It took a 2-hour meeting back in 2004 with a stately transwoman named Madhu for me to realise that my “holistic” comprehension of gender was in fact, profoundly flawed .
Madhu is a ‘Hijra’, part of India’s transgender community, comprised of transpeople, eunuchs, intersex persons and other sexual minorities. At the time of our first interaction, I was an undergraduate student based in the city of Chennai, capital of the South Indian state of Tamil Nadu. Madhu was a spokesperson for her community in the city, often dealing with student groups and NGOs to tackle the slew of problems with which the community was constantly grappling. A recurring theme was finding gainful employment outside of prostitution – something into which Hijras were often coerced, owing to rampant hiring discrimination based on their sexuality.
The first time I met Madhu was at a meeting organized by a student group at my university with herself and some of her colleagues, to discuss these problems at length and derive viable solutions.
Madhu’s personality was as vibrant as her bottle-green sari and the large vermillion bindhi on her forehead. It was not long until we were engrossed in her story. With candour, she recounted how she had never felt at home in her formerly male body, a sense that began revealing itself to her more acutely from the start of her teenage years. When she told her family of her desire to physically transition into a woman, they disowned her. She then fled from her village to the city, and underwent the excruciating pain of non-medical castration, nearly facing death in the process.
A week after our heart-to-heart with Madhu, I had a thought. At the start of the meeting, I realised, I had made mental references to Madhu as ‘him’ and ‘he’ — but as the meeting concluded, Madhu was forever after, ‘her’.
I wondered: was it possible that I had pre-existing, implicit biases towards Madhu (and perhaps all transwomen), that caused me to think of her as a man, even prior to meeting her? What were the implications of these biases, and could I remedy them?
Years later, I found clues to these questions in an ostensibly unlikely place: the world of Behavioral Science.
Behavioral Science, Democratized
We make 35,000 decisions each day, often in environments that aren’t conducive to making sound choices.
At TDL, we work with organizations in the public and private sectors—from new startups, to governments, to established players like the Gates Foundation—to debias decision-making and create better outcomes for everyone.
Why My Bias Against Madhu Matters
The past decade has witnessed several behavioral studies on the toxicity of gender biases, which manifest themselves in several ways. With Madhu, my biases began with perceiving her as male instead of female, since my instincts sought to associate her with her assigned gender at birth, which was male.
Yet, given that I bore Madhu no ill will from the start of our meeting, did my implicit biases even matter?
A host of literature on implicit biases and gender suggests that the answer to my query is an unequivocal ‘yes’. Moreover, this ‘yes’ applies to biases against people across the gender spectrum, with implicit biases adapting themselves and donning different costumes to suit different gender identities.
A 2017 article describes a survey on transmen hailing from workplaces across the USA, who have had the experience of working first as women, and then as men . Their description of gender bias and discrimination in the workplace is complex and exceedingly layered. To begin with, many transmen reported that they were treated better in the workplace after physically transitioning into men, as compared to their earlier experiences as women. A quote from one respondent is telling: “As a male, people assume that you know what you’re talking about. As a female, they assume that you probably don’t.” Still other respondents described instances of discrimination only after colleagues realised they were transmen. The survey strikes at the core of how implicit biases are as multidimensional as the very gender identities against which they are prejudiced. The consequences of these biases are as diverse as they are dire: from choosing not to hire someone due to their gender identity, to paying them less than they would a cisgender straight male, or excluding several staff from having access to a toilet whilst at work.
Implicit biases are similarly pernicious for women in the workplace. A study on hiring bias against women for jobs requiring a certain degree of mathematics expertise finds that “both male and female hiring personnel hired men twice as often as they hired women — despite similar outcomes on the math tests.” The researchers attribute part of this bias to interviews, where women are more likely to play down their successes than their male counterparts. However, the study rightly notes, “If ability is self-reported, women still are discriminated against, because employers do not fully account for men’s tendency to boast about performance”.
