I think “correlation vs causation” is hard to study in these cases.
Take Eric Harris. For what we know, he was on an antidepressant at the time of the shooting. However, Harris’s and Klebold’s criminal behavior started way earlier than his treatment. In January of 1998, the duo broke into a van and stole computers and other electronics from it. Arrested, they were sent to a diversion program and in the context of the program, referred to anger management classes. After that, Harris saw a psychiatrist and was
started on an antidepressant.
Eric kept diaries for two years before the shooting and while IMHO they do trace slow descent into madness, he already comes across as angry and manipulative in the beginning of 1998, over a year before the Columbine massacre and way, way before any treatment.
So to me, while the correlation between Harris being on an antidepressant and the shooting exists, the context is different: the medication was prescribed for preexisting criminality, manifesting itself in felony charges that were dropped post-diversion. These charges happened 1 year and 4 months before the Columbine. (And in a county different from JEFCO, with less affluent school, both boys would have ended in a juvenile). Whether the choice of the drugs was right or wrong is subject to discussions, but the behavior was there.
(Dylan Klebold’s records are sealed so anyone’s guess is as good as mine.)
The fact that other school shooters could be on antidepressants or other “psych meds” as you say, only indicates their significant pre-existing behavioral issues.
Of course, what you post only highlights the need for a multivariate analysis of teenage aggression and criminality, with all multiple contributing factors.