Skip to main content

This website is no longer being updated. Visit Dartmouth Now for all news published after June 7, 2010.


Dartmouth News
>  News Releases >   2007 >   September

Coping with overflowing classrooms in Kenya

Dartmouth College Office of Public Affairs
Posted 09/05/07 • Rebecca Bailey • (603) 646-3661

Photo of Pascaline Dupas
Pascaline Dupas (photo by Joseph Mehling '69)

As American children begin another year of school, debate is renewed over optimum class size. Is the national average of  21 to 24 students per class too high? Should we be aiming for, as organizations like the National Education Association advocate, numbers in the high teens?

This debate might draw a rueful smile in Kenya, where classes in the public schools routinely near or exceed 100. However, just reducing class size isn't enough to improve student performance on standardized tests, suggests a working paper coauthored by Pascaline Dupas, Dartmouth assistant professor of economics, titled "Peer Effects, Pupil-Teacher Ratios, and Teacher Incentives: Evidence from a Randomized Evaluation in Kenya." Rather, it's a question of the type of teacher, the way students are grouped, and the greater involvement of local school committees.

On one hand, the high class sizes are good news: The nation's 2003 ban of fees for public school brought 1.7 million more children into public school in the 2002-05 period, an increase of 30 percent, the paper states. However, the federal government hasn't increased accordingly the number of teachers hired through its Teachers Service Commission (TSC) to be dispatched to the nation's schools, resulting in swollen classes, especially in the first few grades.

Dupas recalled a typical school she and her colleagues visited. "We went into one first grade class where there were no desks whatsoever," she said. "Kids were on the floor packed next each other. The kids were very cute, holding their notebooks and looking very happy to be there, and the teacher was very nice, but the extent to which the teacher could teach the children to read and do math was very limited. It was day care in the most rudimentary way." In other schools, there aren't enough classrooms, and first grade might be held under a tree, the blackboard leaned up against the tree trunk, she said.

Photo of Kenyan student
Matumbufu Baptist Primary School, Bukembe, Bungoma District, Kenya (photo by Aude Guerucci)

The study by Dupas and colleagues from MIT and Harvard University ran from 2005 to 2007 and looked at 210 schools in a region of Kenya, randomly selecting 70 to serve as a control and to receive no additional funds for hiring teachers. The remaining 140 received World Bank funding from the nongovernmental organization International Child Support Africa to hire "contract" teachers—fully trained professionals who aren't a part of the TSC and are paid a quarter of what civil service teachers are paid. The additional teachers were assigned to first grade in the study's first year and to second grade in the following year, allowing the classes in those grades to go from around 80 pupils to around 40. In half of those schools, the first graders were "tracked," or divided into two groups according to their skills. In the other 70, the students would be randomly assigned to the two teachers.

In addition, half of the nontracked and tracked schools were given funds to institute "School-Based Management" (SBM), in which parents on the local school committee monitored, evaluated and decided on the rehiring of the contract teacher. (The committees couldn't extend that scrutiny to TSC teachers because their hiring, assigning and evaluation is all done by the TSC itself.)

At the end of the second year, randomly chosen students from the second grades in all 210 schools were given a standardized test in math and language skills.

According to the test results, class-size reduction alone didn't improve student performance. Students with TSC teachers in the big classes in the comparison schools performed about the same as students with TSC teachers in the nontracked schools with reduced class-size. The improvements came when the students were taught by contract teachers—perhaps for the simple reason that the contract teachers spent more time in the classroom, teaching, as opposed to, say, resting in the school's tea room, according to data gathered on unannounced spot checks the study conducted.

Photo of Kenyan classroom
Matumbufu Baptist Primary School, Bukembe, Bungoma District, Kenya (photo by Aude Guerucci)

Still greater improvements occurred when students were tracked-in contrast to long-standing educational findings from elsewhere in the world that suggest that all students do better when placed with high-performing peers. Even students with TSC teachers performed better, and those teachers were more likely to be in class, teaching, during the spot checks.

Further improvements occurred at the SBM schools, the paper says. Interestingly, the TSC teachers in those schools were more likely to be in class, teaching, rather than leaving their class to be combined with that of the contract teachers. "The most likely explanation ... is that the SBM initiative emphasized the responsibility of the contract teacher with respect to the specific class to which she was assigned, and thus made it more difficult for the ... civil service teachers in those schools to use the extra contract teacher to relieve themselves of their own duties," the paper reasons. "The contract teachers in these schools had a greater incentive to please the school committee and less of an incentive to please the other teachers and the headmaster."

Dupas's coauthors, who also are her colleague researchers at MIT's Abdul Latif Jameel Poverty Action Lab, are Esther Duflo, MIT's Abdul Latif Jameel Professor of Poverty Alleviation and Development Economics and director of the lab; and Michael Kremer, Harvard's Gates Professor of Developing Societies. The standardized tests were designed by Matthew Jukes, assistant professor of education at Harvard.

Dartmouth has television (satellite uplink) and radio (ISDN) studios available for domestic and international live and taped interviews. For more information, call 603-646-3661 or see our Radio, Television capability webpage.

Recent Headlines from Dartmouth News: