Determinants of Scientific Knowledge Creation

The research choices of scientists are driven in part by what is more or less “publishable.” This project aims to understand what kind of research is more or less publishable by studying the selection process of manuscripts submitted to scientific journals. Working with an editorial board of a high-profile journal, we use text processing techniques to characterize the claims made in the submitted manuscripts and how different types of claims are then winnowed at each step stage of review, from initial desk decisions by editors to thorough reviews by external experts. The empirical exercise may uncover biases in the selection process that would then be addressed through experiments with the editors.