Monday, September 14, 2015

Addressing Gender Bias in the SDSS Collaboration

Today’s guest bloggers are Sara Lucatello and Gail Zasowski. Sara is a staff astronomer at INAF - Padova Astronomical Observatory, in Italy. Her research focuses on understanding, through the study of stellar populations, the formation and chemical evolution of the Milky Way. Gail is an NSF Astronomy Postdoctoral Fellow at Johns Hopkins University in the US, where she studies the Milky Way's ISM and stellar populations; she also runs a space camp for middle school students and works to support public outreach and astronomy diversity programming. Both are active members of the Sloan Digital Sky Survey and its Committee for the Participation of Women in SDSS.
In July 2012, at the yearly meeting of the Sloan Digital Sky Survey Collaboration, a slide was shown that outlined the current leadership structure for the fourth generation of the SDSS, which was due to get underway in 2014. Of the 26 positions shown on the slide, only one was filled by a woman.
Many collaboration members in the audience were visibly taken aback, looking around to see who else was surprised. Soon the Sloan Foundation (the largest single source of funding for the SDSS) echoed the surprise and unease, and explicitly mentioned the lack of women in the leadership as a concern in their feedback during the process of granting funding for SDSS-IV.
The issue of gender balance is of course a hot topic for those watching STEM groups and STEM culture. A number of measures have been put in place over the last several years to promote diversity and reduce the effects of explicit and implicit bias. These actions have met with varying degrees of success, as discussed frequently on this blog and elsewhere. Such initiatives are, however, generally implemented at the institutional or funding agency levels. But in modern astronomy, work is increasingly done in large collaborations, which bring together many -- sometimes hundreds of -- scientists from a variety of nationalities and backgrounds. These also face the challenge of dealing with a wide range of backgrounds when it comes to diversity issues, but usually without the infrastructure of, say, a university, to log problems and enforce solutions. Therefore, it becomes crucial to find new approaches to identifying issues and devising new practices aimed at creating a climate of fairness and inclusiveness, regardless of sex, gender identity, ethnicity, race, sexual orientation, economic background, and religion.
With this challenge in mind, in late 2012, the SDSS-III management created the Committee for the Participation of Women in SDSS (CPWS). The initial members were carefully chosen to include both men and women, and to span a range in career stage and seniority. The CPWS was charged to evaluate the gender balance within the collaboration and the leadership, assess the general inclusivity and perceived climate, and make recommendations to get rid of any widely perceived barriers to the participation of collaboration members.
We note that this was a genuine gesture on the part of the SDSS management. That is, despite having initially been inspired by the concerns of the Sloan Foundation, the management was truly concerned about the presence of biases against women and other underrepresented groups; it was interested in ensuring an inclusive scientific climate within the Collaboration (which includes approximately 500 participants from more than 18 countries in North and South America, Europe, and Asia). We emphasize this because we feel it highlights part of the problem with these increasingly ubiquitous large collaborations -- even with genuine awareness and goodwill on the part of many individuals, there is often no structure in place for turning that awareness into best practices applied evenly and fairly across the board for all collaboration members.
As a first step in establishing such a structure for the SDSS, the CPWS conducted both a demographic survey and a series of interviews of collaboration members, spanning all ages and levels of leadership responsibility. CPWS made some recommendations to the SDSS management to improve the hiring practices and to encourage more people to apply for leadership positions. These recommendations came at a critical time -- just under a year before SDSS-IV would begin full operations, while many positions within the SDSS infrastructure and the individual projects (MaNGA, APOGEE-2eBOSS, TDSS, SPIDERS) were being filled.
In mid-2014, the CPWS conducted another, more extensive, demographic survey, followed by a round of interviews with the PIs from the primary SDSS-IV projects. The aim was to take a snapshot of the incipient SDSS-IV and establish a baseline in terms of gender balance[1], both in general and for the leadership. This was to be the first of a series of annual surveys, in order to monitor the Collaboration in terms of climate, diversity, and equal career opportunities.
One critical goal of this survey was to get a sense of what sorts of people were in leadership roles, whether those were officially recognized by the collaboration or not. To that end, we asked people to self-identify as leaders, based on a deliberately broad definition[2], with separate questions about the official recognition (e.g., having a job title) or salary status of their position.
About 50% of the active SDSS members responded to the survey, most located in North American or European institutions. About two thirds of the sample are faculty members or research scientists, while about one third are postdocs and graduate students. The fraction of women, both in the general sample and among members who are in self-reported leadership positions, is about 25%. This number is consistent with the fraction in the US astronomical community, but considerably higher than that in the IAU, which is only 15%.
While this indicates that the SDSS has been successful in recruiting a leadership that is representative of its overall composition (and of the US astronomical community), the aggregate numbers tell just one part of the story.
The details make it clear that there is still a long road ahead to optimize gender balance and equalize career opportunities. For example, it is true that SDSS women have increased their presence in the high ranks of the org chart (compared to the situation described at the beginning of this post), and actually take on self-identified leadership positions at similar rates as men. However, upon closer inspection, the data clearly indicate that women very disproportionately assume roles related to education and outreach, while technical or exclusively scientific positions at the top are almost entirely male-dominated. Full details on the SDSS-IV demographic survey are available in the  complete CPWS report.
Another informative dimension was added by interviews with the PIs of the SDSS-IV projects and the SDSS-IV director on the topic of hiring practices. Because of financial and/or time constraints, several of the top level positions were either inherited from SDSS-III or filled, without open calls, by people already networked into the project. Interestingly, the data suggest that openly advertised positions resulted in a higher fraction of female hires. However, the PIs anecdotally reported to have encountered reluctance from female scientists to volunteer or apply for leadership positions at a higher rate than male scientists, even when explicitly encouraged to join or apply. This suggests that taking on the responsibilities and commitments of these positions is not perceived as worthwhile or feasible for women scientists, who are very likely to already be more over-committed than their male colleagues.
The 2015 edition of the SDSS Demographics Survey is currently underway. Though largely similar to last year's questionnaire, some changes were made to enable reliable tracking of demographics across time. Now, we more carefully explore the issues described above and assess whether members -- of any gender -- hesitate to apply and/or assume leadership positions, and what the underlying reasons may be.
We hope that these efforts will help make the SDSS collaboration an inclusive and fair environment, where every member is afforded equal opportunity to hone and use their skills as scientists, putting forward their best possible work for SDSS and for the community at large.
Progress has been made on some but not all issues, and this effort is certainly an ongoing one, with many people taking part. And as many SDSS members go on to work in other large collaborations, we hope that this self-reflection and internal enforcement of fair practices becomes the norm in these incredibly productive and diverse groups.
[1] We note that the survey gathered information on more than just gender -- data were also collected on racial/ethnic identity, geographical location, career level, etc. These statistics will also continue to be tracked.
[2] We defined a leadership role as any "whose tasks and responsibilities include making decisions that affect other people and the SDSS, organizing regular project discussions or meetings, professional mentoring, or influencing/directing others in their tasks."