Study Evaluates Peers' Effect on Achievement
Researchers at a California think tank released an unusual study last week that tracked the progress of 141,000 students in San Diego public schools over three years and teased out some of the factors that were key to their academic success.
The effort—which researchers say has implications for urban K-12 schools nationwide—yielded some surprises. The study by the San Francisco-based Public Policy Institute of California found, for instance, that students' peers had a more consistent effect on achievement than whether their teachers had advanced degrees.
Julian R. Betts, the lead researcher on the study, said teacher qualifications tended to have less overall impact on learning because they affected elementary and secondary school students differently.
Having a teacher with an advanced degree or more classroom experience significantly boosted students' test scores in middle and high school, according to the study. But the same qualifications mattered less for younger students. In elementary school, the data suggests, students whose teachers had emergency licenses or who had taught for a year or less made slightly more academic progress than those who had credentialed teachers with 10 years' experience.
Having smaller classes, on the other hand, seemed to be more important to students' reading achievement in elementary school than it was in high school.
"The message we seem to be getting is, different types of spending matter differently at different levels of schooling," said Mr. Betts, who is an economics professor at the University of California, San Diego, and a senior fellow at the institute. His co-authors on the study are Andrew C. Zau and Lorien A. Rice.
"That means," he added, "that one-size-fits-all policies are less likely to succeed."
Eric A. Hanushek, an economist who independently reviewed the study, concurred. "I think what we're starting to do is find that schools and teaching are a lot more complicated than we had thought in the past—at least from a policy standpoint."
Both researchers said the San Diego study is important because it adds to a new body of large- scale education policy research built with data on individual students, rather than entire grades. Similar research is under way in North Carolina, Tennessee, Texas, and other states.
Researchers expect the student-level data will enable them to analyze more precisely why some students make more academic progress than others.
For their three-year study, the PPIC researchers looked at students' gains on standardized reading and mathematics exams from the 1997-98 through 1999-2000 school years in San Diego, California's second-largest school district.
Across the board, in both subjects and at all levels of schooling, they found that students made greater test-score gains during years when other students in their grades started out with higher exam scores.
According to researchers' calculations, an elementary school student switching from a low-performing grade-level peer group to a high-performing one could expect to see a 9 percent test-score improvement.
To a lesser degree, the same held true when students were in high-performing classrooms. The classroom peer effects were stronger, though, in elementary school than in secondary school—probably because students switch classes more often in higher grades, the study said.
The effects were particularly clear for students learning English as a second language, who made twice as much progress in grades and classrooms with high-scoring peers.
"We think what's going on here is that, if this year you have a slightly better cohort of students around you, other students do teach you math," said Mr. Betts. "And, if teachers at a school realize, for example, that the grade 5 cohort is unusually strong, that might actually encourage them to push their students to a higher level."
Mr. Betts said the findings could have some implications for efforts aimed at busing students to achieve racial integration or tracking them into classes based on their academic abilities.
A less obvious implication, though, could be that schools benefit more than they think from programs aimed at their poorest-performing students, according to Mr. Betts.
"If you can find targeted interventions that can help students who are lagging behind, that may spill over and benefit all of the students in a grade," he said.
Researchers divided the schools in the study into five quintiles, based on the percentage of students in those schools poor enough to qualify for federally subsidized meals. In keeping with other research, they found large achievement gaps between poor and more affluent schools that started in 2nd grade and narrowed slightly by 11th grade.
A 5th grader attending a San Diego school in the highest socioeconomic quintile, for example, reads at the same level as a 10th grader enrolled in a school in the poorest group. To narrow those wide gaps, the study concludes, policymakers should expand public preschool programs for low- income children.
Vol. 23, Issue 2, Pages 10-11Published in Print: September 10, 2003, as Study Evaluates Peers' Effect on Achievement