If a broad, well-designed, and generous voucher program shows no advantages a decade in, then obviously that’s more telling . . . .

Not necessarily. Note that there are a few studies indicating that vouchers may put competitive pressure on public schools to improve or else. If you had a broad and generous voucher program (rather than a pittance of a program), then if there is indeed pressure from competition, that effect would presumably be greater. Thus, if it turns out that, after 10 years, a voucher program had improved performance in both public and private schools, it might not look like the voucher students had improved relative to the public school students (indeed, it’s conceivable that voucher students could get relatively worse even if everyone was improving in an absolute sense!). Voucher researchers are currently trying to figure out a way to correct for this potential complication, but it’s worth noting.

His previous post did link to the RAND book that reviewed voucher evidence, but only to quote a single paragraph on Pat Wolf’s recent study of the federally funded voucher program in DC. Anyone who reads just a few pages further in that book will be told — precisely contrary to what Ezra says — that “the most compelling evidence available on the achievement effects of vouchers” is a series of randomized experiments from New York, DC, Dayton, and Charlotte showing that black kids improved by “one-third of a standard deviation — fairly large in terms of most educational interventions, equal to about one-third of the average racial gap in achievement in the country.”