That's interesting -- but it certainly conflicts with macro-view analysis of the data[0][1][2] (I could link more, but there are more research links in the first linked thread and I do not wish to waste your most valuable time).
It seems to me that the only way to explain this is the lack of publishing of damaging videos by the youtube account -- i.e. the usual thing you need to pay attention to in the social sciences -- a) who chooses to release the data, b) the conditions under which they choose to release the data.
On the side of the YouTube channel -- based on the patreon channel, it seems to me that they are most likely not part of the police, or the judicial system. Thus they must obtain body camera footage from freedom of information requests.
This means that that information request is subject to filtering on the side of law enforcement, who can almost arbitrarily choose what videos they want to release -- while there are guidelines present, they really only apply if you can prove them being broken, which... due to the nature of FoI requests... you can't...
So, just as a baseline, legal advice holds that data that is currently being pursued legally should be held off from release to a general audience, and court cases can take many years to be processed. So it seems to me that that is one main reason that videos that show misconduct by police offers would be failed to be released, or would otherwise be redacted.
On top of that, you have whether officers themselves working in the police administration care to release the footage. It seems reasonable that an officer may be subject to workplace-based social pressure, and not wish to release footage of wrongdoing by one of his coworkers, it also seems reasonable that in some cases, they might indeed feel departmental pressure to not release footage that displays such wrongdoing, so that the department as a whole does not come under flak. You have cases in the UK where officers themselves deleted videos that would prove wrongdoing on their part[3]. Either way, this is inherently impossible to prove either way.
And then you have whether or not the police officer was recording at all, regardless of what regulations state. There have been a few cases recently where police officers thought that their body camera was off, and used that time to break the law[4][5]. Indeed, in some states, it's entirely up to the officers whether to turn them on in the first place, based on what they consider as "an incident".
And then finally, as a youtube channel accepting donations, they are heavily incentivized to draw "engagement" and game the algorithm, so what they release is not just going to be based both on the political opinions of those within the organization, but also will heavily cater to whatever established audience they have, to ensure that each video is liked and that they gain subscribers, so they can drive donations to keep on doing what they are doing.
So to me it seems that this isn't as nearly cut-and-dry as you assume to think that it is. At the very least, a random youtube channel that releases police video, cannot be thought of as a proper or correct sample from which to draw correctly proportioned information from -- as we can see, there are many reasons why it would misrepresent the number of cases of each involved. While research in this area is perhaps uncomfortable for people to accept, it broadly shows that police -- at least how they behave at the moment -- are universally flawed. I myself would prefer to trust the data.
It seems to me that the only way to explain this is the lack of publishing of damaging videos by the youtube account -- i.e. the usual thing you need to pay attention to in the social sciences -- a) who chooses to release the data, b) the conditions under which they choose to release the data.
On the side of the YouTube channel -- based on the patreon channel, it seems to me that they are most likely not part of the police, or the judicial system. Thus they must obtain body camera footage from freedom of information requests.
This means that that information request is subject to filtering on the side of law enforcement, who can almost arbitrarily choose what videos they want to release -- while there are guidelines present, they really only apply if you can prove them being broken, which... due to the nature of FoI requests... you can't...
So, just as a baseline, legal advice holds that data that is currently being pursued legally should be held off from release to a general audience, and court cases can take many years to be processed. So it seems to me that that is one main reason that videos that show misconduct by police offers would be failed to be released, or would otherwise be redacted.
On top of that, you have whether officers themselves working in the police administration care to release the footage. It seems reasonable that an officer may be subject to workplace-based social pressure, and not wish to release footage of wrongdoing by one of his coworkers, it also seems reasonable that in some cases, they might indeed feel departmental pressure to not release footage that displays such wrongdoing, so that the department as a whole does not come under flak. You have cases in the UK where officers themselves deleted videos that would prove wrongdoing on their part[3]. Either way, this is inherently impossible to prove either way.
And then you have whether or not the police officer was recording at all, regardless of what regulations state. There have been a few cases recently where police officers thought that their body camera was off, and used that time to break the law[4][5]. Indeed, in some states, it's entirely up to the officers whether to turn them on in the first place, based on what they consider as "an incident".
And then finally, as a youtube channel accepting donations, they are heavily incentivized to draw "engagement" and game the algorithm, so what they release is not just going to be based both on the political opinions of those within the organization, but also will heavily cater to whatever established audience they have, to ensure that each video is liked and that they gain subscribers, so they can drive donations to keep on doing what they are doing.
So to me it seems that this isn't as nearly cut-and-dry as you assume to think that it is. At the very least, a random youtube channel that releases police video, cannot be thought of as a proper or correct sample from which to draw correctly proportioned information from -- as we can see, there are many reasons why it would misrepresent the number of cases of each involved. While research in this area is perhaps uncomfortable for people to accept, it broadly shows that police -- at least how they behave at the moment -- are universally flawed. I myself would prefer to trust the data.
[0]: https://twitter.com/equalityAlec/status/1571898316295643136
[1]: https://resistancelab.network/our-work/taser-report/index.ht... (Disclaimer: I know some of the people who worked on this. Interesting Note: 118 cases of taser use against children, 8 against children aged under 11)
[2]: https://www.theguardian.com/us-news/2022/jul/28/hunted-one-i...
[3]: https://www.policeconduct.gov.uk/sites/default/files/Documen...
[4]: https://www.thedailybeast.com/baltimore-cops-turned-off-body... (I admit this is rather a famous case!)
[5]: https://www.miamiherald.com/news/local/community/broward/art...