The first charter school opened 25 years ago. Since then, the charter sector has grown quickly: there are now over 6,000 charter schools operating in 43 states.  For a long time, the debate over charter schools has revolved around the simplistic question of whether they are better or worse than traditional public schools. One rationale for charter schools is that they will introduce an element of market competition to the education sector. Another view is that they can serve as a lab for innovation, allowing educators to experiment with new ways of organizing schools. Charter schools that are successful will thrive, while unsuccessful charters will be forced to exit the market.
A number of new research studies are beginning to investigate some more nuanced questions with regard to charters. Just last week, the annual conference of the Association for Education Finance and Policy featured new research on topics such as the importance of charter organization type, the characteristics of charter schools associated with effectiveness, charter student outcomes beyond standardized test scores. 
A key challenge for this research is to account for the subtle differences between students who choose to attend charters and otherwise similar children who attend traditional public schools. Following an existing literature on teacher and school value-added, this recent charter school research controls for student demographics and baseline academic achievement in order to estimate the effectiveness of individual schools. While this approach has clear limitations, evidence suggests that such value-added estimates of charter schools are actually quite similar to the school quality measures obtained via lottery-based experimental studies. 
Using rigorous non-experimental methods, a 2013 study of charters in 16 states by the Center for Research on Education Outcomes found that average charter school effectiveness increased overall, due in large part to closures of poorly performing schools. 
Two subsequent studies examine the evolution of the charter school sector over time in Texas and North Carolina in greater depth. They investigate how the student composition and academic achievement of charters have changed over time, what drives charter school entry and exit, and the factors that are associated with charter school success.
The Texas study finds evidence that charter schools are evolving in a productive way, and offers additional insight into some potential mechanisms. At the beginning of the study period in 2001, there was substantial variation in quality across charter schools and, on average, charter schools in Texas were less effective than traditional public schools. By 2011, charter schools were roughly equivalent to traditional public schools in terms of their ability to raise academic achievement. The authors show that this change was due primarily to improvement at the bottom tail of the charter school quality distribution. A number of the lowest-performing charters closed voluntarily or were shut down by an authorizer. Those charter schools that remained saw modest, but consistent, improvements in effectiveness. And, the more successful Charter Management Organizations expanded over this period, spawning more effective charter schools.
In conjunction with this improvement, the authors find that student mobility into and out of charters declined over this period, perhaps suggesting that parents were more satisfied with charter schools. At the same time, the students transferring into charters were increasingly higher achieving with fewer disciplinary problems. The authors claim that the improvement in student “quality” cannot fully explain the growing effectiveness of charter schools in the state, although this claim is very hard to prove definitively. While the value-added models utilized by the authors control for the prior student achievement, the increasingly positive selection into charters almost certainly brings more students with hard-to-measure positive attributes. Thus, it seems likely that the changes in student quality and reductions in student mobility were both cause and consequence of improving charter quality.
Finally, the authors provide some intriguing suggestive evidence that the improvement in charter school quality is associated with the growing prevalence of charter schools that adhere to a “No Excuses” approach that focuses on strict discipline, high expectations, and increased learning time.
Another study of charter schools in North Carolina comes to some of the same conclusions. Examining the evolution of the charter school sector in North Carolina between 1999 and 2012, the authors find changes in the mix of students served, the distribution of student performance, and the quality of the student-school match. At beginning of their sample period, the average effectiveness of charter schools was below that of traditional public schools. By 2012, the average charter school was equally effective as the average traditional public school in reading and slightly more effective in math. Moreover, they find that charter students are more likely to remain in the school from year to year.
As in Texas, the North Carolina researchers find that charters in that state are attracting more capable students over time, as measured by prior achievement and attendance. They find that charters now are serving fewer poor and minority students than in prior years. Charters were always more racially segregated than traditional public schools in North Carolina, and they are becoming more so over time. Unlike the authors of the Texas paper, however, the authors of the North Carolina paper contend that the improvement in charter school effectiveness is due to the increasingly positive selection of students into charter schools.
But like in Texas, my read of the evidence is that this question remains unresolved. The authors of the North Carolina study attempt to control for hard-to-measure permanent characteristics of students who attend charters by estimating what is known as student “fixed effect” models, which involves measuring how student performance changes as students switch between the charter and traditional sectors. The key assumption here is that the unobserved, time-varying factors associated with the student’s decision to switch sectors (e.g., getting bullied in current school) are not systematically associated with the change in student performance. Another assumption is that the set of “switchers” are representative of all students who attend charters. I suspect that neither of these assumptions holds in practice.
I don’t expect that the next generation of charter school research will (nor should it) provide a single definitive answer to what is a complex and nuanced issue. But I am heartened to see the research literature evolve to focus on deeper questions. Moving forward, I hope to see even more research that investigates the mechanisms underlying charter school effectiveness.
— Brian A. Jacob
Brian Jacob is a Nonresident Senior Fellow at the Brookings Institution and the Walter H. Annenberg Professor of Education Policy, Professor of Economics, and Professor of Education at the University of Michigan.
This post originally appeared as part of Evidence Speaks, a weekly series of reports and notes by a standing panel of researchers under the editorship of Russ Whitehurst.
The author(s) were not paid by any entity outside of Brookings to write this particular article and did not receive financial support from or serve in a leadership position with any entity whose political or financial interests could be affected by this article.
2. For a complete program of the conference, see: http://aefpweb.org/conferences/program. For example, researchers from the Consortium of Chicago School Research presented preliminary work trying to understand the long-term outcomes of charter high schools in Chicago. Another paper focused on how the organization type (CMO vs. EMO vs. independent) influenced effectiveness among Indiana’s charter schools. Yet another paper investigated principal compensation among New Orleans charter schools.
3. Deming, David J. (2014). “Using School Choice Lotteries to Test Measures of School Effectiveness.” American Economic Review 104, no. 5 (May): 406-11; Abdulkadiroğlu, Atila, Joshua D. Angrist, Susan M. Dynarski, Thomas J. Kane, and Parag A. Pathak. (2011). “Accountability and flexibility in public schools: Evidence from Boston’s charters and pilots.” Quarterly Journal of Economics 126, no. 2: 699-748; Fortson, Kenneth, Natalya Verbitsky-Savitz, Emma Kopa, and Philip Gleason (2012). “Using an Experimental Evaluation of Charter Schools to Test Whether Nonexperimental Comparison Group Methods Can Replicate Experimental Impact Estimates.” Washington, DC: Institute for Education Sciences (April); Bifulco, Robert (2012). “Can nonexperimental estimates replicate estimates based on random assignment in evaluations of school choice? A within-study comparison.” Journal of Policy Analysis and Management, 31, 729-751.