Beyond the Numbers on Gender and Research

Gender diversity in the sciences is often discussed in terms of numbers: build research teams with more women on them and innovation will come. That may be true; research suggests it probably is. But a new paper seeks to push science’s gender diversity conversation beyond just composition of teams, to research methods and research questions -- along with how to manage each in different disciplinary and organizational settings. It also suggests that in paying more attention to research methods and questions, research teams might diversify as a result.

“To realize its full potential, gender diversity needs to be supported by careful stewardship and management techniques across four interdependent domains -- from research teams to the broader disciplines in which they are embedded to research organizations and ultimately to the larger societies that shape them through specific gender norms and policies,” reads the study, published in Nature: Human Behavior. “Understanding how these domains interact -- i.e., how policies and practices in one domain shape developments in the other domains -- is crucial to maximizing the benefits of diversity for science.”

Londa Schiebinger, John L. Hinds Professor of History of Science at Stanford University, and the paper’s senior author, said this week that “we in North America and Western Europe have not been entirely successful increasing the numbers of women in science, despite our many efforts.” And one reason for that is the limited focus on issues of participation, she said, at the expense of considering “how science is done.”

Speaking for herself and her colleagues, Schiebinger said it’s “our hypothesis that attending to diversity in research methods and diversity in questions asked may also lead to greater diversity in research teams.”

She noted that in a separate analysis of 1.5 million medical papers, she and colleagues saw a link between diversity in participation and methods, specifically between author gender and attention to gender- and sex-specific analysis of phenomena. Causality could not be established, Schiebinger said, but her team is now launching a large empirical study to understand relationships between diversity and creativity and innovation.

Numbers Still Count

The new paper does not suggest ignoring team composition, which it says has important implications for scientific diversity by virtue of cognitive diversity -- which research, in turn, suggests can heighten creativity and the search for new solutions. Indeed, more science-specific research on this topic is needed, the paper argues, and those few studies focusing on science tend to count citation rates, publication productivity and patents, as opposed to other measures of impact. Namely, the paper suggests societal relevance, which the National Science Foundation already considers in funding proposals.

Another reason that this kind of gender diversity shouldn’t be ignored, the paper says, is that gender-diverse teams may encounter higher levels of conflict than more homogenous teams, and that should be carefully managed.

Methods

Still, approaching gender diversity through diversifying research methods, including through research and analysis that considers effects by specific genders or sex, referred to as "GSA," merits more attention, the researcher argue. Schiebinger’s Gendered Innovations project, which includes such case studies as breaking the female postmenopausal paradigm, is one example.

“Integrating GSA into research design can lead to new insights that enhance the external validity and precision of scientific research with human outcomes,” the paper argues. Beyond osteoporosis, which also affects many men, well-known examples of GSA include heart disease in women, pregnant crash-test dummies, machine translation, genetics of sex determination and water infrastructure for sustainable development.

As to evaluating how gendered methodological approaches might influence research outcomes, the paper argues that granting agencies’ polices can be examined to see whether GSA is a funding criterion. Peer-reviewed publications can be analyzed for GSA, as well, the paper suggests, noting that the European Commission has evaluated publications for GSA by scientific field and country. Between 2010 and 2013, the highest proportion was found in the social sciences (7 percent of publications), health sciences (4 percent) and humanities (3 percent) with the natural sciences and engineering showing 0 percent. The Nordic countries, which are recognized as global leaders with respect to societal gender equality, were top performers.

Future research might also develop methods to evaluate the value of GSA to society, in terms of human well-being and economic impact, the paper says, since “Doing research wrong costs lives and money.” The researchers note that between 1997 and 2000, 10 drugs were withdrawn from the U.S. market because of life-threatening health effects, eight of which posed greater health risks for women than for men.

Research Questions

A third approach to gender diversity concerns diversity in research questions, the paper says, evaluated by how the “entrance of new actors into scientific disciplines -- whether women into traditionally male disciplines, such as biomedicine, or men into traditionally female disciplines, such as nursing -- influences research priorities and agendas,” or vice versa.

