Papers - Michael Kremer

Joshua D Angrist, Eric Bettinger, Erik Bloom, Elizabeth King, Michael Kremer

Cited by*: 124 Downloads*: 50

Colombia's PACES program provided over 125,000 pupils from poor neighborhoods with vouchers that covered approximately half the cost of private secondary school. Since many vouchers were allocated by lottery, we use differences in outcomes between lottery winners and losers to assess program effects. Three years into the program, lottery winners were 15 percentage points more likely to have attended private school, had completed .1 more years of schooling, and were about 10 percentage points more likely to have finished 8 th grade, primarily because they were less likely to repeat grades. The program did not significantly affect dropout rates. Lottery winners scored .2 standard deviations higher on standardized tests. There is some evidence that winners worked less than losers and were less likely to marry or cohabit as teenagers. On average, lottery winners increased their educational expenditure by about 70% of the value of the voucher. Since winners also worked less, they devoted more total resources to education. Compared to an equivalent expansion of the public education system, the voucher program increased annual government educational expenditure by about $24 per winner. But the costs to the government and to participants were probably much less than the increase in winners' earnings due to greater educational attainment.

Paul Glewwe, Michael Kremer, Sylvie Moulin, Eric Zitzewitz

Cited by*: 42 Downloads*: 16

This paper compares retrospective and prospective analyses of the effect of flip charts on test scores in rural Kenyan schools. Retrospective estimates that focus on subjects for which flip charts are used suggest that flip charts raise test scores by up to 20 percent of a standard deviation. Controlling for other educational inputs does not reduce this estimate. In contrast, prospective estimators based on a study of 178 schools, half of which were randomly selected to receive charts, provide no evidence that flip charts increase test scores. One interpretation is that the retrospective results were subject to omitted variable bias despite the inclusion of control variables. If the direction of omitted variable bias were similar in other retrospective analyses of educational inputs in developing countries, the effects of inputs may be even more modest than retrospective studies suggest. Bias appears to be reduced by a differences-in-differences estimator that examines the impact of flip charts on the relative performance of students in flip chart and other subjects across schools with and without flip charts, but it is not clear that this approach is applicable more generally.

Paul Glewwe, Nauman Ilias, Michael Kremer

Cited by*: 21 Downloads*: 25

Advocates of teacher incentive programs argue that they can strengthen weak incentives, while opponents argue they lead to teaching to the test.' We find evidence that existing teacher incentives in Kenya are indeed weak, with teachers absent 20% of the time. We then report on a randomized evaluation of a program that provided primary school teachers in rural Kenya with incentives based on students' test scores. Students in program schools had higher test scores, significantly so on at least some exams, during the time the program was in place. An examination of the channels through which this effect took place, however, provides little evidence of more teacher effort aimed at increasing long-run learning. Teacher attendance did not improve, homework assignment did not increase, and pedagogy did not change. There is, however, evidence that teachers increased effort to raise short-run test scores by conducting more test preparation sessions. While students in treatment schools scored higher than their counterparts in comparison schools during the life of the program, they did not retain these gains after the end of the program, consistent with the hypothesis that teachers focused on manipulating short-run scores. In order to discourage dropouts, students who did not test were assigned low scores. Program schools had the same dropout rate as comparison schools, but a higher percentage of students in program schools took the test.

Michael Kremer, Edward Miguel

Cited by*: 16 Downloads*: 19

We examine social learning using data from a program that promoted use of deworming medicine in Kenyan schools. These drugs kill worms in the body; although people are soon reinfected, treatment interferes with the cycle of transmission, generating positive externalities. Individuals randomly exposed to more information about deworming drugs through their social network were significantly less likely to take the drugs and more likely to believe the drugs are "not effective." This finding is consistent with the hypothesis that those exposed to the program had overly optimistic prior beliefs about net private drug benefits. The combination of strong social effects and extensive social networks among teenagers implies that a "child-to-child" public health approach focused on teenagers will speed social learning. There are large differences between social effect estimates relying on experimental variation (negative estimates) and nonexperimental methods (positive estimates).

Michael Kremer, Dan Levy

Cited by*: 16 Downloads*: 9

This paper examines a natural experiment in which students at a large state university were randomly assigned roommates through a lottery system. We find that on average, males assigned to roommates who reported drinking in the year prior to entering college had one quarter-point lower GPA than those assigned to non-drinking roommates. The 10th percentile of their college GPA is half a point lower than among males assigned non-drinking roommates. For males who themselves drank frequently prior to college, assignment to a roommate who drank frequently prior to college reduces GPA by two-thirds of a point. Since students who drink frequently are particularly influenced by frequent-drinking roommates, substance-free housing programs could potentially lower average GPA by segregating drinkers. The effect of initial assignment to a drinking roommate persists and possibly even grows over time. In contrast, students' college GPA is not influenced by roommates' high school grades, admission test scores, or family background. Females' GPAs are not affected by roommates' drinking prior to college. Overall, these findings are more consistent with models in which peers change preferences than models in which they change endowments.

