|
 
What the Research Reveals About Charter Schools
CER Report
September 16,2003
THE CENTER FOR EDUCATION REFORM WHAT HAT THE RESEARCH ESEARCH REVEALS EVEALS ABOUT BOUT CHARTER HARTER SCHOOLS CHOOLS September 2003 of the Studies Summary and Analyses The Center for Education Reform 1001 Connecticut Avenue, NW Suite 204 Washington, DC 20036 Phone: (202) 822-9000 Fax: (202) 822-5077 www.edreform.com What the Research Reveals About Charter Schools Introduction Three years ago, the Center for Education Reform (CER) conducted a review of all existing research on charter schools, using studies and reports published from the mid-1990s through the fall of 2000. By the end of that review, CER had discovered a trend: of the 53 studies that met the standard for objectivity, the overwhelming majority – 50 – had determined that charter schools were living up to their mandate to be innovative, accountable, and successful. In the 36 months that have passed since CER first published What the Research Reveals, the number of operating charter schools in the nation has expanded from slightly over 2,000 to nearly 3,000. In that same period, the body of charter school research has also grown, standing now at a total of 98 reports having been issued since 1995 –– 45 more than were in the original report. The third edition of What the Research Reveals About Charter Schools builds on the first. It contains all the report summaries found in the first two editions, but adds overviews of major research published during the past two years. Where it does not differ from the first two is in the trend that runs through it: charters are doing the job they were designed to do, with 88 major reports now showing that charter schools are improving education for America’s kids. Jeanne Allen President WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 2 The Studies 2003 1) Apples to Apples: An Evaluation of Charter Schools Serving General Student Populations (Jay Greene, Greg Forster and Marcus Winters, Manhattan Institute): July 2003 Findings: This study compares charter schools with the traditional public schools most likely to have similar student populations. The authors hope to improve on studies that frequently look only at raw achievement data without accounting for student populations. Nationally, they found that charter schools serving “general student populations” (meaning schools not targeted to specific groups) outperformed the geographically-nearest traditional public schools by about three percentile points in math and two points in reading. In addition, charter students in Texas and Florida significantly outpaced the nearest traditional public schools – by seven to eight percentile points in Texas and six points in Florida. A large percentage of the nation’s charters – those targeting specific groups of students – were excluded from this analysis in order to compare schools with populations similar to traditional public schools. 2) Catching the Wave: Lessons from California’s Charter Schools (Nelson Smith, Progressive Policy Institute): July 2003 Findings: This study analyzes extant research on charter schools in California and finds that charters are performing as desired. The authors find that California’s charters are expanding learning opportunities especially for at-risk students, and are showing faster improvements on the state’s Academic Performance Index (API) than traditional public schools. In addition, they are producing innovations that are being adopted by many traditional school districts, though evidence of a charter competition-induced “ripple effect” appears inconsistent, with some districts changing and others seemingly unaffected. The report ends by calling for an expansion of charters and lowering the restrictions on them. 3) Charter School Operations and Performance: Evidence from California (RAND Corporation): June 2003 Findings: California’s charter schools produce levels of reading and math achievement on par with the state’s conventional public schools, despite employing teachers with fewer credentials, suffering from pronounced facilities shortages, and receiving less money per pupil. In addition, the authors find that charters that started from scratch outperform traditional public schools; conversion charters perform on par with traditional public schools; and charters with a home-based learning component lag behind traditional public schools. This third point is qualified, however, because it is very likely that students schooled at home often have disabilities for which the study could not control. Demographically, charter enrollment is roughly equivalent to traditional schools, with larger African-American, smaller Hispanic, and proportionately equal white populations attending charters. