Race to the Big Lab: Gender Disparities in Large Team Collaboration and Its Impact on Early Academic Careers

Paper Title

Race to the Big Lab: Gender Disparities in Large Team Collaboration and Its Impact on Early Academic Careers

Publication Info

• Topic area: Gender disparities in academic collaboration and career development.
• Keywords: Gender inequality, large-team collaboration, early-career researchers, social capital, network centrality, publication productivity, Computer Science, academic networks, synthetic difference-in-differences, survival analysis.

Background and Problem

• Problem / challenge: Limited understanding of how large-team collaborations impact early-career researchers' development, particularly regarding gender disparities in access to such collaborations.
• Significance: Large-team collaborations are critical for career visibility, resource access, and social capital accumulation, yet unequal access may perpetuate systemic inequalities in academia.
• Motivation and related work: Prior studies have explored academic productivity and collaboration but have largely overlooked the role of team size as a competitive resource and its implications for gender disparities. This paper addresses this gap by investigating the causal effects of large-team collaboration on career outcomes and gender-based access inequalities.

Solution

• Proposed approach: Synthetic Difference-in-Differences (SDID) and survival analysis to evaluate the impact of first-time large-team collaboration on early-career researchers' network centrality, productivity, and gender disparities.
• Novelty:
  1. Empirical evidence linking large-team collaboration to enhanced network centrality and productivity for early-career scholars.
  2. Identification of gender disparities in access to large-team collaborations as a mechanism of systemic inequality.
  3. Application of SDID to mitigate selection bias and improve causal inference in academic collaboration studies.
  4. Subfield-specific analysis to assess heterogeneity in collaboration patterns and gender dynamics.
• Procedure and key techniques:
  • Construction of collaboration networks using SciSciNet data (2000–2004 cohorts).
  • Definition of network centrality and productivity metrics, including weighted neighborhood centrality and citation-weighted publication counts.
  • SDID models to estimate causal effects of large-team collaboration on centrality and productivity.
  • Survival analysis and Cox proportional hazards models to evaluate gender disparities in access to large-team collaborations.
  • Robustness checks using alternative event definitions, productivity measures, and subfield analyses.

Results

• Concrete findings:
  • First-time large-team collaboration increases network centrality by approximately 4–7 points and annual publication counts by 0.75–1 articles across cohorts.
  • Citation-weighted productivity increases by 3 additional RCI-weighted articles per year after the event.
  • Men researchers are 16% more likely than women to access large-team collaborations during their early careers.
• Advantage over baselines: SDID models demonstrate significant post-event treatment effects while addressing selection bias, showing consistent benefits across cohorts and genders.
• Experiments / evaluation:
  • Data from SciSciNet (72,113 authors, 7.3 million publications).
  • Metrics: network centrality, publication counts, citation-weighted productivity.
  • Subfield analysis in Human-Computer Interaction, Data Science, and Computer Security.
  • Robustness checks with alternative event definitions and productivity measures.
• Limitations and future work:
  • Focus on Computer Science limits generalizability to other fields.
  • Gender classification based on names may oversimplify diverse gender identities.
  • Binary treatment of large-team collaboration overlooks nuanced collaboration dynamics.
  • Future work should incorporate richer datasets, altmetrics, and qualitative evidence to refine findings.

Summary

This study demonstrates that first-time large-team collaboration significantly enhances early-career researchers' network centrality and productivity, providing structural advantages and career visibility. However, it also reveals persistent gender disparities, with men being 16% more likely to access large-team collaborations. Using SDID and survival analysis, the paper highlights large-team collaboration as both a career accelerator and a mechanism of inequality. Findings underscore the need for institutional and design interventions to promote equity and inclusion in academic collaboration, particularly in labor-intensive fields like Computer Science. Future research should explore broader disciplinary contexts and refine methods to capture diverse collaboration dynamics and gender identities.