Nature Jobs has a special section on participation of women and under-represented minorities in science and engineering. Some excerpts from this report about the US:
In 2003, 51% of the US population was female and more than 25% of the population was from a minority group under-represented in science: African Americans, Latinos and Native Americans. Women earned well over one-third of the science and engineering doctorates awarded in 2003–04 and African American and Latino doctorates have steadily increased during the past ten years. [...]
But women hold fewer than one-third of all science and engineering faculty posts, and just 18% of full professorships. For minorities, the numbers are below 10% and 6.7%, respectively. When the numbers are dissected at the disciplinary level, many fields find they are doing far worse in hiring talented women and minorities than should be expected, given the numbers of doctorates they award to those groups. [...]
This article looks into one of the major causes for under-representation of women and minorities: BIAS.
Women and minorities must both deal with implicit bias, a problem that is well-documented in the social-science literature, but one that has garnered little attention from the science sector until recently. Dean describes the problem of implicit bias in these terms: "People are most comfortable with people who think and look like themselves."
This type of bias cuts across all divides and has been shown to affect everything from basketball refereeing calls to hiring practices. In addition, a strong gender bias has been found in workplace scenarios, with both men and women consistently overrating men and underrating women in job qualifications (see Virginia Valian's chapter in Why Aren't More Women in Science? (eds S. J. Ceci and W. M. Williams); American Psychological Association Press, 2006).
An interesting case study accompanies this article; it's about 55 chemistry chairpersons who attended a workshop on gender disparity in their field:
Before the workshop, when participants were asked why women were not being recruited, hired and retained in their departments, the participants blamed factors largely beyond their control: too few women in the applicant pool, losing females to other departments and no money for recruiting both members of a couple. After the training on implicit bias, participants were more likely to admit to a lack of commitment or downright opposition to hiring female faculty members, says Geraldine Richmond, a chemist at the University of Oregon in Eugene who is evaluating the workshop's impact.
Participants left with a commitment to implement at least two items within their departments or institutions, such as doubling the number of female applicants in the next faculty search, or advocating subsidized childcare. And the participants agreed to evaluate the effectiveness of their efforts in the future.