[For votes to count, referees must reasonably explain why they voted as they did. Thus, please explain your vote. If you voted to publish pending minor changes, specify each change, why it is needed, and, possibly, how it should/could be done.]
The authors made a point, in the early part of the paper, to indicate how traditional qualitative research concerns, like ensuring access to study populations and perspective bias, were either replaced with new methodological challenges (i.e. the need to gain access to videos from police) or mitigated in new ways (using grounded theory methods to create code sheets). I did not find the later to be new information either though. For example, police researchers have long had to find ways to both work with, and gain the trust of, police officers and police administrators (along with potentially other city officials) to study a range of things. In the past they may have been looking for access to police use-of-force reports, some other police records, or asking to actively ride with police to observe them. While videos certainly give researchers a new tool for data collection, they do not change the process that is needed to build enough trust to gain access to, or successfully work with, police officers or police records (whether video, audio, live, or written documents). The same could be said for using a grounded theory method to creating code sheets. While this is excellent advice, it is not new. Researchers have the same concerns when observing live events and may use the same technique to create code sheets or check sheets to reduce the negative impact of perspective bias. While videos do allow for replay, that benefit will be somewhat obvious to future researchers, and not requiring of an article.
The authors rely heavily on one project that they did in Newark, NJ. While this study sounds like it was very well done methodologically, and probably very interesting, one study, in one location, may not give enough perspective for a piece that attempts to guide others. There were details throughout that I noted were not universally true. For example, depending on the agency and camera types some officers wear cameras on their chests, others on their ears, and others still on their shoulders. There are many different types of cameras out there. Some are now attempting to correct the missing audio in the initial buffering window of 30 seconds. Some cameras do not live stream, but they can be downloaded to a secure cloud storage center and then video do not have to be downloaded to an external hard drive to review. Researchers can simply be given access (either full or with restrictions) to the database. Some policing agencies are much more open to research in general, especially if they have a relationship with a local university. Size of the agency may also be a factor there. While I understand that the authors are not suggesting that all of their experiences are universally true, many of the observations in this paper are based on those experiences. If those observations are not universal, then a paper about how to address them, is also not universal.