Mapping gender diversity at MIT

A trio of researchers has created and published a data visualization map that examines trends in undergraduate gender diversity at MIT. The big reveal is heartening: Over the past 20 years, MIT’s female undergraduate population has risen to nearly 50 percent of total enrollment and such growth has been sustained across almost every department and school.

Professor of aeronautics and astronautics Karen Willcox, researcher Luwen Huang, and graduate student Elizabeth Qian devised an interactive map to show these aggregate trends, and much more. The tool, using data from the MIT Registrar’s Office, allows users to explore gender diversity on a class-by-class and department-level basis, to see links between classes, such as prerequisite requirements, and to conduct keyword searches to reveal variations in related subjects across MIT.

“MIT should be proud of the leadership it has shown,” says Willcox. “The positive trends in gender equity are not seen in just one or two departments, but literally across the spectrum of science, engineering, arts, humanities, social sciences, management and architecture. One of the unique features of our tool is that it provides insight at the subject level, going deeper beyond aggregate statistics at the major level. We hope that this will be a basis for data-driven decisions — for example, by understanding what about a particular subject’s pedagogy makes it appeal to a more diverse audience.”

The map appears as a series of discipline-based ball-and-stick clusters, with each node representing a class. The size of the node indicates the class’s total enrollment. The color of a node, ranging from teal (fewer women enrolled) to salmon (more women in enrolled), represents the percentage of women in a particular class, and helps to illustrate how diversity has changed over time.

For example, in a slice across classes in MIT’s Department of Electrical Engineering and Computer Science (EECS) in 2006, the nodes appear as light and darker teal, indicating enrollments of less than 25 percent women. Fast forward to 2016, and the same slice has node colors all in shades of salmon, indicating female enrollments of 35 percent or more. In part, this change is a reflection of the steady increase in total female EECS majors, particularly over the past six years. However, since the analysis is conducted at the class level, this change is also a reflection of more women from other majors taking computer science classes.

“It is gratifying to see the change in composition of our EECS student body,” says Anantha Chandrakasan, former department head of EECS and now dean of the School of Engineering. “While it is true that we have had a dramatic increase in [computer science and engineering] majors, female enrollment has nearly tripled in the past five years. It’s a useful model for us to consider as we are improving gender equity across the school.”

Willcox credits the positive momentum in EECS to several different elements, saying, “anecdotal evidence suggests that the pedagogical reform undertaken by EECS in 2008 has played a large role.” She also points out the important role of leadership, namely Chandrakasan’s support of studies such as the EECS Undergraduate Experience Survey and his commitment to programs such as the Women in Technology Program and Rising Stars, an effort to bring together women who are interested in careers in academia.

Enrollments in the Department of Mechanical Engineering have achieved similar gender parity. This is especially impressive given that the national average of female undergraduate majors in the field is 13.2 percent. Willcox again highlights the efforts made by another leader, Mary Boyce, the first woman to head that department from 2008 to 2013 and now dean of engineering at Columbia University. The results of an internal study announced in June, suggested that the department's ongoing proactive approach — revamping the curriculum, enhancing recruitment efforts — played a part in their success.

“The map, of course, cannot reveal specific causes of changes in gender diversity, but it does provide a place to begin a conversation,” says researcher Luwen Huang, who is an expert in visualization design. “The interactivity of the map was designed to encourage the user to explore, discover connections across classes, and ask questions.”

The researchers caution that looking at department-based data only provides one view. In the case of EECS, a deeper dive shows that introductory programming classes have historically had high female enrollments, but that finding may be deceptive. "When you look at introductory courses like 1.00 (Engineering Computation and Data Science) and 6.00 (Introduction to Computer Science and Programming), you see high levels of female enrollment,” Willcox explains. “That’s not because there are more women in those fields, but likely because women might lack the preparation and/or the self-confidence to skip introductory classes.”

Biannual surveys of MIT undergraduates and other internal reports seem to bolster such a supposition, suggesting that women at MIT may experience negative stereotyping and feel less confident than their male counterparts. Lower or higher female enrollment in certain classes and departments may also be due to a variety of other factors, from job prospects to the influence of peers to level of interest in the subject matter.

The data and tool provide a starting point to begin such analysis and to take potential actions. Being open about data, sharing data, and being data-driven are valuable forcing mechanisms, says the team, and a hallmark of MIT’s ethos of transparency. Further, having a visual map of gender diversity across MIT, they say, is literally eye opening.

"This map provides ample evidence that our efforts to enroll a diverse undergraduate class have had a dramatic impact on MIT,” says Ian A. Waitz, vice chancellor and the Jerome C. Hunsaker Professor of Aeronautics and Astronautics. “However, while these demographic trends are impressive, they are not sufficient. We must continue to work hard to create an inclusive, welcoming environment for all.”

de MIT News http://ift.tt/2y3c91c