Charter movement’s U-shape

Charter schools in California tend to be bipolar, with disproportionate numbers of very high and very low performing schools, according to a newly released analysis by the California Charter Schools Association.

In duplicating a pattern that it found last year, in its first “Portrait of a Movement” annual report, the Charter Schools Association renewed its call for local authorizers to focus attention on academic achievement and to not renew the poorest performers on the state’s Academic Performance Index. CCSA has identified 29 schools – about 5 percent of the state’s eligible charters – that fail to meet the minimum academic criteria of three measures; CCSA is recommending closure of the 10 that are up for charter renewal by the end of this school year. (Small schools and schools classified as serving highly mobile and at-risk students, known as ASAM schools, were excluded from the list.)

One of CCSA’s metrics, which is highlighted in the Portrait, is the Similar Students Measure, or SSM. It predicts a school’s API after factoring in student demographics – including family income, parent education level, mobility, ethnicity, and percent English Language Learner and Special Education students – and then plots whether schools exceed or fall short of the prediction. It’s the SSM, a schoolwide counterpart to the value-added metric that projects a teacher’s impact based on students, that shows concentrations of charter schools at the upper and lower ends in comparison  with non-charter schools. By taking into account the student body served, particularly in schools with large proportions of disadvantaged students, the SSM gives a richer picture than the raw API score alone, according to the Association.

Charter schools form a 'U' based on their predicted API scores, with large clusters of low- and high-performing schools, compared with district schools, when defined in 5 percentile groups.  Click to enlarge. (Source: 2012 Portrait of a Movement, page 4.)
Charter schools form a 'U' based on their predicted API scores, with large clusters of low- and high-performing schools, compared with district schools, when defined in 5 percentile groups. Click to enlarge. (Source: 2012 Portrait of a Movement, p. 4.)

Based on three years of student testing of 789 charter schools, the SSM showed that one out of eight charter schools (12.7 percent, or 100 schools) fell in the bottom 5th percentile of their predicted API, compared with only 4.2 percent, or 312 schools, of non-charter schools (see graph and chart). If they had performed purely as predicted, only 39 charters would have been in the bottom 5th percentile.

Nearly one out of five (19 percent, or 150 schools) fell in the bottom 10th percentile of predicted API scores, compared with 9.1 percent, or 673 schools, of non-charters.

At the other end, one out of seven charters (14.7 percent, or 116 schools), fell in the top 5th percentile band and more than one out of five (21.8 percent, or 172 schools) were in the top 10th percentile. Had they performed as predicted, again only 39 charter schools would have been in the top 5th percentile. By comparison, 4 percent of non-charters (295 schools) are in the top 5th percentile and 9.1 percent (673 schools) are in the top 10th percentile.

It’s the “U” pattern of top and bottom schools, so visible on a graph, that CCSA wants to turn into a “J.”

The percentage of charter school students in the highest performing charter schools (13.4 percent of those tested in 2011) far exceeded the percentage attending the lowest performing charters (7.5 percent). Source: 2012 Portrait of a Movement, page 6. Click to enlarge.
The percentage of charter school students in the highest performing charter schools (13.4 percent of those tested in 2011) far exceeded the percentage attending the lowest performing charters (7.6 percent). Click to enlarge. (Source: 2012 Portrait of a Movement, p. 6.)

By another measure, it’s already happening. Nearly twice as many charter students attend the top 5th percentile of schools, based on their predicted API scores, than attend the bottom 5th percentile of charter schools: 30,350 students or 13.4 percent of charter students attending the highest-performing schools, versus 17,115 or 7.6 percent of charter students attending the lowest-performing schools.

The 'U' becomes more of a 'J' when the measure is the percentage of students in high versus low performing charters, based on their predicted API scores. That's because fewer students are enrolled in the least performing schools. Click to enlarge. (Source: 2012 Portrait of a Movement, page 5.)
The 'U' becomes more of a 'J' when the measure is the percentage of students in high- versus low-performing charters, based on their predicted API scores. That's because fewer students are enrolled in the lowest-performing schools. Click to enlarge. (Source: 2012 Portrait of a Movement, p. 5.)

