Measuring Geographic Sentiment toward Police Using Social Media Data

Abstract

Using Twitter messages published online from October 2018 to June 2019, and opinion mining (OM) technology, the current study analyzes the geographic sentiments toward police in 82 metropolitan areas within the United States. Building on the frameworks of the neighborhood social contextual models, the construct validity of “sentiment toward the police” is assessed via its relationship with the features of various metropolitan areas. Results of the regression analysis indicate that the violent crime rate, racial heterogeneity, and economic disadvantage significantly affect sentiment toward the police. Our results suggest that opinion mining of social media can be an important instrument to understand public sentiment toward the police.