Los Angeles charter schools struggle to lift achievement, by Douglas Lee Lauen

Which city has more charter school students than any other? New York? New Orleans? Washington, D.C.? These hotbeds of charter school activity do not hold a candle to Los Angeles, which has about 140,000 students in charter schools, more than double the enrollment of New York City. Also, charter schools in L.A. vary a great deal in their relationship to the district administration. Some are entirely new schools (start-ups), while others emerge from existing public schools (conversions). In addition, charters can remain tied to the district in many ways (affiliated), or they can establish a great deal of autonomy (independents). This makes L.A. a great place to examine differences across these charter school types and a great place to move beyond the now rather tired question of “are charter schools better?” to the more interesting question of “under what conditions do particular charter schools flourish?”

In our article Positioning Charter Schools in Los Angeles: Diversity of Form and Homogeneity of Effects we examined the three most prominent types of charters serving elementary grades between 2002 and 2008: independent start-ups, independent conversions, and affiliated conversions. We asked two empirical questions. First, what types of families and students attend particular kinds of charter school organizations? We found substantial differences in the background characteristics of charter and traditional public school students. Charter school students were less likely to be Black, Latino, LEP, special education, and low income and were more likely to be White, academically gifted, high achieving, and have more highly educated parents. For example, about 12 percent of the parents of traditional public school students attained a college degree or higher, compared with 35 percent of the parents of charter school students. We also found that one type of charter – affiliated conversions – serves a much more advantaged student population than either independent start-ups or independent conversions. The percentage of highly educated parents (college degree or higher) in these three types of schools is 71 percent, 25 percent, and 6 percent, respectively.

Second, across this segmented field of charter schools, do differing types of charters display varying effects on achievement growth over time? Despite serving a more advantaged student population than traditional public schools in LAUSD, charter effects on student test score growth were unimpressive. We examined charter school effects on test score growth overall, by charter type, and across four different cohorts of students, only for those students who remain in a charter or traditional public school during the time series. We report no statistically significant positive effects of attending a charter school on achievement growth. For the first three cohorts studied, charter school effects on test score growth were negative and significant. For the last cohort studied, the effect was negative, but not statistically significant. We find some small differences across charter types, but none of the charter school enrollment effects on test score growth for any cohort were positive among any of the three types examined. In conversion charters, the type of school with the most advantaged student population, the effects on test score growth were negative and significant for all four cohorts.

Now, the inevitable caveats. Our results come from a non-experimental design, so a number of alternative explanations for our findings remain plausible. The first is selection bias. Our estimates would be downwardly biased if students with less growth potential selected into charter schools. Students observed in a charter school in second grade, however, have substantially higher second grade scores, on average. So it appears that selection into charter schools is more likely positive rather than negative. Given higher levels of second grade scores, the downward trend among charter students probably reflects the difficulty that charters have had in raising test scores, not that students with less potential for growth have chosen to attend a charter school.

A second alternative explanation for our findings is differential attrition. It could be that students with high potential for growth leave charters, or students with low potential for growth leave traditional public schools, or both. If either or both conditions hold, growth rate differences between the two sectors could converge. After some empirical tests, we conclude that selective attrition is a possible explanation for the findings reported here.

A third threat is the curse of newness. Charter schools could be less effective simply because they are newer schools suffering growing pains. We find some weak evidence that charter schools are improving.

Finally, we must stress that these results cover only the period from 2002–2008 within LAUSD and are not generalizable to other cities or other time periods in the L.A. district. Moreover, our study uses student-level data with complete coverage for affiliated conversions, substantial coverage for independent conversions (86%), but only partial coverage of independent start-ups (60%). It is possible that findings based on student-level data with more complete coverage of the latter two categories could arrive at different conclusions. Unfortunately, such data were not available to test this claim.

So what do these findings suggest for the charter school debate? Proponents of charter schools, often drawing from neoclassical logic, argue that parents search across a (presumably) diverse market in demand of higher quality schools that better suit their family’s preferences. Yet critics of charter schools, often informed by a class-conflict perspective, claim that charters and market arrangements act to further stratify students and families along ethnic or social-class lines. Family demand under this account is structured institutionally a priori, sorting low-income students into less effective schools, rather than exerting competitive pressure to raise quality. How student or family selection unfolds — what kinds of children select what kinds of schools — is assumed to interact with the relative effectiveness of chosen schools under both neoclassical and class conflict theoretical perspectives.

We know that many parents actively search for school options, especially in urban centers like L.A. where campuses vary in size, safety, and the value-added effects on student learning. But as individual parents enter such a mixed market of diversifying schools, perhaps seeking to maximize their own utilities on a number of dimensions, they enter an organizational field populated by a colorful variety of schools. These schools then failed to lift their children’s performance, relative to peers attending regular public schools. So, even when parents behave like (neoclassically) rational actors, they play on a field populated by politically structured types of organizations that are not equally inviting of all students. Moreover, during the early years of charter school expansion in LAUSD, most charters — especially affiliated conversion charters serving the most advantaged students — struggled to improve student outcomes. This raises uncomfortable questions about charter school quality and accountability in Los Angeles, particularly for those schools most closely affiliated with the district. Apparently, at least in the early years of charter schools in L.A., school choice has not had significantly positive effects on the achievement of those who left traditional public schools. Therefore, we conclude that charter schools in this locale have done little to promote educational opportunity.