Differences in Math Eyed More Closely
Two years ago, a team of researchers from the University of Illinois at Urbana-Champaign turned conventional wisdom on its head with a study suggesting that, once students' socioeconomic differences are taken into account, public school students perform as well as, or even better than, private school pupils on national math tests.
Now, in a follow-up study published November's issue of the American Journal of Education, the researchers offer two potential explanations for their findings: teacher credentials and reform-oriented math instruction.
Using data from the 2003 National Assessment of Educational Progress tests in 4th and 8th grade mathematics, the researchers found that public school students were more likely to be taught by certified teachers, and more likely to be in classes focusing on reform-oriented approaches, and that those differences could be statistically linked to better math achievement.
Those two variables were among more than a dozen possibilities that the researchers tested. Only the math-instruction and teacher-credential variables showed statistically significant relationships.
Sarah Theule Lubienski, an associate professor of curriculum and instruction at the university, was the lead researcher on the study. Her co-authors are her husband, Christopher Lubienski, an associate professor of organization and leadership, and doctoral student Corinna Crawford Crane.
Ms. Lubienski said the correlation between math instruction and achievement was "not surprising" because the NAEP tests are aligned with the National Council of Teachers of Mathematics' standards for teaching the subject.
One early critic of the first study was Paul E. Peterson, a professor of government at Harvard University, who argued that the Lubienskis' measures of student poverty in private schools were flawed. He reanalyzed the same data using different measures and found that private school students outperformed their public school counterparts.
The Lubienskis didn't change their measure for the new analysis. But two studies conducted since using different methods and databases back up the researchers' findings.
Vol. 28, Issue 16, Page 6