Seeing is Believing: How Video of Police Action Affects Criminal Justice Beliefs

• In Event: Public Opinion and Law Enforcement

Abstract

It is well established that the news media both shapes and distorts how citizens view issues of crime and justice Media coverage in the U.S. is at best loosely related to local crime conditions (Graber 1980; Lowry et al. 2003) and often reinforces stereotypes of black criminality (Gilliam et al. 1996; Gilliam and Iyengar 2000) Yet the growing availability of video footage of citizens interactions with the police, recorded by both individuals and police departments, may fundamentally change this process. By providing a more direct account of what happened, these videos may lead citizen to rely less their on preconceived notions of race and crime and make them more likely to agree on what happened and what flaws exist in the current system.

In this paper, we explore the consequences of viewing instances of police brutality on citizens attitudes about crime, justice, and politics more generally, in two ways. First, using an online sample, we randomly expose subjects to three conditions: 1) a news report describing the use of police force on a black citizen; 2) the same report plus actual video footage of the interaction from a police body camera; 3) and just the video footage of the interaction. Relative to a control group viewing an unrelated, non-political video, we expect that subjects in all conditions will be less trusting of the police, and government more broadly. We expect these effects to be moderated by subjects' race (Peffley et al. 1996) and their prior beliefs about the relative fairness of the criminal system and their tendencies to attribute crime to systemic vs individualistic causes (Peffley and Hurwitz 2010), but we expect such heterogeneity will decrease with the prominence of the video footage.

Second, using the audio from 25,000 body camera videos from the state of Washington, we are developing a classifier that will differentiate between aggressive and non-aggressive police-citizen encounters using the audio from coded body camera videos. This data will be merged with another collection of both local and national news coverage, to explore what types of interactions get covered and whether interactions that contain actual footage are covered differently than those which do not. This portion of the study will increase the external validity of the aforementioned experiments. If citizens are found to rely less on their preconceived notions of race and crime when exposed to actual body camera footage, then the degree to which those videos actually reach the public is of considerable import. In the conclusion, we explain how these data can also be used to study framing effects, something we will discuss in the present paper but will not fully test until the next study in our broader research agenda.

Authors

• Paul Franz Testa, Brown University

• Bryce Dietrich, University of Iowa