Charters: Better than or just different from district schools?

New data show stark differences in who enrolls -- and remains -- in charter schools

A TRIFECTA OF initiatives focused on raising the charter school cap, including Governor Baker’s legislative proposal, a potential ballot initiative, and a class action lawsuit filed on behalf of Boston parents, has brought more heat than light to the charter school debate. The underlying argument of pro-charter school advocates can be boiled down to “charters are better than public schools while serving the same, or similar, student populations.”

In 2010, the Legislature passed An Act to Close the Achievement Gaps, which attempted to address acknowledged inequities in enrollment practices between public schools and charter schools related to under-enrollment of English language learners, low-income and special needs students. The law attempted to rectify the charter school practice of leaving empty seats unfilled in later grades by not accepting students after the initial enrollment grade.

A study released in October, “Who is Being Served by Massachusetts Commonwealth Charter Schools,” commissioned by the Massachusetts Association of School Committees, analyzed five years of Department of Elementary and Secondary Education (DESE) data since the law’s enactment and found that charter school enrollment practices do not support the pro-charter argument of “better than.” Rather, the data suggest that charter school enrollment – especially in Boston, Springfield, Worcester and Gateway Cities – is “different from” that of sending public schools.

Charter schools have chosen to be judged almost solely by their performance on MCAS and use results to support their “better than” claims. For this reason, policy makers should pay attention to the impact of enrollment practices on testing outcomes. Enrollment loss appears to be planned for and built into charter business strategy and contributes to higher test performance.

Certain student demographic characteristics are established predictors of testing outcomes. Therefore, looking at who enrolls, who stays and who leaves helps to understand charter outcomes. In general,

  • Regular education students outperform special education students.
  • English speakers outperform non-English speakers.
  • Middle- and high-income students outperform low-income students.
  • Females outperform males.

Charter schools often describe themselves as “schools of choice.”  It appears that parents, especially in urban districts, are choosing both to enroll and subsequently to remove their children from their charter schools. The data suggest that parents of 25 percent of K-4 students in urban charters take their children out of their schools before they complete the school’s program. In middle grades, the parents of 40 percent of the students in urban charter middle schools are choosing to remove their children. Parents of charter high school students are choosing to remove their sons and daughters in even higher percentages, 50 percent in urban charters and 45 percent in non-urban schools during the course of the school program.

Charters Under-Enroll Low-Income Students

Approximately 33 percent of all students attending publicly financed district and charter schools are low-income. The single greatest predictor of student performance on standardized tests is family income.

Skinner chart 1On average, urban charters tend to enroll approximately the same percentage of low-income students as urban district schools. However, this is not true of all charter schools. In reference to the 2014 state test results, the Mystic Valley Charter School in Malden website states, “Mystic Valley Students Ace MCAS. Mystic Valley Regional Charter School students once again scored at the top of all Massachusetts students on the most recent MCAS 10th Grade ELA exams. Ninety-nine percent of the Mystic Valley sophomores scored either Advanced or Proficient on the ELA portion of the exam. Not one member of the sophomore class received a Failing grade on any of the three MCAS test areas.”


The level of poverty at this school compared with the Malden public schools is telling.  From 2010 to 2014, an average of 60 percent of students in the Malden district public schools are identified as low-income. Over the same period, an average of 20 percent of Mystic Valley Charter School students are poor, or one-third the rate of Malden.

Barnstable’s Sturgis Charter School website states, “U.S. News and World Report recently announced the Best High Schools Ranking for 2015. Sturgis received a gold medal and is ranked #1 high school in MA, #6 charter school in US, and #32 high school in US.”

The Barnstable public schools five-year poverty rate is about 34 percent and the Dennis-Yarmouth public schools rate is about 42 percent. Both are above the state average. Sturgis consistently has about 7 percent of its students identified as low-income — a fraction of the low-income percentage in these two sending districts.

Charter Schools Under-Enroll English Language Learners

Slightly more than 4 percent of Massachusetts students are English language learners; the vast majority are enrolled in urban public schools. While overall there has been an increase in the percentage of ELLs in charters since 2010, much of this can be attributed to two out of the 52 urban charters: MATCH Community Day in Boston, with an 82 percent ELL population, and Community Day Gateway in Lawrence, at 79 percent.

Skinner chart 2The number of students tested by ACCESS is a good measure of how many ELL students are enrolled since all ELLs must take this state-mandated language proficiency assessment. Most other charter schools do not enroll equivalent percentages of ELL students. Holyoke and Springfield have well above the state average of ELLs. Yet the charter schools in these two Gateway Cities enroll lower percentages of ELLs than the district averages based on ACCESS testing taking numbers. Holyoke public schools tested 28 percent of all students (1,560 of 5,571); Holyoke Community Charter tested 7 percent (51 out of 702 students); Paulo Freire Charter tested 8 percent (12 out of 152).

Similarly with Springfield, the district school system tested 17 percent of all students (4,397 of 25,862). Baystate Charter tested 8 percent (13 of 161); Martin Luther King, Jr. Charter tested 10 percent (39 of 373); Sabis Charter tested 2 percent (45 of 1,574 ); Veritas Charter tested 13 percent (21 of 156 ).

