California charter school students outperform district school ‘twins’ in national study

A student at Rocketship Public Schools in San Jose works on a math problem.

Credit: Andrew Reed / EdSource

Charter school students in California significantly outperformed similar students in nearby traditional public schools in reading and scored about the same in math, according to a comprehensive study of pre-pandemic test results of charter schools nationwide, released earlier this month.

The gap between California charter and district schools in reading achievement has widened since the first report 14 years ago by the Center for Research on Education Outcomes at Stanford University. There has been steady progress in math as well; charter schools’ scores were significantly behind in the first study in 2009.

This was particularly true when the difference in test scores was translated into learning gains or losses for multisite charter management organizations operating in California. For example, when compared with similar students in district schools, students in Los Angeles-based Alliance College-Ready Public Schools gained the equivalent of 107 days in additional learning, about 40% of a year. Students in Bay Area-based Rocketship Public Schools gained three-quarters of a year in additional learning days in math, based on CREDO’s methodology.

CREDO designated 32 California charter management organizations as “gap-busters,” whose average achievement exceeded the state average, and whose historically disadvantaged students showed growth that was strong or stronger than their non-disadvantaged peers in the same schools; 22 CMOs excelled in both math and reading,  including ACE Charter Schools in San Jose, Para Los Niños in Los Angeles, and King-Chavez Neighborhood of Schools in San Diego.

Charter schools are independently run public schools. Managed by nonprofit boards of trustees, they are not bound by some requirements of the state education code while being held to many of the same accountability, teacher credentialing and testing mandates.

About 1 in 9 California public school students — 685,000 out of 5.85 million students — attended one of about 1,300 charter schools in 2022-23. Although California has the largest number of charter schools and students in the nation, CREDO found their performance gains, relative to similar district school students, were smaller than charter students in a dozen of the 29 states covered by the study, plus Washington, D.C., and New York City. Charter school students in a dozen states also outperformed their peers in math; in California, the difference was not significant. The differential was biggest in Massachusetts, Illinois, Rhode Island and New York state charter schools.

Charter school and similar district students in all other states performed about the same, except for one state where charter school students performed worse in reading (Oregon) and three states where they performed worse in math (Oregon, South Carolina and Ohio).

Progress over time

CREDO’s report, released earlier this month, is its third tracking charter school students since 2009. The latest covered 2014-2019 and included four years of test results.

“Between the 2009 and 2023 studies, against a backdrop of flat performance for the nation, the trend of learning gains for students enrolled in charter schools is both large and positive,” the report said.

As in its past studies, CREDO’s analysis found wide differences in which students benefited from attending a charter school in California. 

Black, Hispanic and low-income students excelled relative to similar students in typical district schools in reading and math, while white students and especially students with disabilities did significantly worse in both math and reading than their peers. English language learners in California charter schools did slightly better than district school peers in reading but significantly better in math.

“We are pleased to see that California’s charter schools are performing particularly well with historically underserved students and improving over time,” said Elizabeth Robitaille, chief schools officer for the California Charter Schools Association. 

Compared with charter management organizations, students attending homegrown, stand-alone charter schools showed more modest results, doing significantly better, though to a smaller degree, than similar students in traditional public schools in reading and about the same in math.  Granada Hills Charter High School, a popular charter school in Los Angeles, is also the nation’s largest charter school, with 4,600 students.

As centrally run organizations, CMOs usually have uniform curriculums and instructional practices, and share a common educational philosophy as well as administrative costs. CMOs like Aspire Public Schools, KIPP, and Green Dot Public Schools have located mainly in urban areas and targeted underachieving, low-income students of color. In California, CMOs comprise 38% of charter school students — about the same proportion of students as the nation. (California Charter Schools Association’s data put the number of students closer to 50%.)

Nationally, CMOs “are producing much of the learning gains we observed,” the report said. The ability of CMOs to scale up their success “puts dozens of CMOs at the forefront of efforts to provide education that is both equitable and effective in moving student achievement.” 

In California, however, some charter management organizationss have scaled back their operations or shifted their growth to the Inland Empire and other states because of declining student enrollment in urban areas, a difficulty hiring teachers and stiff resistance from authorizing school districts to further charter school expansions.

