Association for Public Policy Analysis & Management

RAND Corporation and Cardiff University are investigating the utility of Twitter data for understanding what types, for whom, and where online hate speech acts as an online signature that predicts offline hate crime. The overarching goals of the research project are to (i) address the lack of reliable knowledge about hate crime prevalence in the U.S. by (ii) identifying and analyzing online hate speech and to (iii) connect the prevalence of online hate speech to offline hate crimes, improving upon the existing data on hate crimes. To that end, the research team are developing a statistical model connecting online hate speech and offline hate crime, using tweets and reported hate crimes. The project uses Los Angeles County as a test case and a possible model for wider application. The current paper will describe the types of hate speech identified during a year-long collection period in LA County, and then present preliminary results from the statistical model examining the relationship between online hate speech and offline reported incidents. It will also discuss the strengths and weakness of existing sources of data on hate crimes, the underlying data collection processes, and potential solutions for overcoming existing data limitations.