Were the social interventions sampled randomly, or were they chosen for their counterintuitive outcomes?

They were chosen, but I'm not sure how one could even sample randomly - where would one look, how would one compile a list, based on what criteria? Most programs haven't been studied at all in any kind of rigorous fashion.

The best I can say is that Rossi, well-respected in the area of social interventions, takes an extremely pessimistic overall view on the effectiveness of interventions.