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 3 4) State University Authorized Charter Schools’ Achievement on the 2003 State English Language Arts Examination (Charter Schools Institute, State University of New York): June 2003 Findings: Results of the 2003 English Language Arts (ELA) exam in New York show charter schools authorized by the State University of New York (SUNY) are making significantly greater fourth and eighth grade language arts progress than are traditional New York public schools. Taken together with the other charter schools in New York – those authorized by the Board of Regents and the Department of Education – 14 out of 15 charters showed grater improvement between 2002 and 2003 than New York’s traditional public schools. The report also reveals that the improvements aren’t just taking place at the top, as more charter students exceeded state standards and fewer performed at “Level 1” (serious academic deficiencies). 5) What Parents Think of New York’s Charter Schools (Duncan McCully and Patricia Malin, Zogby International/Manhattan Institute): June 2003 Findings: When asked to grade their schools, 42 percent of New York charter school parents assigned their school an “A” grade overall, compared to only 21 percent who gave the same grade to their child’s previous school (65 percent reported that their child previously attended a public school). 51 percent gave their charter an “A” for its quality of instruction. In answer to a question about what their previous school did better than their charter, the most common reply (33 percent) was “nothing.” 6) Delaware Charter Schools Sixth Annual State Report (Delaware Department of Education): June 2003 Findings: This collection of reports on the state’s eleven charter schools offers no statewide analysis, but provides data for each school. All schools taken together, this report shows that students at Delaware’s charter schools do at least as well as the students in the their host districts, that charter students are consistently making achievement gains, and that charter parents are very satisfied with the education their children are receiving. Because no overall demographic or other state data is provided, no in-depth comparisons were conducted for this report. 7) The Performance of California Charter Schools (Margaret Raymond, CREDO/Hoover Institution): May 2003 Findings: Charter schools at the elementary level outpace traditional elementary schools according to 1999-2001 data based on state Academic Performance Index (API) scores. (2002 data is not used because the tests involved have changed). When charter elementary schools are compared only with elementary schools in districts with charters, their superior rate of growth is even higher. When compared to “local competitor” schools (schools in the same areas) there are no statistically significant growth scores differences, but this is attributed to the fact that California charter schools serve higher proportions of low-performing students. Charter middle schools lagged significantly behind both traditional middle schools and “local competitors,” but there are too few schools in the sample to make it statistically sound. Charter high schools have growth scores nearly twice the rate of conventional high schools or “local competitors.” Finally, the authors found that teacher credentialing has only a slight impact on API scores, while there is a significant correlation between API scores and school size, with smaller being better. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 4 8) North Carolina’s Charter Schools Pass Accountability Test (The League of Charter Schools): April 2003 Findings: Scores on the end-of-grade ABC (North Carolina’s accountability system) tests show that students at the state’s charters are achieving at roughly the same level as their peers in traditional public schools. In reading, traditional public school students scored an 87.6 percent composite while charter students scored 86.7 percent. In math, traditional publics scored 91.3 percent and charters hit 89.1 percent. Included in the charter scores were nine schools that would be identified as “alternative schools” which, were they traditional publics, would not have been included in score tallies. As charters, however, they were included. 9) The State of Charter Schools in Colorado 2001-02 (Colorado Department of Education): March 2003 Findings: The state’s annual report on charter schools suggests that when controlling for race and reduced-price lunch eligibility, Colorado’s charters do slightly better than the state’s traditional public schools in reading, writing, math and science. Colorado’s charters tend to serve fewer minorities and poor kids, but also have a high percentage of schools that specifically target at-risk students. Finally, a finding that especially stands out is that a significantly higher percentage of charters than traditional public schools received accountability ratings of "excellent" (20 percent vs. 7 percent), though a slightly higher percentage were also "unsatisfactory" (7 percent vs. 2 percent). 10) Veteran Charter Schools Outperform Non-Charters on API (Michael Agostini, Charter Schools Development Center): March 2003 Findings: This analysis of base 2002 API scores shows that California charters that have been in existence for five or more years outperformed non-charter public schools and younger charters. Charters in existence for five or more years had an average API score of 708, versus 689 for all publics and 667 for all active charter schools. (The goal for a school is a score of 800.) The report does not measure longitudinal growth (how well individual students progress over time), but when compared to other California studies this one reinforces the notion that the longer a child is in a charter school, the more progress he or she makes. 