You're assuming that the civilians are acting under the assumption that the police are wearing body cameras. But why would you assume this? Most police nationwide do not wear body cameras. For most of recent history, the police in those districts did not wear body cameras. When the study was started, obviously the police themselves knew they were wearing cameras, and hence, would have acted differently because they know they were being watched. But the vast majority of civilians do not follow local-police-news closely enough to assume that the police are wearing body cameras.

> The problem with this line of reasoning is that it assumes the public even knows how to tell if police officers are wearing body cameras.

I don't see how it assumes that at all. I think the gist is that police behave in a way that produces less complaints when they are conscious that their actions are being recorded.