While the study did control for factors like socioeconomic status, race, and job, I think it's inherently difficult to control for the idea that somebody is muc

While the study did control for factors like socioeconomic status, race, and job, I think it's inherently difficult to control for the idea that somebody is much more likely to carry a gun when they perceive their risk of being attacked as higher. For example, if they frequently interact with dangerous people in their personal lives. I would not be surprised to learn that, say, people who get restraining orders against their exes are more likely to be murdered by their exes than people who don't. Does that mean that getting a restraining order is dangerous? Or does it mean that people in dangerous situations tend to get restraining orders?