“Women Leaders still need to be Trailblazers”

Harvard professor Iris Bohnet on why organisations’ inherent biases must be eradicated before we can truly improve gender diversity

Ever felt so flattered and grateful that you’d been offered a new role that you forgot to negotiate your salary? Don’t worry – it even happens to Harvard professors. “It was quite a surprise to me when I was offered a job at Harvard as an assistant professor in 1998,” says behavioural economist and Harvard Kennedy School professor Iris Bohnet. “I didn’t negotiate my first salary. I’m a typical woman.”

Bohnet became a full professor in 2006, and later director of the Women and Public Policy Program, and did a better job at the negotiating table this time around – even though her knowledge about biases still held her back. But she decided to do something about it, realising that her expertise in behavioural economics could be applied to the persistent problem of how to achieve better gender diversity in organisations. Her book, What Works, was published last year. Bohnet talks to People Management about the power of nudge theory, why HR needs to be more scientific and what held back Hillary Clinton’s presidential campaign.

Are men and women just as biased as each other?

Absolutely. On the one hand, the fact that everyone is biased is bad news – because it means everyone is biased. But I often argue it’s good news, too, because it’s not about pointing fingers at anyone. Research shows that our own demographic characteristics are less important [in determining our attitudes towards genders] than who we see. Seeing really is believing: if we don’t see male preschool teachers or male nurses, and if we don’t see female engineers or female CEOs, we don’t associate these jobs with men and women respectively.

How easy is it to overcome bias?

Biases aren’t hard-wired, but it takes a long time for our mindsets to change. What ‘feels’ right, what your ‘gut instinct’ tells you, the thing you’ve always done – that’s often bias. Given that our minds are stubborn beasts in many ways, we have to make it easier for people to do the right thing.

Where does nudge theory come into this?

In the mid-2000s, many behavioural scientists started to think differently about how we can apply our knowledge, arguing that we shouldn’t just use it to describe how people are, or to describe problems that we have, but to prescribe how we can overcome some of the reasons people make mistakes. Nudge, by Richard H Thaler and Cass R Sunstein, argued that, in our collective toolbox, besides hard instruments such as incentives or regulations, and soft instruments such as information, are middle-ground tools: nudges. These are typically more effective than raising awareness and less costly than incentives.

At what stages in women’s careers are biases most apparent?

It’s hard to generalise, because the issues differ by industry. But there is this common pyramid where we have more diversity at entry levels, and then we lose women as we climb up the hierarchy ladder. Thinking about where we start to lose women is crucial for all organisations, both for equality and productivity. You have to ask: are women opting out? Where do they go? Often companies find that, although it looks like many women opt out for family reasons, a large proportion go to work for a competitor.

However, not every industry benefits from the same talent pipeline. If the talent pool is 50-50, is your company hiring 50 per cent men and 50 per cent women? Once that problem is addressed, then you can look into why you might not be promoting proportionately.

What interventions can companies make to design gender equality into their processes and structures?

The beauty of nudges it that they’re the low-hanging fruit. Take, for example, a hiring problem. We can easily de-bias job advertisements in a few minutes, using software that highlights ‘masculine’ and ‘feminine’ adjectives.

Unstructured interviews are also important. The cost of creating one is not super high; you can use one of the tools out there, or create your own list of questions, and that might just take an afternoon of HR’s time. Then, ideally, you would keep track of the questions that are particularly predictive of future performance, and drop those you realise don’t correlate with performance.

But HR should bear in mind that these interventions must be treated as experiments. New drugs are tested with a group that receives the treatment and one that gets a placebo; we can apply this methodology to testing a new hiring protocol – by using that only with a treatment group, and having a control group too. HR needs to help to create learning organisations; initiatives can’t just be ‘done’ or ‘adopted’ – you have to keep measuring what works and what doesn’t.

How can HR professionals think more scientifically?

I think we should apply the same kind of rigour that we have in our marketing and engineering departments to our HR departments. It strikes me as a little ironic that we spend more money on understanding whether a particular product or colour appeals more to men or women than on experimenting with HR processes – even though human capital is our most valuable resource.

This does require a bit of a cultural shift. I think I can safely predict that, in 10 years, most HR departments will be people analytics departments. Many people in HR realise that human judgement is flawed, and that a bit of technology can help them to make better decisions.

Is organic change more effective than government interventions?

There are pros and cons. I have been impressed by the UK’s ability to increase gender diversity on FTSE 100 boards, from about 12 per cent in 2011 to about 26 per cent now. That came about through a coalition of private and public sector leaders, politicians and interest groups working together to make it happen – in little more than four years. It was organic, but also orchestrated. There was a lot of nudging going on, a lot of backroom diplomacy, with chairmen saying: ‘We did it, so why can’t you?’

Although there’s been some criticism that many of these roles have come at non-executive director level, science tells us that often it’s more important that the pilot – and this was the pilot – shows success rather than solves very big problems. It’s one of the few examples demonstrating that you can promote change in a relatively short amount of time, without introducing quotas.

Did gender bias play a role in Hillary Clinton’s electoral defeat?

There is evidence suggesting that women struggle with what we call the competence-likeability dilemma. It’s hard for women to be perceived as competent and likeable at the same time. Clinton definitely struggled with that. You could see her go back and forth between being the commander-in-chief and power posing, and being a mother figure and using inclusive language.

The campaign format isn’t conducive to women’s success; very competitive environments are just not helpful to a woman’s voice. I mean that both figuratively and literally: it is more likely that women will be perceived as screaming, being shrill, or seen as getting emotional when we talk loudly. This is gender bias at work.

The other problem is that we make big inferences about female leaders based on a very small sample. It’s harder to learn from a small sample because the information is noisier. Take Margaret Thatcher, for example – she became a role model and now we think that female politicians have to be tough, they have to be married and they have to act like her.

Sadly, I fear the competence-likeability problem is going to stick with us for a while. The women who will bring leaders up to a critical mass still have to be trailblazers.

Iris Bohnet

Swiss-born Iris Bohnet became an assistant professor at the Harvard Kennedy School in 1998. After getting tenure in 2006, she was appointed director of the Women and Public Policy Program in 2007. Bohnet is also co-chair of the Global Future Council on Behavioural Sciences of the World Economic Forum.

(reproduced from People Management magazine February 2017)

 

facebooktwittergoogle_pluslinkedin