| The under-representation of women in most science, technology, engineering, and mathematical (STEM) disciplines is a troubling phenomenon. Many studies demonstrate that there is still a significant gender bias, especially at higher career levels. It is the major concern of the entire scientific community to elucidate the systemic factors contributing to the gender discrepancy in career performance and advancement in STEM disciplines. However, the complex nature of the problem, lack of data, and the prevalence of qualitative methods make the task highly challenging. Fortunately, the rise of complex system research and advancements in information technology provide us exciting possibilities. With bibliographical data on millions of scientific publications at hand, we can now conduct data collection and analysis on a unprecedentedly large scale. In this dissertation I report a systematic, quantitative investigation of the key structural factors that lead to the current gender discrepancies in STEM. We first fully characterize the academic careers of thousands of researchers in terms of productivity, impact, and collaboration patterns. Using statistically-sound methods, we precisely quantify the gender differences and identify factors that emerge only from a macroscopic perspective. Our analyses reveal for the first time the roles of resource requirements and academic career-choice risk on gender differences in publication rate and impact, as well as field-dependent scientific collaboration patterns. Our findings have significant policy implications for achieving diversity at the faculty level within the STEM disciplines. To accurately measure scientific research impact, we design an objective evaluation system for scientific research impact based on a thorough understanding of the citation process and rigorous statistical framework. In contrast to previous heuristic solutions, our measure is particularly reliable and is robust against manipulation. |
PDF Full Text Request