Thanks to the generous support from the NSF grant #1933803 Social Dynamics of Knowledge Transfer Through Scientific Mentorship and Publication and complementary funding from the Science of Science and Computational Discovery Lab, we are pleased to announce micro-grants funding for six projects, each receiving $ 2,000. We were blown away by the quality, breadth, and ambition of the proposals. Congratulations to all!
About the grants
The goal of these micro-grants is to create momentum for the project members to get started on a larger scale project that could result in a publication or a larger grant.
List of funded grants
Rethinking Validity in Data-driven Science of Science
- Caifan Du, The University of Texas at Austin
- Silvia Gutiérrez, Leipzig University
The increasing use of data sources generated and collected by a third party raises a concern for validity in data-intensive research. We propose a study to investigate data sources with high utility to the Science of Science community, intentionally selected for gaining a more holistic understanding of how different types of third-party data sources could affect validity in research and how researchers could thus mitigate their possible threats to validity primarily through research design.
Identifying Bibliometric Features Associated with the Use of Questionable Research Practices
- Nicholas Fox, Center for Open Science
- Olivia Miske, Center for Open Science
Published research in psychology is not as replicable as desired. One reason for failures to replicate is the existence of false positive findings in the published literature. Using questionable research practices (QRPs) is one way these false positive findings may enter the literature. This project seeks to better understand the bibliographic “fingerprint” of papers written by authors who have self-identified as QRP users. If differences exist between papers written with QRP-using authors and those without QRP-using authors, those differences could be used as potential targets for interventions designed to reduce the use of QRPs in the future.
Enabling interactive exploration of rich scientific multigraphs with Neo4j Bloom
- Stanislava Gardasevic, University of Hawaii
- Alec McGail, Cornell University
- Jodi Schneider, University of Illinois Urbana-Champaign
- Manika Lamba, University of Delhi
Science has been described as a “complex, self-organizing, and evolving network of scholars, projects, papers and ideas” (Fortunato et al., 2018, p. 2). Tools for exploring the vast data available regarding science and its development could help researchers explore and ultimately recount the history and development of academic fields, help new students navigate the complex web they are to engage with as new researchers, or help to identify fraudulent scientific claims. Current state of the art projects in this area address visualization tools and mapping of science by providing static, broad pictures of specific domains (e.g. Börner, 2010; Chen, 2013). By contrast we seek to provide general queries that could be reused to explore any dataset representing a scholarly network. We also wish to build a prototype tool which makes navigating science more independent, creative, and fun. In this document we propose a series of workshops which develop interactive visualizations of complex multimodal scientific relationships, paired with a user interaction study which assesses to what extent these visualizations help new PhD students to get acquainted with their intended field of study in an engaging way.
Making the tough shot–studying successful topic switch strategies using large-scale scholarly databases
- David Janku, Seznam
- Alexander J. Gates, Northeastern University
- Mengyi Sun, University of Michigan, Ann Arbor
Mounting evidence suggests that switching research topics can be harmful for scientific careers (1-3). Paradoxically, several studies also found that the most successful researchers have evolving research interests and that breaking away from the research topics of their Ph.D. advisors is a key step for building their own success (4, 5). That the most successful scientists make the toughest move suggests they must have done something special. We plan to integrate several databases of scholarly activities, such as the Microsoft Academic Graph (MAG)(6), the “MENTORSHIP” database(7), and PubMed(8), to identify the spectrum over which individuals switch topics and explore the potential factors that differentiate those who encountered difficulty and those who ultimately achieved success. Our research has important implications for researchers’ selection of topics, especially for early-career scientists. For the whole scientific community, identifying factors that encourage individual scientists to explore more diverse topics could potentially accelerate scientific discovery.
Citation Contact Tracing: Classification of Retracted COVID-19 Articles
- Eunice Chan, Western University
- Alicia Takaoka, University of Hawai‘i at Hilo
This project builds on the work done at Science of Science Summer School by the COVID-19 Citation Retraction Team. Forty-six retracted COVID-19 citations in the in the areas of medicine and technology were found on PubMed. This dataset was used to determine if a bipartite network was present, identify network connectivity among retracted authors, visually display commonly used keywords, and evaluate the papers in which these articles were cited in order to determine if there was a propagation of misinformation. In the next iteration of this project, articles on Google Scholar will be used to determine a discipline of publication network, a retracted paper network, and a citation network to identify the possible propagation of misinformation.
Science in Motion: Tracing the reconstruction of social and semantic communities over time
- Robert Ward, Georgia Institute of Technology
- Alec Mcgail, Cornell University
- Aaron Mendon Plasek, Columbia University
Science is constructed twice, first when it is written and again when it is referenced by others. Each new paper stakes a claim for itself in the scientific landscape and redefines the historical positions of the papers and communities it cites. We seek to develop tools to empirically measure both how the words of scientific texts, whether they be abstracts, journal articles, conferences proceedings, and other textual sources, change their individual meanings (as measured via embeddings) with respect to each other over time. We aim to use word change to document disciplinary change within coteries of practice, and how ideas and methods circulate within these coteries. We also seek to explore how tracking word change over time can serve as a counterfactual to citation analysis.