Coming late to controversy

In my usual spaced-out and self-centred way, I appear to have missed out on a blogwar-of-sorts about the IIT-JEE. Usually, I would’ve shrugged and moved on, but this war was about gender bias in the JEE! You can’t expect me to resist that! The fact that this is actually being discussed gives me this huge high, so forgive the Tipsiness of this post, and read on!

The consensus appears to be that Abi has lost the blog war (I’m not going to recap it for you here) for a lack of supporting data. Vivek found data showing that there are very few women in engineering courses, JEE or not, and so, he concluded, the idea that the JEE is biased is rubbish.

Now, apart from about two years of JEE-coaching that I have almost-completely blanked out of my mind, and which was more than ten years ago anyway, I don’t know much about the JEE. However, the biases inherent in an examination must be proved in relation to the examination itself, and data in relation to admissions indicates very little about the bias of the exam.

Let me explain. One of the first exams I ever set included a long and complicated fact situation, to which a very simple and obvious principle of law had to be applied, and the question required arguments on both sides to be presented. I thought it was a rather easy question, but it left some students weeping, and me wondering what had gone wrong. The answer lay in the question. The fact situation I used involved a stock exchange, and while it didn’t need students to know anything about the stock market, many of them (in the fourth year of a five-year law course) just blanked out when they saw the facts. They could not relate to the fact situation, it was unfamiliar, and on top of the existing exam tension, it was just plain SCARY! The fact scenario was so unfamiliar that they couldn’t get beyond it to focus on the actual principles involved, which they knew and could’ve easily applied.

What I learnt from that paper was that there are always biases in a paper that influence how students perform. In may case, it was my comfort with corporate/commercial law contexts that led me to set the question in a stock exchange, not realising how many students would be intimidated by it. In relation to entrance exams, this translates into a bias in the ‘gatekeeper’, and therefore, a denial of opportunity. An article in the Yale Journal of Law and Feminism[1] explored the gender bias of the LSAT in great depth,

documenting how the LSAT decreases women’s and (particularly) minorities’ admission opportunities in the 1990’s, even compared to men and Whites who had similar accomplishment levels over four years of college. As a way of studying the impact of current definitions of merit, I compare present admission practices with two admission models based on undergraduate grade-point averages. Either alternative admission model results in the admission of about two thousand more women to ABA schools, and would create overall gender parity in legal education

The impact of ‘current definitions of merit’, as reflected in the choice of questions, the model of test administration and many many other things that make up the test, can only be measured in comparison to other admission models. Women in my classes do better on research projects than on multiple-choice examinations. They do better on exams that allow collaborative work than on those that require individual work. But Kidder actually identifies four sources of exam bias: stereotype threat, speededness and differential guessing, the gendered effects of subject matter selection and biased questions.

Isn’t it time we examine our entrance exams in relation to these, and figure out whether they’re biased? As Vivek pointed out very effectively, we have ‘missing women’ in science, and there are many many many social biases that have led to this. But that is no reason to assume that our exam papers themselves are not tainted by these biases. It’s time we take a hard look at the choices we make and the biases we reveal when we arm our ‘gatekeepers’, isn’t it?

[1] William C. Kidder, “Portia Denied: Unmasking Gender Bias on the LSAT and its Relationship to Racial Diversity on Legal Education”, 12 Yale J.L. & Feminism 1