Fall 22

Journal of Child and Family Studies (2018) 27:3132–3140 https://doi.org/10.1007/s10826-018-1143-z

INVITED REVIEW

Charter School Enrollment Effects: A Review of Results from Recent Large-Scale Studies

Luke J. Rapa1 ● Antonis Katsiyannis1 ● Robin Parks Ennis2

Published online: 17 June 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract Responding to calls for increased accountability regarding performance of students in public schools and concerns over the capacity of public schools to improve outcomes, school choice has become a mainstream and often controversial issue in public education; it is also a priority of the current federal government administration. Given the mixed evidence of effects on academic performance of charter school enrollment, along with the common perception that charter schools fail to meet the needs of special populations of students—including students with disabilities and English language learners (ELLs)— this study examined the performance of these subgroups across traditional public schools and charter schools. Specifically, we reviewed five recent large-scale studies that assess the impact of charter schools on academic performance paying particular attention to how students with disabilities and students identified as ELLs fare in charter schools relative to their peers served in traditional district schools.

Keywords Charter schools ● Enrollment effects ● Students with disabilities ● English language learners

Introduction

Responding to calls for increased accountability regarding performance of students in public schools and concerns over the capacity of public schools to improve outcomes, school choice has become a mainstream and often con- troversial issue in public education. This matter of school choice is possibly more controversial than ever, particularly given the funding priorities and interests of the current administration and its Secretary of Education, Betsy DeVos. Lauded as one means of addressing performance concerns and stirring competition in order to improve public schools, charter schools—publicly funded and generally governed by a group or organization under a legislative contract (or charter) with the state, district, or other entity—allow par- ents the ability to choose a school they believe to be well- suited to meet the needs of or benefit their child(ren) (Thomsen 2016).

Minnesota passed the first charter school law in 1991, and the nation’s first charter school was established in 1992. Charter schools enjoy a certain level of autonomy; however, they must meet specific accountability standards outlined in the charter (Thomsen 2016; see also, Garda 2012). A charter school’s autonomy typically involves release from state and district regulations governing traditional public schools, such as those involving staffing, curriculum, and budget decisions (Clark et al. 2015). Yet, as publicly funded entities, charter schools are subject to the same account- ability standards as traditional public schools.

In 2015–2016, there were 6855 public charter schools across 43 states and the District of Columbia, a substantial growth since 2000–2001 when there were 1993 public charter schools (only seven states, Kentucky, Montana, Nebraska, North Dakota, South Dakota, Vermont, and West Virginia, are currently without charter legislation). Simi- larly, the number of students enrolled in public charter schools rose from 0.4 million in 2000–2001 to 2.8 million in 2015–2016, a 2.4 million student increase (with the greatest growth occurring at the elementary level). In 2015–2016, the percentage of charter schools as a percen- tage of total public schools was 7%, and the percentage of students enrolled in charter schools, as a percentage of total enrollment in public schools, was 5.7%.

* Luke J. Rapa [email protected]

1 Clemson University, Clemson, SC, USA 2 University of Alabama at Birmingham, Birmingham, AL, USA

12 34

56 78

90 () ;,:

12 34 56 78 90 (); ,:

California has the largest number of students enrolled in charter schools, at 568,800 students, representing 9% of all students in the state. The District of Columbia, with 35,800 students, has the highest percentage of public school students enrolled in charter schools at 43% (National Center of Education Statistics 2018; see also, McFarland et al. 2017). This dramatic growth, in both the number of charter schools and the number of students attending charter schools and schools, over the last 15 years—particularly since 2009—may be attributed to the Race to the Top competition, which was part of the American Recovery and Reinvestment Act of 2009 (Clark et al. 2015). Conse- quently, given the explosive growth of charter schools, it is necessary to examine whether these schools improve out- comes for enrolled students relative to traditional school peers, particularly among underrepresented groups and students with disabilities. First, we address the issue of participation by underrepresented groups and students with disabilities. Second, we briefly discuss studies focusing on differential effects on academic achievement between charter school students and peers in traditional public schools with an emphasis on recent large-scale studies, paying particular attention to the performance of students with disabilities and students identified as English language learners (ELLs). Third, we discuss implications for improved practice and additional research.

Student Body Composition in Charter Schools—Minority Enrollment and Students with Disabilities

From 2000 to 2015, the student body composition attending charter schools mirrored that of public schools in general (see Table 2). Over those 15 years, Hispanic enrollment grew from 19% to 32% and Asian enrollment grew from 3% to 4% whereas White enrollment decreased from 43% to 33%, Black enrollment decreased from 33% to 27%, and American Indian/Alaska Native enrollment decreased from 2% to 1%.

