To Protect and To Serve? Analyzing Entity-Centric Framing of Police Violence

Framing has significant but subtle effects on public opinion and policy. We\npropose an NLP framework to measure entity-centric frames. We use it to\nunderstand media coverage on police violence in the United States in a new\nPolice Violence Frames Corpus of 82k news articles spanning 7k police killings.\nOur work uncovers more than a dozen framing devices and reveals significant\ndifferences in the way liberal and conservative news sources frame both the\nissue of police violence and the entities involved. Conservative sources\nemphasize when the victim is armed or attacking an officer and are more likely\nto mention the victim's criminal record. Liberal sources focus more on the\nunderlying systemic injustice, highlighting the victim's race and that they\nwere unarmed. We discover temporary spikes in these injustice frames near\nhigh-profile shooting events, and finally, we show protest volume correlates\nwith and precedes media framing decisions.\n