There are a couple problems with those studies. The first is mental health issues are a very wide umbrella. This includes anxiety, depression, and things like

There are a couple problems with those studies. The first is mental health issues are a very wide umbrella. This includes anxiety, depression, and things like eating disorders which all parties can agree are probably not associated with increased rates of violence. But if we limit it to diagnoses like bipolar or schizophrenia I'm sure there are increased risks of violence. My mother (who had schizophrenia) for instance was fine 364 days a year. But on the 365th she'd try to claw my grandmother's eyes out because she thought she was the devil. And EMTs and other people who aren't being called out to deal with my mother on the 364 days a year she's fine. The other issue with those studies is they control for a host of other factors to get the conclusion it's not correlated with violent crime. If you just look at a bland, we separate people into two buckets, mentally ill and non mentally ill. Which bucket commits more violent acts per capita, the mentally ill do. You have to do a lot of statistical correction before you found that mental illness isn't correlated with violence.