4 ms·
The only way to fairly evaluate a program is on the difference between it and a control intervention of some sort. Like, you can't just claim that the best stud
by ThrustVectoring 9y ago
The only way to fairly evaluate a program is on the difference between it and a control intervention of some sort. Like, you can't just claim that the best students must have had the best teachers. A fair chunk of data supports the claim that people have gotten sober while in AA, which is significantly different than the claim that AA helps people. Unless you've got comparisons with control groups, "good work being done there" is disputable.
- warcher 9y agoAA is all volunteer, run by people who are doing their level best to help with a monstrously difficult problem, with demonstrable success. I do think they got some things right. I don't think their work is finished, and they do need to examine their methods and the efficacy of the same.
- ThrustVectoring 9y agoOh, I completely agree. I don't think AA is run by bad people or anything - my disagreement is with the level of institutional support given to AA. Decisions come with opportunity costs, and the cold logic of prioritization means that sometimes the best thing to do is to give less money to "people who are doing their level best to help with a monstrously difficult problem, with demonstrable success".