Eric Premack, founder and CEO of the Charter Schools Development Center in Sacramento, which advises primarily stand-alone charter schools, said he disagrees with CREDO’s “unstated assumption that test scores are everything.” 

 “Many charter schools, especially CMO-managed ones, care a lot about test scores and focus on instructional methods designed to boost them,” he said.  “Many (other) charter schools, however, are formed by folks who hate standardized testing and aren’t particularly concerned about their scores.  Thus, the generally high relative scores are all the more remarkable.”

Robitaille had a different take. “We’ve done reports looking at ways charter schools support students’ socio-emotional needs and redefine college and career readiness. Test scores are not the only determining factor of effectiveness, but certainly are important.”

Other studies of academic performance have compared test scores of charter schools and neighboring schools or compared charter school students with those students who were denied admission through a lottery. For its comparison, CREDO paired more than 1.8 million charter school students with “virtual twins” – students with similar racial, ethnic, income, grade and other student characteristics at district schools that charter school students would have attended. CREDO said that more than 80% of tested public school students were included in its data set.

“As a former researcher/methodologist,” Premack said, “I appreciate their novel approach.  The ‘virtual twin’ methodology is creative and presumably helps to significantly reduce the potential for selection bias and other comparability issues.”

Critics in California and other states have discounted comparisons, charging that charter schools weed out poor-performing students or discourage low-performing students from enrolling. CREDO said if there were cherry-picking, then charter school students would show higher academic performance when they enroll. By following students’ growth over time, CREDO’s analysis “found the opposite is true: charter schools enroll students who are disproportionately lower achieving than the students in their former traditional public school,” it said. 

Then, by measuring individual students’ growth over time — a method of accountability that California has not yet adopted, unlike most states – CREDO is providing “critical information to help us understand the relative effectiveness of schools in helping students grow,” Robitaille said.

Losses and gains in days of learning

CREDO converted the differences in scores on standardized test scores between charter and district school students in each state into gained or lost days of learning, based on a 180-day school year.

Nationally, charter school students outpaced their peers in district schools by 16 additional learning days in reading and six days in math in this year’s report.

 California charter schools outpaced their peers in district schools by 11 days in reading and four days in math; the latter gain was not statistically significant, CREDO said.

Hispanic students in California charter schools showed the largest growth, relative to district school peers. Special education students showed the most learning loss.

California CMOs showed 19 days of additional learning in reading; stand-alone charter schools achieved seven extra learning days. CMOs’ 10 days of growth in math were not statistically significant, CREDO said; nor was the one day of extra learning by stand-alone charter schools. 

Macke Raymond, CREDO’s founder and director, suggested that how charter schools are authorized in states could be one factor in variations of performance. In California, school districts are charged with approving charter schools and overseeing their performance. Until Gov. Gavin Newsom revised the approval and renewal process through legislation in 2019, the State Board of Education heard all appeals of denials. In other states, like Massachusetts, the state or other entities besides districts, such as universities, play a more active role in working with and holding charters accountable. 

“I was surprised when I looked at California’s results. There is still a proportion that is doing worse than the local district school option,” Raymond said. “That says to me that authorizers are not doing their job or are cognizant of low-performing schools and choose not to act.”  

In its initial 2009 study, CREDO found that charter school students nationally did significantly worse than their virtual twins in district schools in both reading (six fewer days of learning) and math (17 days behind). The only bright spots in California then were significant growth in reading for Black students and positive gains in both reading and math for English learners.

By the second study, in 2013, charter school students nationwide were doing nearly as well in math and exceeded their district school peers in reading. The positive trend continued in the latest study.

The one area of severe poor performance was in online charter schools, which constituted 6% of charter schools nationally before Covid and a tiny proportion in California; 73% of the 214 online schools in the study did worse in reading and 90% did worse in math. Students in those schools had the equivalent of 58 fewer days of learning in reading and 124 fewer days in math — enough to lower the overall growth of charter school students nationwide by six learning days in math and reading.

CREDO did not break out the data for California online charter schools. But previous EdSource articles pointed to poor performance during the years of the study by students at the two largest online operations — K12, which is operated by California Virtual Academies, and California Connections Academy.