In 2014, regarding attendance in high poverty schools (75% of students or more qualify for free or reduced-price lunch), 33% of students attended charter schools vs. 24% attending public schools (National Center of Education

Statistics 2018). Indeed, families zoned to more racially/ ethnically and economically diverse schools are more likely to participate in school choice (Kristie et al. 2015). How- ever, school choice may lead to increased segregation. Schools in Durham, for example, were more segregated by race and class as a result of school choice programs vs. having students attending their geographically assigned schools (Bifulko et al. 2009).

Regarding students with disabilities, in states with charter laws, this population comprised 10.62% of total enrollment in charter schools and 12.46% of total enroll- ment in traditional public schools. Charter schools that operate as independent districts enrolled 11.5% students with disabilities vs. 9.74% in charter schools that operate as part of a district. Yet, enrollment patterns vary by type of disability. Charter schools reported higher percentages of enrollment of students with specific learning disabilities, autism, and emotional and behavioral disorders and lower percentages of enrollment of students with developmental delays and intellectual impairments. Charter schools also served students with disabilities in more inclusive settings and used disciplinary suspensions proportionally with tra- ditional public schools (Rhim and Kothari 2018). There were 137 charter schools that focused primarily or entirely on students with disabilities (Rhim et al. 2015). To ensure access, the U.S. Department of Education (2014) issued a Dear Colleague letter to school districts on requirements under federal civil rights laws and the applicability to charter schools under Section 504 of the Rehabilitation Act of 1973 (U.S. Department of Education 2014). Specifically, eligibility criteria for entrance to charter schools must be nondiscriminatory and cannot have the effect of excluding students based on having a disability (see also, Dunn et al. 2018; United States Government Accountability Office 2017).

Academic Outcomes Associated with Charter School Enrollment

Most charter schools are open enrollment schools, allowing any student within district boundaries or the state to enroll given space availability. Critics, however, argue that these schools draw students and resources away from traditional

Table 1 Change in number of charter schools and charter school enrollment: 2000–2001 to 2015–2016

Year Number of charter schools

Fall enrollment in charter schools

Charter schools as a percent of total public schools

Charter school enrollment as a percent of total fall enrollment in public schools

2000–2001 1993 448,343 2.1% 1.0%

2005–2006 3780 1,012,906 3.9% 2.1%

2010–2011 5274 1,787,091 5.3% 3.6%

2015–2016 6855 2,845,322 7.0% 5.7%

Journal of Child and Family Studies (2018) 27:3132–3140 3133

public schools and suggest that inadequate oversight will lead to charter schools with suspect effectiveness (Clark et al. 2015). Broadly, performance of students attending charter schools in reading and mathematics has been mixed. For example, according to the Center for Research on Education Outcomes at Stanford University (CREDO; Cremata et al. 2013), the examination of student outcomes across 27 states revealed that the fraction of charter schools outperforming traditional schools was 25% in reading and 29% in math; the fraction that performed worse was 19% in reading and 31% in mathematics. Further, comparison of outcomes in the 16 states included in CREDO’s original 2009 study and their 2013 study reflected gains in reading and math performance for Hispanic, Black, and English learners as well as those with disabilities in reading.

Gleason et al. (2010) reported that the performance of students attending charter middle schools performed simi- larly to peers in math or reading, attendance, grade pro- motion, and student conduct. However, students attending urban charter middle schools performed better in math; charter schools outside large urban areas, serving fewer low income students, and those serving higher achieving stu- dents had negative effects on reading and mathematics. Similarly, Zimmer et al. (2009) found performance of stu- dents in charter middle and high schools was similar to those in traditional public schools; charter schools did not have an effect on achievement on nearby public schools. In contrast, charter school students were more likely to grad- uate and go on to college (in Chicago and Florida). Regarding elementary schools, students in charter schools experienced small but statistically significant drops in per- formance (Imberman 2011).

Clark et al. (2015) also reported mixed results in per- formance across schools and students; however, more positive impacts were observed for more disadvantaged students and more negative impacts were observed for more advantaged students (terms used by original authors). Similarly, Betts and Tang (2011) reported that charters under-perform traditional public schools in some locations, grades, and subjects, and out-perform traditional public

schools in other locations, grades, and subjects based on a meta-analysis. Non-significant effect sizes were reported for middle school reading and for high school math and read- ing; studies that focused on urban areas had larger effects than studies that examine wider areas. Further, KIPP charter middle schools had positive effects of 0.096 and 0.223 for reading and math respectively; New York City and Boston charter schools also generally reported higher performance gains than charter schools elsewhere.

Further, Zimmer et al. (2009) reported that charter schools did not have an effect on the achievement of nearby public schools. In contrast, an analysis of statewide school- level data of Michigan schools from 1994 to 2004 found that charter competition had an increasingly negative impact on student achievement and school efficiency in Michigan’s traditional public schools (Arsen and Ni 2012; Ni 2012; Carr and Ritter 2007). In contrast, positive performance effects were reported in Florida (Sass 2006) and Texas (Booker et al. 2005).

