Diversity, Exclusion, and Networking in Research Communities

Aliens researching!
Image credit: People illustrations by Storyset [https://storyset.com/people]

Breaking the ice and networking

Having attended the ACM CHI Conference on Human Factors in Computing Systems (a premier international conference on Human-Computer Interaction) several times since 2019, I observed that networking can be challenging for students and junior researchers. They (including myself) often struggle with figuring out how to initiate conversations, who to talk to, or what topics to discuss.

Power dynamics can be intimidating for juniors, such as graduate students or recent Ph.D. graduates, making it hard for them to initiate conversations with a professor or an industry veteran. So, senior members and organizers may play an integral role in engaging with junior community members to promote diversity and inclusion. When seniors initiate the interaction, it can ease the awkwardness and break the ice.

During my most recent conference attendance, I made an open call for students to meet with me. To my surprise, many responded with enthusiasm. Without my proactive invitation, few would have approached me.

In this vein, recently, I discovered a website (adplist.com) where people offer free mentorship, with many senior members willing to engage in conversation with juniors. I find this inspiring and hope the site will continue to grow, and more people will sign up as mentors and mentees. Services like these can help bridge the gap between experienced individuals and newcomers.

Exclusivity of research

While some argue that parties are the conferences’ “meat” — yes, academics do party! — and are the breeding ground for collaborations and job hunting, I’ve rarely heard my peers claim substantial benefits from these events. Perhaps the dynamics differ at more senior levels.

I enjoy the social aspects of conferences, as face-to-face interactions can significantly influence potential collaborations. Yet, over the years, I’ve found it increasingly difficult to approach unfamiliar faces. Some CHI events are often invitation-only (I have heard similar stories from other conferences), fostering, by design, a culture of exclusion, possibly inadvertently.

Those unable to attend in-person conferences face additional substantial disadvantages — and exclusion. A quick analysis of 59 A* conferences in 2023 shows that most were held in countries with rigorous visa applications. High conference fees present another hurdle for those without considerable scholarships or supervisor grants. Consequently, those attending these conferences are much more likely to encounter research leaders in their respective fields. The lack of such opportunities can negatively impact career growth.

Only a very small set of countries get a chance to host a flagship computer science conference (only in 2023).

Power of connections

The significance of networking and personal connections became apparent when I noticed colleagues with similar experiences advancing more quickly. Although I felt excluded due to my limited connections, those with the “right” connections benefited from a sense of inclusion.

It took me roughly five years to penetrate some niche research communities as an associate chair or being part of a program committee, possibly due to my struggles with networking and the exclusive culture of research. Though, I’m still unsure whether I’m in or not!

Over the years, I wrote numerous external reviews for conferences, workshops, and journals. However, my applications for starting to become a program member only gained acceptance after I became a senior research scientist at Bell Labs and gained two years of post-Ph.D. experience, and had several papers accepted at reputable venues. It’s unclear which achievement carried the most weight!

Research shows that research lacks diversity!

Research studies corroborate my personal experiences. My friend, Naomi Saphra, wrote a short article discussing the segregation of Chinese and American citation networks, showing that authors often cite those from their own countries. This isn’t limited to citations but extends to the general operation of the research machinery.

Similarly, our recent research revealed a lack of diversity in a conference championing fairness and transparency in AI research (the ACM FAccT Conference on Fairness, Accountability, and Transparency). Many studies presented at this conference are based exlusively on Western datasets.

This issue isn’t unique to this conference; others, like CHI, share similar issues — lack of diversity. These patterns extend beyond datasets to other fundamental elements such as grants, scholarships, postdoc positions, Ph.D. programs, collaborations, partnerships, visas, and more.

World wide distribution of FAccT papers. ratio(𝜓𝑠𝑐) showing over-represented(𝜓𝑠𝑐 >1) and under-represented countries (𝜓𝑠𝑐 < 1) based on the datasets under study between 2018 — 2022. The papers ratio 𝜓𝑠𝑐 is the number of papers from a given country 𝑐 (where the study participants are from) over the number of all papers in FAccT, divided by 𝑐’s population over the world’s population. Countries in light gray (𝜓𝑠𝑐 = 0) did not have participants in FAccT between 2018 — 2022, while those in darker shades of blue and red indicate under- and over-represented countries, respectively. See the full analysis and paper [here](https://dl.acm.org/doi/10.1145/3593013.3593985).

What can the research community do?

Despite my criticisms, I genuinely appreciate the research communities’ openness and self-reflection (thus, writing this post). Therefore, I’d like to propose the following to improve inclusivity and diversity in the research community:

  • Promote mentorship programs (like adplist.com) and recognize these efforts on resumes. Incorporating such initiatives into conference programs could be beneficial.
  • Encourage international collaborations between institutions, which would require public and industry funders to champion such partnerships. Grant programs requiring partnerships across multiple countries may prove helpful!
  • Require researchers to complete a standard form detailing their data and participants at the time of submission. This would facilitate the production of beautiful and informative visualizations and reports for each conference and, hopefully, one day, for all scientific fields.
  • Advocate for more in-person, authentic interactions with people for the design and development of research and technology. Besides data collected in labs and crowdsourcing services, engaging with people directly can be priceless.
  • Diversify conference locations. Countries like Rwanda, Turkey, or the UAE could be more accessible to more attendees. Local exposure to different cultures (and food!) can provide researchers with enriching experiences while supporting local economies. After all, an influx of a few thousand visitors will benefit a developing country’s economy more than a developed country!

Endnote: The views shared in this post are shaped by my personal experiences within the research domains I associate myself with, namely, privacy, security, and human-computer interaction. The five-year timeline delineated here starts in 2018, marking the beginning of my Ph.D. journey. Though, my exposure to research began a year prior to this, during my tenure as a research assistant.

Mohammad Tahaei
Mohammad Tahaei
Research Lead

Research Lead