Paper short abstract

This paper examines how research infrastructures produce asymmetric and discriminatory representation through SDG classification in bibliometric databases and name-based demographic inference tools such as NamSor.

Paper long abstract

This paper examines algorithmic representational asymmetry in research infrastructure through two linked case studies: SDG-related classification in major bibliometric databases and name-based demographic inference tools used in research policy. The first study analyzes how SDG classifications in Web of Science, Scopus, and OpenAlex represent sustainability research, showing that these systems systematically privilege Global North institutions, economic superpowers, and topics aligned with dominant publication patterns, while overlooking marginalized groups, poorer countries, and other SDG-relevant populations. The second study assesses NamSor, a tool used to infer gender, country of origin, and ethnicity from names, and shows both its discriminatory limitations and its selective usefulness in policy contexts. Together, the two cases reveal a common problem: classification tools presented as neutral infrastructure can encode uneven visibility, misrecognition, and structural bias. Rather than simply measuring existing inequalities, these systems help produce them by shaping which researchers, populations, and themes become legible in data-driven research governance. By bringing bibliometric classification and onomastic inference into the same analytical frame, the paper contributes to STS debates on infrastructure, classification, epistemic injustice, and the politics of algorithmic representation/visibility. It argues that representational asymmetry should be understood as a central feature of contemporary research infrastructures, with important consequences for research evaluation, diversity monitoring, and science policy.