The value of connecting unacquainted friends at work shouldn’t be understated, especially for women. In this role, you’re exposed to new information, ideas, and advice, your performance soars, and even your prospects of promotion can improve. But like so many other workplace phenomena, why do men benefit more from brokering these new connections than women?

Researchers have previously tried to answer this question by viewing this process of brokering from a structural angle. Performance is determined by where the individual sits in the network and not according to the individual themselves. The broker is seen, perhaps understandably, to hold a position of power within the network and so variations in their performance are explained by examining how others react to them. For instance, earlier studies have suggested that women who are making these connections perform worse than men in broker networks because these types of networks are typically masculine. The person in the position to make connections is associated with the male stereotype of power and dominance. Women who take on powerful, dominant roles are penalised for violating gender stereotypes that prescribe that women should be communal. As a result, female brokers become the target of animosity, and their reputation suffers. But this explanation doesn’t tell the full story.

My colleague Ajay Mehra and I conducted three separate studies examining the performance of women and men in friendship networks. We found that women who perceived themselves to be brokers in a friendship network not only performed worse than men under the same conditions, but worse than women in networks where everyone is directly connected to each other. It didn’t matter if the women were actually friendship brokers, only that they believed they were.

In contrast, the men’s performance remained constant across both network types. Men have traditionally benefited from open networks where contacts are connected to you but not to each other, because they’re associated with stereotypically male characteristics like status, power, and control. The imbalance is due to a “stereotype threat” which is a pattern that induces performance anxiety in women about how they’ll perform connecting friends to each other. In turn, the anxiety taxes their working memory, reducing their focus on the task at hand. It stirs a fear of being judged and attacks their cognitive ability to problem-solve and self-regulate in interactions with others.

The real kicker is that women often aren’t consciously aware of the threat while it affects their performance by undermining their ability to fully capitalise on the benefits of having new information from their network at their fingertips. It’s a self-fulfilling prophecy that means women who bridge disconnected friends in their network actually do, on average, perform poorer than men. With such strong potential to succeed in this role, we must level the playing field for women so they, too, can benefit from holding such an advantageous position in social networks. I believe there are three ways to do this.

Women need to understand how this stereotype threat can negatively affect their performance simply by knowing it exists. Individuals try to compensate for the negative effects of this anxiety with extra effort. While this might be beneficial with a simple task, in knowledge-intensive businesses such as professional and financial services firms, hospitals, and universities, where problems tend to be complex and ill-defined, it’s not accomplished as easily.

We need to highlight the baselessness of this stereotype and ensure women are aware it exists, along with its triggers. When situational cues signal the stereotype threat, women will experience uncertainty about their performance and unconsciously search for external hints that confirm this fear.

Women need to be given support to reframe the perceived threat as a constructive opportunity. A lot of the successful women I’ve spoken to reframe the broker role in more stereotypically feminine terms. They approach making those connections as a chance to generate opportunities such as bringing people together for collaboration and developing junior staff. More often than not, their actual behavior as brokers is no different from that of a man’s. They are just thinking about it differently.

While being able to connect people puts women in a position of power and dominance, it also clashes with the stereotypical expectations of women that persist. It is no surprise that the effect of this stereotype threat is more pronounced in industries that are stereotypically masculine or where women are underrepresented like mathematics, engineering, and negotiations. And women represent just 29% of the college-educated U.S. STEM workforce.

Breaking down these disparities will involve reframing the act of making connections and emphasizing collaboration rather than masculine traits such as self-interested behavior. Doing so will mean that both men and women will be able to maximise their investment of being a broker within a friendship network.

Raina Brands, PhD, is an assistant professor of organisational behaviuor at London Business School, where she studies gender bias in organisations through the lens of social networks.