The tl;dr being in the mid 1900s when somebody actually DID try to measure the results of a mentor program for at-risk youth, they found the impact was slightly negative.
Actually fixing societal problems is a really hard problem and what actually works is often very counterintuitive. I think the whole field could do to let the data guide them better (also difficult because so many outcomes are "make people feel better", and most people are pretty bad at actually knowing how they are or will be feeling).
The tl;dr being in the mid 1900s when somebody actually DID try to measure the results of a mentor program for at-risk youth, they found the impact was slightly negative.
Actually fixing societal problems is a really hard problem and what actually works is often very counterintuitive. I think the whole field could do to let the data guide them better (also difficult because so many outcomes are "make people feel better", and most people are pretty bad at actually knowing how they are or will be feeling).