Recent Large-Scale Studies

Given the common perception that charter schools fail to meet the needs of special populations of students, including students with disabilities and ELLs (Government Accountability Office 2012), we were particularly interested in examining the performance of these subgroups across traditional public schools and charter schools. Thus, we did not intend this review to be a comprehensive literature review of all published empirical studies examining the effects of charter school enrollment on student performance. Instead, our aim here was to review recent large-scale stu- dies that assess the impact of charter schools on academic performance, paying particular attention to how students with disabilities and students identified as ELLs fare in charter schools relative to their traditional district school peers.

Though not without limitations or free from criticism (e.g., Hoxby 2009; Maul 2015), a series of reports pub- lished by the CREDO represents the most recent,

Table 2 Demographic composition of traditional public schools and charter schools: 2000–2001 to 2015–2016

White Black Hispanic Asian/Pacific Islander

American Indian/Alaska Native

Two or more races

Year TPS Charter TPS Charter TPS Charter TPS Charter TPS Charter TPS Charter

2000–2001 61.2% 42.7% 16.9% 33.2% 16.6% 19.4% 4.2% 2.9% 1.2% 1.8% — —

2005–2006 57.4% 40.5% 16.9% 32.1% 19.8% 22.4% 4.6% 3.6% 1.2% 1.4% — —

2010–2011 53.1% 36.2% 15.5% 28.9% 22.9% 27.3% 5.0% 3.7% 1.1% 0.9% 2.3% 2.9%

2015–2016 49.9% 33.1% 14.7% 26.8% 25.6% 31.7% 5.4% 4.3% 1.0% 0.7% 3.4% 3.4%

Change 2000–2001 to 2015–2016 –11.3 –9.6 –2.2 –6.4 +9.0 +12.3 +1.2 +1.4 –0.2 –1.1 +1.1 +0.5

TPS traditional public school

3134 Journal of Child and Family Studies (2018) 27:3132–3140

substantial, large-scale data on the effects of charter school enrollment on academic performance. In the past 5 years, CREDO has published two substantial, large-scale national studies, including the National Charter School Study (Cre- mata et al. 2013) and the Urban Charter School Study (CREDO 2015). More recently, in the latter half of 2017 alone, CREDO published three large-scale, state-specific studies, including Charter School Performance in New York (CREDO 2017a), Charter School Performance in New York City (CREDO 2017b), and Charter School Performance in Texas (CREDO 2017c).

Uniquely, while this series of studies addresses charter school performance broadly, it also includes data that affords the examination of charter school effects for stu- dents with disabilities and those identified as ELLs. We know of no other recent (i.e., within the last 5 years) large- scale datasets or studies that allow for the examination of the performance of these subgroups in charter schools vs. traditional public schools. As such, our review draws on these reports to assess the effects of enrolling in charter schools for these groups in particular. We begin with a brief overview of each CREDO study, and then highlight dif- ferences in academic performance—specifically for students with disabilities and those identified as ELLs—based on enrollment in charter schools vs. enrollment in traditional public schools.

National Charter School Study

The National Charter School Study (Cremata et al. 2013) was built on the prior study by CREDO entitled Multiple Choice: Charter School Performance in 16 States (CREDO 2009). Updating and expanding that analysis, the National Charter School Study reported on charter school perfor- mance for 27 states that, taken together, comprised more than 95% of the charter school student population at the time. Demographically, of the 1,704,418 students included in the charter school sample, 54% were students in poverty, 8% were identified as students with disabilities, and 9% were identified ELLs. The ethnic-racial composition of the charter school sample was 35% White, 29% Black, 28% Hispanic, 3% Asian/Pacific Islander, and 4% Other. School- and student-level data were compiled in collaboration with each represented state’s department of education, and the resultant dataset included almost 80% of tested public school students in the United States. The study included data of students enrolled in charter schools and traditional public schools from 2006–2007 through 2010–2011.

In order to compare the academic performance of stu- dents enrolled in charter schools with those enrolled in traditional public schools, this study utilized a virtual con- trol record analysis technique, which matched each charter school enrollee with a virtual comparison composite

enrollee of the traditional public school. In so doing, the aggregate academic performance (i.e., growth) of students enrolled in charter schools could be compared with those enrolled in traditional public schools in order to determine the effect of charter school enrollment. Calculated effect sizes were converted into “days of learn- ing” in order to determine, in practical significance, what the overall charter school effect was on student performance.

In sum, enrollment in a charter school was associated with an effect size of 0.005 SD (n/s) growth in math and an effect size of 0.01 SD (p ≤ .01) growth in reading. Con- verted into a “days of learning” difference, students enrolled in a charter school experienced 7 additional days of learning growth in reading as compared to their traditional public school peers. Math growth, on the other hand, was equal between students enrolled in a charter school and their traditional public school peers. That is, charter school enrollees experienced no additional days of learning growth in math.

