Reflecting on Social Bias: Challenges and Opportunities for Computational Social Science

Computational social science (CSS) is an evolving field with transformative potential to expose and address pressing societal challenges. Yet, because CSS, much like other scientific disciplines, operates within structural contexts addressing social bias remains a critical issue requiring attention in the field. Social bias, defined as the systematic discrimination and stereotyping of groups and individuals, relates to CSS in multiple ways. Drawing on literature, we critically examine these connections across three perspectives: understanding, detecting, and mitigating social bias. First, we provide a foundation for understanding how social bias shapes both CSS as a research field and its methodology. Second, we highlight how social bias can be detected not only within the CSS methodology itself but also through CSS applied to broader societal domains. Finally, moving forward, we propose strategies to mitigate social bias in the CSS methodology and CSS as a research field. With our contribution, we seek to demonstrate the breadth of ways in which social bias and CSS intersect and to chart paths toward a more equitable and robust CSS.