Twenty-nine charters have fewer than 10 ELLs tested in each grade so data is unavailable from DESE, indicating an ELL enrollment that is well below what is typical for Gateway districts. Seven are in Boston, where over 30 percent of the students in the district public schools are ELLs. Another 14 charters have ACCESS data for only one or two grades offered, including an additional nine charters in Boston.

Charter Schools Under-Enroll Special Education Students

Skinner chart 3Special education services must address the learning needs of individual students and are to be provided in the “least restrictive environment.” Approximately 66 percent of students with disabilities (SWD) in district public schools are provided services through full inclusion in regular education classrooms; 20 percent through partial inclusion; 9 percent in substantially separate programs; and 6 percent are in out-of-district placements.

Close to 90 percent of special needs students in charters were enrolled in full inclusion programs. Students with moderate to severe disabilities are either not enrolled or under-enrolled when compared with their host and sending districts.

Students with the most challenging learning needs requiring other than full inclusion programs most often fall into one of six categories: autistic, developmentally delayed, emotionally impaired, intellectually impaired, or hearing and/or visually impaired: 98 percent of the 51,334 students in these categories are enrolled in district public schools or out-of-district placements.

Charters Enroll and Retain More Female Students

A 2014 American Psychological Association study that reviewed 308 studies between 1914 and 2011 found that, “Despite the stereotype that boys do better in math and science, girls have made higher grades than boys throughout their school years for nearly a century.” The authors found that while boys “typically score higher on math and science tests, females have the advantage on school grades regardless of the material.” Therefore, schools that enroll higher percentages of female students should, on average, have better results than those with higher percentages of male students.

Males require more academic support than females. Two-thirds of all special needs students are male. Males are twice as likely to be suspended from school as females. In 2013-14, district public schools suspended 3 percent of all females and 7 percent of all males. Charter schools suspend 7 percent of females and 13 percent of males.

Skinner chart 4Further, in some charters, male, low income, special needs, and ELL students are suspended at higher rates. In 2013-14, Berkshire Art & Technology Charter Public School had a 35 percent suspension rate: out of every 100 student, 35 were suspended at least once. City on a Hill had a 33 percent rate and Roxbury Prep a 47 percent rate. In all schools, males were suspended more than females; low-income students, ELLs, and special needs students were also suspended at high rates.

As the chart below illustrates, Berkshire Arts, a grade 5-12 school, has enrollment loss such that only 64 students completed to grade 12 in five years. At City on a Hill, a grade 9-12 school, only 239 of the 587 ninth graders were enrolled in grade 12. At Roxbury Prep, a grade 5-8 school, only half of the 6th graders are still enrolled in grade 8.

Skinner chart 5

The gender differential in suspension rates results in fewer male students enrolled in charter schools. In 2013-14, approximately 20,000 more males attended district public schools than females; charters enroll and retain more females. Charter enrollment can be divided into three grade spans:

  • Elementary School: Kindergarten to grade 4. While enrollment remains relatively constant across the kindergarten to grade 4 span at non-urban charters, the urban charter kindergarten to grade 4 female enrollment loss is 24 percent and the male loss is 34 percent. Three out of ten students enrolled in kindergarten have left by grade 4.
  • Middle School: Grades 5-8, with the highest enrollment in grade 6 when some charters begin enrollment. For every 10 students enrolled in grade 6 in an urban charter school, four are gone by grade 8, a loss of 40 percent. Non-urban charters show enrollment growth over grades 6-7 with a slight decline in grade 8.
  • High School: Grades 9-12. Over twice as many females are enrolled as males in all grades. Urban charters enroll fewer students than their non-urban counterparts and have higher enrollment losses. Female enrollment loss from grade 9 to grade 12 is 43 percent and for males is 52 percent. For every ten students enrolled in grade 9, almost five
  •  leave by grade 12.

Skinner chart 6

 Impact of Enrollment Practices on MCAS Results

On the 35 grade-level administrations of ELA and math MCAS tests from 2010 to 2014, females outperformed males in every grade in English language arts, often by 10 percentage points or more. In mathematics, females outperformed males on 30 administrations, performed the same on two administrations, and placed behind males on three.

The impact of these practices on charter school MCAS results is best illustrated by looking at tested enrollment results from charters themselves.

At Sturgis Charter School in Barnstable over the five administrations of MCAS from 2010 to 2014, 846 students took the grade 10 tests:

  • 470 females (56 percent).
  • 376 males (44 percent).
  • 124 were low-income (15 percent).
  • 98 were students with disabilities (12 percent).
  • Two were English language learners (0.2 percent).

Skinner chart 7

At Boston Collegiate Charter School over the five administrations of MCAS from 2010 to 2014, 265 students took the tests:

  • 158 females (60 percent).
  • 107 males (40 percent).
  • 103 were low-income (39 percent).
  • 45 were students with disabilities (17 percent).
  • None were English language learners.
Meet the Author

Charter school demographic data do not support the pro-charter argument that more schools are needed because charters are “better than” public schools at educating the same or similar students. Charter schools enroll and retain more females than males; fewer special needs students and virtually no students with moderate to severe disabilities; fewer English language learners; and, in some cases, fewer low-income students. In short: fewer males, fewer English language learners, fewer low-income students, and fewer special needs students and most with mild learning needs results in better test performance.

Kathie Skinner is the head of Skinner Research, an independent consulting firm. She is the former director of the Center for Education Policy and Practice at the Massachusetts Teachers Association. She can be reached at