Urban Charter School Study

Extending analyses carried out in the National Charter School Study, the Urban Charter School Study (CREDO 2015) explored the academic performance outcomes of charter school students vs. traditional public school students within 41 designated urban regions across 22 states. In many cases, the demographic profiles of the charter schools samples from across the 41 urban areas approximated or matched those of their traditional public school counter- parts. However, in many other cases they did not. For example, for 25 of the 41 urban regions included in the study, there was a greater than 5 percentage point difference observed in the percentage of students in poverty within the traditional public school setting vs. the charter school set- ting, meaning within an urban region either the traditional public schools or the charter schools were enrolling a meaningfully higher percentage of students in poverty. Fifteen of the 25 discrepant regions reported a higher per- centage of students in poverty in the traditional public school setting vs. the charter school setting. Across regions, percentages ranged from 45 to 99% of students in poverty in traditional public schools compared to 11–93% of students for charter schools.

For 4 of the 41 regions, there was a greater than 5 per- centage point difference observed in the percentage of students with disabilities, with 3 of those 4 regions reporting a higher percentage of students with disabilities in the tra- ditional public school setting vs. the charter school setting. Across regions, percentages ranged from 1 to 21% of stu- dents with disabilities in traditional public schools com- pared to 2–27% of students for charter schools.

Journal of Child and Family Studies (2018) 27:3132–3140 3135

For 10 of the 41 regions, there was a greater than 5 percentage point difference observed in the percentage of students identified as ELLs, with 8 of those 10 regions reporting a higher percentage of students identified as ELLs in the traditional public school setting vs. the charter school setting. Across regions, percentages ranged from 0–30% of students identified as ELLs in traditional public schools compared to 0–34% of students for charter schools. (Note: see the source publication for complete demographic information across the 41 urban regions).

The Urban Charter School Study drew on administrative data of students enrolled in charter schools and traditional public schools from 2006–2007 through 2011–2012. As with the National Charter School Study, this study utilized the virtual control record analysis technique to match charter school enrollees with traditional public school enrollees for the purposes of comparison and calculated effect sizes were converted into “days of learning” in order to determine what the overall charter school effect was on student performance.

Broadly, in terms of effects, enrollment in a charter school within an urban context was associated with an effect size of 0.06 SD (p ≤ .01) growth in math and an effect size of 0.04 SD (p ≤ .01) growth in reading. Converted into a “days of learning” difference, students enrolled in a charter school experienced 28 additional days of learning growth in reading as compared to their traditional public school peers. Math growth was even more substantial for students enrolled in a charter school as compared to their traditional public school peers, with charter school students experiencing 40 additional days of learning.

Charter School Performance in New York

The study of Charter School Performance in New York (CREDO 2017a) examined charter school performance for schools operating in the State of New York. This study utilized data of students enrolled in charter schools and public schools in the State of New York from 2011–2012 through 2015–2016. Demographically, of the 106,352 stu- dents included in the charter school sample, 76% were students in poverty, 15% were identified as students with disabilities, and 6% were identified as ELLs. The ethnic- racial composition of the charter school sample was 6% White, 58% Black, 32% Hispanic, 2% Asian/Pacific Islan- der, and 1% Native American. Mirroring the approach used in the National Charter School Study and the Urban Charter School Study, this study also utilized the virtual control record analysis technique to match charter school enrollees with traditional public school enrollees for the purposes of comparison. Similarly, calculated effect sizes were con- verted into “days of learning” in order to determine what the overall charter school effect was on student performance.

In terms of effects, enrollment in a charter school within the State of New York was associated with an effect size of 0.11 SD (p ≤ .01) growth in math and an effect size of 0.06 SD (p ≤ .01) growth in reading. Converted into a “days of learning” difference, students enrolled in a charter school experienced 34 additional days of learning growth in reading as compared to their traditional public school peers. Math growth was even more substantial for students enrolled in a charter school as compared to their traditional public school peers, with charter school students experien- cing 63 additional days of learning.

Charter School Performance in New York City

The Charter School Performance in New York City (CREDO 2017b) study examined charter school perfor- mance for schools operating in New York City. This study utilized administrative data of students enrolled in charter and public schools in New York City from 2011–2012 through 2015–2016. Demographically, of the 84,179 stu- dents included in the charter school sample, 76% were students in poverty, 17% were identified as students with disabilities, and 6% were identified ELLs. The ethnic-racial composition of the charter school sample was 4% White, 56% Black, 36% Hispanic, 2% Asian/Pacific Islander, and 1% Native American. Mirroring the approach used in the National Charter School Study and the Urban Charter School Study, the Charter School Performance in New York City study also utilized the virtual control record analysis technique to match charter school enrollees with traditional public school enrollees for the purposes of comparison. Calculated effect sizes were converted into “days of learn- ing” in order to determine what the overall charter school effect was on student performance.

