Scientific authorship by gender: trends before and during a global pandemic

The study investigates whether the COVID-19 pandemic altered gender disparities in scientific authorship. Pre-pandemic, notable gender imbalances existed in many disciplines (e.g., surgery, computer science, physics, mathematics) and roles (e.g., senior and invited authors). The pandemic disrupted scientific work (lab closures, remote work, increased teaching), and increased domestic responsibilities (childcare, eldercare) that disproportionately affect women, potentially amplifying disparities. Prior evidence on pandemic impacts is mixed across fields and outcomes (submissions vs. publications). The authors hypothesized that the pandemic would increase male relative productivity and reduce women’s authorship shares. To test this, they analyzed corresponding-author gender for submissions to a broad portfolio of IOP journals across multiple disciplines and countries using self-identified gender, comparing pre-pandemic to pandemic periods and examining variation by region and discipline.

The paper summarizes 38 prior studies examining gender disparities in scientific authorship around COVID-19, spanning numerous disciplines and journals. Methods to infer author gender varied: algorithms based on first name and country, assessments of names, pronouns, photos, databases (e.g., Social Security data), and authors’ knowledge. Findings were inconsistent: several studies reported reduced women’s authorship during the pandemic (e.g., biomedical research, Elsevier-wide analyses), others found no change, and some observed increases in women’s contributions. Only two prior studies incorporated self-identified gender in any capacity; most relied on inference algorithms that can be inaccurate, especially for some name origins (e.g., Chinese). This study addresses gaps by using a large multi-journal dataset with self-identified gender and focusing on corresponding authorship.

Data: Manuscript submissions to Institute of Physics (IOP) Publishing from January 2019 through July 2021 (N=119,592). For each submission: journal, corresponding author’s self-identified gender (male, female, non-binary, or prefer not to say/no response), and submission month/year. Journals (N=57) were categorized into disciplines per IOP: astronomy and astrophysics (2), bioscience (15), environmental science (9), interdisciplinary (7), materials (22), mathematics (7), physics (29); some journals mapped to multiple categories. Geography and periods: Country of the corresponding author’s organization was available for 119,515 submissions (77 lacked country and were excluded from country analyses), spanning 152 countries. Regions followed UN classifications (Africa with five subregions, Asia with five, Europe with four, Latin America and the Caribbean with four, Oceania, and Northern America). Pandemic start date was defined per country as the date when cumulative confirmed COVID-19 cases reached ≥50 (Oxford COVID-19 Government Response Tracker). Because exact submission days were unavailable, the midpoint of each month was assigned as the submission date. Analysis: Primary outcome was the distribution of corresponding authorship by gender over time and across geography and disciplines, restricted analyses to submissions with specified gender for percentage calculations (N=99,114 overall; N differs by stratification). A difference-in-difference framework compared monthly submissions by gender between pre-pandemic and pandemic periods, with fixed effects for time and adjustments by country or journal. Statistical analyses were conducted in R 4.1.0 and Microsoft Excel.

