Embed Notice

@lauren What if you plot the prison time against breaches and you determine there is 0 correlation- that they are completely independent variables.

What I'm getting at is "the problem of punishment"- that you see some outcome, 1, which you want to be 0 (no breaches). To attempt to coax that variable back to 0, you introduce some independent variable, punishment. It has no effect on the outcome, but now you have breaches *and* people in jail.