Monday, July 30

Study: NYC charter schools outperform public schools

A new study on New York charter schools provides compelling evidence that charter school students see greater test score gains than they would have in regular public schools. The study, written by Harvard Economics Professor Caroline Hoxby and the National Bureau of Economic Research's Sonali Murarka, can be downloaded here. (Be forewarned: the study is 81 pages long, but the Executive Summary outlines the main findings in two pages.)

Hoxby and Murarka find that charter schools raise student performance by about .09 standard deviations in math and about .4 standard deviations in reading. We can put those numbers in terms of our standardized tests: On New York's standardized tests, students receive scores from 1 (not meeting standards) to 4 (exceeding standards); a score of 3 is considered proficient. Charter schools raise math scores about 12% of the distance between a score of 1 and a score of 2, or between 2 and 3, etc. For reading scores the gain is about3.5% of a performance level. Keep in mind that these gains are in addition to whatever gains the students would have made in ordinary public schools.

The findings in Hoxby and Murarka's study are particularly convincing because they use truly random sampling to compare charter school students and regular students. The notorious challenge in evaluating any educational policy is that it is often extremely difficult to separate the effects of the policy from other correlations. For example, if we tried to compare the effects of private school and public school, we certainly couldn't just compare standardized test scores of public school students with those of private school students, since the students differ in many ways besides the school they attended. If private school scores were 20% higher than public school scores, for instance, some of that difference might be a result of the differences in education quality, but some of it may come from differences in parental education or socioeconomic status. Even small class size, which many parents and educators have long believed to be good for student achievement, has been very difficult to measure how beneficial small classes are, since students often differ in other ways as well. (If parents who tend to be more concerned for their children's education tend to demand small classes more often than other students, and if those concerned parents also read to their kids more, for instance, some of the test score differences between small-class and large-class students may come from differences in how much they're read to at home.)

So how can we accurately measure the effects of some educational policy? If students are randomly assigned to either "treatment groups" (those with the effect) and "control groups" (those without), we can compare those two groups without worrying about other factors that might bias the results. Fortunately, New York City charter schools provide an ideal opportunity for such a comparison. In New York, when more students apply for a charter school than can be admitted, they are selected by lottery-- a random sample! So Hoxby and Murarka compared the scores of students who were selected for admission with those who entered the lottery but were not selected. This type of comparison is typically considered so convincing that researchers will weigh the results of one good random-comparison study more heavily than any number of other, non-random results.

Finally, the study notes that charter school advocates would say the true benefits of charter schools are even greater than the findings in this study suggest. Since one of the supposed benefits of charter schools is that they bring competition for regular schools, advocates would say that charter schools actually improve the quality of education in regular public schools, as those schools are forced to improve to keep students from leaving for charter schools. If that's true, even the non-charter school students, i.e. the "control group," would have benefited from the effects of charter schools.

1 comment:

Seth Pearce said...

I have always liked the idea of charter schools. This study is more evidence that they are effective.
The perspective of how charter schools foster competition in public schools is interesting and different.