The study found that students in the program did generally score higher. The reporters were confused by the fact that the findings for the whole group of students were not statistically significant at the prescribed cut-off. The researchers were only 91 percent certain (statistically) that the better performance of voucher-program students was due to the program rather than chance, and they had to be 95 percent certain. They did find statistically significant positive findings for some subgroups of students.
Compounding this error, the reporters then quote an education researcher saying, “We have no evidence that vouchers work.” This too is incorrect.
There have been ten analyses of random-assignment voucher program experiments (random-assignment being the gold-standard of testing treatment effects). All ten demonstrate positive voucher effects, 9 out of 10 find statistically significant effects for at least some subgroups, and 8 out of 10 find statistically significant effects for the whole voucher group.
And the parents involved are extremely happy with it and think their kids are safer. And the vouchers cost a third or less than what is spent in public schools. Oh, and these programs are all small and some highly regulated, which limits their effectiveness.
But if we use vouchers, we can’t hire more union members!