SYLLABI AND THE SOCIALIZATION OF GRADUATE STUDENTS

This study investigated gender representation in syllabi—that is, the proportion of assigned readings that are female authored. Syllabi socialize PhD students (henceforth, “graduate students”) into academia, conveying not only academic content but also implicit and explicit messages about what constitutes model work—and which scholars do that work. Thus, gender representation in syllabi affects how future scholars view and engage with academia.

Similar to studies that document a gender citation gap, we expected to find a gender syllabus gap. Several mechanisms could lead instructors to under-assign work by women. The first mechanism is path dependence. Scholars designing syllabi experience significant time constraints. They tend to assign some of the same readings that they themselves read as graduate students and to seek out relevant syllabi from other instructors. They also may rely on classic works and “elite readings,” in which gender gaps in citation counts are largest (Zigerell 2015). Second, instructors are likely to assign work by well-known scholars; male authors, in general, are likely to occupy more central locations in scholarly networks. However, network effects would lead female instructors to assign more female-authored work than men, given gender homophily.

Third, implicit gender biases could lead instructors unconsciously to favor male-authored readings. Beginning at a young age, individuals adopt the gender stereotypes that women are less brilliant than men and less capable academics (Bian, Leslie, and Cimpian Reference Bian, Leslie and Cimpian2017; Leslie et al. Reference Leslie, Cimpian, Meyer and Freeland2015; Williams, Phillips, and Hall Reference Williams, Phillips and Hall2014). Scholars have found evidence of gender bias in academia in evaluations of scholarly work, letters of recommendation, and certain hiring practices (Knobloch-Westerwick and Glynn Reference Knobloch-Westerwick and Glynn2013; Krawczyk and Smyk Reference Krawczyk and Smyk2016; Lee and Ellemers Reference Lee and Ellemers2015; Madera, Hebl, and Martin Reference Madera, Hebl and Martin2009; Rivera Reference Rivera2017). One study observed hiring practices that favor women (Williams and Ceci Reference Williams and Ceci2015).

Because instructors draw on top journals to create syllabi (Teele and Thelen Reference Teele and Thelen2017), we adopted women’s publication rates in 10 top journals as a benchmark. We expected underrepresentation to be particularly pronounced in first-author positions. Typically, political scientists use alphabetical author order. However, when author order is not alphabetized, we expected male authors to appear earlier in the author list, indicating greater importance.^{Footnote 3} We hypothesized that:

We expected that instructors from underrepresented groups assign a higher proportion of female-authored readings.^{Footnote 4} Both men and women of color, as well as white women, are likely to be more aware of barriers to demographic representation, due to both personal experience and informal networks (Brink and Benschop Reference Brink and Benschop2014; McDowell, Singell, and Stater Reference McDowell, Singell and Stater2006). As a result, underrepresented individuals may deliberately diversify syllabi. For example, recent studies found that female scholars cite more female-authored work (Maliniak and Powers Reference Maliniak and Powers2013; Mitchell, Lange, and Brus Reference Mitchell, Lange and Brus2013). We hypothesized that:

We also expected that older female instructors assign more women’s research. Scholars who went through graduate school decades ago likely had little exposure to female-authored research. However, motivation and networks may matter more for older women. Studies show that women socialized during the second wave of feminism are distinctly feminist, relative to subsequent generations (Schnittker, Freese, and Powell Reference Schnittker, Freese and Powell2003). In addition, older female instructors may have larger gender-homophilous networks, including the mentoring of junior women, making them more aware of women’s publications. We hypothesized that:

Finally, instructors socialized in more gender-equal environments may assign more female-authored readings due to greater awareness of diversity and perhaps lower gender bias. Indeed, the gender citation gap is smaller in more gender-equal countries and those with higher human development (Sugimoto, Ni, and Larivière Reference Sugimoto, Ni and Larivière2015). Nonetheless, gender inequality in academia does not simply mirror societal gender inequality; rather, national academic institutions structure female academics’ opportunities and challenges (Bain and Cummings Reference Bain and Cummings2000; Husu Reference Husu2000; Le Feuvre Reference Le Feuvre2009). Although many European countries have lower indices of gender inequality than the United States, hierarchy within European universities may create barriers to female instructors’ advancement. Relative to other countries’ political science associations, the American Political Science Association (APSA) “has the longest history of institutionalizing multiple forms of diversity within its organizational structure, beginning in the wake of the Civil Rights Movement and the Women’s Movement” (Abu-Laban, Sawer, and St-Laurent Reference Abu-Laban, Sawer and St-Laurent2018, 15). Hence, European-trained instructors may assign fewer female-authored readings.^{Footnote 5} We hypothesized that:

DATA AND METHODS

We collected syllabi through a multiphase process. First, 301 syllabi were obtained from an IRB–registered survey that sampled APSA members teaching in US-based PhD programs; these respondents also answered demographic and institutional questions. Second, contributed collections and online searches yielded another 160 syllabi. Third, graduate-student project affiliates at top-50 programs collected 450 syllabi on our behalf. (See online SI: Methods for more details about our data collection and a discussion of the representativeness of each data-collection component.)

The resulting dataset—GRADS—comprises data at the levels of (1) syllabi, (2) instructors, and (3) assigned readings. We processed readings through a combination of hand-coding and machine-learning to parse the different components (e.g., author name and title). To code author gender, we first matched given names against a list of known political scientists whose names we expected would be coded incorrectly. We then matched remaining given names against existing gender datasets, including those based on US and UK censuses and social media data (see online SI: Methods for a full discussion). Instructor gender was coded by examining online biographies and using the names dataset.^{Footnote 6} Appendix table 1 compares the GRADS instructor sample to APSA members and faculty in PhD-granting departments in the United States. Although our sample is whiter, older, and more male than the APSA membership, it is comparable to the population of PhD-level instructors.^{Footnote 7}

SUPPLEMENTARY MATERIAL

To view supplementary material for this article, please visit https://doi.org/10.1017/S1049096519001239.