School choice and achievement: the Ohio charter school experience.

• Author: Gray, Nathan L.
• Position: Report

K-12 education policy has recently received much scrutiny from policyanakers, taxpayers, parents, and students. Reformers have often cited increases in spending with little noticeable gain in test scores, coupled with the fact that American students lag behind their foreign peers on standardized tests, as the policy problem. School choice, specifically charter school policy, has emerged as a potential remedy. School choice is hypothesized to have both participant and systemic (sometimes called competitive) effects. This article concentrates on the latter by using a novel design not used before in studies of this subject. School level data from Ohio are analyzed to estimate if traditional public schools potentially threatened by charter schools respond with positive test score gains. Specifically, an exogenous change to the education system in 2003 provides a natural experiment to examine potential systemic effects. Results indicate that the threat of charter schools seems to have had a small positive effect on traditional public school achievement.

Evaluating the Systemic Effects of Charter Schools

Great concern exists that poorly educated citizens will be unable to compete in a more technical globalized economy. Walberg (2007: 5) points to the fact that education and wages are highly and positively correlated, and thus citizens "without advanced mathematics ... in high school are unlikely to succeed in the hard sciences and engineering." Similar conclusions follow logically for other disciplines. Thus, U.S. students appear to be at risk of losing in competitive labor markets (Hanushek 1998). The concerns expressed by these two researchers exemplify the policy problem: lower educational quality causes the United States to fall behind other countries in math and science, leading to lower wages for American workers and a decline in economic productivity.

The lack of any substantial connection between resources and school quality (Hanushek 1998) and the potential consequences for the nation if educational failure continues have caused citizens, policymakers, and researchers to examine a large number of potential educational reforms that go beyond altering resource levels. The nation has debated and implemented a number of such reforms, ranging from longer school days, year-round schools, and performance pay for teachers, to school choice policies. Unlike reforms that focus directly on mandating changes to the way schools are operated, such as seat time requirements or curriculum changes, school choice is a form of systemic reform. Such policies focus on altering the structure of public schooling itself, thereby altering incentives in ways that are believed to lead individuals to alter behaviors.

School choice policies can take on a myriad of forms including inter- and intra-district traditional public school choice, magnet schools, vouchers, and charter schools. To date, vouchers and charter schools, mainly by virtue of their more recent incorporation into the reform debate, have been less studied than other choice policies. These two policy alternatives have quite different characteristics, and generalizing effects from one to the other is difficult. The focus of this article is on charter schools--public schools of choice with fewer state restrictions but additional accountability for results (Finn, Manno, and Vanourek 2001).

Introducing school choice into the educational system through charter schools is hypothesized to yield two distinct types of effects on academic achievement: participant effects and systemic effects. Participant effects involve those students actually attending charter schools. Studies addressing participant effects attempt to estimate whether students who attend a charter school benefit from having done so. Systemic effects, by contrast, concentrate on the impact charter schools have on traditional public school performance. In particular, it is hypothesized that by introducing charter schools, traditional districts will be forced or incentivized to compete for students empowered to choose among several education suppliers, creating positive incentives for improvement in a market-like environment. Thus, the key question is: Do schools in districts facing the threat of charter school location near them respond positively with regard to academic performance?

There is a growing literature on the systemic effects of school choice, but many researchers fail to fully investigate those effects because the variables used to measure the systemic effects are often endogenous and poor indicators of competition (1) (Merrifield 2001). This article avoids those two limitations by using an exogenous threat variable to measure the systemic effect. Specifically, the state of Ohio provides for a unique opportunity to study the systemic threat of charter schools. The Ohio charter school law changed in 2003 providing a natural experiment. Prior to 2003, charter schools could locate only in certain school districts. From 2003 onward, the law expanded the geographical areas in which charter schools could locate and expanded the number of potential charter school authorizers and operators. Schools in poor performing districts suddenly became much more susceptible to having charter schools locate near them. That policy change created an exogenous shock that allows one to rigorously test whether academic achievement in a school facing the threat of charter school infiltration improves at a rate different than in schools not so threatened.

Evaluating the systemic effects of charter schools is important. Society puts enormous emphasis on education; it is the lifeblood of a free society and a thriving economy. Policies concerning education, therefore, deserve a thoughtful, appropriate, and thorough review as to their effects. The nation spends billions of dollars to educate K-12 students. Taxpayers, parents, policymakers, and students need to know the impact of charter school policy on traditional public schools.

Over the last two decades researchers and advocates have debated the benefits and deficiencies of school choice. Proponents of choice policies contend that allowing families to match students with schools will enhance learning for 'all children. Opponents fear those policies may foster greater class segregation, drain resources from traditional public schools, and benefit only the most advantaged students. Furthermore, the answer to the question of the systemic effects of charter schools is vitally important to the larger school choice debate. If schools do not at least feel a threat from competition, then much of the theory surrounding charter schools and other accountability measures are seriously flawed and policies should change.

Literature Review

Two important facets of K-12 systemic effect studies emerge from a review of the existing literature and both deal with the quality of the independent variable of interest-competition. First, many measures of this variable are endogenous. Second many of those used are not reliable measures of competition. (2) Generally, the endogeneity issue in charter school studies stems from the non-randomness of charter location. That is, the poor quality of a traditional public school may attract charter schools just as the threat of rain attracts people to carry umbrellas. With regard to systemic effect measures, researchers have used a variety of variables to capture competition ranging from those of poor quality to those of higher quality. However, many of those measures technically fail to capture the true essence of competition.

One consequence of an endogenous independent variable is that it potentially confuses causation. There is reason to believe charter operators will choose to locate their school near poorer performing traditional public schools in an effort to attract as many students as possible. The theory is that those students in the poorest performing schools are probably the unhappiest and will take advantage of a nearby charter school with zero tuition. If researchers fail to control for enough variables affecting the location decision, they may wrongly conclude that charter schools cause poor achievement in traditional public schools when actually the charter schools may have been purposely located near public schools that are already performing poorly.

Ni (2007) argued that the endogenous variable issue could be resolved using a fixed effects (FE) transformation. She suggested that the time invariant characteristics of a school determine the likelihood of charter alternatives locating near a traditional public school. According to Ni (2007: 16), "The FE estimator overcomes the non-randomness of charter school location by implicitly controlling for the unobservable time invariant school characteristics that influence its likelihood of facing charter competition." That claim, however, may be unreliable because it depends on the cause of the endogeneity. Ni's contention is true only if the cause of endogeniety is time invariant. But such an assumption is difficult to support because very few characteristics of schools are actually time invariant. Nevertheless, many researchers use the fixed effects transformation to capture time invariant characteristics to ensure coefficients in regressions are as accurate as possible.

The most widely accepted practice for minimizing the adverse effects of an endogenous independent variable requires obtaining an instrumental variable (IV). Bettinger (2005) and Imberman (2011) represent the two studies that address the charter location endogeneity problem with an 1V approach. Bettinger employed an IV strategy using two instruments: the proximity of a charter school to the nearest of 10 authorizing universities in the state, and a Herfindahl index for race. Imberman also used two instruments--the size and vacancy of large buildings--to predict the number of charter schools in a given area.

Betttinger...