- Gender disclosure and overall composition: Of 119,592 submissions, 68.0% were from male corresponding authors (81,381), 14.8% female (17,655), 0.1% non-binary (78), and 17.1% unknown (20,478). Among submissions with specified gender (N=99,114), 82.1% were male, 17.8% female, and 0.08% non-binary. Non-disclosure varied by journal (1.9% to 32.9%) and region (13.4% in Oceania to 17.8% in Northern America). - Overall pre- vs pandemic: Women’s share rose from 16.5% pre-pandemic to about 18.7–18.8% during the pandemic; men’s share fell from 83.5% to 81.1–81.2%. Both men and women submitted more manuscripts per month during the pandemic than pre-pandemic, but the pre-existing upward trend in submissions slowed for both genders. Difference-in-difference estimates indicated a statistically different influence of the pandemic on men versus women overall (with time fixed effects and accounting for country or journal), though no significant shifts were detected when stratified by region or discipline. - By region (percent female, pre → pandemic; change): Africa 19.7% → 19.1% (−0.7%); Asia 15.9% → 18.3% (+2.4%); Europe 17.5% → 20.0% (+2.5%); Latin America & Caribbean 15.5% → 16.4% (+0.8%); Oceania 17.3% → 18.9% (+1.5%); Northern America 17.7% → 19.7% (+2.0%); Overall 16.5% → 18.7% (+2.2%). Men remained the majority in all regions and subregions. - By country: Among countries with ≥30 submissions (N=87), men outnumbered women in all. Highest female representation: Ghana (38.1%) and Tunisia (35.5%). Lowest: DPRK and State of Palestine (0%). - By discipline (percent female, pre → pandemic; change): Astronomy & astrophysics 10.3% → 10.3% (0.0%); Bioscience 19.3% → 22.0% (+2.6%); Environmental science 19.5% → 21.5% (+2.1%); Interdisciplinary 10.1% → 15.2% (+5.1%); Materials 17.2% → 19.9% (+2.7%); Mathematics 11.5% → 13.2% (+1.7%); Physics 14.8% → 16.3% (+1.5%). Every individual journal had a majority of male corresponding authors (63.7–91.5% male). - Non-binary authorship: Extremely small shares (typically 0.0–0.2% by region/discipline); 0.1% overall among specified genders, limiting analysis. - Trend dynamics: Although women’s percentage increased during the pandemic, the pre-pandemic trajectory of increasing women’s submissions slowed, implying women’s share would likely have been higher absent the pandemic.

The findings contradict the initial hypothesis that the pandemic would reduce women’s authorship relative to men. Women’s submissions increased more than men’s during the pandemic overall and across most regions and disciplines, except Africa (regional decrease) and no change in astronomy/astrophysics. Nevertheless, men remained the dominant corresponding authors across all strata, and the pre-pandemic upward trajectory of women’s authorship slowed during the pandemic, suggesting pandemic-related headwinds for women despite absolute gains. Comparisons with prior studies reveal mixed literature; discrepancies likely arise from methodological differences: focus on corresponding vs. first/all authors, discipline coverage, submissions vs. accepted/published articles, definitions of pandemic windows, and especially the use of self-identified gender here versus inferred gender in most prior work. The results highlight persistent systemic gender imbalances and suggest that pandemic effects may differentially impact facets of productivity beyond submissions (e.g., acceptance, teaching, mentorship).

Using a large, multi-journal dataset with self-identified gender, the study shows that women’s share of corresponding-author submissions to IOP journals increased during the COVID-19 pandemic compared to pre-pandemic levels, but the pre-existing trend of rising women’s authorship slowed. Male dominance persisted across regions, countries, journals, and disciplines. The work underscores the complexity of pandemic impacts on scientific productivity and calls for further research into acceptance outcomes, authorship roles (first/senior), career stage, caregiving responsibilities, and other forms of scholarly work (teaching, mentorship, reviewing).

- Gender not specified for 17.1% of submissions; analyses assume similar disclosure behavior across periods, which may not hold. Non-binary authors comprised ~0.1%, limiting analysis of gender-diverse groups. The gender options were limited (male, female, non-binary), whereas gender identity is a continuum. - Lack of data on author characteristics (e.g., age, career stage, position, parental status), which may mediate pandemic impacts. - Submission dates available only to month/year; pandemic start defined by country-specific thresholds (≥50 cases) is an approximation; substantial heterogeneity in policies and institutional responses across and within countries is not captured. - Focus on corresponding authors may better reflect senior roles but is not equivalent to first/senior/all authorships; findings may not generalize to other authorship positions. - Analysis covers submissions, not acceptances or publications; results may differ for acceptance rates and citations. - Journal portfolio emphasizes STEM fields (particularly physical sciences), potentially limiting generalizability to other disciplines (e.g., humanities) and subfields. - Country-level analyses excluded submissions without country or with unclear pandemic period; some journals map to multiple categories, complicating discipline-level attribution.