11) Findings from the 2002 Survey of Parents With Children in Arizona Charter Schools: How Parents Grade Their Charter Schools (Lewis Solmon, Human Resources Policy Corporation): March 2003 Findings: In a survey administered during the 2001-2002 academic year, 66.9 percent of Arizona charter parents assigned their charter schools an “A” or “A+” rating, up from 64 percent a year earlier. Moreover, the percentage of parents giving charters such a grade far exceeds the 38 percent of Arizona’s traditional public school parents who gave their schools similar grades in May 2000 (the last time the state conducted such a survey). In addition to the data, researchers found a strong correlation between high grades and parental interest in a school’s academics. Academics were also found to be much more influential in grading than were items such as after school programs, technology, or building quality. Finally, arguably the most important finding in the report is that “parents are clearly capable of soundly assessing the quality of a given school,” indicating that parents tend to make informed decisions about where to have their children educated. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 5 12) Illinois Charter School Annual Report (Illinois State Board of Education): January 2003 Findings: As in the two previous years, in 2003 the Illinois State Board of Education found that a majority of the state’s charters are doing better than similar schools in their home district, but not all are exceeding district scores. The greatest change occurred in Chicago, where seven of the city’s charter schools recorded better average scores on the Illinois Scholastic Achievement Test (ISAT) than did other district schools, while only three charters scored below the district average. However, although children in charters come in performing at lower levels of achievement on the Prairie State Achievement Exam (PSAE), because the state’s test does not yet measure growth from year to year, its results cannot be used to make absolute comparisons. Illinois’ charters and traditional public schools in their districts serve roughly equal minority and low-income populations. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 6 2002 13) Annual Charter School Report 2001-02 School Year (Office of Student Learning and Achievement, Georgia Department of Education): October 2002 Findings: Georgia’s charter schools serve slightly lower percentages of minority and socio-economically disadvantaged students than traditional public schools, but slightly higher percentages of special education and English for Speakers of Other Languages (ESOL) students. Achievement data shows charters doing as well or better than traditional public schools academically. On almost all standardized tests charters recorded gains, exceeding state scores on the Criterion-Referenced Competency Tests (CRCT) for 4th, 6th, and 8th grade Reading and Mathematics, as well as on the Middle Grades Writing Assessment for grade 8. In addition, in all subject areas of the Georgia High School Graduation Tests (GHSGT), charters had greater percentages of students passing on their first try than did traditional public schools. 14) Strengthening Pennsylvania’s Charter School Reform: Findings From the Statewide Evaluation and Discussion of Relevant Policy Issues (Gary Miron, Christopher Nelson and John Risley, The Evaluation Center, Western Michigan University): October 2002 Findings: Pennsylvania’s charters are making noticeably greater academic achievement gains than similar traditional public schools. They do this while serving disproportionately higher percentages of traditionally under-served students than their host districts, including minorities (63 vs. 54 percent) and free lunch qualifiers (56 vs. 53 percent). However, the authors also point out that charters serve an appreciably smaller percentage of students with disabilities. The teacher attrition rate is also significantly higher in charters than traditional Pennsylvania public schools, though charter teachers are much more likely to report having autonomy in the classroom than are teachers in traditional public schools. Finally, the authors saw little evidence that traditional public schools are changing in response to competition from charters. This report, however, has little scientific data to support that claim. 15) Utah Charter School Evaluation (Center for the School of the Future): October 2002 Findings: This state-mandated study of Utah’s charter schools finds that the eight schools in existence long enough to be measured (three years) serve larger percentages of minority and at-risk students than Utah’s traditional public schools. It also finds no evidence of “creaming,” pointing out that half of the charters examined serve predominantly at-risk and disabled students, and the four remaining schools show no signs of taking the best students. The authors also indicate that there is no indication of a ripple effect, largely because only 1/800th of the state’s students attended charter schools. Academic achievement between charters and traditional schools is found to be comparable, with charter elementary schools generally outperforming traditional elementary schools, but secondary charters being surpassed by traditional high schools. This is reported with a strong qualification that no information about academic improvement was available. