Nicole Mideo presented “Letters of recommendation: data on gender bias”. Here is her summary:

We learned in a previous BREWS discussion about implicit bias and it’s clear that reference letters are an important part of progressing up any career ladder, so for this discussion we asked whether implicit bias could impact the quality of letters being written for males versus females.

We began with a discussion of the Trix & Psenka (2003) paper which qualitatively compared reference letters written for male and female applicants for faculty positions at US med schools. Letters for female applicants were shorter, included more grindstone words (e.g., “dependable”), fewer repeated standout words (e.g., “excellent”), included more ‘doubt raisers’ and gender terms.

Given BREWS’s emphasis on data, I collected my own, running a bunch of reference letters written by me and other professors in EEB through an online implicit bias detector. The estimated bias in letters I have written spanned from quite heavily female-biased language to quite heavily male-biased language. This was true for the full dataset too. With the help of John Stinchcombe, I analysed bias in these letters using mixed effects models, with a random effect of professor (anonymised) and fixed effects of gender, stage, prof gender, and all interactions. Two surprises emerged. (1) There was no significant effect of the gender of the person the letter was being written about. (2) There was a significant effect of the stage of that person. Letters for undergrads (specifically ones who the letter writer only knew from lecture courses) were more female-biased than letters for undergrads who had worked in the lab, while letters for grad students, postdocs, and faculty were further towards the male-biased end of the spectrum.

These results got me digging into the guts of the algorithm. Standout words, ability words (e.g., “intelligent”), and research words (e.g., “data”) are all categorised as male-biased, while grindstone words and teaching words (e.g., “course”) are categorised as female-biased. The algorithm counts the instances of these types of words in a letter, looks at the difference in numbers of male and female biased words and divides this difference by the total number of “gendered” words to get an estimate of bias. So, the stage results make a lot of sense. If a letter writer knows a student only from a lecture course, then there are likely to be a lot of teaching words that lead to “female-biased language”.

The algorithms cite a Schmader et al. (2007) study as inspiration, which statistically compared letters written for male and female applicants for faculty positions in chemistry and biochemistry. The results showed a significant reduction in standout adjectives in letters for females, but all other differences were non-significant or marginal, e.g., letters for females didn’t contain significantly fewer research or ability words nor did they contain significantly more teaching or grindstone words. Despite this, the online algorithms still assume a gender bias in all of those categories of words. This seems problematic and we discussed rerunning the analysis with only the standout adjectives as being ‘gendered’.

Finally, we discussed what evidence exists that the words categorized as female-biased result in weaker letters, which seems to be the underlying operating assumption; the Trix & Psenka study, however, only looked at letters for successful applicants! We felt that the specific job could alter the value of word categories (e.g., teaching words will be very valuable for applicants to teachers’ college), and that different fields will have different ‘cultures’ of letter writing.

Overall, we agreed that assessing implicit bias in reference letter writing requires more data and rigorous analysis.

Chelsea Rochman presented on “Women mentors and their contribution to gender composition in science”. Here is her summary:

We talked about the role of mentors in their contribution to gender equity. We specifically talked about the role of women mentors.

We began the discussion talking about these readings in the Atlantic about women mentors and bosses acting as “Queen Bees” (“Why do women bully each other at work?“, “Why women get criticized for being candid at work“, and “The Myth of the Queen Bee“).

We also read and dug into some work by Denon Start and Shannon McCauley about gender composition of academic research groups and a publication by Ellemers et al., 2004 called: Underrepresentation of women in science: differential commitment or the queen bee syndrome?

In short, the Atlantic articles discussed how women can sometimes act as Queen Bees:

This term was first defined by G.L. Staines, T.E. Jayaratne, and C. Tavris in 1973. It describes a woman in a position of authority who views or treats subordinates more critically if they are female. This phenomenon has been documented by several studies.

The question we wanted to ask is how this affects gender equity. Does this contribute to the underrepresentation of women in science?

Overall, we found that in the past women seemed to have a negative bias about women and their ability to succeed in academia. Over time, this seems to be disappearing as more women enter and succeed in academia. In addition, it seems that some gender bias in academia occurs at the applicant phase – suggesting that the more women apply, the more women will succeed. The good news – as more women enter academia and mentor women in academia, we get closer to achieving gender equity.

