Certainly the literature on poverty programming describes extraordinary experimental design programs in the United States. Obviously their early head start programs, for instance, led the way in terms of identifying types of supports for poor kids--integration for school dropouts, zones where they're really tackling school dropouts in certain American inner cities. But what happens is that you'll have these fabulous experimental designs, one-off programs, but they've never been able to scale them up or ramp them up, partly because of the scale involved or the lack of other public infrastructure that would support these types of initiatives.
On April 15th, 2008. See this statement in context.