How to get more women to work in tech

Lucia Del Carpio, INSEAD Assistant Professor of Economics, and Maria Guadalupe, INSEAD Associate Professor of Economics and Political Science

Role models and reassurance about potential success can impact women’s career decisions.

‘What made you decide to join your profession?’ When people are asked this question, they often give answers such as: “I liked such-and-such subject in school”, “I was inspired by so-and-so” or “It pays well”. In most cases, there is an unspoken assumption: They also believed that they had a reasonably good shot at succeeding in this career.

According to the classic Roy model, individuals choose their occupation based on a rational decision process. They try to maximise their income, taking into account their skills and relative rewards in different sectors. It’s all about optimisation: People go where they believe they will have the greatest comparative advantage.

In the case of women, assumptions that certain sectors do not welcome them may limit their willingness to join. Consider the coding profession in Peru, where women represent only 7 percent of the workforce. Is this under-representation strictly a matter of personal preference, or is it at least partially due to stereotypes and social norms that dissuade women from choosing this sector? Could some women shun a career simply because they don’t believe they could thrive in it?

In two field experiments in Latin America, we observed that advertisements that included a female role model and also corrected misperceptions about women’s ability to pursue a career in tech significantly increased the number of suitable female applicants to a software coding bootcamp.

A bias-correcting intervention

We partnered with a non-profit organisation that helps young women from low-income backgrounds become digital coders via an intensive five-month training programme. This international organisation was interested in increasing application rates.

In our first experiment, which took place in Peru, we created two versions of the programme registration webpage. Both included basic information about the training as well as the application form. However, in one version, we added a message to correct biased perceptions and beliefs about the role of women in tech. This short message had three core components:

1. It asserted that women can be successful in the tech field.
2. It emphasised that the training gave access to a network of women in the sector.
3. It told the story of a recent graduate (role model).

Though we did expect our intervention to have an impact, we were stunned by the results. Reassurance about career prospects more than doubled the number of applicants: 15 percent of women exposed to our bias-correcting message applied for training vs. only 7 percent of those who visited the basic webpage.

In addition, when we compared the top 50 candidates in each group, those from our treatment group garnered significantly higher scores on cognitive and tech-specific ability tests administered during the selection process.

What makes the needle move

We then ran a second experiment, this time in Mexico, to figure out which of the three components of our bias-correcting message resonated most with applicants. To do this, we created four versions of the training provider’s registration webpage. The control one included the full message. In each of the three other versions, we skipped one component: One didn’t mention how women can be successful in tech, one didn’t emphasise the existence of a peer network and one didn’t feature a role model.

Based on this second experiment, we concluded that the presence of a role model and the assertion that women can be successful in tech are most important. Their absence reduced applications by 23 and 18 percent, respectively, vs. the control group. (The version that didn’t mention the peer network yielded about the same application rate as the control one.)

Deflecting self-segregation due to biases

By now most of us have heard that women earn 79 cents on the dollar compared to men. While it’s true that women with full-time, year-round jobs make less money than their male counterparts, most of this gap isn’t due to straight-up sexism by discriminatory employers. To a larger extent, it reflects the fact that women tend to work in lower-paid, female-dominated sectors.

Although not a panacea (and none exists), one way to lessen the gender wage gap, over time, would be to nudge more women towards better-paying careers. The tech industry is only one example. Not all women will be interested in such careers and that’s fine. The idea is to ensure that gender norms and self-segregation biases do not unnecessarily steer away women, notably those who do show interest and aptitude, from certain careers.

Our results indicate that in a male-dominated sector, simple tweaks to the ‘selling message’ can significantly impact women’s career decision making. HR departments, policymakers and schools may want to take note. Meanwhile, all of us should cultivate an awareness of how social identity cues and prescriptions can have subtle yet far-reaching consequences on women’s behaviours.

Lucia Del Carpio is an Assistant Professor of Economics at INSEAD.

Maria Guadalupe is an Associate Professor of Economics and Political Science and the Academic Director of the Randomized Control Trials (RCT) Lab at INSEAD.