4b. Determining your controlled terms: lessons from WITNESS and Berkeley Copwatch

Berkeley Copwatch is a police monitoring group based in Berkeley, California. For over thirty years, the organisation has monitored the police, filming policing misconduct and offering information to local citizens on their rights. Over the last three decades, the volunteers leading this work have created a rich record of violations by police in the area. 

WITNESS is a non profit organisation based in the United States that helps people use video and technology to protect and defend human rights. WITNESS collaborated with Berkeley Copwatch around the development of a People’s Database and archive.

The problem #

Maintaining records and organising them in a systematic way to support the work of Berkeley Copwatch began to pose a challenge. As the organisation matured, it became clear that it was necessary to raise the standard of information gathered and the ability to use it in support of various policy campaigns. They “needed to be able to find the footage and aggregate incidents in order to demonstrate patterns of abuse…Patterns of enforcement required that we be able to substantiate our concerns with aggregated footage/data and be able to gather all of this information together in an organised and comprehensive way, sometimes quickly” (Berkeley Copwatch).

The solution #

To address this challenge, Berkeley Copwatch partnered with the international human rights organisation WITNESS to develop the People’s Database for Police Accountability. The aim of this partnership was to streamline the workflows within Berkeley Copwatch, develop protocols for data collection and organisation, and strengthen their platforms for collecting, organising and analysing data. As part of this process, WITNESS and Berkeley Copwatch developed a set of tools to support community-based police accountability groups in collecting, organising and storing documentation of police encounters. The goal of the tools was to conduct investigations more quickly and effectively, coordinate justice campaigns, and challenge unjust laws, policies and police practices. 

One such resource is the Data Dictionary + Controlled Vocabulary, which was developed to define the fields used in Berkeley Copwatch’s People’s Database for Police Accountability. It serves as a reference for Berkeley Copwatch to ensure consistent cataloguing and accurate interpretation of the data. The controlled vocabulary was developed to ensure all terms used in the People’s Database for Police Accountability are clearly defined with meanings and uses, agreed-upon by all those in their organisation. 

How they did it #

Step 1: Clarify the list of terms needed #

WITNESS and Berkeley Copwatch agreed that the value of a controlled vocabulary is to establish a set of words/terms whose meanings and uses were agreed-upon by their group so that they could catalogue and find information efficiently and consistently. They knew that they needed the terms of this vocabulary to design and configure a database. 

Next, they revisited the questions they wanted their data to answer: 

  • Which videos show which incidents involving which officers?
  • Which officers most often use force?
  • What kind of force are officers using?
  • What are the histories of incidents for each officer?
  • Who else was present at an incident? What did they do?
  • Are officers deliberately interfering with the right to watch? Which officers?
  • Do officers conduct searches based on people’s race?

They kept these prioritised research questions in mind so that they could strategically design their controlled vocabulary to contain terms which would be able to answer these research questions.

Step 2: Plan the process #

Next, Berkeley Copwatch thought about who would be using the controlled vocabularies and in which way (e.g. as a dropdown list when filtering information about x). They considered how a member of their team entering information into the database would need to be able to select the right term to describe an incident, while another member of their team researching in the database would need to know what search terms can return the information they are looking for. Then, they considered how their controlled vocabulary would change over time, and who would maintain control of their controlled vocabulary.

Step 3: Organise and define the terms #

Keeping in mind their research questions, Berkeley Copwatch compiled a list of terms that would serve to answer them. For example, one of the research questions the Berkeley Copwatch team wished to address was “Who else was present at an incident? What did they do?” This question is related to two key entities in their data model: an incident entity, and a participant entity. In the relationship called ‘Participant at Incident’ that connects a participant entity to an incident entity, there is a value called ‘Role’. This is a dropdown list of terms that describe the potential role that a participant might play in an incident. 

To determine what these terms would be, the team drew on their past interactions with participants in police interactions and estimated what potential roles participants and bystanders might occupy. In this way, the terms below were selected based on the experiences of the Berkeley Copwatch team and volunteers. 

Definition: Participant Role in Incident

ComplainantA person alleging police misconduct or crime happened and is pursuing further action. (Can be the same as subject of the stop, although the cataloguer should select ‘Subject’ if this is the case).
CopwatcherAn observer trained by or affiliated with a copwatching organisation, and who is practising copwatching principles of non-violence and noninterference.
Legal ObserverA person specifically trained to gather facts at the scene of an incident/protest/strike, etc. For example, the National Lawyers Guild. 
MediaA person acting in their capacity as a journalist, either with press credentials or who self-identifies as a member of the press including freelance and independent journalists. Can include people who report through social media.
Relative or Friend of SubjectA person connected with the subject of the incident, such as a family member, friend, partner, etc. Useful for follow-up actions.
SubjectA person who is the focus of the incident that the record documents.  Can include victims, survivors, people stopped or detained. Unidentified subjects should be described in the narrative.
WitnessA person who observed the activity that led to the stop OR the police encounter with the subject, and may be willing to serve as a material witness to the incident in the future.
Witness: DocumenterA person unaffiliated with a copwatching organisation who documented, filmed, or recorded the incident. Useful for tracking down additional documentation of the incident. 
(Source: Data Dictionary + Controlled Vocabulary, page 75-76)

This list fulfils the characteristics of a well constructed controlled list as described in our resource on Determining your controlled list of terms. Firstly, the list is exhaustive. It includes all potential roles a person might occupy during an incident based on the experience of Berkeley Copwatch volunteers. Additionally, the definition of each potential role is carefully distinguished from the other terms. Where there might be overlap, the vocabulary offers instructions on the appropriate term to use (e.g. “Can be the same as the subject of the stop, although the cataloguer should select ‘Subject’ if this is the case”). 