“This approach measures links between changing gender demographics and changes in research questions and priorities,” the researcher add. “It rests on the assumption that social norms and expectations cultivate variations in the interests and perspectives of women, men, and gender-diverse individuals, and that increased variation in interests and perspectives can broaden agendas and discoveries.”

No prior research has considered such a link, the paper says. And, to be sure, any such research would require “careful attention to confounding factors, such as changing societal gender ideologies, changing funding priorities, the role played by social movements in altering scholarly agendas, and differences in the social and intellectual organization of disciplines.”

Yet historical examples do suggest that women’s headway into traditionally male-dominated disciplines has coincided with expanding research agendas, the paper says. As surge in women entered medical schools in the 1980s and 1990s, attention to traditionally underresearched areas of women’s health, such as heart disease, breast cancer, urology and autoimmune disease, increased. More women entering primatology in the 1970s and 1980s also coincided with breakthroughs debunking traditional "sex-based stereotypes about primate behavior," according to the paper. Social science and history have their own examples.

Which came first, the openness of disciplines to new questions or the increase of women in these fields, the paper asks, recalling Schiebinger’s hypothesis that questions and methods might influence team composition. The answer: influence presumably flows in both directions. But “we need carefully designed longitudinal studies, comparing stepwise developments over time, to resolve these questions empirically,” the paper says. “Recent advances in computational text analysis show great promise in quantifying the effects of gender diversity on research questions.”

Context Matters

Beyond those three questions, context also matters, according to the paper, which looks at management of each within four distinct but interdependent realms: research teams, disciplines, research organizations and societies at large.

Regarding teams and reducing potential conflict, the article says research suggests that members’ beliefs about the potential benefits of gender diversity are “crucial for positive performance outcomes.” One experimental study found, for example, that gender-diverse teams persuaded to believe in the benefits of group diversity were better at solving complex problems based on heterogeneous information than gender-diverse teams persuaded to believe in the benefits of group homogeneity. Beyond openness to diversity as a condition for success, the paper says that teams also should be encouraged to collaborate based on members’ expertise, and to follow processes that turn “individuals into team players.”

Four processes in particular have been shown to mitigate potential conflicts in gender-diverse groups, the paper says: team identification, or members’ compliance with team norms and conventions; transparent team processes, or teamwork characterized by clear coordination and communication; team efficacy, or team members’ belief in their combined collaborative ability to solve a given problem; and openness to experience.

Disciplines

Research teams function within “larger disciplines, each with unique norms and cultures,” the paper says. And encouraging gender integration in male-dominated disciplines “may reduce tokenism and negative stereotypes that lead to in-group and out-group biases,” so increasing the likelihood of seeing positive effects of team diversity. To that point, the paper cites a study of 52 scientific research collaborations at a U.S. university, which found that women researchers were more productive and likely to contribute toward team goals in disciplines where women were better represented. Similarly, “ghettoization,” where women are grouped into lower-ranked jobs or less prestigious subfields, must be avoided. In medicine, for example, it says, “men dominate general surgery, neurosurgery and orthopedics, while women dominate gynecology, pediatrics and family medicine. This underlines how horizontal stratification can hinder realizing the benefits of gender diversity for research outcomes.”

Ensuring good career and funding opportunities in both male- and female-dominated subareas and recognizing the “value of non-mainstream perspectives may help to reduce such ghettoization effects.”

Research Organizations and Scholarly Societies

Organizations may support gender diversity by cultivating inclusive research climates, such as by promoting equality with respect to salaries, promotion and workloads. Diversity flourishes in work settings where employees “feel free to openly express aspects of their social identities, whether this be sexual orientation or diverse gender identities,” the paper says, and diversity can be supported “through democratic approaches to decision-making that encourage wide-ranging insights and viewpoints.” Organizational leaders also need to devise practical strategies to introduce researchers to relevant GSA, the paper underscores. It should be part of the core scientific curriculum. And evaluation practices that consider metrics beyond citation counts and other bibliometrics can help.

Societies also can help promote a more nuanced sense of gender diversity by developing gender norms that promote equality and by developing policies linking team diversity, GSA and diversity in research questions to funding success.

