Paper short abstract:

This project investigates how police-officer involved homicide (POIH) data gaps relate to semantic inconsistencies in naming and classifying POIH homicides. Definitions and expressions by which these homicides are referred to greatly vary among agencies and activists groups.

Paper long abstract:

Newspaper headlines throughout 2014-15 reported a spate of killings of unarmed African Americans by police officers around the United States. The revelation of this disturbing trend of racially targeted violence exposed the difficulty of tallying the extent of these killings, as there exists a significant gap in data on the number of police officer-involved homicides (POIH) across the US. States are not required by law to report killings by law enforcement, and when states do report, their data is collected and interpreted differently across states and local jurisdictions.

The incompleteness and lack of standardization of national police-officer involved homicide (POIH) data provided the starting point for this research. This paper draws on our previous assessment of several POIH datasets for Los Angeles County collected by federal agencies, local organizations and activist groups. (Currie et al., under review). We investigated these datasets applying Geoff Bowker's (1994) technique of "infrastructural inversion," and drew insights from what Dalton & Thatcher (2014) define as "counter-data actions" through the organization of a civic hackathon.

Through the hackathon, we found that in many cases, the gaps in the POIH data were related to semantic inconsistencies in naming and classifying instances of POIH, which render a life taken incommensurable with the data point that represents it. Our paper ultimately details how involving communities in critical analysis of POIH data can lead to the development of projects that promote social justice and act as a countervailing force to the rationalization of complex events processed as discrete POIH data points.