Research Article
Depicting Recommendations in Academia: How ResearchGate Communicates with Its Users (via Design or upon Request) About Recommender Algorithms
@INPROCEEDINGS{10.1007/978-3-030-77417-2_1, author={Luciana Monteiro-Krebs and Bieke Zaman and Nyi-Nyi Htun and S\~{o}nia Elisa Caregnato and David Geerts}, title={Depicting Recommendations in Academia: How ResearchGate Communicates with Its Users (via Design or upon Request) About Recommender Algorithms}, proceedings={Data and Information in Online Environments. Second EAI International Conference, DIONE 2021, Virtual Event, March 10--12, 2021, Proceedings}, proceedings_a={DIONE}, year={2021}, month={6}, keywords={Academic Social Networking Sites Algorithmic transparency Recommender systems ResearchGate Human-Computer Interaction New materialism GDPR}, doi={10.1007/978-3-030-77417-2_1} }- Luciana Monteiro-Krebs
Bieke Zaman
Nyi-Nyi Htun
Sônia Elisa Caregnato
David Geerts
Year: 2021
Depicting Recommendations in Academia: How ResearchGate Communicates with Its Users (via Design or upon Request) About Recommender Algorithms
DIONE
Springer
DOI: 10.1007/978-3-030-77417-2_1
Abstract
Academic social networking sites (ASNSs) are increasingly using recommender systems to deliver relevant content to their users. Meanwhile, the profound opacity of the algorithms that filter information makes it difficult to distinguish among the elements that may suffer and/or exert influence over the interactions within ASNSs. In this article, we investigate how ResearchGate communicates with its users about the recommender algorithms used in the platform. We employ a walkthrough method in two steps (interface analysis and company inquiry using the General Data Protection Regulation (GDPR)) to investigate ResearchGate’s communication strategies (via the design of the platform or upon request) regarding the use of recommender algorithms. The results show six main entities (Researcher, Institution, Research project, Publication, Job and Questions), along with the large amount of metadata involved in the recommendations. We show evidence of the mechanisms of selection, commodification and profiling, and in practice demonstrate the mutual shaping of the different actors (users, content, platform) in their interactions within the platform. We discuss the communication strategy of the company to shy away from providing details on automated profiling.