Similar to the effects reported for the State of New York, enrollment in a charter school within New York City was associated with an effect size of 0.11 SD (p ≤ .01) growth in math and an effect size of 0.04 SD (p ≤ .01) growth in reading. Converted into a “days of learning” difference, students enrolled in a charter school experienced 23 addi- tional days of learning growth in reading as compared to their traditional public school peers. Math growth for stu- dents enrolled in a charter school also outpaced that of their traditional public school peers, with charter school students experiencing 63 additional days of learning.

Charter School Performance in Texas

The Charter School Performance in Texas (CREDO 2017c) study represents an evaluation of the performance of charter schools operating in the State of Texas. This study utilized the data of students enrolled in charter and public schools in Texas from 2011–2012 through 2014–2015.

3136 Journal of Child and Family Studies (2018) 27:3132–3140

Demographically, of the 240,191 students included in the charter school sample, 72% were students in poverty, 7% were identified as students with disabilities, and 20% were identified ELLs. The ethnic-racial composition of the charter school sample was 17% White, 21% Black, 60% Hispanic, 6% Asian/Pacific Islander, 1% Native American, and 3% Multi-racial. Mirroring the approach used other studies, this study also used the virtual control record ana- lysis technique to match charter school enrollees with tra- ditional public school enrollees for the purposes of comparison. Finally, calculated effect sizes were converted into “days of learning” in order to determine what the overall charter school effect was on student performance.

In terms of effects, enrollment in a charter school within the State of Texas was associated with an effect size of 0.01 SD (n/s) growth in math and an effect size of 0.03 SD (p ≤ .01) growth in reading. Converted into a “days of learning” difference, students enrolled in a charter school experienced 17 additional days of learning growth in reading as compared to their traditional public school peers. Math growth for students enrolled in a charter school was equal to that of their traditional public schools peers.

Outcomes for Students with Disabilities and English Language Learners

Broadly, students with disabilities underperformed their counterparts without identified disabilities, regardless of enrollment in traditional public schools or charter schools. However, students with disabilities who were enrolled in charter schools generally performed better than their tradi- tional public school peers with disabilities, with fairly consistently reported higher growth in math and in reading (see Table 3). Converted into a “days of learning” differ- ence for charter school enrollees, growth for students with disabilities ranged from 9 to 51 additional days of learning in math and 11 to 23 additional days of learning in reading.

The only exceptions to this trend were in reading (0 days of learning, or no difference), as reported in the National Charter School Study (Cremata et al. 2013), and in math (40 fewer days of learning) and reading (29 fewer days of learning) as reported in the Charter School Performance in Texas (CREDO 2017c). In short, students with disabilities appeared to perform equally well or outperform their peers with disabilities when considered nationally, when con- sidered in urban settings, when in the State of New York, and in New York City, but not in the State of Texas.

Similar to students with disabilities, students identified as ELLs generally underperformed their peers who are not identified as ELLs, regardless of enrollment in traditional public schools or charter schools. Compared to the effects of charter enrollment for students with disabilities, however, students identified as ELLs who are enrolled in charter

schools appeared to have more varied academic growth compared to their ELL peers enrolled in traditional public schools (see Table 4).

In the National Charter School Study (Cremata et al. 2013), ELLs enrolled in charter schools had 36 additional days of learning in both math and reading as compared to their traditional public school counterparts. Contrarily, in the Urban Charter School Study (CREDO 2015), ELLs enrolled in charter schools saw no difference in learning as compared to their traditional public school counterparts (0 days of learning difference in math or reading). In Texas, students identified as ELLs enrolled in charter schools performed worse than their traditional public school counterparts, with 11 fewer days of learning in math and reading. In the State of New York, ELLs enrolled in charter schools had 46 additional days of learning in math and 11 additional days of learning in reading. Within New York City, a similar effect was observed in math, with an additional 40 days of learning for ELLs enrolled within charter schools. However, no differences were observed for ELLs enrolled in charter schools in reading.

