Setting the Data Standard
Accountable Now makes an impact by sharing data people can use. Law enforcement agencies (LEAs) across the country use different processes and terminology for tracking police use of force. In order to be useful, datasets need standard categories and matching definitions. LCCHR and NORC work together to harmonize use of force data so it is uniform and users can make comparisons and find patterns.
Harmonizing Police Use of Force Data
Local LEAs across the United States record police use of force differently. NORC harmonizes LEA's data by combining them into a single dataset with matching definitions. With harmonized data, we can compare datasets from multiple cities and uncover patterns.
LEA use of force datasets contain multiple categories of information. The first step in harmonizing datasets from different LEAs was to create a single data model with specific categories and definitions. Next, LCEF and NORC experts created standards for each category and mapped data from each LEA onto a common set of values across the datasets. NORC then "normalized" the defining three key components:
Incident – each use of force that an officer applied. Incident data often includes the date and location, what led to the incident, and what occurred during the incident.
Civilians – the characteristics of any individual involved (other than the officer), like their age, race and gender. This can also include details about whether the individual sustained an injury, whether they had a criminal record, or whether charges were filed from the incident.
Force – the description of tactics, weapons and the amount of force used and can vary across LEAs. Read the Definitions in the next section to see descriptions of types of force.
LCEF and NORC defined standard categories of use of force for the LEA datasets:
Baton: baton strikes, baton open modes, PR-24s, expandable batons.
Canine/k9: K-9 deployments, bites, canine contact, K-9 inflicted injuries.
Chemical spray/pepper spray: OC spray, pepperball saturation, CS foggers, CS spray, CN gas, pepper spray, other chemical agents.
Chokehold/neckhold: neck holds, LVNR.
Firearm: discharge of handgun, shotgun, rifle, service weapon.
Hard hands: punches, elbow or closed fist and hand strikes, kicks, head butts, active resistance (elbow, feet, hands, knee), push.
Improvised impact weapon: flashlights, non-traditional impact weapons, other impact weapons.
Officer presence/verbal commands: verbal commands, Combat stances.
Projectile: impact munition, 40mm Less lethal, bean bags, shotgun less lethal, pepperball impact.
Soft-hands: grabs, holds, joint locks, elbow grip, wrist grip, shoulder grip, pressure point compliance, hand controlled escort, handcuffing, hands, defense tech.
Takedown: handcuffing take down, body weight leverage, leg sweep, tackles.
Tasers: CED drive stun, CED prongs, CEW.
Vehicle: vehicle pursuit, vehicle as weapon, pursuit immobilization technique.
Weapon display includes: baton display, Taser display, Taser display arc, firearm display.
Other: foot pursuit, ballistic shields, knife, flash bangs, legal interventions, force specified as "Other".
Race and ethnicity are not captured consistently across the initial set of LEAs that were harmonized. Austin is the only LEA among the initial datasets to capture race and ethnicity separately. So that we could compare all the datasets, we classified individuals whose ethnicity was reported as Hispanic within the Austin data as Hispanic/Latino regardless of the race reported.