Michael Kremer, Edward Miguel, Rebecca Thornton

Cited by*: 16 Downloads*: 13

We report results from a randomized evaluation of a merit scholarship program for adolescent girls in Kenya. Girls who scored well on academic exams had their school fees paid and received a cash grant for school supplies. Girls eligible for the scholarship showed significant gains in academic exam scores (average gain 0.12-0.19 standard deviations) and these gains persisted following the competition. There is also evidence of positive program externalities on learning: boys, who were ineligible for the awards, also showed sizeable average test gains, as did girls with low pretest scores, who were unlikely to win. Both student and teacher school attendance increased in the program schools. We discuss implications both for understanding the nature of educational production functions and for the policy debate surrounding merit scholarships.

Esther Duflo, Pascaline Dupas, Michael Kremer, Samuel Sinei

Cited by*: 15 Downloads*: 26

We report results from a randomized evaluation comparing three school-based HIV/AIDS interventions in Kenya: 1) training teachers in the Kenyan Government's HIV/AIDS-education curriculum; 2) encouraging students to debate the role of condoms and to write essays on how to protect themselves against HIV/AIDS; and 3) reducing the cost of education. Our primary measure of the effectiveness of these interventions is teenage childbearing, which is associated with unprotected sex. We also collected measures of knowledge, attitudes, and behavior regarding HIV/AIDS. After two years, girls in schools where teachers had been trained were more likely to be married in the event of a pregnancy. The program had little other impact on students' knowledge, attitudes, and behavior, or on the incidence of teen childbearing. The condom debates and essays increased practical knowledge and self-reported use of condoms without increasing self-reported sexual activity. Reducing the cost of education by paying for school uniforms reduced dropout rates, teen marriage, and childbearing.

Michael Kremer, Sylvie Moulin, Robert Namunyu

Cited by*: 13 Downloads*: 9

Kenya's education system blends substantial centralization with elements of local control and school choice. This paper argues that the system creates incentives for local communities to build too many small schools; to spend too much on teachers relative to non-teacher inputs; and to set school fees that exceed those preferred by the median voter and prevent many children from attending school. Moreover, the system renders the incentive effects of school choice counterproductive by undermining the tendency for pupils to switch into the schools with the best headmasters. A randomized evaluation of a program operated by a non-profit organization suggests that budget-neutral reductions in the cost of attending school and increases in non-teacher inputs, financed by increases in class size, would greatly reduce dropout rates without reducing test scores. Moreover, evidence based on transfers into and out of program schools suggests that the population would prefer such a reallocation of expenditures.

Paul Glewwe, Michael Kremer, Sylvie Moulin

Cited by*: 8 Downloads*: 51

Although there is intense debate about the effect of increased expenditure on education in developed countries, there is widespread consensus that provision of textbooks can substantially increase test scores in developing countries. This paper evaluates a program through which a Dutch non-profit organization provided textbooks to 25 rural Kenyan primary schools that were chosen randomly from a group of 100 candidate schools. After one school year, average test scores did not differ substantially between program and comparison schools. However, for those students in the top quintile of the distribution of initial academic achievement, the program raised test scores by at least 0.2 standard deviations.

Michael Kremer, Edward Miguel

Cited by*: 7 Downloads*: 13

The history of foreign development assistance is one of movement away from addressing immediate needs and toward focusing on the underlying causes of poverty. A recent manifestation is the move towards sustainability,' which stresses community mobilization, education, and cost-recovery. This stands in contrast to the traditional economic analysis of development projects, with its focus on providing public goods and correcting externalities. We examine evidence from randomized evaluations on strategies for combating intestinal worms, which affect one in four people worldwide. Providing medicine to treat worms was extremely cost effective, although medicine must be provided twice per year indefinitely to keep children worm-free. An effort to promote sustainability by educating Kenyan schoolchildren on worm prevention was ineffective, and a mobilization' intervention from psychology failed to boost deworming drug take-up. Take-up was highly sensitive to drug cost: a small increase in cost led to an 80 percent reduction in take-up (relative to free treatment). The results suggest that, in the context we examine, the pursuit of sustainability may be an illusion, and that in the short-run, at least, external subsidies will remain necessary.