Exploring Diversity in HCI Through Decades: Who’s In, Who’s Out, and Who’s Still Missing?

Abstract

A central question in Human-Computer Interaction (HCI) is: How do we shape technology, and how does technology shape us in return? [1]. This invites a critical follow-up: Who is “we”? Who has been represented in the field, and who remains absent? To meaningfully engage with diversity, equity, and inclusion, we must examine whose voices have historically informed HCI research. This study addresses that gap through a large-scale content analysis of CHI proceedings as HCI’s most influential venue. I explore how often age, gender, and health have been represented over time, and how these intersect with study purpose, methodology, phase, technology, and users. The aim is to map trends, highlight underrepresented populations, and critically reflect on the field’s trajectory.

I analysed CHI proceedings from the midpoint of each of the past three decades: 2005, 2015, and 2024 using a two-step approach using Atlas.ti. First, I performed keyword searches in paper titles and abstracts to deductively identify references to age (eg. child, elder), gender (eg. man, woman), and health (eg. impairment, ability). This provided a quantitative overview of representation over time. In the second step, informed by the grounded theory approach, I conducted inductive manual coding of the identified papers across five analytical dimensions: purpose, methodology, phase, technology and user. This framework enabled a cross-sectional view of how diversity is positioned in HCI research.

CHI has grown rapidly from only 93 publications in 2005 to 1057 in 2024. But has its attention to diversity grown as well? The preliminary results suggest that age was the earliest and most consistent focus, especially children: 5 papers in 2005, 11 in 2015, and 39 in 2024, mainly around learning & education. Most were exploratory and didn’t always involve the target group. Older adults were barely represented, typically appearing in evaluation phases, not design. Gender appeared meaningfully in 2015. By 2024, there was growth in LGBTQ+ studies and work on women's reproductive health and digital privacy. However, these areas remain niche. Though Feminist HCI is often cited, its influence appears largely theoretical. Health papers rose from just 5% of CHI papers in 2005/2015 to around 20% in 2024. Most focus on physical or sensory disabilities, especially visual impairment. Mental health is still under-explored, with few studies on depression or anxiety. Intersectional approaches remain rare with only 45 studies investigating at least two diversity dimensions in 2024. As the use of AI-based symptom checkers is rising and are often trained on male-centric data, we can miss how marginalised groups experience illness, reinforcing diagnostic bias (e.g. heart attack, autism).

This study assumes inclusive design starts with inclusive research: who gets represented shapes who gets considered. Technologies remain designed for the young, tested on the old, and built on narrow assumptions thus risking reproducing social inequities. While findings are indicative rather than exhaustive, as limited by keyword scope, abstract focus, and single-researcher coding, they highlight important gaps. 

References

[1] D. L. Cogburn, "HCI in the so-called Developing World," Interactions, vol. 10, no. 2, pp. 80–87, 2003. doi: 10.1145/637848.637866.