Who am I, and who are you, and who are we? A Scientometric Analysis of Gender and Geography in HCI

Title of the Paper

Who am I, and who are you, and who are we? A Scientometric Analysis of Gender and Geography in HCI

Paper Information

• Research Domain: Scientometric analysis of gender and geographic diversity in Human-Computer Interaction (HCI)
• Keywords: Gender, Geography, Diversity, Research Equity, Conference Authors, HCI, Academic Productivity, COVID-19

Research Background and Issues

• Identified Problems or Challenges:

  • Gender and geographic diversity among authors and participants at HCI conferences are imbalanced, with pronounced disparities at the intersection of gender and region.
  • Female researchers are significantly underrepresented at conferences compared to male researchers, particularly at prestigious conferences like CHI.
  • Geographic location (conference venue and authors’ countries) significantly impacts academic participation, favoring researchers from North America and Europe, while participation from Latin America, Africa, and the Middle East remains extremely low.
  • The HCI community has focused on diversity dimensions (e.g., gender, geographic location) in external user studies, but internal diversity issues within the community are largely overlooked.

• Significance:

  • Gender and geographic diversity are widely recognized as critical in academic communities: studies show that diverse teams enhance innovation quality.
  • Conference publication and participation are key metrics for academic excellence, and inequitable distribution of these opportunities may lead to systemic unfairness for certain groups.
  • Lack of comprehensive gender and geographic representation may limit the global applicability of HCI research and its universal impact on user groups.

• Research Motivation and Related Work:

  • Existing literature has explored the impact of diversity factors within user groups in HCI, but there is limited focus on the diversity of authors and participants within the community itself.
  • There is currently a lack of systematic scientometric analyses quantifying gender and geographic representation at HCI conferences.

Proposed Solution

• Methods or Solutions:

  • Employ scientometric methods to conduct historical and comparative analyses of author data from six major HCI conferences (CHI, DIS, UbiComp, CSCW, OzCHI, and NordiCHI) between 2015 and 2021.
  • Analyze the distribution of conference authors from gender and geographic perspectives, including relationships between author gender, workplace location (country/region), and conference venue.
  • Investigate the impact of online conferences (during the pandemic, 2020-2021) on gender and geographic participation, comparing data with in-person conferences.

• Innovations:

  • First systematic integration of gender and geographic intersectionality into scientometric analysis of HCI conferences.
  • Comprehensive analysis of the impact of international and regional conferences on participation by researchers from different geographic areas, including gender-neutral and nonbinary authors.
  • Propose experimental solutions (e.g., regional conference fee adjustments, quota systems, transparency reporting) to address inequalities in gender and geographic backgrounds.

• Implementation Steps and Key Techniques:

  1. Data Collection and Cleaning:
     • Extract conference data from the ACM Digital Library, including authors’ countries, genders, and affiliated institutions.
     • Use manual verification and automated tools (e.g., the "SexMachine" algorithm) to identify gender information.
  2. Data Processing and Analysis:
     • Conduct statistical analysis of gender distribution among first authors and co-authors.
     • Examine the impact of conference venues on the geographic distribution of authors.
  3. Special Issue Analysis:
     • Study the effect of the COVID-19 pandemic on equity in online/in-person conferences.
     • Explore participation patterns across different cultural backgrounds and regions (e.g., Australia, Nordic countries, Asia).
  4. Statistical Validation:
     • Apply log-linear tests and chi-square tests to evaluate interactions between factors.

Research Findings

• Specific Results:

  1. Gender Representation:
     • Female authors’ participation rates are higher at lower-prestige conferences (DIS, CSCW) and lower at high-prestige conferences (CHI, UbiComp).
     • Women have higher participation rates at locally organized conferences, but their representation significantly decreases when conferences are held far from their work regions (e.g., in Asia).
  2. Geographic Representation:
     • Researchers from North America and Northern Europe participate in international conferences at rates exceeding their population proportion.
     • Researchers from Asia, Latin America, Africa, and the Middle East are significantly underrepresented.
     • Online conferences led to a notable increase in participation by researchers from Asia and Australia, particularly at CHI.
  3. Gender-Neutral and Nonbinary Authors:
     • Explicitly identified gender-neutral or nonbinary authors constitute a very small proportion (7 authors explicitly identified).
     • An increase in nonbinary authors was observed in 2021, but the proportion remains significantly below expectations.

• Comparison with Existing Research:

  • This study corroborates preliminary observations in other literature regarding gender inequality in the HCI community, while providing quantitative data to support these findings.
  • Comparative analysis reveals that certain conferences (e.g., CSCW and DIS) perform better in gender diversity, whereas conferences like UbiComp exhibit the most severe issues.

• Limitations and Future Directions:

  1. Limitations:
     • Data samples from certain regions (e.g., Africa, the Middle East) are too small to allow comprehensive analysis of their impact.
     • Diversity beyond gender (e.g., race, sexual orientation) is not explored.
     • Automated algorithms for gender identification in Asian names are not fully reliable.
  2. Future Directions:
     • Expand data scope to include long-term tracking of academic participation in underrepresented regions like Africa and Latin America.
     • Incorporate exploration of diversity in race, sexual orientation, and other identity backgrounds.
     • Study interactions between gender, career seniority, and academic collaboration networks.
     • Conduct in-depth investigations into gender-neutral and nonbinary authors, including social and cultural barriers.

References

This paper provides a rich set of references covering diversity research, structural analyses of the HCI community, and studies on the impact of the COVID-19 pandemic on academic participation.