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 7 16) Evaluation of Connecticut Charter Schools and the Charter School Initiative (Gary Miron and Jerry Horn, The Evaluation Center, Western Michigan University): September 2002 Findings: Charter school students in Connecticut are typically drawn from traditional public schools (89 percent), have disproportionately large minority populations (41 percent black, 26 percent Hispanic), and travel an average distance to school that is more than two times farther than their nearest traditional public school (4.78 compared to 2.18 miles). In addition, though charter students’ raw scores on such tests as the Connecticut Mastery Test (CMT) and the Connecticut Academic Performance Test (CAPT) are lower than their host districts, their rate of improvement is much greater. 17) Michigan Public Charter Schools See MEAP Scores Rise Faster than Regular Public Schools (Kirk Johnson, Mackinac Center for Public Policy): September 2002 Findings: Between 2000 and 2001 Michigan charter school students improved their scores on the Michigan Education Assessment program exams at a considerably faster pace than did their public school peers. In 4th grade reading, charters saw a 43 percent increase in the number of students passing. Traditional public schools saw only 10.1 percent improve. In 4th grade math charters outpaced other public schools 28.3 to 0.5 percent. Finally, 7th grade reading scores at charters outpaced traditional publics by 55 to 28.8 percent. 18) South Carolina Charter Schools: Five-Year Evaluation Report (Weaver B. Rogers, Ph.D. and Associates): July 2002 Findings: Because of the small number of charter schools and students in South Carolina, this report offers no definitive conclusions about charters there, including achievement data. It reveals, though, that more than 40 percent of charter students are minorities and over a third receive free lunches. In addition, 94 percent of charter school teachers are certified, and charters tend to have consistently smaller classes in subjects like math and English. 19) Texas Open-Enrollment Charter Schools Fifth-Year Evaluation (Texas Association of School Boards): July 2002 Findings: Charter schools in Texas serve significantly larger African-American (41 percent vs. 14 percent) populations than traditional public schools and more than a third serve populations that are composed of more than 75 percent at-risk students. Reflective of this service to historically underserved populations, Texas charters produce significantly lower raw scores on the Texas Assessment of Academic Skills (TAAS) than traditional public schools, and a much larger percentage of charters are identified as “low performing.” However, when individual student achievement data is analyzed, the researchers found TAAS passing rates for charter students increasing quickly, and students who stay in charters longer have greater gains than students newly arrived at charter schools. 20) Texas Charter Schools: Do They Measure Up? (Matt Moore, National Center for Policy Analysis): June 2002 Findings: Though charters in Texas struggle in their first couple of years, for students who remain in them for at least two consecutive years achievement improves at a much faster rate than at other public schools. For instance, for such students in open enrollment charters, Texas Learning Index (TLI) scores increased 1.6 points in reading and 2.8 points in math, versus 1.4 and 1.5 points, respectively, in traditional public schools, between 1999 and 2000. In addition, this study reports that much of Texas charters’ initial score lags can be explained by the typically under-served populations with which they WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 8 work; 39.7 percent of charter students are African American, though they comprise only 14.4 percent of traditional public school students. In general, charters also cater disproportionately to at-risk kids. 21) Annual Report to the Governor, The Temporary President of the Senate, and the Speaker of the Assembly on the Status of Charter Schools in New York State in the 2000-01 School Year (The New York State Department of Education): May 2002 Findings: Charter schools in New York serve populations that are 68 percent African American and 18 percent Hispanic. Valid achievement data is not available in this report because no longitudinal scores were recorded to demonstrate progress, and for several of the charters the scores represented first year, baseline data. 22) State of Charter Schools 2000-01 (Colorado Department of Education) March 2002 Findings: In general, with the exception of the 9th grade Reading assessment, Colorado’s charter schools surpassed state averages on the 2000-01 Colorado State Assessment Program (CSAP). Charters also outscored economically and ethnically “matched” public schools, indicating that charters did not surpass state average scores simply because they have, in some cases, fewer minority students and students eligible for free and reduced price lunches. The report also notes that since 1997, the percentage of charter students eligible for free or reduced lunches has increased by 40 percent. 