Thanks to private-school choice—vouchers, tax-credit scholarships and education savings accounts—this year nearly half a million children in 29 U.S. states and the District of Columbia will attend schools their parents selected.

Critics of school choice often argue that low-income families lack the knowledge or ability to choose meaningfully between schools. Worrying that parents will be taken advantage of or make poor decisions, they oppose choice programs or favor onerous testing requirements to prove they are effective.

New studies on school choice in Colombia and Barbados, however, suggest families know something that tests can’t detect. These two countries, with per capita incomes a quarter and a third of America’s, respectively, can teach us a lot about how the most economically disadvantaged families choose schools.

Stanford’s Eric Bettinger and his research team found that students who won a lottery for a voucher in Colombia were 17% more likely to complete high school on time than students who lost the lottery. The study, released in July, used a method of random assignment to compare apples to apples. So it isn’t because of selection bias that lottery winners earned 8% more than lottery losers by the time they turned 33. It’s because their parents were allowed to choose schools that were better fits for their children.

The positive effects on earnings were even larger for female students (11%) and students who applied to vocational schools (17%). For a single educational intervention, these are substantial increases. The researchers conclude that vouchers “greatly increased [a low-income child’s] chance of transitioning to the middle class.”

Likewise, a rigorous 2018 study revised a few months ago found school choice boosted social mobility in Barbados. Researchers Diether Beuermann and Kirabo Jackson compared the outcomes of more than 7,000 students who had scored right above and below an arbitrary cutoff that Barbados used to determine whether they could enroll in their parents’ preferred school. The study found that attending schools chosen by parents improved student well-being significantly, based on an index of educational attainment, occupational rank, earnings and health.

The results are mixed, however, when it comes to test scores. Two earlier evaluations of the same school-choice program in Colombia, published in the American Economic Review, found it increased test scores and educational attainment substantially. By contrast, the Barbados study found no effect of school choice on test scores, despite the long-run gains in real-life outcomes. This is the latest in a series of studies finding disconnects between effects on test scores and other outcomes—income, high-school graduation, college enrollment, college completion and more—for which tests are supposed to be a proxy.

If test scores aren’t reflecting the long-run outcomes that we care about most, then our thinking needs to change. As the Barbados study concludes, “parents may be rational to prefer schools that have no short-run test-score impacts.”

Parents see more than test scores. Several surveys of parents participating in school-choice programs find that instruction in religious values, morality and character is among the top reasons they select a given school. They want schools that teach their children how to be not only good students but good people. That means inculcating skills and behaviors such as impulse control, conscientiousness and grit—what used to be called “character education.” Unfortunately, character education is generally watered down or absent in traditional U.S. public schools.

Character education may help explain why studies of school-choice programs find they reduce teenage pregnancy and crime. In Colombia, female voucher students were 18% less likely to give birth as a teenager, and males were 32% less likely to father a child by a teenage partner. In Barbados, teenage girls were 59% less likely to give birth. Likewise, a 2019 study of Milwaukee’s voucher program found it reduced paternity suits by 38% and reduced convictions in drug-related crimes by 53% and property-damage crimes by 86%. Staying out of trouble and graduating from college don’t guarantee success in life, but they greatly increase the odds.

As parents know, kids are more than test scores. The evidence suggests that even the least advantaged families tend to do a better job than standardized tests at identifying schools that produce the outcomes that matter. Parents know better than do the critics who doubt they can choose the right schools for their children.