Skip to main content Skip to navigation

FAQs

Is their a fee to participate in this study?

No.

Do we need to blur out the officers’ faces in the videos? Check our own state law?

No. We annotate in the cloud and retain CJIS compliance for interacting with the secure footage. The CSI Lab has an evidence.com portal. Data is shared via the “agency partner” feature within the cloud management evidence system. If your agency does not use Axon, we would discuss an alternative to retain full security and confidentiality around the footage. The lab is a secure research site and all research assistants receive a background check and complete confidentiality agreements.

What is the privacy language? Do you share the videos with the public or share you are working with us?

All data is confidential, and what we annotate from the footage lacks personally identifiable information. For officer identification, we create a matching ID. If anyone were to obtain our dataset, it predominately contains 0/1s and time points of when events occur. As we are not collecting nor disclosing personally identifiable information in our data collection, in WA, the agency is not required to disclose the lab is analyzing the footage. The same is true for our partners in Idaho.

Is their strong indemnification language? Footage for research only, thought of no good deed goes unpunished.

The MOU/DUA does not allow us to identify the agency where the data originates. As we are working with several agencies, this provides an additional layer of protection.

What specific topics are you evaluating?

This depends on the agency research question. All projects in the lab are guided by what the agency wants to learn from their data. Current ongoing projects focus on use of force, de-escalation and non-escalation, training adherence, and stress and emotion.

How do you know this mapping/evaluation works?

As we deconstruct the data into an objective timeline of what transpired, and these data undergo several levels of verification, we are confident our process works. We can send more of our published works or refer you to an agency that has agreed to answer questions about their collaboration with the CSI Lab.

How will the reasonable beliefs of the officer and community member (individually) be factored into the mapping?

We do not focus on what we believe occurs. Rather, the coding is on objective events. They either occurred or did not within the video. If we are unable to determine if an event occurred because of camera angle or bad sound, it is not coded. We annotate for every person in the incident, of which we have a video record. This allows us to document what transpired, accounting for all parties.

How will the officer’s observations (those caught on and off camera) be factored into the mapping?

We indicate when a camera is turned off. We accept there are events that are not captured, and therefore accept that what we report is our objective assessment of what transpired during the interaction for which we have a video record.

How do you know what observations were not caught on camera?

We use the calls for service log to cross-reference the video. If an agency wants to check for compliance activations, we can automate that mapping process. In many ways, it depends on what the agency wants to learn from the footage. As this is a collaboration, we work with each agency to determine the process that will work best.

How will the observations of the community member be factored into the mapping?

We do not currently accept video from community members, though we are developing a solution allowing us to annotate footage from a community members perspective. If the question is asking about actions taken by the community member, we annotate events for all participants. If there are two officers, one suspect, and three bystanders, we are annotating what each person is doing in the video.       

How will environmental factors be factored into the mapping?

We annotate for key environmental factors that are contextually relevant. In total, we annotate for over 150 events, which capture if the event occurred, where in the interaction did the event occur, and what if any changes are observed.

How does the nature of the call (dispatched information) get factored into the mapping?

If the agency agrees to provide this information, we map it to the incident. The most common request from agencies is to identify differences between officer-initiated and those initiated from dispatch.

What is the length of the study/research?

The research is ongoing, and we do not intend to stop processing footage. Currently, we have 15,000 hours of processed footage. Each department is enrolled until they decide to end the partnership. During this time, we work with the agency to use their BWC footage to answer key research questions.

How much BWC footage would you need from XXX?

This depends on what XXX wants to learn from the footage. Before we begin development on the project, we have an onboarding call to discuss timelines and the research process. Given the changes we are making in response to COVID-19, we are upfront that some projects may take a year to complete because we are currently behind on processing data for 9 months. For example, we have close to 800 use of force videos to process, which is for a priority project.