Conducting a Systematic Social Observation of Body-Camera Footage: Methodological and Practical Insights

Vijay Chillar, Eric L. Piza, & Victoria A. Sytsma (2021)

Journal of Qualitative Criminal Justice & Criminology

*This study was funded by the Charles Koch Foundation, Policing and Criminal Justice Reform program

Key Takeaways

• Increased use of video recording technologies enables researchers to observe human behavior in its natural setting
• Data generated by such technologies facilitate a Video Data Analysis (VDA) framework, which focuses on situational dynamics and behaviors to understand how people act and interact
• Body-worn cameras (BWCs) provide valuable insight into the situational dynamics of police-citizen interactions
• VDA of BWC footage can help overcome limitations of traditional qualitative methodologies
• There are certain challenges associated with VDA of BWC footage that researchers will need to navigate

Research Summary

Video recording technologies provide researchers with modes for observing human behavior in its natural setting. These data sources facilitate a video data analysis (VDA) methodological framework, which focuses on situational dynamics and behaviors to understand how people act and interact. Body-worn cameras (BWCs) are an emerging technology within this framework and allow researchers to overcome various challenges that are inherent in qualitative research. However, the use of BWC footage can present a host of new challenges for researchers to address. This article incorporates a practical example detailing the application of a VDA framework using BWC footage in a research project on police use of force in Newark, NJ.

VDA can overcome concerns about access and individual protections because it does not require direct interaction with subjects, making it preferable in some contexts. This is especially so when protections are taken to prevent researchers from collecting identifying information of research subjects. In the current study, Newark Police Division (NPD) policy requires that officers notify civilians that they are being recorded, alerting them their actions can be viewed in the future. However, civilians are not agreeing to have the recording used for research purposes in these moments. As such, researchers should always ensure the principle of beneficence (i.e., the study’s  benefits outweigh the potential risks) is met.

Researchers have finite time to scan an interaction, record relevant variables, and make sense of them in traditional participant observation research. In contrast, the use of BWC footage affords researchers unlimited time which provides opportunity to rigorously interpret the context and setting. The permanent record also navigates problems associated with observer and participant recall error. However, perspective bias has the potential to skew how the viewer interprets such footage. In the current study, video is recorded from only the officer’s perspective. Since NPD requires all officers to wear BWCs, we benefited from having various vantage points of a single event. However, to ensure consistent coding, SSOs should be preceded by in-depth footage reviews, the operationalization of analytical dimensions (i.e., the content of interest), and the creation of data collection protocols and code sheets.

Technical limitations can present a number of challenges to this line of research. In our study, the BWC vendor did not enable video streaming within its software. This required all videos to be downloaded from a central server to an external hard drive kept in a secure location at police headquarters. The inability to view footage remotely precluded the participation of all authors in the coding. To maintain a team approach to coding, the authors regularly met to discuss coding decisions and answer any questions the main coder raised. However, the process would have been much more efficient if the BWC system enabled video streaming. Further, the first 30 seconds of audio were muted in all videos, making communication between officers, dispatchers, suspects or involved parities (which could have contained valuable data for the study) inaudible at the very beginning of each use of force event.

Lastly, BWC footage is time consuming to view, and interactions between humans can be chaotic, fast-paced, and nuanced. Taking careful note of visual cues and context is important, as is having no illusions about the time commitment required to conduct the research. It took approximately 91 hours to download all videos for our study sample, and about 300 hours to systematically view and code the videos. Researchers interested in using BWC footage as a data source should be mindful of the upfront time and resource requirements.