Patterns of Force: A Comparison of Data Sources on Officer-Involved Homicide Rates
Open Access
- Author:
- Fry, Sarah V
- Graduate Program:
- Sociology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 19, 2024
- Committee Members:
- Eric Plutzer, Outside Unit & Field Member
Scott Yabiku, Professor in Charge/Director of Graduate Studies
Richard Felson, Major Field Member
John Iceland, Major Field Member
Eric Baumer, Chair & Dissertation Advisor - Keywords:
- police
use of force
homicide by police
justifiable homicide
structural equation modeling
sem
organizational complexity
organizational control
law enforcement
police brutality
mapping police violence
fatal encounters
the counted
supplementary homicide reports
shr
uniform crime reports - Abstract:
- In this dissertation, I develop a model for measurement of homicide by police that draws on multiple sources of data and then utilize this measurement model to examine potential city/agency-level predictors of homicide by police. The issue of data choice is of key importance to research on geographic variations in homicide by police. I utilize data from one official data source on homicide by police (the Supplementary Homicide Reports) and three media-based data sources (Fatal Encounters, Mapping Police Violence, and The Counted) as indicators of the latent, “true” 2015 to 2016 homicide by police rate. I then use this measurement model to examine various potential predictors of the homicide by police rate including city-specific and agency-specific factors. I then compare the model predicting latent rates of homicide by police to models where only one data source is used as the outcome. A key finding regarding city-specific predictors is that a city’s percent Black and Black-White housing segregation are associated with lower homicide by police rates. Some key findings from the agency-specific predictors tested are that more restrictive vehicle pursuit policies, requiring more new recruits to engage in community policing training, and authorizing the use of chemical weapons (like tear gas) are associated with lower homicide by police rates while having more specialized unit types and authorizing officers to use soft projectiles or neck restraining techniques (such as chokeholds or vascular neck restraint) are associated with higher homicide by police rates. In terms of the findings related to measurement, I find that all four of the data sources on homicide by police appear to relate to the same underlying factor. However, one data source (Mapping Police Violence) is very strongly related to the estimated latent homicide by police rate and models using it are nearly identical to the more complex model using all four indicators of the latent rate.