women sitting around a conference table having a discussion
The tech industry has a gender diversity issue, and it's getting worse PHOTO: wocintechchat.com

Facebook COO Sheryl Sandberg once said, “The world has gone social. And women are more social than men.” 

This isn't merely Sandberg's opinion. Sandberg has access to copious Facebook data to back up her claims. It's therefore somewhat ironic that the very company that provides the platform women have embraced employs more than twice as many men as women

Tech Has a Gender Diversity Problem ... and It's Getting Worse

The gender diversity at high tech firms like Facebook has actually grown worse over the past several decades. Google’s diversity woes are well documented and have rightly attracted the attention of not only regulators, but research institutes like Stanford’s Peace Innovation Laboratory. Questions are being raised about why Silicon Valley is so awful to women.

We know this: the business world is becoming more interconnected. Silicon Valley is associated with pervasive industry-wide networking. 

Our network research results show women are better networkers than men, both outside and inside the enterprise. The only place their superiority is not evident is in the ranks of senior management. Why is this so?

Unconscious Bias Hampers Gender Diversity Efforts  

I interviewed Stanford’s Mark Nelson and Rosanna Guadagno recently to launch a collaborative effort addressing gender diversity issues.

The Stanford Peace Innovation Laboratory (PIL) has been researching how digital media can facilitate engagement across diverse social groupings. As part of this, Peace Innovation Labs around the world are building a standard data framework to capture pervasive real-time mediated-social interaction data. This framework would be used to facilitate more peaceful outcomes, by design.

PIL's current focus though is on the gender diversity issues of its neighbors in Silicon Valley. Guadagno had previously published research on the gender divide in Facebook use, and found women used Facebook to develop and maintain longer-term relationships, whereas men tended to be more task- and achievement-focussed in their use. 

We shouldn’t be surprised, as this is consistent with role expectations beyond the digital environment. The question here is whether woman’s preferences for networking might change when the context changes from an "outside of work" social context to a more task-oriented, male-dominated enterprise context?

Nelson’s interest in the real-time interactions data captured by the SWOOP social analytics platform stemmed from his observations that at times, the diversity policies created to combat gender bias could do more harm than good. The interviews and surveys many organizations use to surface gender bias often suffer from an unconscious bias by their respondents.

Nelson felt the only way to uncover unconscious gender bias was to look at actual real-time behavioral data on operational engagement — in other words, how men and women really interact in the workplace.

Collaborative and Networking Behavior, By Gender

Nelson said, “Stanford Peace Innovation Lab is excited about tools with capabilities for doing real time, high resolution measurement and analytics in the real world, rather than only in a lab environment.” 

We started our collaboration by directing our enterprise social networking benchmarking to benchmark gender performance within enterprises. We tested this by looking at the online networking patterns of employees in a large financial institution over a period of six months. 

We were able to identify individual collaboration patterns of active staff members on the online collaboration platform who were identified by gender. We identified over 7,500 staff members, who fell in an almost equal male/female gender split.

The results were stark. Women surpassed men in collaboration in nearly every dimension. It appeared the female networking behaviors did translate to the enterprise.

Networking Dimensions

Of the 23 collaborative benchmarking measures we make at an individual level, 12 showed a statistically significant difference between men and women. We normalized the measures for comparative purposes. 

As seen in the figure above, women led in all dimensions. However, for one of those dimensions, "%Broadcaster," a higher score is seen as a negative, making this the only dimension where men outperformed women.

A Cross-Gender Networking Analysis

A look at cross-gender interactions found some further differences:

From / To

  Women

  Men

Women

67 percent   

58 percent  

Men

33 percent

42 percent

Women to Women connections were denser than Men to Men connections. Women reached out to male colleagues far more (58 percent) than men reached out to women (33 percent). The following diagram compares the women-only network with the men-only network:

comparison of women's networking habits and men's in the workplace

The red lines show reciprocal relationships. The results support Guadagno’s findings on gendered social networking interactions outside the enterprise. In other words, the networking behavior of women did not change when they were working. 

Said Guadagno, 

“This data is fascinating in that it illustrates the power of big data to demonstrate meaningful differences in the ways in which men and women treat each other in the workplace. I would love to see how these differences in communication relate to work outcomes such as productivity, creativity, as well as raises, promotions, etc. Furthermore, these results are also consistent with the framework we have developed at Stanford Peace Innovation Lab to use mediated interaction to facilitate positive, pro-social engagement across difference boundaries.”

The Gender Paradox

Returning to the gender paradox: why does it exist?

We don’t have a clear answer, but here are some theories:

  • Networks have yet to supplant the task-driven hierarchies that favor men. Time will correct the imbalance as networks continue to replace hierarchies as the preferred organizational model
  • Our sample set was not structured by seniority levels, so perhaps these stronger women’s networks only exist at lower levels of the hierarchy, and therefore their potential influence may be aligned with a more union-style movement
  • Women who rise to senior management ranks find they need to network more actively with men, to progress their careers, at the expense of helping other women to attain more senior ranks. We had previously identified this with women on boards
  • Unconscious bias of men in senior ranks against helping women reach similar levels. Some organizations have deployed quota systems  to combat this bias, but as Nelson observed previously, such policies can also have the reverse effect, by suggesting senior woman achieved their positions on the basis of something other than merit.

Is Behavioral Change Within Reach?

We believe the unconscious bias against women in senior roles is the main culprit. If we surveyed all the men in our sample, we suspect few would believe they chose to interact with men more than women by close to a 2:1 ratio. We agree with Nelson — it’s only through the provision of real-time interactions data that we can address such unconscious bias.

As Nelson said,

“The ability to measure and analyze what’s happening in any organization, right down to the individual level, in real time, is a big step toward making a previously invisible world visible. This not only allows us to measure and benchmark exactly what isn’t working. More importantly, it lets us test for interventions that measurably reduce gender bias in the workplace, by increasing positive, value-creating behavior across gender boundaries. This will eventually provide organizations with a growing suite of thoroughly proven tools to improve positive behavioral outcomes across group boundaries. 

"As we face repeated, egregious examples of discrimination against women in the tech industry in Silicon Valley, tools like this will enable tech entrepreneurs — we call them peace entrepreneurs — to build technology interventions driven by quantitative behavioral data that measurably increase positive engagement across difference boundaries.”

This is not simply about designing organizational level diversity policies — in many cases they already exist. This is about holding a mirror up to both men and women, showing them the nature of their day to day interactions. 

Once individuals can see how they're interacting, they will be in a far better position to change the way they interact. Perhaps only then will we gain the behavioral changes that the policy makers have been seeking for decades now.