http://www.thelancet.com/journals/lancet/article/PIIS0140-6736%2811%2961851-5/fulltexthttp://promoteacceptance.samhsa.gov/pub ... ?printid=1
For that last link, I find these particularly compelling
“Research has shown that the vast majority of people who are violent do not suffer from mental illnesses (American Psychiatric Association, 1994).”
Sure, but if the vast majority of people, period, do not suffer from mental illnesses, mentally ill people could still be more violent on average than non-mentally-ill people.
“. . . [T]he absolute risk of violence among the mentally ill as a group is still very small
What's the absolute risk of violence among the non-mentally-ill?
In a 1998 study that compared people discharged from acute psychiatric inpatient facilities and others in the same neighborhoods, researchers found that “there was no significant difference between the prevalence of violence by patients without symptoms of substance abuse and the prevalence of violence by others living in the same neighborhoods who were also without symptoms of substance abuse (Steadman, Mulvey, Monahan, Robbins, Applebaum, Grisso, Roth, and Silver, 1998).”
Which essentially says that violence occurs at the same rate in either population, which is not at all what "more likely to be victims of violence than perpetrators" seems to be trying to imply.
And, of course, you're excluding substance abuse, which may have a different effect on the mentally ill.
All that being said, treating "mentally ill people" as some sort of homogenous group seems silly. Which mental illness they have would presumably play a part.