Making Use of Multilevel Modeling in Sexual Assault Research: Successes, Challenges, and Considerations - CANCELLED

• In Event: Advances in Modeling and Causal Inference

Abstract

Police officers play an important role for sexual assault survivors seeking help from the criminal legal system. One of their main responsibilities is to investigate the crime to determine if there is enough evidence to arrest a suspect. Research guided by focal concerns and ecological theories indicate police arrest decisions are likely influenced by legal and extralegal factors at multiple levels of analysis, yet most work has focused on incident-level factors. Further, almost all research on police arrest decisions has been conducted at a single level of analysis. My work uses multilevel modeling to remedy this. Specifically, I created a two-level logistic model to examine the effect of incident (level 1), police department (level 2), and community (level 2) factors on arrest decisions in over 20,000 sexual assault cases across the U.S. Data were pulled from three 2020 publicly available data sources: the National Incident-Based Reporting System (NIBRS), the Law Enforcement Management and Administrative Statistics (LEMAS), and the U.S. Census Bureau. I will discuss the main methodological successes and challenges in this project, present select findings, and discuss implications for using more diverse statistical methods in future research.

Author

• Erin Hoffman, DePaul University