23) Charter Schools in New York: A New Choice in Public Education (Charter Schools Institute: State University of New York): March 2002 Findings: This report finds that SUNY-sponsored charter schools initially start out with students who are much worse academically than student populations in the traditional public school districts from which they come. For instance, the average charter school enrolled a population scoring only a 95.3 across all tests on the state’s School Performance Index (SPI). The average sending district scored an SPI of 114.8 on the same tests for the same year. (The major cities discussed are New York City, Rochester, Albany and Buffalo.) In addition, the authors identify a charter “ripple effect,” citing the implementation of new education programs in districts hosting large numbers of charter schools; advice issued by the New York State School Boards Association urging its members to view themselves as competing for students; and the conversion of five New York City district schools into charters. 24) California Charter Schools Serving Low SES Students: An Analysis of the Academic Performance Index (Simeon Slovacek, Antony Kunnan and Hae-Jin Kim): March 2002 Findings: This study compares charter and non-charter public schools serving high numbers of children from poor socio-economic backgrounds, and concludes that poor children in charter schools perform better than children in non-charter public schools. In schools with at least 50 percent of their students in the Free and Reduced-price Lunch Program, charter schools’ scores on California’s Academic Performance Index (API) improved by 22 percent, versus 19 percent for non-charter public schools. In schools with 75 percent Free and Reduced Lunch participation, charter schools’ scores improved by 28 percent, four percentage points better than non-charter schools. The study also discounts as factors in student achievement such items as school day length, type of instruction, and length of time a school has been open. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 9 25) Making the Grade: Comparing DC Charter Schools to Other DC Public Schools (Mark Schneider and Jack Buckley, FOCUS): March 2002 Findings: Using a survey of parents and econometric statistical analysis, this study compares Washington, DC charter schools to traditional DC public schools. The authors find that DC’s charters have teachers that are better regarded by parents; have cleaner, more stimulating school buildings; and are given higher overall grades by parents, even when controlling for non-school factors that often bias opinion surveys. However, there is no achievement data available in this report. 26) Illinois Charter School Annual Report (Illinois State Board of Education): January 2002 Findings: As in the previous year, the 2002 Illinois State Board of Education found that a majority of the state’s charters are doing better than similar schools in their home district, but not all are exceeding district scores. Data also show that Illinois’ charters and traditional public schools in their districts serve roughly equal minority and low-income populations. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 10 2001 27) 2001 New Jersey Charter Schools Evaluation (KPMG): October 2001 Findings: This study shows that charter students in New Jersey are making faster progress than students in their districts of residence on assessments such as the math section of the Elementary School Proficiency Assessment (ESPA) and the language arts portion of the Grade Eight Proficiency Assessment (GEPA). Also highlighted are charters’ disproportionately large minority population, smaller average class sizes (19 vs. 21 students), greater instructional time (on average, 40 minute longer days and years that are 25 days longer than districts of residence) and better faculty attendance rates. Finally, it focuses on the "extremely high" parental and student satisfaction rates for charter schools. 28) U.S. Department of Education: Challenge and Opportunity: The Impact of Charter Schools on School Districts (John Ericson and Debra Silverman, RPP International): June 2001 Findings: Charters do affect how school districts behave, confirming the “ripple effect” theory of charters. More than half of the school districts became more customer service oriented, increasing the frequency of their communications with parents as a result of competition from charter schools. As a result of charter school growth, most districts implemented new educational programs, made changes in educational structures in district schools, or created new schools with programs that were similar to the local charter schools. After examining 49 districts with the highest critical mass of charters in five states, the study concludes that districts do make positive changes in their educational services and district operations as a result of charter schools. 