Climate and Accountability

Edmund Bertschinger

MIT has done a lot to improve the experience of women and others underrepresented in STEM (science, technology, engineering, and math) yet inequities of experience persist. Indications include:

• Only 23% of MIT faculty are women, up from 21% in 2013.
• Of the 914 female undergraduate respondents to the 2014 Community Attitudes on Sexual Assault survey, 284 reported experiencing sexual harassment, rape, sexual assault, and other unwanted sexual behaviors while at MIT.
• Responses to the 2016 and 2017 Quality of Life Surveys indicate that MIT women systematically report having to work harder than their peers or colleagues to be taken seriously compared with men. Half of female faculty respondents agreed with this assertion compared with only 17% of male faculty respondents. Similar results hold for both tenured and untenured women.

These quantitative results are put in more personal terms by many people who have shared their stories with me during my recently concluded service as the inaugural Institute Community and Equity Officer (ICEO). From the female undergraduate and graduate students who are excluded from full participation by male peers, to the staff members who are put down by supervisors, to the faculty members who are frustrated by others talking over them, there are real problems calling for recognition, understanding, and solution. The largest component of this is gender harassment, defined in the recent National Academies Consensus Study Report “Sexual Harassment of Women: Climate, Culture, and Consequences in Academic Sciences, Engineering, and Medicine,” as verbal and non-verbal behaviors that convey hostility, objectification, exclusion, or second-class status about members of one gender.

Creating a Climate Dashboard

Conversations with many people suggest that one’s social identity – especially gender, class or socioeconomic status, sexual orientation, gender identity, and disability status, affect one’s experience of MIT. In order to assess this as fully as possible using existing Quality of Life Survey data, the ICEO and Institutional Research created a publicly accessible Climate Dashboard for MIT. Based on almost 13,000 responses during 2012-13 and 11,500 responses during 2016-17, with about a 50% response rate to these surveys, the dashboard summarizes the experience of groups defined by role (e.g., graduate student, postdoc, support staff on main campus, research staff at Lincoln Laboratory), gender, race, sexual orientation (self-identified in the survey), tenure status, and more.

Because racial tensions have drawn so much attention nationally during the last five years, I expected that MIT data would show that the largest differences of experience arise from differences in race or ethnicity. However, that was not the case. Gender, class (as indicated by role at MIT), and sexual orientation each account for more variation in the survey responses than does race.

The effects are multiplicative: the most marginalized groups on campus are female, LGBTQ graduate students (and, presumably, students of color, although the numbers are too small to parse this finely). Existing MIT surveys do not adequately assess gender identity or disability status. Yet individual stories of queer students, staff of color, and people with disabilities show that they generally experience a different, and often less supportive, MIT than I do as a straight white male able-bodied senior faculty member.

How was the dashboard created? The choice of survey questions (items) to analyze for the dashboard was made iteratively. A preliminary qualitative study in 2013-2014 presented in the ICEO Report identified five major themes for concern: unconscious bias and micro-inequities, discrimination or harassment based on social identity, abrasive conduct, sexual harassment, and excessive stress. The sexual harassment topic was excluded from the dashboard because it was investigated separately for students in the 2014 CASA survey and we have no such data for postdocs and employees. Quality of Life Survey items were chosen to sample the other topics. In addition, a principal component analysis of a large bank of survey items was carried out to identify which groups of questions had the most explanatory power. When several survey items were strongly correlated, analysis of variance was used to select the single survey item with the most explanatory power. The list was refined further as different subsamples were investigated, combining information about social identities, role and work unit. For 2012-13, a total of six survey items was used for employees and postdocs and five for students. New survey items were added for 2016-17, resulting in seven for employees and postdocs and eight for students (with the earlier items repeated allowing measurement of changes with time). Full details about the design and use of the dashboard appear elsewhere.

Back to top

The dashboard summarizes a lot of information. Surveys do not provide a controlled experiment nor are they sufficiently precise to answer every question. With these caveats, here are some additional findings from the surveys, beyond the fact that gender effects are large:

1. Based on their sense of fair treatment, the most privileged groups are undergraduates, graduate students, and faculty, while the least privileged are service staff, administrative staff, and support staff. Although some people may be surprised to find students feeling treated most fairly, others will recognize that even at a technical institute, administration and faculty are most solicitous of students (especially undergraduates).
2. Students respond much more negatively than other groups to “Taken seriously” (i.e., they are much more likely than employees to agree that “I have to work harder than some of my peers to be taken seriously.” (Students were asked this question only in 2017.) This might be ascribed to impostorism, i.e., the fear of being revealed to be incompetent despite high ability and achievement. However, the large gender differences in this item suggest that not being taken seriously may be due more to the environment than to the individual. As a female graduate student asked me, “Why do you call it impostorism when we are treated as though we really don’t belong?” A similar point was made this summer in the New York Times.
3. The largest gender differences occur for Other Instructional staff (mostly lecturers and instructors), faculty, and research staff at Lincoln Lab. This result affirms reports I have heard from women in these roles.
4. Compounding effects of social identity (i.e., intersectionality) are obvious in the climate dashboard. Combining gender and sexual orientation produces the most positive (for heterosexual men) and negative (for LGBTQ women) climates. (Gay men who are faculty or are research staff at Lincoln Lab are also among the top-ranked groups for a positive climate, revealing the complications of intersectionality.)
5. Between 2012-2013 and 2016-2017, the range of mean responses grew for each category of social identity. In other words, mean differences grew when comparing men and women, whites and underrepresented minorities, heterosexual and Lesbian/Gay/Bisexual or Unsure/Other (this selection is mainly a proxy for queer, transgender, and other non-binary gender identities). In particular, the experience of women became worse on average for all groups except faculty, while the experience of men improved in most groups (with exceptions for administrative staff, postdocs, and research staff at Lincoln Lab). The growing disparity of experience should concern everyone.
6. In 2012-2013 service staff had the worst average climate rating of any group at MIT; in 2016-2017 they had the best, and were the most satisfied group. Why? One reason may be the SEIU contract negotiated in 2016, which resulted in a substantial pay raise for janitors. (Service staff also had the lowest survey response rate of any group, 30% in 2016.)
7. Postdocs declined from 4th to 9th place (of 11 groups) in the climate rankings over the four years between surveys. They experienced the most negative change. This may be related to increasing financial stresses combined with other challenges identified in the National Postdoc Survey.

What is next? The reader may ask why we didn’t extend the analysis prior to 2012-2013, why we didn’t include certain groups like people with disabilities, or how their own department ranks. For the first two, we lack good survey data; MIT has been changing its survey questions over time, and the only long-term baseline we have is overall satisfaction. By this measure, there has been substantial overall improvement during the last 20 years, but the analysis by demographic groups is available only in recent surveys. We haven’t asked about respondents’ disabilities in the Quality of Life Surveys to date. We do have information on respondents’ work units (academic department or otherwise), but for most units the number of respondents is small and does not allow for dividing by demographic group given our requirement of 15 or more responses. The ICEO Report showed that department or work unit is, like role and social identity, a source of considerable variation in the climate for inclusion. This information is available to the senior administration and is being used to help improve department climates. I hope that, going forward, MIT will continue to ask the subset of questions used in this dashboard so that it can be kept current, and will seek additional demographic information about respondents. In particular, respondents should be given the option to specify both gender identity and sexual orientation using common designations as well as disability status.

Recognizing climate challenges is only a beginning. Concerns must be understood and addressed, progress assessed, and the cycle repeated. Having observed these issues play out over more than 30 years on the faculty, I believe that three ingredients are necessary for sustaining change: committed leadership, upward pressure from the community, and internal and external accountability measures. The third ingredient has not always been effectively utilized, but we can change that, now.

Accountability

Since 2010, various groups in the MIT community have created sets of recommendations for making MIT more equitable, inclusive, and diverse, ranging from the 2010 Report on the Initiative for Faculty Race and Diversity to the 2016 Report on the Status of Undergraduate Women at MIT. The Academic Council Working Group was created in response to two 2015 reports (from the Black Students’ Union and the Black Graduate Students Association) with 18 recommendations. Nine MIT reports about equity and inclusion presented during 2010-2016 made a total of 177 recommendations. Most of the reports have received little attention. In my view, we don’t need many more recommendations, but we do need attention focused where community members have already invested so much effort (see “upward pressure from the community,” above).

To this end, the ICEO hosts a Recommendations Scorecard summarizing progress on each of the 177 recommendations. The average completion rate per report ranges from 22% for the Recommendations of LBGTQ+ Students and Communities at MIT to 57% for the BSU recommendations. As of this summer, the overall completion rate is 39%, indicating that much work remains to be done. As more progress is made, the scorecard will be updated. I hope that this internal accountability measure will help guide committed leadership to continue improving the experience of all people at MIT.

In the end, internal accountability measures rarely suffice to sustain change. Leadership changes, students move on, employees retire, new initiatives shift our attention. However, issues of equity, diversity, and inclusion do not disappear. MIT prides itself on solving problems, but sometimes it needs help. All faculty who have participated in the Corporation Visiting Committee process recognize the value of an external evaluation and accountability measure. Indeed, the MIT Corporation is interested in the topics of this article. But the Visiting Committee process focuses on the research and teaching of academic departments, and it cannot be counted on to have the expertise to deal with complex issues of equity and inclusion.

Fortunately, another mechanism exists: the STEM Equity Achievement (SEA) Change initiative of the American Association for the Advancement of Science. SEA Change provides a LEED-like certification for institutional efforts to promote equity, diversity, and inclusion in colleges and universities, focusing on the STEM disciplines. Participation requires agreeing to a set of guiding principles, conducting an evidence-based self-assessment, and developing a plan to make progress. The institutional application is substantially smaller than a Visiting Committee binder but requires participation of a cross-functional team with access to and support of key stakeholders (e.g., MIT senior leadership and Deans). Once the university as a whole has an entry-level certification, individual STEM departments can apply for their own rating. I believe it is very important for MIT to show leadership among universities by participating in this initiative sooner rather than later or not at all. The data presented above show a compelling need to hold ourselves accountable. I urge faculty to call on their leaders to begin the SEA Change process in 2018.

	Back to top
	Send your comments