Discussion

This study focused on the review of recent large-scale studies carried out by CREDO, as these studies represent

Table 3 Difference in “days of learning” for students with disabilities enrolled in charter schools

Growth in TPSa

Growth in charter schoola

“Days of learning” difference for charter school students

National charter school study (2013)

Math n/a 0.02* +14

Reading n/a 0.01 0

Urban charter school study (2015)

Math n/a 0.01* +9

Reading n/a 0.02* +13

Charter school performance in New York (2017)

Math −0.18* −0.09* +51

Reading −0.15* −0.11* +23

Charter school performance in New York City (2017)

Math −0.19* −0.11* +46

Reading −0.18* −0.16* +11

Charter school performance in Texas (2017)

Math −0.08* −0.15* −40

Reading −0.19* −0.24* −29

n/a indicates data are not available given the particulars reported in the source study

TPS traditional public school aGrowth reported in standard deviations

*p ≤ .01

Journal of Child and Family Studies (2018) 27:3132–3140 3137

best data currently available to assess the impact of charter school enrollment on student academic performance, and because these studies include data specifically allowing for the exploration of charter school enrollment effects on students with disabilities and ELLs. The CREDO studies, though not without detractors (e.g., Hoxby 2009; Maul 2015), were carried out as the result of extensive colla- boration and the data collection efforts made by CREDO and manifold individual state departments of education. In a sense, this represents a meaningful portion of what we know about charter school enrollment effects at this point in time.

As with other reviews of the effects of charter school enrollment on student performance, this review, which focused on recent large-scale studies, reveals mixed effects of charter school enrollment overall. Across large-scale studies conducted in the past 5 years, there was a broad range of enrollment effects for the general population, with varied results by locale and subject area. In reading, gains associated with charter school enrollment ranged from approximately 7 additional days of learning nationally (Cremata et al. 2013) to 34 additional days of learning in New York State (CREDO 2017a). In math, however, no gains of learning were observed nationally (Cremata et al.

2013) or in Texas (CREDO 2017c). Yet, somewhat sub- stantial gains were seen in urban contexts, in New York State, and in New York City. In these areas, gains in math performance ranged from 40 additional days of learning in urban contexts (CREDO 2015) to 63 additional days of learning in New York State (CREDO 2017a) and New York City (CREDO 2017b). In brief, taking the results of recent large-scale studies into account, there is evidence that stu- dents in urban charter schools, charter schools in New York State, and charter schools in New York City tend to benefit from charter school enrollment. In Texas and nationally, however, the effects of charter school enrollment appear to be more modest or null.

For students with disabilities, those who were enrolled in charter schools generally performed better than their tradi- tional public school peers with disabilities, with fairly consistent higher growth reported in math and reading. With the exception of enrollment in a charter school in Texas (CREDO 2017c), students with disabilities who enrolled in charter schools performed as well or better than their counterparts in traditional public schools, in both math and reading. Broadly, with these data in view, this suggests that charter schools are not de facto underserving students with disabilities, in whole or across the board. Instead, charter schools appear to be meeting the learning needs of students with disabilities at least as well as their traditional public school counterparts.

On the other hand, students identified as ELLs who are enrolled in charter schools appeared to have more varied academic growth compared to their ELL peers enrolled in traditional public schools. Early indicators, as drawn from the National Charter School Study (Cremata et al. 2013) indicated that charter schools might be supporting ELLs better than their traditional district counterparts, but more recent studies—including the study focused on urban charter schools (CREDO 2015) as well as those carried out in New York State (CREDO 2017a), New York City (CREDO 2017b), and Texas (CREDO 2017c)—reveal more mixed findings. There was no effect of charter school enrollment for ELLs in urban settings (CREDO 2015). In Texas (CREDO 2017c), ELLs in charter schools fared slightly worse than their counterparts in traditional public schools. ELLs enrolled in charter schools in New York State and New York City performed better than their counterparts in math, while effects were modest or null in reading. In contrast to supporting the needs of students with disabilities, charter schools appear to be less adept at con- sistently ensuring the academic growth of their ELL stu- dents. What’s more, in some cases ELLs appear to be better served in traditional public schools (e.g., CREDO 2017a).

Given the paucity of publicly available data about the performance of students enrolled in charter schools—par- ticularly students with disabilities and ELLs—the results of

Table 4 Difference in “days of learning” for identified English language learners enrolled in charter schools

Growth in TPSa

Growth in charter schoola

“Days of learning” difference for students in charter school

National charter school study (2013)

Math n/a 0.05* +36

Reading n/a 0.05* +36

Urban charter school study (2015)

Math n/a 0.04 0

Reading n/a 0.07 0

Charter school performance in New York (2017)

Math −0.07* 0.01 +46

Reading −0.08* −0.06* +11

Charter school performance in New York City (2017)

Math −0.07* 0.00 +40

Reading −0.09* −0.10* 0b

Charter school performance in Texas (2017)

Math −0.06* −0.08* −11

Reading −0.12* −0.14* −11

n/a indicates data are not available given the particulars reported in the source study

TPS= traditional public school aGrowth reported in standard deviations bThe reported difference between growth in TPS and Charter School is not statistically significant

*p ≤ .01

3138 Journal of Child and Family Studies (2018) 27:3132–3140

this review bespeak the need for more extensive national data to be collected, compiled, and made available, in order to facilitate more comprehensive examination of the effects of charter school enrollment. It is important that such data become available for researchers, policy makers, and prac- titioners alike. We hope that regardless of the setting, educators will seek to use effective teaching practices, with evidence to support their utility with all students, in parti- cular those with disabilities and ELLs.

