> Isn't trying a plausible sociological intervention then measuring its effectiveness a win regardless of how effective it ended up being?

Only if the value of whatever smidgen of knowledge you gain (assuming you do actually gain any) outweighs the much-more-than-smidgen cost of the intervention to all the people who weren't helped, and in a significant number of cases were harmed, by it.

As you can see, I don't think whatever benefits are gained actually do outweigh the costs. Which means that no, continuing to try such interventions is not a win.

The lesson we should be learning from all this is to stop trying to do so many things top down. That whenever you see some kind of large scale statistic that you think is not what it should be, the last thing you should do is concoct a large scale intervention to try to change it. Or even concoct a small scale experiment to "test" the idea, with the intention of scaling it up if it seems to work. Still less should you sacrifice a poor goat who did nothing to deserve it, just to underscore the point that your intervention won't work.