Observational studies focusing on the Scandinavian countries, for instance, found that policies favoring large research grants, such as “centers of excellence,” reinforce gender inequalities by reallocating resources to “prestigious mega-projects with few women,” the paper says, addressing that last criterion. Instead, to fully harness the potential benefits of diversity, funding agencies should share resources across numerous projects and focus areas rather than concentrating them on a few large-scale grants, the paper argues.

‘About Excellence in Science’

Some critics of team composition-based gender diversity initiatives argue that science should be an absolute meritocracy, and that such initiatives are therefore meaningless and may even hurt science. The new multipronged approach makes that position harder to defend, since it makes an extremely explicit argument as to how gender diversity -- beyond numbers or, more pejoratively, quotas -- benefits science.

“Integrating sex and gender into research design is about excellence in science,” as has been demonstrated in areas including biomedicine, Schiebinger reiterated. The National Institutes of Health require that all publicly funded research consider sex as a biological variable, the Canadian Institutes of Health Research have similar policies, and the European Commission encourages sex and gender analysis across all fields of research, where relevant, she said.

So these issues go well beyond diversity, in that “using excellent research methodologies -- which include attending to sex and gender analysis, where relevant -- is about creating the best possible science.”

Anita Williams Woolley, associate professor of organizational behavior and theory at Carnegie Mellon University, in her own research has found that gender diversity in terms of composition of teams increases group problem-solving ability. She said she “absolutely” agreed with the new paper’s conclusions, since research on problem solving, “in science and otherwise, underscores the importance of diversity for innovation.”

In business settings, she said, “we always advise that the diversity of [a] team engaging in innovation needs to mirror the diversity of the population they seek to serve.” And because science is “supposed to be trying to solve the problems faced by society, when scientific teams lack the diversity present in society, their perspective on which problems to solve or how to solve them is going to be narrower than it should be.”

Woolley said that goes for gender, as well as other types of diversity.

Diversity in composition doesn’t guarantee diversity in questions or methods, she added, but it does raise probability of success in achieving other forms of scientific diversity, “since diverse team members are much more likely to bring the different perspectives that open the team up to different approaches.”

Musa Al-Gharbi, a Ph.D. candidate in sociology at Columbia University and a core member of Heterodox Academy, which seeks to increase viewpoint diversity in academe, said Schiebinger’s and her colleagues’ paper highlights one of the “key ideas undergirding” his group -- namely that “one's identity commitments inform research at a fundamental level, influencing what lines of research people are drawn to, how social problems are defined, the methods used to understand a given phenomenon, the ways data are interpreted and how any conclusions are ultimately communicated.”

There is no single objectively "right" way to make any of those determinations, he said, “but rather, an undefined range of legitimate approaches. A researcher's positionality inevitably mediates this decision-making process.”

When everyone in a field shares the same background, values or commitments, he said, that field will suffer from “significant biases, blind spots and distortions.” And that’s just as true in fields such as physics or biology as it is in sociology or political science, he said.

It’s “just a feature of social cognition.”

The American Geophysical Union is among the disciplinary organizations that have addressed gender diversity in the sciences, including by updating its ethics policy to include harassment as a kind of research misconduct. Eric Davidson, president, declined to comment on any specific research study but said that the union “firmly and enthusiastically believes that science improves, and science’s ability to serve society improves, when it is supported by a diverse and inclusive work force.”

He added, “The questions we ask and the ways we attempt to answer them are enriched by a diverse community of researchers. The challenges our global society faces in the future cannot be solved if we don’t have a diverse and inclusive community of researchers ready to lead the way.”

For Mathias Wullum Nielsen, assistant professor of political science at Aarhus University in Denmark and the paper’s lead author, “gender diversity is already an important characteristic of science” in that “women, men and gender-diverse people already share labs, work spaces and equipment in most disciplines.” So his paper is more about “understanding how to best support the possible benefits of gender diversity for science,” he said. And the three-pronged approach he and his team propose has the “potential to drive scientific discovery and innovation,” with each piece likely reinforcing the others.

Nielsen said he hoped that the context-based framework also will assist research leaders, universities, funding agencies, industries and governments “in realizing the possible benefits associated with diversity.”