The fact that implicit biases are so varied, and often individualised, further convolutes matters, as they afflict even those who see themselves as without prejudice. The use of the pejorative phrase, “That’s so gay” makes for an apt example . A study on using the phrase shows that the harm runs deep. While the phrase is blatantly derogatory and causes immediate unpleasantness for the target, there are also long-term effects, since saying “that’s so gay” exacerbates the perpetrator’s implicit bias toward gay people, perpetuating a cycle of more acute bias and discrimination.
Judging by the literature, therefore, someone like Madhu who hails from a sexual minority has precious little chance of going a day without encountering some form of bias- be it while waiting for a bus at the station, or being screened for a job interview.
What makes matters worse is that the subject of implicit bias does not apply solely to gender — it has a series of implications, an apt example being the vast literature on biases and race, with consequences ranging from race-based hiring discrimination to the higher rates of incarceration for people of colour.
There Is Always Hope: How to Combat our Biases
The story so far seems desperately bleak. Yet, as recent research into debiasing has shown, there are tactics for combatting our biases that offer hope.
A study by Broockman and Kalla (2016) suggests that the key to alleviating biases against transgender individuals could be as simple as looking to engage with perspectives from ‘the other side’ (i.e., those who hold discriminatory views toward transgender individuals) . Their study involved canvassers actively seeking to engage with voters who held anti-transgender viewpoints by knocking at their doors and engaging in brief conversations. Their report states, “[h]ere, we show that a single approximately 10-minute conversation encouraging actively taking the perspective of others can markedly reduce prejudice for at least 3 months… A randomised trial found that these conversations substantially reduced transphobia, with decreases greater than Americans’ average decrease in homophobia from 1998 to 2012.” The gender identity of the messenger did not change these results.
Another debiasing attempt came from Morewedge et al . The authors designed an experiment which involved a training video on biases, followed by a video game designed to “elicit and mitigate” specific biases- a tactic they deemed largely effectual.
The more we learn about implicit biases, the more they seem to reflect the dual-systems approach put forth by Kahneman (2003), which gained prominence in his Thinking Fast and Slow. Kahneman’s contention is that people are subject to two distinct modes of thought, dubbed System 1 and System 2. System 1, which can be broadly construed of as intuition, is instinctive, forming instantaneous impressions based on heuristics, which can then lend themselves to cognitive (and implicit) biases.
The AI Governance Challenge
The real hope in combating biases like my own against Madhu during our first meeting, lies in System 2, which involves active contemplation, or ‘slow thinking’. This type of thought, Kahneman argues, can override the heuristics of System 1 when they lead to subpar decisions — thus preventing a heuristic from becoming a bias.
All in all, we are still in the early stages of our understanding of both, the ways in which implicit biases shape our behavior, and the ways we can actively combat them. As is often the case, a sound place to start is to actively challenge our own prejudicial perceptions when they lead us to conclusions that are harmful to those around us.
At the very least, we owe it to the millions just like Madhu, who should not have to wage war on discrimination each day simply to be themselves.
 Names have been changed for this article, to protect the privacy of individuals.
 A transgender person whose gender identity is male, and was female at birth.
 Nicolas, G., & Skinner, A. L. (2012). “Thats So Gay!” Priming the General Negative Usage of the Word Gay Increases Implicit Anti-Gay Bias. The Journal of Social Psychology, 152(5), 654-658. doi:10.1080/00224545.2012.661803
 Broockman, D., & Kalla, J. (2016). Durably reducing transpho2bia: A field experiment on door-to-door canvassing. Science, 352(6282), 220-224. doi:10.1126/science.aad9713
 Morewedge, C. K., Yoon, H., Scopelliti, I., Symborski, C. W., Korris, J. H., & Kassam, K. S. (2015). Debiasing Decisions. Policy Insights from the Behavioral and Brain Sciences, 2(1), 129-140. doi:10.1177/2372732215600886
About the Author
Namrata Raju is currently pursuing a Master in Public Administration degree at the Harvard Kennedy School. Before this, she worked for 7 years on consumer behaviour research, predominantly in the MENA region and other emerging markets.