Rebecca Schalkowski & Rebecca Batstone presented “The relevance of socioeconomic background in academia”. Here are their summaries:

Rebecca S.: In my part of this talk I summarized different studies on factors relating to socioeconomic status which affect a child’s chances for academic success starting from the earliest development through the end of high school. These include occupation, income and education, which are all known to affect a child’s school achievements (American Educator Spring 2012, reviewed Sirin, 2005), IQ (Duncan et al, 1994), likelihood to do well in high school (Palardy, 2008) and attend college (Conley, 2001). Differences in household wealth as defined by the above factors have further been associated with affecting reading achievement (Aikens and Barbarin, 2008), math achievement (Chen et al., 1996), working memory (Noble et al., 2005), and the ability to regulate emotions and thought processes (Evans and Rosenbaum, 2008). The reason for families of lower socioeconomic status to suffer these effects as given by Daniel T. Willingham (American Educator, 2012) derive from lower access to opportunities. These opportunities can be classified by the three types of capital a person or family can have or lack: financial (e.g. books, tutors), human (e.g. skills & knowledge through education and experience of adults surrounding children) and social capital (e.g. connections or networks with people who have financial or human capital). This situation causes both lower resources and higher chronic stress to people from lower socioeconomic groups (Klerman, 1991, Conger et al., 1994) which will in turn lead to decreased academic performance, caused by physiological, psychological and economical disadvantages, causing them to drop out of high school up to 5 times more frequently and college (National Center for Education Statistics, 2008; Langhout, Drake, & Rosselli, 2009).

Rebecca B. focused on two recent large-scale studies: the first (Chetty et al. 2017, “Equality of Opportunity Project”) examined the socioeconomic background of students attending elite colleges in the states (spoiler alert: mostly rich kids), and the second (Clauset et al. 2015, Science Advances) examined a major predictor for who ends up landing a faculty position, namely, the prestige of candidate’s alma mater.

The Chetty et al. (2017) dataset comes from a project entitled the “College Mobility Report Cards”, and includes data from over 30 million students born in the US between 1980-1991 who graduated from colleges in the US between 1999-2013. The data compares the student’s income ranking in their early thirties and that of their parents to see whether attending a particular college was associated with the student’s upward mobility (going from a low-income bracket to a higher-income bracket). Two main findings were discussed: first, students attending “Ivy League” colleges (e.g., U. Chicago, Stanford, MIT), are 77 times more likely to come from families in the top 1% income distribution compared to the bottom 20% income distribution, indicating elite colleges are clearly failing at recruiting students from diverse socioeconomic backgrounds. Second, students from low-income families who attend elite universities receive earnings post-graduation equal to those from higher-income families, meaning that these colleges successfully “level the playing field” when it comes to financial success post-graduation, and also highlights that there is little cost to colleges for admitting students from low-income families. Unfortunately, student access to top colleges from the bottom 10 to 40% income distributions has not changed since 1999-2013, and funding for students to attend these elite colleges has declined 18% nation-wide since 2008, making it even less likely socioeconomic diversity will improve in the near future.

The second dataset (Clauset et al., 2015) consisted of 19,000 tenue-track or tenured faculty from 461 North American departmental or school-level academic units in three disciplines: computer science, business, and history. The goal of this project was to determine the factors that influence faculty hiring. One such factor emerges through “faculty hiring networks” – collective assessments whereby both the candidate being hired and the institution hiring must make a positive assessment of one another’s quality (e.g., based on teaching and research programs). The authors used faculty hiring networks to construct “social prestige” rankings for each institution, whereby institutions that disproportionately succeed in placing faculty and hire candidates from higher-ranked programs are characterized as being more prestigious compared to others. The authors found that across the disciplines examined, there exists a systematic bias in terms of who ends up getting a faculty placement. Only a quarter of the institutions included in the dataset are responsible for producing 71 to 86% of all tenue track faculty, and the size of the placements are not merely reflecting the size of the unit. The authors also found only 9 to 14% of faculty are placed at institutions with a higher prestige ranking than their doctorate, indicating steep prestige hierarchies, whereby less prestigious institutions hire candidates who graduated from more prestigious institutions in order to bolster their own prestige. As a result of this trend, most PhD’s slide down the prestige scale when they actually land a faculty position, and interestingly, women slide further down the scale compared to their male counterparts from the same institutions. Finally, more prestigious institutions also tend to be more central, well connected, and hold a more influential network position, which fosters the free exchange of ideas, and emphasizes the benefit of landing a position in such an institution. Linking back to the first dataset discussed, social inequality present at early academic levels (i.e., during undergrad) may be carried forward and even amplified at later academic stages. Programs that increase representation of students from diverse socioeconomic backgrounds at prestigious institutions are therefore extremely important to buffer against social inequality at latter stages, whereby the merit of a candidate is strongly influenced by the prestige of the university they graduated from.