Source-linked AI summary
The role of gender in scholarly authorship
Jevin D. West, Jennifer Jacquet, Molly M. King, Shelley J. Correll, Carl T. Bergstrom
TL;DR
Gender inequity in scholarly authorship may persist in subtler forms than overall publication counts suggest. Using a large, citation-connected JSTOR corpus, the paper examines gender representation across authorship positions and finds continuing disparities, especially in single-authored and last-author publications. The analysis identifies these patterns but does not determine the mechanisms producing them.
Problem
Gender disparities may appear in authorship order and single-authored publishing even when broader academic indicators suggest improving equality.
Method
The study analyzes gendered authorship trends across disciplines and time using JSTOR papers categorized with citation-network methods.
Results
27.2% of authorships from 1990–2012 are women versus 15.1% from 1665–1989, yet women represent only 26% of single-authored papers since 1990 and remain underrepresented as last authors.
Takeaways & Limitations
Gender disparities in scholarly authorship remain through differences in overall representation and prestigious author positions despite gains in first-author parity.
Takeaways & Limitations
The analysis delineates gendered authorship patterns but cannot uncover the mechanisms producing the observed disparities.
Abstract
from arXiv · showhide
Gender disparities appear to be decreasing in academia according to a number of metrics, such as grant funding, hiring, acceptance at scholarly journals, and productivity, and it might be tempting to think that gender inequity will soon be a problem of the past. However, a large-scale analysis based on over eight million papers across the natural sciences, social sciences, and humanities re- reveals a number of understated and persistent ways in which gender inequities remain. For instance, even where raw publication counts seem to be equal between genders, close inspection reveals that, in certain fields, men predominate in the prestigious first and last author positions. Moreover, women are significantly underrepresented as authors of single-authored papers. Academics should be aware of the subtle ways that gender disparities can appear in scholarly authorship.
I. INTRODUCTION
Gender inequities persist in higher education despite mixed trends in publication and hiring outcomes. The study examines gender composition and authorship order across disciplines because prestigious author positions matter for academic careers.
- Women remain underrepresented among full professors and earn lower average salaries than men in comparable positions.Women represent one-quarter of full professors and earn 80% of men's average salary in comparable positions.
- Publication differences vary by discipline: men publish more in chemistry and mathematics, women publish more in electrical engineering, and no significant differences appear in three other fields.
- The study tracks authorship order because first and sometimes last author positions influence hiring, promotion, and tenure, particularly in scientific fields.
- Prior medical-literature studies find women underrepresented as first and last authors, although first-author disparities have recently declined while last-author disparities persist.
- A network-based community-detection approach categorizes papers hierarchically, enabling comparisons of gender representation across fields of different sizes and scales.
The JSTOR corpus
The study uses JSTOR's extensive scholarly archive and restricts analysis to a citation-connected subset that can be hierarchically categorized.
- 8.3 million JSTOR documents span the sciences and humanities from 1545 to early 2011, including 4.2 million research articles.
- The JSTOR network dataset contains 1.8 million documents that cite or are cited by other corpus documents and are therefore amenable to network analysis.
- 94% of the network dataset belongs to one giant citation-network component in which articles are mutually reachable through citation trails.
- The analysis uses the network dataset because citation information enables hierarchical categorization, and its identified authors are 21.9% female versus 20.8% in the full corpus.
Mapping the hierarchical structure of scholarly research
The paper maps scholarly disciplines as a hierarchy rather than a single structural layer by applying the hierarchical map equation to citation networks.
- Citation relationships form a network whose topology can map the structure of science.
- The map equation has proven effective for mapping scientific-network structure.
- The hierarchical map equation is used to represent disciplines, fields, and subfields across multiple structural levels.
Determining gender of authors
Author gender is inferred from first names using Social Security Administration records, with ambiguous or insufficiently represented names left unidentified.
- Social Security Administration records provide annual top-1000 names for boys and girls born in the United States from 1880–2010.
- An author's gender is assigned only when the first name is associated with one gender at least 95% of the time.Names such as Mary and John qualify, whereas ambiguous names such as Leslie and Sidney do not.
- The method assigns genders to 6,879 unique first names: 3,809 female and 3,070 male.
- Among 3.6 million authorships, 2.8 million have full first names, covering 1.5 million papers, and 73.3% of those authorships receive confident gender assignments.
Gender and authorship order
The analysis compares overall gender composition with representation in each authorship position, including first, second, third, and last authorship. It also considers last authorship separately because it can signal principal-investigator status in some fields.
- The study calculates gender composition for all papers and separately for each authorship position.Positions include first, second, third, and subsequent authorships.
- Last authorship receives separate attention because it can denote the principal investigator or group leader in some fields.This interpretation is especially relevant for papers with at least three authors, such as in molecular biology.
- Authorship-position analysis is designed to reveal disparities that overall gender composition can miss.
Authorship and author order
Across the JSTOR network dataset, women’s representation varies by time, subfield, authorship position, and paper authorship structure. Overall gains coexist with persistent underrepresentation in single-authored papers and prestigious authorship positions, while some trends are affected by identification and collaboration patterns.
- 21.9% of gender-identified authorships in the JSTOR network dataset were women, but representation varied across time, fields, and authorship positions.
- The apparent increase in women’s representation may partly reflect easier identification as authors increasingly used first names instead of initials.
- Even sociology’s relatively balanced gender composition concealed dramatic differences among its subfields.
- Women became less underrepresented as first authors after 1990, but remained underrepresented as last authors, a prestige position in the biosciences.The last-author gap does not hold in every field; mathematics shows even representation across positions at low frequency because author order is typically alphabetical.
- Multi-authored papers increased over time while single-authored papers declined, alongside an increase in women’s representation.Some authorship-order patterns may therefore reflect changes in collaboration and gender composition over time.
IV. DISCUSSION
Although women’s representation in authorship has improved, gender disparities persist in single-authored work and prestigious author positions. These patterns vary across disciplines, and the analysis cannot identify the mechanisms producing them.
- Women’s representation varies by discipline: molecular biology shows first-author overrepresentation and last-author underrepresentation, whereas sociology shows underrepresentation in both positions.In mathematics, alphabetical ordering corresponds to neither under- nor over-representation at first or last author positions.
- Because publication and authorship position affect hiring, promotion, and tenure, underrepresentation in prestigious positions can affect women’s representation among faculty.
- 27.2% of authorships from 1990–2012 were by women, compared with 15.1% from 1665–1989.
- Since 1990, women represented only 26% of single-authored papers in the JSTOR dataset.
- The difference between total female authorships and first authorships has remained below 2% since the 1960s, while the discrepancy for last authorships remains above 5%.First-author parity has improved, but women remain underrepresented among last authors.
- The analysis identifies gendered authorship patterns but does not determine the mechanisms producing the disparities.The authors discuss possible explanations but state that their data do not establish which mechanisms operate.
- The study’s large, cross-disciplinary dataset shows that progress toward gender equality does not eliminate disparities in intellectual authorship positions.