Media portrayal and location behavior: Crime reporting and house prices in U.S. cities

Media representations of cities shape public perception of them. However, the extent to which these media-driven perceptions subsequently influence individual decisions about where to live has remained underexplored. This study leverages advanced Natural Language Processing (NLP) techniques to capture media coverage of crime and compare it to actual reported crime. We show that variations in media attention to crime across U.S. cities help explain local house price dynamics. Media portrayal of crime explains house price dynamics better than FBI crime rates, although both measures nonetheless complement each other. Our results call for more attention to behavioral and cognitive explanations of urban growth and decline and, methodologically, our approach contributes to the development of “digital urban studies” scholarship.