Each definition is well-exemplified with a description of how a term would be applicable in specific situations or incidents. Additionally, terms are equal in granularity and specificity focusing on roles at the individual level and they are listed in alphabetical order.

The Berkeley Copwatch and WITNESS team understood the importance of defining their terms with contextual relevance. For instance, the table below shows the controlled vocabulary definitions for specific Berkeley-area neighbourhoods. Recording these definitions in such clearly defined terms was necessary because, although local residents may understand intuitively what the terms ‘People’s Park’ or ‘The Hills’ refer to, an outside user will not. 

Definition: Neighbourhoods

People’s ParkIncidents that occur within or adjacent to People’s Park, or involve people who are members of the People’s Park community.
SouthsideIncidents that occur south of campus, but not on Telegraph Avenue. Southside is east of Shattuck Ave, north of Woolsey, south of Bancroft.
NorthsideIncidents that occur north of Hearst Ave, east of Sacramento, but not the Hills. This includes the North Shattuck area and the North Berkeley BART Station. 
West BerkeleyIncidents that occur west of MLK Jr. Way south of Hearst Ave. 
North Berkeley (NW)Incidents that occur north of Hearst Ave. and west of Sacramento.
The HillsIncidents that occur in residential areas north or east of Northside.
MarinaIncidents that occur on Frontage Road or west.
The ElmwoodIncidents that occur east of College Ave, north of Alcatraz, and south of Clark Kerr campus. 
Downtown ShattuckIncidents that occur north of Dwight, east of MLK Jr. Way, and south of Hearst Ave.
Telegraph AveIncidents that occur on Telegraph Avenue. 
Lorin DistrictIncidents that occur north of Alcatraz, east of Sacramento, west of Adeline, and south of Russell. This includes the Ashby BART station.
CampusIncidents that happen on or adjacent to UC Berkeley main campus & Clark Kerr campus.
(Source: Data Dictionary + Controlled Vocabulary, page 57)

To determine the list of terms to be included in their controlled vocabulary, the Berkeley Copwatch and WITNESS project team drew on the 30+ years of experience of the group. They considered those terms that they already used regularly in their work for which they had no written definitions and all those terms which could pose confusion to those outside the organisation. 

Once they had gathered their terms, the Berkeley Copwatch team refined, completed and documented them. Some of the definitions are drawn from existing legal standards for policing in Berkeley, CA. For example, terms describing the ‘Use of Force’ by police (e.g. armoured vehicle, baton, canine, flash bang, handgun) use definitions employed by the Berkeley Police Department. Other term definitions were developed based on the experiences of the Berkeley Copwatch volunteers. For example, terms describing the ‘Result of Stop’ (e.g. arrest, citation, injury/death, property confiscation) use definitions developed based on the experiences and observations of the police watch group.

Step 4: Test the terms #

For Berkeley Copwatch, testing their terms was an ongoing process, as opposed to a standalone step, and part of the reason they wanted to include clear language that the terms could change over time. 

Step 5: Document the controlled vocabulary #

The Data Dictionary + Controlled Vocabulary was created by WITNESS and Berkeley Copwatch to accompany the People’s Database. It serves as a reference for those utilising the database to understand what terms are held within and how these terms are defined. This document is an exemplary model for other groups working on human rights documentation projects. By integrating the controlled vocabulary into the data dictionary, WITNESS and Berkeley Copwatch made sure that all explanations of their data model and terms could be found in a central, accessible place. They distinguish the terms of the controlled vocabulary by highlighting them in bright yellow. In alphabetical order, the vocabulary presents first the term that will be uniformly used in the database in one column, then the agreed-upon definition in the second column. 

The Data Dictionary + Controlled Vocabulary was last modified in May 2020, but WITNESS and Berkeley Copwatch emphasise that it is ever a work in progress and will adapt over time. This adaptation results from continued brainstorming, testing, tinkering, re-trying and re-evaluating conducted by Berkeley Copwatch volunteers as they continue to collect and organise new information.

References #

Berkeley Copwatch. People’s Database. (N.d.). Retrieved February 17, 2022, from https://www.berkeleycopwatch.org/people-s-database

The People’s Database for Community-based Police Accountability: A Berkeley Copwatch + WITNESS initiative. WITNESS Media Lab. (October 15, 2020). Last accessed January 27, 2022 from https://lab.witness.org/berkeley-copwatch-database/

Data Dictionary + Controlled Vocabulary. WITNESS and Berkeley Copwatch to accompany the People’s Database. (May 12, 2020). Last accessed January 27, 2022 from https://library.witness.org/product/data-dictionary-and-controlled-vocabulary/ 

Powered by BetterDocs