A recent study outlined in the Harvard Business Review says that the informal shortlist hiring managers put together tends to create gender biases before the formal process even begins. This is at least partially because such a list, often even more influential than the more formal list, is drawn from people the manager knows, who are more likely to be men the higher up one goes in the hierarchy.

After conducting a series of experiments on the matter, the study came to the conclusion that this effect can be at least partially mitigated by expanding the number of people on the short list. In one experiment, people were asked to name three people who'd be good for a role in an action film and then, when they were done, asked them to name three more. In another, people were asked to consider candidates for a job at a tech company and then, just like in the last one, come up with three more names after that.

In both cases, the expanded lists contained far more women than the initial one. In the case of the action movie experiment, doubling the list from three to six generated 33 percent more women; in the case of the tech company experiment, the expanded list had 44 percent more women. The researchers believe this is because the more time people spend brainstorming about a given problem, the more likely it is that the ideas they generate will diverge from the status quo.

Expanding the list, though, only seems to increase gender diversity in areas where men tend to dominate. When running a similar experiment asking parents to select three, then three more, role models for their young child, the researchers found the gender of the child affected list composition. When considering a boy, subjects did indeed choose more female role models on the expanded list in a similar manner to the other two experiments; but when considering a girl the researchers found the expanded lists had fewer female role models than the short list. This is because, the researchers believe, a role model for girls is already seen as a female role, and so making the list longer did not generate more women.