The effort to broaden data collection and availability is especially important at the present moment, given the pro- choice, pro-voucher, privatization policies of the current administration and its Secretary of Education. The proposed Department of Education 2019 President’s Budget includes $500 million to expand and otherwise support public charter schools—an increase of more than 45% from the 2017 appropriation and of 25% from the 2018 appropriation (U.S. Department of Education 2018). Despite some positive trends in the performance of students in charter schools— including students with disabilities yet, to a lesser extent, ELLs—more comprehensive data are needed in order to assess the effects and effectiveness of charter schools. Indeed, such data are needed so that we can further explore for whom and under what conditions charter school enrollment may have positive effects on student perfor- mance and growth.

Given the growing number of public school students and dollars participating in charter schools, future research in this area is warranted. As mentioned, scholars have raised concerns about the analysis used by CREDO, the largest source of information currently available. For example, Maul (2015) raised several concerns about the conclusions drawn from the analysis, questioning the use of the match pair identification process rather than propensity score matching, a practice more commonly used in educational research. Similarly, Hoxby (2009) echoes Maul’s concerns that the matching used was unclear, with researchers failing to report what variables were used to match students in charter schools to their traditional public school controls. Maul also notes that the used of days of learning as a metric or analysis is not sufficiently justified. Related to outcomes, Maul notes that the studies’ labeling their findings as sub- stantial lacks creditability, as effect sizes reported only account for <0.1% of the variance in test scores. Similarly, Hoxby notes a statistical error that could have given the charter schools analyses a negative bias towards zero that is greater than the margin of error, thus making the small differences even less meaningful. Finally, Maul notes that external factors, such as factors motivating parents to select charter schools, may play a role in the outcomes of charter school students and may contribute to those marginal dif- ferences. This is a factor that is difficult to measure, but it cannot be overlooked. Future researchers seeking to better

understand the impact of charter schools should consider these significant limitations and build experimental studies that allow for more accurate interpretation.

Conclusion

Given the mixed evidence of effects on academic perfor- mance of charter school enrollment, along with the common perception that charter schools fail to meet the needs of special populations of students—including students with disabilities and ELLs—the explosive growth of charter schools and public funding (including increased federal support) appear to be unwarranted. Also, given the limited research in this area, it is a challenge to recommend specific attributes associated with effective charter schools. There- fore, we caution policy makers to consider available infor- mation about charter school effectiveness and demand that charter schools not only establish and realize ambitious performance targets for students enrolled. Policy makers should expect the same improved performance of students attending traditional public schools as well. We further recommend that the National Center for Education Statistics track performance of students attending charter schools— including the performance of those identified as students with disabilities or ELLs—so researchers can attend to the attributes differentiating effective from non-effective charter schools, and their traditional public school counterparts, regarding key metrics pertaining to participation rates, social and behavioral indicators, and academic performance.

Compliance with ethical standards

Conflict of Interest The authors declare that they have no conflicts of interest.

References

Arsen, D., & Ni, Y. (2012). The effects of charter school competition on school district resource allocation. Educational Administration Quarterly, 48(1), 3–38.

Betts, J. R., & Tang, Y. E. (2011). The effect of charter schools on student achievement: A meta-analysis of the literature. National Charter School Research Project. Seattle, WA: Center on Rein- venting Public Education.

Bifulko, R., Ladd, F. L., & Ross, S. L. (2009). Public school choice and integration evidence from Durham, North Carolina. Social Science Research, 38, 71–85.

Booker, K., Gilpatric, S., Gronberg, T., & Jansen, D. (2005). The effect of charter schools on traditional public schools. Washington, DC: National Center for the Study for the Privatization in Edu- cation, Teacher College, Columbia University. http://www.tc. columbia.edu/ncspe/working-papers/OP104.pdf.

Carr, M., & Ritter, G. (2007). Measuring the competitive effect of charter schools on student achievement in Ohio’s traditional

Journal of Child and Family Studies (2018) 27:3132–3140 3139

public schools (No. 146). New York: National Center for the Study of Privatization in Education, Columbia University.

Clark, M. A., Gleason, P. M., Clark Tuttle, C., & Silverberg, M. K. (2015). Do charter schools improve student achievement? Edu- cational Evaluation and Policy Analysis, 37, 419–436.

Center for Research on Education Outcomes. (2009). Multiple choice: Charter school performance in 16 states. Stanford, CA: CREDO at Stanford University.

Center for Research on Education Outcomes. (2015). Urban charter school study. Stanford, CA: CREDO at Stanford University.

