Usable Knowledge is a trusted source of insight into what works in education — translating new research into easy-to-use stories and strategies for teachers, parents, K-12 leaders, higher ed professionals, and policymakers. Usable Knowledge is produced at the Harvard Graduate School of Education by Bari Walsh (senior editor) and Leah Shafer (staff writer). Contact us at email@example.com.
Making Incentives Work
Scholarships have greater impact when given to kids who are told they are high performers, new study finds
If you run a small school in a poor city, and you’re trying to boost outcomes for your students by awarding scholarships to encourage attendance and achievement, do you give the scholarships to the best students, or to the neediest?
As it turns out, both pathways have beneficial effects on enrollment and attendance, finds Associate Professor Felipe Barrera-Osorio, a global economist who evaluates educational policies throughout the developing world, particularly those that incentivize schooling through cash transfers. But only one type of scholarship — the merit-based one — led to improvements in actual performance, regardless of how gifted the students in the need-based group were.
The findings, pulled from a study that Barrera-Osorio conducted in Cambodia, suggest that the way incentives are described to their recipients has a greater effect on outcomes than previously understood. The results are detailed in a new paper [PDF] coauthored with Deon Filmer of the World Bank, to be published in a forthcoming edition of the Journal of Human Resources.
“There is a long and important debate about scholarships — whether you allocate them based on merit or based on poverty,” Barrera-Osorio says. “There’s a tension between equity and efficiency there. If you allocate based on merit, you will probably gain more efficiency, because the students who receive the scholarship will take more advantage of it. If you allocate based on poverty, there are more benefits in terms of equity, since you’re redistributing resources.”
Barrera-Osorio and his team created three groups of schools in Cambodia and randomly assigned one group as control, one as providing merit-based aid, one as providing need-based aid. They collected baseline information about school demographics, poverty, and performance.
He told the third- and fourth-grade students in the merit-based sample that he’d be giving a scholarship to the best students in the class. (Half of each selected class received it; half did not.) He did the same thing for the similarly aged poverty-based sample, telling students they’d be receiving a scholarship to help alleviate financial burdens.
He then compared the kids who received both types of scholarships with kids in the control group who would have been eligible for one or the other type.
He found that both scholarships “induced higher participation in education,” increasing the probability that students would reach sixth grade. But only the students in the merit group increased their test scores.
The most interesting finding, said Barrera-Osorio, came when he dug deeper into the performance of each sample group. In the group targeted for merit-based scholarships, the students who received those scholarships were both poor and not poor. And in the group targeted for the poverty-based scholarship, some students were high-achieving, some not.
“That meant I could look at what happened to a kid who was told he was getting the scholarship because he was academically strong, but who was also poor, versus another kid who was told he was getting the scholarship because he was poor, but who was also good in school.
“What we found is that it didn’t matter what your poverty level was in the merit group. Both of types of kids, poor and not poor, saw improved results in test scores.
“And in the poverty treatment group, it didn’t matter if you were a high or low achiever at baseline. You didn’t produce any results. So the student who is hearing ‘you’re smart’ does something. The student who is hearing ‘you’re poor’ doesn’t do anything.”
According to Barrera-Osorio, these findings — the idea that the impact of the program depended upon how the intervention was described to participants — suggest that in order to balance equity and efficiency, a two-step targeting approach might be optimal: first, target low-income individuals, and then, among them, target based on merit.
Get Usable Knowledge — Delivered
Our free monthly newsletter sends you tips, tools, and ideas from research and practice leaders at the Harvard Graduate School of Education. Sign up now.