Lucie Kovarik
Pronouns: she/her
Research Mentor(s): Jinseok Kim
Research Mentor School/College/Department: Institute for Social Research / Information
Program: UROPF
Session: Session 3 (11:00am – 11:50am)
Authors:
Abstract
This research project investigates the gender productivity gap as well as sampling issues and entity issues. The proposed research project goals include promoting fair evaluation of women in science by accurately labeling data and evaluating machine learning methods. The first aim of the project is to create large-scale labeled data to train algorithmic models to merge the same author entities split under different names. The second aim is to implement the entity disambiguation models learned from aim number 1 on author names recorded in PubMed that indexes research papers in biomedicine. Thirdly, we aim to analyze the disambiguated PubMed data from aim number 2 and demonstrate how the correct identification of female scholars can lead us to different understanding of research productivity and impact of female scholars in biomedicine. The methodologies being used are data organization and verification and machine learning coding. A database of researchers was combed through for mistakes and verifications. We analyze our findings through a disambiguation system. We hope to find how name changes are an important factor to creating an inequality between male and female researchers. Our project will yield a database and line of code that haven’t been created before and form conclusions and discussion based on the current workings. Our main conclusions are that name changes affect the fair evaluation of women in science as female researchers as the bibliographic data services typically don’t account for the potentiality of that occurring. These ideas relate to the big picture because this level of scrutiny and data ambiguity determination can compound and lead to more widespread code and database accuracy to prevent the inequality of male and female researchers. The social benefit of this research can be related to any female researcher regardless of name change status.
