Repeating a successful intervention 20 times before failure means you can dramatically boost the success to failure rate.
In this context that would mean running the same intervention at scale 30 times without each run impacting the next. Maybe that is possible, but it seems unlikely to me especially if the intervention has a half life on the order of a single generation as life experience and social norms change.
The initial run is usually a relatively small scale. What works in NYC may not work in LA but it’s likely to work in many locations.
The question I was raising was locality though, it was temporal. Such experiments take time to conduct, and running a couple dozen or more times takes even more time.
The question I was raising was whether said results would be useful beyond a decade or two into the future, not across different cities today.
If my hunch were to price right, anything we'd learn there would have a short half live and may not hold out past, say, a generation or two.
I can say anecdotally that my grandparents would respond differently to many situations and interventions than I would. They grew up during the Great Depression and he flew in the war, those two events alone very much impacted how they viewed things for the rest of their life.
Looking at those only a generation behind me and even they seem to respond very differently than I'd expect those in my generation would.