(seulement en anglais)
"Gender equality must rest on a moral argument. It just is the right thing to do. Full stop," Professor Iris Bohnet says in her book What Works: Gender Equality by Design (Harvard University Press), presented at UNESCO on 20 April 2016. Built on new insights into the human mind, What Works shows how behavioral design can help us improve our societies, often at shockingly low cost and surprisingly high speed. It points out dozens of evidence-based interventions that could be adopted right now and demonstrates how research is addressing gender bias, improving lives and performance.
Wide Angle invites you to read a part of Professor's Bohnet introductory remarks.
As late as 1970, only 5 percent of musicians performing in the top five orchestras in the United States were women. Today, women compose more than 35 percent of the most acclaimed orchestras, and they play great music. This did not happen by chance. Rather, it required the introduction of blind auditions. The Boston Symphony Orchestra was the first to ask musicians to audition behind a screen, and in the 1970s and 1980s most other major orchestras followed suit. When they did so, usually in preliminary rounds, it raised the likelihood that a female musician would advance by 50 percent and substantially increased the proportion of women hired.
To change this, no great technological feat was required, just awareness, a curtain, and a decision. Or, more precisely, a design decision. A simple curtain doubled the talent pool, creating amazing music and transforming what orchestras look like.
Design is everywhere
There is no design-free world. Organizations have to decide how to search for and select future employees. How they advertise open positions, where they post the job openings, how they evaluate applicants, how they create a short list, how they interview candidates, and how they make their fi nal selections are all part of choice architecture.
Much like interior designers or landscape architects, behavioral designers create environments to help us better achieve our goals. They do not define goals, but they help us get there. Behavioral design goes beyond law, regulation, or incentives, although it acknowledges that these are and will remain important. But they do not always work. Based on 41 million observations for the population of Denmark, for example, research shows that tax subsidies have only a tiny impact on savings. Such incentives require people to take action and respond— which 85 percent of Danes fail to do.
In contrast, behavioral designs that do not rely on people reacting to incentives but instead employ automatic mechanisms— such as automatic employer contributions to retirement accounts—do much better. They substantially increased the amount of money retirees have available. We do not always do what is best for ourselves, for our organizations, or for the world— and sometimes, a little nudge can help.
Design is everywhere. We constantly make choices about how to present information, structure interviews, or create teams, and we live day in, day out with the consequences of those choices. By changing the design, we change the outcome: good design can lead to positive outcomes — nudge by nudge. We begin by uncovering the root causes for certain behaviors and designing interventions accordingly. These root causes include one difficult truth: no one is immune from biases.
Stereotypes lead to unintentional discrimination
A few years ago I entered a day-care center at my workplace, Harvard University. I had our young son in my arms. Like millions of parents who have taken their child to a caregiver for the first time, I was extremely anxious. One of the first teachers I saw was— a man. I wanted to turn around and run. How could I entrust this man with the most precious thing in my life? He did not conform to my expectation of what a loving, caring, and nurturing preschool teacher looked like. My reaction was not based on a conscious thought pro cess, but rather on something deep in my gut. Was I being sexist? I fear the answer is yes.
Thankfully, I overcame my biased snap judgment, the teacher proved great, and he became a trusted caregiver. But to this day my gut reaction bothers me. Only about 10 to 20 percent of the elementary school teachers in the United States and many other countries are male. These men face an uphill battle. Just as in orchestras, there is likely an untapped talent pool of elementary school teachers. What is more, society’s failure to draw on that pool of talent matters. A 2015 study by the Organisation for Economic Co-operation and Development (OECD) finds that at age fifteen, boys are 50 percent more likely than girls to lack basic proficiency in reading, mathematics, and science. The presence of male role models can impact what boys believe possible and important for themselves: seeing is believing.
Stereotypes serve as heuristics— rules of thumb— that allow us to process information more easily, but they are often inaccurate. What is worse, stereotypes describing how we believe the world to be often turn into prescriptions for what the world should be. Much psychological research shows that we cannot help but put people (and other observations) into categories. It rarely is a conscious thought process that informs our thinking about demographic groups. Rather, when we learn the sex of a person, gender biases are automatically activated, leading to unintentional and implicit discrimination.
Bias hurts counterstereotypical individuals across gender, race, class, ethnicity, nationality, or caste. Simulations show that even a tiny bias in performance evaluations can lead to huge disparities in representation at the highest levels. Assuming the typical corporate pyramid structure where only a few make it to the top, and holding every thing else constant, one simulation found that a bias accounting for only 1 percent of the variance in evaluation scores led to only 35 percent of the discriminated-against group being represented at the top. Without the bias, each group would have held 50 percent of these seats.