Center for Research on Education Outcomes. (2017a). Charter school performance in New York. Stanford, CA: CREDO at Stanford University.

Center for Research on Education Outcomes. (2017b). Charter school performance in New York City. Stanford, CA: CREDO at Stan- ford University.

Center for Research on Education Outcomes. (2017c). Charter school performance in Texas. Stanford, CA: CREDO at Stanford University.

Cremata, E., Davis, D. H., Dickey, K., Lawyer, K., Negassi, Y., Raymond, M. E., & Woodworth, J. L. (2013). National charter school study. Stanford, CA: CREDO at Stanford University.

Dunn, M., Katsiyannis, A., & Ryan, J. B. (2018). Charter schools and students with disabilities: legal and practice considerations. Intervention in School and Clinic, 53, 252–255.

Garda, R. A. (2012). Disabled students' rights of access to charter schools under the IDEA, Section 504, and ADA. Journal of the National Association of Administrative Law Judiciary, 32, 517–542.

Gleason, P., Clark, M., Tuttle, C., & Dwoyer, E. (2010). The Eva- luation of Charter School Impacts (NCEE 2010-4029). Washington, DC: U.S. Department of Education, Institute of Education Sciences, National Center for Education Evaluation and Regional Assistance. https://ies.ed.gov/ncee/pubs/20104029/.

Hoxby, C. M. (2009). A serious statistical mistake in the CREDO study of charter schools. Stanford, CA. http://credo.stanford.edu/ reports/memo_on_the_credo_study.pdf.

Imberman, S. A. (2011). The effect of charter schools on achievement and behavior of public school students. Journal of Public Eco- nomics, 95, 850–863.

Kristie, J. R., Larsen, E. S., & Hausman, C. (2015). School choice & social stratification: How intra-district transfers shift the racial/ ethnic and economic composition of schools. Social Science Research, 51, 30–50.

Maul, A. (2015). Review of Urban Charter School Study 2015. Boulder, CO: National Education Policy Center. http://nepc. colorado.edu/%20thinktank/review-urban-charter-school.

McFarland, J., Hussar, B., de Brey, C., Snyder, T., Wang, X., Wilk- inson-Flicker, S., & Hinz, S. (2017). The condition of education 2017 (NCES 2017-144). U.S. Department of Education. Washington, DC: National Center for Education Statistics. https:// nces.ed.gov/pubs2017/2017144.pdf.

National Center of Education Statistics. (2018). Public charter school enrollment. https://nces.ed.gov/programs/coe/indicator_cgb.asp.

Ni, Y. (2012). The impact of charter schools on the efficiency of traditional public schools: evidence from Michigan. Economics of Education Review, 28, 571–584.

Rhim, L. M., Gumz, J., & Henderson, K. (2015). Key trends in special education in charter schools: A secondary analysis of the civil rights data collection in 2011–2012. New York, NY: National Center for Special Education in Charter Schools. http://static1. squarespace.com/static/52feb326e4b069fc72abb0c8/t/564109d3e 4b027fb791d1964/1447102931798/crdc_full.pdf Google Scholar.

Rhim, L. M., & Kothari, S. (2018). Key trends in special education in charter schools: asecondary analysis of the Civil Rights Data Collection. https://static1.squarespace.com/static/52feb326e4b06 9fc72abb0c8/t/5a9556768165f588f48664fe/1519736443922/280 272+NCSECS+Full+Report_WEB+%281%29.pdf.

Sass, T. (2006). Charter schools and student achievement in Florida. Education Finance and Policy, 1(1), 91–122.

Thomsen, J. (2016). 50-state comparison: Charter school policies. Denver, CO: Education Commission of the States. Retrieved 27 September 2016, http://www.ecs.org/charter-school-policies/.

United States Department of Education, Office for Civil Rights. (2014). Dear Colleague letter: Charter schools. https://www2.ed. gov/about/offices/list/ocr/letters/colleague-201405-charter.pdf.

United States Department of Education. (2018). Department of Edu- cation Budget Tables. https://www2.ed.gov/about/overview/ budget/tables.html.

United States Government Accountability Office. (2012). Charter schools: Additional federal attention needed to help protect access for students with disabilities (GAO-12-543). http://www.gao. gov/assets/600/591435.pdf.

United States Government Accountability Office. (2017). Private school choice: Federal actions needed to ensure parents are notified about changes in rights for students with disabilities (GAO-18-94). https://www.gao.gov/products/GAO-18-94.

Zimmer, R., Gill, B., Booker, K., Lavertu, S., Sass, T. R., & Witte, J. (2009). Are charter schools making a difference? A study of student outcomes in eight states. Santa Monica, CA: RAND Corporation. https://www.rand.org/pubs/research_briefs/RB9433. html.

3140 Journal of Child and Family Studies (2018) 27:3132–3140

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.