A German science blogger has been trying to collect some data regarding world wide school shootings (or killing sprees on schools in general, not necessarily with guns) . He makes a very interesting point in that he says, that there are (luckily) not enough school shootings to make a statistical statement of any significance.
Regarding the Answer
Let's say 90% of school shooters have been on anti-psychotics, until now there is not enough data to say, that it is not a coincidence. And another red flag would be if there had been made no distinction between causation vs. correlation.
- Maybe most school shooters are unhappy and that's why they get prescription drugs/anti-psychotics.
- Maybe prescription drugs/anti-psychotics lead to school shootings.
Such a statement without putting it into context is highly suspicious in any case.
Regarding the Data
The actual Data of mentioned science Blogger, Marcus Anhäuser, consists of, to this point, 65 entries . Most of the data comes from different Wikipedia articles some more from articles and sites about school shootings and violence at school. Check out link  for more details, try Google translate if you can't read German.
Anhäuser doesn't make any conclusions at this point, just pointing to some emerging patterns. (I summarize his statements, this is not specifically targeted against the statement that anti-psychotics are a reason for school shootings)
- Almost all killing sprees seem to happen in the USA, Germany and China.
- In China less people get killed, maybe because the actors use melee weapons.
- It seems like this is a recent phenomena (starting in the 70s), but maybe that's just because data from earlier incidents is missing
Finally he emphasizes, that this data is definitively not conclusive, since his list is incomplete.
A Google Maps visualization gives an overview over the different incidents , generated from the data in .