Severity of punishment
https://nij.gov/five-things/pages/deterrence.aspx
https://www.nap.edu/read/18613/chapter/7
I think Riot may actually have supporting evidence as well. According to Riot, their least severe punishment is their most effective.
14 Comments
In the first link, they also contribute the results to the consistency and promptness of the punishment. When people were charged to be in prison for 30/60/90 days for violating parole, they often didn't have the room for them so it was unreliable. On lowering the sentence it was much more likely they had the room to send someone to jail, resulting in a much more consistent punishment.
The second link also goes over the certainty of punishment being the best increase to deterrence.
their least severe punishment is their most effective.
No. According to Riot their first punishment is the most effective. Riot's statistics would reveal literally no usable information on which punishment is the most effective, since anything past the first punishment is no longer a random sample, as anybody who can be reformed by a 10-game chat restriction will never actually receive a 14-day ban to test whether this would also help.
What's more: Chat restrictions are not just less severe than bans, they're an entirely different form of punishment. If you really want to look at the effect severity has, you'd have to compare timed bans of different length. Fun fact: Riot did use to give out 1-day 3-day and 7-day bans. They abandoned them when they had noticed that those had basically not effect whatsoever, compared to a 14-day ban.