Investigating Academic Graph‐Based Factors behind Funding Success in National Institutes of Health

Abstract

While major funding agencies are striving for diversity and fairness, the mechanisms behind funding success have yet to be fully elucidated. Existing studies reveal valuable evidences about the effect of the applicant's individual attributes, e.g., gender and age, on the funding success. However, the relationship between funding success and academic activities, e.g., collaborator's characteristics, remains underexplored. This work collects massive scholarly data from open academic graphs and public data about National Institutes of Health awards to investigate the effect of various academic graph‐based factors on the “K to R” success. Leveraging a heterogeneous graph model for predicting the “K to R” success, we regard the gain in the model performance of a factor as a proxy variable for the magnitude of its effect on the “K to R” success. Our preliminary results suggest that interest by peers in the applicant's research and the timing of the interest are strongly correlated with the outcome. Meanwhile, the applicant's social connections, e.g., their collaborators, can also contribute to the outcome.