Mittwoch, 4. Mai 2016

Interview: Prof. Iris Bohnet, Harvard University

Iris Bohnet, Professor of Public Policy, is a behavioral economist at Harvard Kennedy School


How do we avoid “unconscious bias“ to build a better society regarding gender equality?

Avoiding unconscious bias is almost impossible. Instead, we have to make it easier for our biased minds to get things right, or put differently, break the link between our biased beliefs and our actions. Awareness of one’s biases certainly is a first step in the right direction but to translate it into behavior, more is required than awareness. Much research suggests that awareness alone is not enough. However to learn about their own biases, people should take the Implicit Association Test

Thousands of people have already taken the test and learned that they, too, were biased against people of certain races, cultures, genders, religions and even looks. For example, we tend to prefer tall men to short men or are more likely to trust more attractive than less attractive people. Of course, the evidence shows that attractive people are not more trustworthy but behave just like everyone else. But our biased minds quite literally cannot “see” this and instead, associate good looks with good behaviors.

Once aware of our biases, we can start to design around them, to keep them from affecting our behavior. For example, organizations might want to blind themselves to the “looks” and more generally, the demographic characteristics of job applicants. New software such as, e.g., APPLIED or UNITIVE, makes it easier for companies to do so.  

The software allows hiring managers to quite literally liberate their minds to focus on talent instead of whether someone looks the part. My book offers 36 designs to de-bias organizational practices and procedures in talent management and elsewhere to level the playing field for everyone.

I do not think we have given equal opportunity a fair chance yet—but behavioral design allows us to move the needle towards more equality in significant ways. Blind auditions on major symphony orchestras in the United States such as the New York Philharmonic or the Boston Symphony orchestra has helped increase the fraction of female musicians from about 5 percent in 1970 to now almost 40 percent. This is in stark contrast to many of the leading European orchestras that have not introduced screens behind which musicians audition and where there still are only about 10 to 15 percent female musicians.

What are the evidence-based interventions your approach of behavioral design suggests to be adopted in a new solution in relation to „conventional“ diversity training programs?


The book contains dozens of such interventions that in addition to being based on evidence are also rather inexpensive and can be adopted rather quickly.  They can be as simple as screen savers and room decorations. For example, a study from the University of Washington demonstrated that by changing the decorations of a computer science classroom from a Star Wars theme to gender-neutral images of nature and art, female students’ associations between women and careers in computer science were strengthened. 

Another study displayed pictures of male and female political leaders on screensavers before participants were due to deliver a speech. They found that women who were exposed to the female role model gave significantly higher rated and longer speeches (the screensavers did not affect men either way).

Sadly, we have not found similar positive effects of diversity training programs. Either, the impact is not measured at all or in the few cases where a rigorous evaluation was conducted, the evidence was not encouraging. While companies are estimated to spend about $8Billion a year on such trainings in the United States alone, at this point, the evidence suggests that this is not money well spent.

But you also want to pay attention to how you craft groups Pushing for diversity can increase the “collective intelligence of groups”, research shows, as the group can take advantage of complimentary perspectives to maximize performance. However, numbers matter. People should avoid creating skewed groups dominated by one gender, race, or social type. 

The lone member of a minority group is more visible and may be reduced to his or her demographic characteristics. Such “tokenism” makes it more difficult for members of the minority to contribute their complementary expertise. Organizations pondering how to best create teams should try to have a “critical mass” of such minority representatives. Although an exact number is hard to determine, useful rules of thumb are at least one-third of a large group or at least three people in absolute numbers. This can help avoid the dynamics of social categorization.

Are there any cultural differences in tackling the gender bias and income inequality in order to improve lives around the world?

The book includes evidence from very different settings, including in local politics in India, healthcare in Rwanda, corporate boards in the UK, or the financial sector in the United States. The research tends to target features of human thinking that are present across the globe, albeit with some nuances. 

People might want to check out the Women and Public Policy Program’s searchable online platform at Harvard, the Gender Action Portal (GAP) where we summarize research for decision makers who might want to tackle such diverse issues as, increasing girls’ school participation and completion rates, doing something about the enormous tragedy of “gendercide,” the sex-selective abortion and neglect of young girls (the UN estimates that about 200 Million women and girls are “missing” due to gendercide) or increasing the fraction of female directors on corporate boards. They would find evidence on all of these questions on GAP.

But I place a lot of emphasis on continuing to measure the impact of interventions whether they are introduced in a law firm in Brazil, a school in Kenya or an online job search platform in China. We will not know whether or not something works unless we measure. And we do not measure nearly enough. 

In many ways, we have to bring the same kind of rigor and scrutiny that we employ in our finance departments to our human resource department. I am quite certain that “big data” or “people analytics” will revolutionize HR in the next ten years. In fact, it is happening already now. One of the tools HR-managers might want to look at is EDGE, software helping companies measure how they do in terms of gender equality. 

Companies being EDGE-certified include Deloitte Switzerland, L’Oreal USA, Banco Compartamos, Mexico as well as the World Bank, among others. In my book, WHAT WORKS: Gender Equality By Design, I suggest that DESIGN stands for D as in data, E as in experimentation and SIGN, as in signpost, where we put up signpost with guide all of us in the right direction.  


Thank you very much.




Iris Bohnet, Professor of Public Policy, is a behavioral economist at Harvard Kennedy School, combining insights from economics and psychology to improve decision-making in organizations and society, often with a gender or cross-cultural perspective.


Her new book „What Works – Gender Equality by Design“ (2016) is just published by Harvard University Press.









Keine Kommentare: