AI Safety

Study finds ideal gender ratio for research teams maximizes citations

Mixed-gender teams outperform same-gender teams by up to 15% in citations.

Deep Dive

Researchers Chengzhi Zhang, Jiaqi Zeng, and Yi Zhao studied NLP and Library & Information Science papers, finding an inverted U-shaped relationship between team gender diversity and citation impact. The ideal balance: one gender makes up 5–15% of authors. Mixed-gender teams gradually earn more citations than same-gender teams, though women remain underrepresented in both fields—especially in NLP.

Key Points
  • Mixed-gender teams achieve 8–12% higher average citations than same-gender teams in NLP and LIS.
  • Optimal gender ratio: minority gender makes up 5–15% of authors (inverted U-curve).
  • Female underrepresentation is more pronounced in NLP than LIS, yet both benefit from diversity.

Why It Matters

Provides data-driven guidance for forming research teams that maximize citation impact and scientific influence.

📬 Get the top 10 AI stories daily