is just to emphasise that coaching cannot offer the sheer breadth in education, which a good school does.

    (d) The design of the JEE paper

    In the early 2000s, a score of even 50 per cent was sufficient for a person to rank in the top 100-400 range. The papers were just too long, so even attempting half was considered a safe strategy.

    About 35 per cent was sufficient to get in.

    This just indicates an overly difficult as well as lengthy examination, which was prone to non-academic traits and abilities: like question selection, speed, tricks, etc.

    Solving a Rubix cube is great training in problem-solving, but solving it in five minutes? Great training for sure, but ultimately worthless and effort which was perhaps better spent on something more productive or creative.

    Some thought needs to be given to the balance between testing for speed, recall and genuine problem-solving ability.

    What if everyone was given double the time and the rank list saw a slower but better problem solver reach the top?

    This is ultimately an academic issue, and it depends on how seriously the IITs are interested in solving this problem.

    Similarly, there are a lot of esoteric topics particularly in the Chemistry portion of the syllabus, unlikely to be present in the board exam syllabus, or ever required in most kinds of engineering.

    The exam should value general principles over specifics, and such content can go.

    The multiple choice questions-based format of the question paper also has serious issues despite the fact that it is quick to grade, and the scores are not prone to subjectivity. But that is an entirely academic issue.

    (e) Data and statistics, please

    The IITs are terribly opaque about their data. Who got in? What board, what region? They release some information now, but it is far from what it should be. Something as simple as what was the Class 10 and Class 12 board of the entrants. Parents and students can plan accordingly.

    Also, some tracking of how different demographics and ranks perform once inside the IIT system. If data shows that girls with 40 per cent in the entrance exam perform as well as boys with 50 per cent in the entrance test, there is perhaps a valid case to have an extraordinary bump-up for them.

    A recent study shared on social media suggested that girls have an average GPA exceeding that of boys by one entire grade point on a scale of ten. A formal analysis of that kind of data can make a more objective case for increasing the percentage of girls.

    Studies from Massachusetts Institute of Technology versus Carnegie Mellon University suggest that the undergraduate performance of girls with 700 in Scholastic Aptitude Test mathematics is similar to that of boys with 730 in the same.

    All I suggest is that statistics of past instances of the JEE be properly dissected, to come up with fair policies, rather than quick and arbitrary fixes like quotas.

    Once the points (a), (b) and (e) are addressed, some of the current unwanted biases will disappear. Remember: everyone is from different state boards, and the out-of-sync structure is a serious issue.

    (c) - Not much can be done on this front and (d) is entirely dependent on whether the IITs wants to give consideration to a balance of attributes there.

    My guess is that with (a), (b) and (c) you will see some improved representations across not just gender but also boards, income groups and geographical location.

    At that point, the exam will turn into something which doesn't coerce people into the coaching centres simply to help them demystify the syllabus. Once the syllabus and academic requirements for the IIT-JEE exam are transparently known to all, kids will be able to study efficiently and confidently even at home. Such kids will also have a shot.

    Right now such kids do not stand a chance at all.

    Till these issues are addressed, we will have permanent make-shift arrangements like quotas.

    This piece was first published in The Learning Point and has been republished here with permission.

    Get Swarajya in your inbox.


    Swarajya Magazine Cover Image