Charter schools are a central component of current efforts to change the face of public education in the United States. Charter schools are publicly financed, but free of many of the regulations that govern traditional public schools, such as those involving staffing, curriculum, and budget decisions. Independent public schools in Australia are similar to charter schools in some respects such as autonomy in staffing and budget decisions.
Advocates of charter schools and independent public schools argue that their autonomy allows them to innovate, test new ideas, and bring competitive pressures to improve traditional public school systems. However, there is much research evidence indicates that they are failing in this regard and a new study published last month adds to this.
The new study examined the achievement of students who were admitted through a lottery process to 33 over-subscribed charter middle schools across 13 US states over a two-year period. It found that, on average, the charter schools lowered student achievement in reading and mathematics, although the difference was not statistically significant.
Attendance at a charter school in the year following the admissions lottery was estimated to have lowered student reading and mathematics achievement by a small amount (.058 standard deviations in reading and .076 standard deviations in mathematics) in the school year immediately following the lottery. It also lowered achievement by similar levels (.070 standard deviations in reading and .065 standard deviations in mathematics) in the second school year after the lottery.
There was also considerable variation in impacts across schools, with some charter schools in the study positively affecting student achievement.
To estimate charter school impacts, the study compared test score outcomes of students admitted to these schools through the randomized admissions lotteries with outcomes of applicants who were not admitted. The students that did not succeed in the lotteries stayed in traditional public schools. The background characteristics of students who were successful in the lotteries were broadly similar to those who were not successful. Results were compared over two years after the lotteries.
The estimated impacts generally did not vary across sub-groups defined by students’ race or gender. However, there were some differences between disadvantaged (as measured by eligibility for free or reduced price lunch) and non-disadvantaged students.
There was no statistically significant difference in the results of disadvantaged students who were successful in the lotteries and those that were not successful. One exception was a significant improvement in mathematics in the second year of attendance at a charter school for disadvantaged students. In contrast, charter schools were estimated to have significantly lowered reading and mathematics results for non-disadvantaged students in both years of attendance.
The authors qualified the findings for disadvantaged and advantaged students. They said that variation in impacts could reflect variation in the effectiveness of the charter schools as well as variation in the effectiveness of the non-charter schools available to students in that area. While the estimates could indicate that charter schools serving more disadvantaged students are more effective than those that serve more advantaged students, they could also indicate that the non-charter schools attended by disadvantaged students are weaker than those attended by non-disadvantaged students, leading to greater opportunity for positive charter school impacts among disadvantaged students. That is, significant performance differences in the traditional public schools attended by disadvantaged and non-disadvantaged lottery losers could influence the estimated impact of charter school attendance. The authors of the study stated that they were unable to rigorously disentangle these two possibilities.
Another caveat to the study is that the charter schools in the study were not randomly selected. The researchers identified eligible charter schools and invited them to participate in the survey. Many declined the invitation and this may have influenced the results of the study, particularly in relation to the findings for sub-groups of students. Among 146 schools deemed to be potentially over-subscribed after an initial survey screen, 49 did not agree to participate in the study. Ultimately, 77 schools agreed to participate in the study, but this was narrowed down to only 33 schools because they remained over-subscribed for the whole period.
The authors also acknowledged that the study included only over-subscribed charter schools that held admissions lotteries, and the impacts for these schools could be systematically more positive than impacts of charter schools that are not over-subscribed.
The new study covers only a small number of charter schools. However, its results are broadly consistent with many other studies that show no significant difference between student achievement in charter schools and traditional public schools. The research evidence to date is that more school autonomy promises much but has failed to deliver.
Trevor Cobbold
Melissa A. Clark, Philip M. Gleason, Christina Clark Tuttle & Marsha K. Silverberg. Do Charter Schools Improve Student Achievement? Educational Evaluation and Policy Analysis. Published online 9 December 2014.