Despite media attention paid to general issues of race and gender, we still know relatively little about the intersection between different social categories— for example, to what degree evidence on white women also applies to African, Asian, Hispanic, or Native American women. Similarly, the research on faltering academic achievement among boys and men, and what to do about it, is relatively young. A series on gender, education, and work in the Economist in the spring of 2015 highlights the challenges poorly educated men in the United States and elsewhere face. They are falling behind not only in school but also in work and society more generally. The series calls for a “change in cultural attitudes”: “Men need to understand that traditional manual jobs are not coming back, and that they can be nurses or hairdressers without losing their masculinity.”
Thankfully, experimentation is becoming increasingly popular. More and more governments are designing policy interventions in collaboration with social science researchers, allowing them to evaluate their impact. Corporations are using advanced technologies and social media to test different marketing strategies and human resource practices. And nongovernmental organizations are running scientifically valid experiments to explore how to decrease homelessness or recidivism most effectively. Still, we should do more. At all levels, we need to create learning environments where people are encouraged to try out something new, possibly fail, and then learn from it.
Gender equality is not just a numbers game
Much has been written about the “business case” for gender equality, and research continues to accumulate. One clear insight is that the answer to what degree closing gender gaps yields economic returns is difficult to determine if outcomes are based on flawed decision processes. Take the example of orchestras. I presume that orchestras benefited from the introduction of blind auditions because curtains allowed evaluators to choose the best performers and build the best team— which also increased the fraction of women.
It is a trivial point but one largely overlooked in the literature. Whether or not the share of women and men in groups, say, corporate boards, is related to company performance does not depend only on the percentage of each gender on the board but also on how the board members, women and men, are chosen, how the boards are organized, and what the rules of engagement and decision making are. Gender equality is not just a numbers game. Numbers matter, but how those numbers came to be and how they work with each other is quite possibly even more important.
A recent study measuring the impact of an increase in the talent pool on the US economy between 1960 and 2008 found that aggregate output per worker had grown by 15 to 20 percent due to an improved allocation of talent. For example, while in 1960 the effective talent pool for doctors and lawyers consisted of white men—94 percent of all doctors and lawyers in the United States were white men— this had changed dramatically by 2008, when the fraction of white male doctors and lawyers had decreased to 62 percent. Casting the net more widely, including women and African, Asian, Hispanic, and Native Americans, had paid off.
A simulation assuming that women are completely excluded from the labor force found that this would lead to income per capita losses of almost 40 percent. Using labor market data for 126 countries from the International Labor Organization (ILO) to calculate the actual gender gaps in workforce participation (as well as in self-employment and pay if available) in various regions of the world, the total income losses are largest (27 percent) in the Middle East and North Africa. In addition, for an increasing number of countries, the talent argument has gained importance as the gender gap in education has reversed and more women than men graduate from college. In the United States, for example, more than half of bachelor’s degrees have been held by women since the mid-1980s, and by the early twenty-first century, almost 60 percent of bachelor’s degree holders were female.
Yet, gender equality is not a magic bullet automatically leading to economic progress. This is why, at the end of the day, the case of gender equality must rest on a moral argument. It just is the right thing to do. Full stop.
Through behavioral design we can move the needle toward creating equal opportunities for female musicians, for male teachers, and for everyone else. Good design often harvests low-hanging fruit, left on the tree not so much because of bad intentions but rather because of the mind bugs that affect our judgment. Behavioral design offers an additional instrument for our collective toolbox to promote change; it complements other approaches focusing, for example, on equal rights, education, health, agency, or on policies making work and family compatible.
But how can we know that a particular design is effective? We can try different strategies and measure their impact. We can examine the effectiveness of behavioral design much like we evaluate the impact of a new drug, running a clinical trial in which people, schools, or even villages are randomly assigned to treatment or control groups. The goal of random assignment is to create groups that are as identical as possible, so that any change in behavior can be attributed to the “treatment.”
Building on what works, behavioral design creates better and fairer organizations and societies. It will not solve all our gender- related problems, but it will move the needle, and often at shockingly low cost and high speed.
* * *
Iris Bohnet, Professor of Public Policy, is the Academic Dean of Harvard Kennedy School and the director of the Women and Public Policy Program at HKS. She is also an associate director of the Harvard Decision Science Laboratory and the faculty chair of the executive program « Global Leadership and Public Policy for the 21st Century » for the World Economic Forums Young Global Leaders.