That’s because the lowest-performing schools are smaller, an indication they may be having a harder time with finances and enrollments (see chart and graph).

Success with low-income students

The Portrait offers a further look at Calfornia’s 987 charter schools:

  • Charters operated by a nonprofit Charter Management Organization – like KIPP, Aspire Public Schools, Rocketship Education, and Alliance for College Ready Public Schools in Los Angeles – were concentrated at the top, with 40 percent of charters operated by a CMO in the top 10th percentile of Predicted API measure. Independent, solo operations tended to be clustered in the lower 10th percentile.
  • While there are high-achieving independent study and virtual or online charter schools, a disproportionate number are clustered at the bottom. Of the 25 schools identified as online charters, eight (32 percent) fell in the bottom 10th percentile, with three or one-eighth in the top 10th percentile. It’s still a small sample, the report notes, so more research is needed.
  • Charters serving primarily low-income children are doing well academically. Nearly a quarter of the 108,000 students tested in schools where at least half of the families qualified for free or reduced lunches attend schools in the top 5th percentile of schools. Thirty-five percent of students attend schools in the top 10 percentile  compared with 11 percent in the bottom 10th percentile of schools.
Data comparing Oakland's charter schools to Oakland Unified schools show that nearly half of charter school students attend schools in the top 10 percentile, far exceeding their predicted API score. Click to enlarge. (Source: 2012 Portrait of A Movement, page 36.)
Data comparing Oakland's charter schools to Oakland Unified schools show that nearly half of charter school students attend schools in the top 10th percentile, far exceeding their predicted API score. Click to enlarge. (Source: 2012 Portrait of A Movement, page 36.)

The Portrait highlights the success of charters in Oakland, where they comprise 23 percent (29 schools) of Oakland’s 126 schools and 19 percent of its 31,700 students. The report credits Oakland Unified’s “active oversight” and its “rigorous charter review process containing clear and transparent standards for approval and renewal.” A surprising 48 percent of charter students in Oakland attend a charter whose Predicted API score fell in the top 10th percentile, compared with 6 percent of students in district schools; only 4 percent of Oakland charter students attend a school in the bottom 10th percentile.

Weeding out lowest performers

The Similar Students Model is one of three metrics that CCSA uses to determine which schools to recommend for charter non-renewal. The other two are the absolute API score (must be over 700) and growth in API (minimum of 50 points over three years). CCSA says it invites charters to make the case why they shouldn’t be on the list by presenting  other data on student achievement. A few have taken them up on the offer, says Jed Wallace, CEO of CCSA. And CCSA says that it visited most, though not all, low-performing charters last year.

There’s been little change over the past decade in the concentration of low-performing charters. If its recommendations were followed, as charters come up for renewal, the number of charters clogging the bottom 10th percentile could be significantly reduced over the next seven years, the organization asserts.

But CCSA’s methodology and its criteria have met resistance, within the ranks of the community and outside of it. Eric Premack, executive director of the Sacramento-based Charter Schools Development Center, criticized the use of self-reported demographic data on income and family education and the focus on API and California Standardized Tests, which he calls a crude measure of a school’s performance.

Premack’s criticism may have resonated where it counts. CCSA spent months negotiating with Assembly Education Committee Chairwoman Julia Brownley to have the minimum API score of 700 and its three-year growth target incorporated into AB 440, on charter revocations.

But after a two-hour meeting with Gov. Jerry Brown, Brownley pulled it and another bill on charter accountability last September, shortly before they were to go to Brown for his signature. No one has said why, but Brown, who started two charter schools in Oakland, made it clear in a veto of a bill sponsored by Senate President pro Tem Darrell Steinberg last year that he wants to de-emphasize API and test scores as a measure of a school’s performance and look at qualitative measures, including school inspections.