Paris’s prestigious postgraduate school Ecole Normale Superieure was in for a surprise this summer when results of its admission exam for literature came out.
A whopping 77% of the successful candidates for the main literature section were women, significantly higher than the 59% average over the past five years, according to ENS admission lists published in the French official gazette.
The reason? Entrance to the elite Latin Quarter-based institution’s program this year was based entirely on a blind, written test that didn’t reveal the candidates’ identities. An oral exam that usually accompanies the written test had to be canceled because of the pandemic. Of the 72 people admitted to the section, 55 were women.
The result didn’t go unnoticed. For Sandra Lapointe, a professor of Philosophy at McMaster University and a research affiliate at Ryerson University’s Diversity Institute and Future Skills Centre in Canada, the outcome was a sign of unconscious biases that can emerge during an oral presentation.
Her tweet on the matter triggered scores of responses, with many concluding that the process was unfair, suggesting a gender bias at the school.
Paris ENS is France’s top school and agregation there is highly competitive. This year, due to COVID they cancelled the orals, only had the (anonymous) written exam. As a result ~80% of successful candidates were women. 8-0. Eighty.
Women usually represent ~40% of ENS agrégés— Prof Sandra Lapointe (@mslapointe) August 19, 2020
For Lapointe, however, the lesson is more nuanced. After all, women did make up a majority of those admitted even before this year.
Instead, for her the case provides the perfect conditions for a small-scale experiment, demonstrating that the school needs to take a closer look at its oral exam.
“Actually, this sends a signal that should make people ask relevant questions,” she said in a phone interview.
The school said in an emailed statement that “it is far from indifferent to the biases and logic of self-censorship and selection that can operate, in particular on the basis of gender and social origins.”
As it does every year, it plans to take a close look at the results, particularly at variables such as gender, geography and social context, it said. Of particular concern to the school is the underrepresentation of women in the Sciences, ENS said. Only 28% of those admitted to the Sciences were female, compared with 62% for Letters.
The school sees the oral exams as a key part of the admissions process in Letters since “they correct biases attached to the written exams,” it said.
For Lapointe, while it’s impossible to draw specific conclusions about bias in the ENS case, “in general, the number of women admitted increases when exams are anonymous, as researchers have shown many times.”
For instance, blind orchestra auditions increase the chances of female musicians being selected. When a screen is used to conceal candidates from the jury during preliminary auditions, the likelihood that a female musician would advance to the next round jumped by 11 percentage points, and during the final round, “blind” auditions increased the likelihood of female musicians being selected by 30%, according to a Harvard University study.
On the other side of the balance, investors preferred pitches presented by male entrepreneurs compared to those made by female entrepreneurs, even when the content was exactly identical, another Harvard study revealed.