Understanding Implicit Ethnic Bias: An Eye-Tracking Hci Study on How Visual Cues Influence Perceptions of Refugees

Publication Date

2-17-2026

Document Type

Article

Publication Title

Iran Journal of Computer Science

Volume

9

Issue

1

DOI

10.1007/s42044-026-00391-8

Abstract

Refugee populations worldwide face biases related to ethnicity, religion, country of origin, gender, and age. This study examines how such biases shift depending on whether participants view refugee identities through images or text labels. Using an empirical human–computer interaction (HCI) approach, we investigated ethnic bias toward Ukrainian and Syrian refugees across three conditions: (1) a cropped face image without any identifying text, (2) a cropped face image with a text label indicating ethnicity, and (3) a full-body image that includes contextual background. Each trial presented a refugee image centered between two artworks—one depicting an intellectual activity and the other a neutral pose. Participants chose the painting they felt best paired with the refugee image. Pupil size, response time (RT), and left/right choices were recorded. No significant bias appeared in the cropped face with no-label condition (ps >.05). When ethnicity was conveyed through text labels, participants tended to favor Ukrainian refugees as shown by significant differences in responses and RTs (ps <.05). In contrast, the results showed a significant ethnic bias in the full-image condition, where participants more often associated Syrian refugees with images depicting intellectual activities as reflected in both responses and pupil size (ps <.005). This study advances HCI research using eye-tracking to measure unconscious reactions to refugee images, exploring how bias changes when identity is conveyed through images rather than text labels, and employing isolated facial images to minimize contextual bias. This work demonstrates how visual cues and labels influence perception, improving methods for bias research and emphasizing the importance of presentation choices in refugee-related media.

Keywords

Cognitive load, Contextual bias, Ethnicity, Eye-tracking, Human–computer interaction (HCI), Media representation, Refugees

Department

Computer Science

Link to Full Text

Share

COinS