29) Navigating Newly Chartered Waters: An Analysis of Texas Charter School Performance (Timothy Gronberg and Dennis Jansen, Texas Public Policy Foundation): May 2001 Findings: Continuing students in charter schools have greater improvement in their Texas Assessment of Academic Skills test scores in both reading and math than do continuing students in traditional public schools. At-risk charters are performing well and have a positive value-added effect relative to traditional public schools. Students moving into a charter, however, often exhibit a first-year drop in Texas Assessment of Academic Skills (TAAS) test scores. Charter schools serve larger percentages of African-American students and at-risk students than do traditional public schools, and the study also notes that charter schools are cost efficient and achieve a given level of student performance at a lower expenditure per student than a comparable traditional public school district. 30) A Study of Charter School Accountability (Paul Hill, Robin Lake, and Mary Beth Celio, Center on Reinventing Public Education, Daniel J. Evans School of Public Affairs, University of Washington): April 2001 Findings: Although charter schools experience start-up periods of confusion, operators quickly learn to maintain the confidence of authorizers, families, teachers and donors in order to focus on providing quality instruction. The study concludes that conventional school district offices concentrate on detailed compliance-oriented oversight, rather than performance-based oversight. Agencies unaccustomed to working with individual schools sometimes deal with charters “by the book” more severely than they treat conventional public schools. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 11 31) The State of Charter Schools in Colorado, 1999-2000: The Characteristics, Status and Performance Record of Colorado Charter Schools (Colorado Department of Education): March 2001 Findings: More than 90 percent of the charter schools use more than one assessment tool to measure different dimensions of student learning. The average score of the charter schools on the Colorado Student Assessment Program exceeded the state average by a significant margin, and also exceeded the scores in “matched” public schools. Charters’ student population is nearly as racially diverse as the state’s public school enrollment. Most charter schools offer a program that serves students continuously from elementary through middle school, from middle school through secondary school, or throughout K-12. Forty-two percent of charters enroll 200 students or less, and the schools offer a diverse array of education programs and instructional approaches. 32) Does Charter School Attendance Improve Test Scores? The Arizona Results (Lewis Solmon, David Garcia, Kern Paark, The Goldwater Institute): March 2001 Findings: This study concludes that the longer a student attends a charter school, the greater the academic gains, with no similar result in district school students. Charter students generally are making greater gains in reading and about the same gains in math as traditional public schools. The research also suggests that mobility within the charter sector is usually better than stability in a district school. Also noted is that charters to do not “cream” the best students, and that charter school students starting with lower average test scores in math and reading were more likely to be classified as special education, were less likely to be gifted, and were more likely to be white and to speak English. The study looked at 40,000 to 60,000 students who spent one to three years in charter schools or who have been enrolled in comparable district schools. The analysis controls for race, gender, grade level, number of days absent, Limited English Proficiency (LEP), years in the district, and gifted/special education identification. 33) Illinois Charter School Annual Report (Illinois State Board of Education): January 2001 Findings: Academically, some charters are doing better than similar grades in their home school district, but not all are exceeding district scores. However, parental satisfaction is high. The report concludes that charter schools have been successful for the students they serve and that, although they have been in existence “for only a brief period of time, they are serving as seeds of change in their local communities.” Data also show that Illinois’ charters and traditional public schools in their districts serve roughly equal minority and low-income populations. WHAT THE RESEARCH REVEALS ABOUT CHARTER SCHOOLS The Center for Education Reform September 2003 12 2000 34) Annual Charter School Report for Georgia, 1999-2000 School Year (Georgia Department of Education): October 2000 Findings: Academic performance at the 20 charter schools operating for at least three years shows slightly better performance between 1997 and 2000. Eight schools showed stable or increased achievement on all tes
|
