Fall 22

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=hpje20

Peabody Journal of Education

ISSN: 0161-956X (Print) 1532-7930 (Online) Journal homepage: https://www.tandfonline.com/loi/hpje20

Charter School Enrollments in Context: An Exploration of Organization and Policy Influences

Natalie Lacireno-Paquet

To cite this article: Natalie Lacireno-Paquet (2006) Charter School Enrollments in Context: An Exploration of Organization and Policy Influences, Peabody Journal of Education, 81:1, 79-102, DOI: 10.1207/S15327930pje8101_5

To link to this article: https://doi.org/10.1207/S15327930pje8101_5

Published online: 18 Nov 2009.

Submit your article to this journal

Article views: 120

View related articles

Citing articles: 1 View citing articles

Charter School Enrollments in Context: An Exploration of Organization and Policy Influences

Natalie Lacireno-Paquet Leadership in Education University of Massachusetts Boston

Using the 1999–2000 Schools and Staffing Survey Public Charter School Questionnaire (U.S. Department of Education, 2000), this article provides an in-depth description of charter schools in the United States and presents multivariate analyses of the enrollment of low-income and minority students in the population of charter schools, taking into account both school organi- zational and state policy characteristics. The results suggest that manage- ment by for-profit educational management organizations (EMOs) can have both positive and negative effects on low-income and minority enrollment depending on EMO type and urban location. Notable were the strong and significant results on the importance of state policy characteristics in terms of shaping charter school enrollments of low-income and minority students.

79

PEABODY JOURNAL OF EDUCATION, 81(1), 79–102 Copyright © 2006, Lawrence Erlbaum Associates, Inc.

I thank Jeff Henig, Joe Cordes, and Sheila Kirby for feedback on earlier versions of the models presented in this article. Thanks also to Ken Wong whose thoughtful comments strengthened this article. This research was supported by a dissertation grant from the Ameri- can Educational Research Association, which receives funds for its “AERA Grants Program” from the U.S. Department of Education’s National Center for Education Statistics, and the In- stitute for Education Sciences, and the National Science Foundation under NSF Grant REC-9980573. Opinions are mine and do not necessarily reflect those of the granting agencies. Any errors are my sole responsibility.

Correspondence should be sent to Natalie Lacireno-Paquet, University of Massachusetts, Leadership in Education, 100 Morrissey Boulevard., Boston, MA 02125. E-mail: Natalie. [email protected]

Researchers have not yet extensively analyzed the 1999–2000 Schools and Staffing Survey Public Charter School Questionnaire data, the first population survey of public charter schools in the United States available to researchers (U.S. Department of Education, 2000). In this article, I use these data to examine charter schools enrollments, paying close attention to the internal organizational characteristics of charter schools and to the policy context in which they operate. This article presents an in-depth de- scription of charter schools in the United States, and it presents multi- variate analyses of the enrollment of low-income and minority students in the population of charter schools. The goal of these analyses is to under- stand the internal school organizational characteristics that lead to sorting at the school level. The multivariate analyses also examine the extent to which the characteristics of state charter school policies influence charter schools’ enrollment of low-income and minority students. In doing so, this article addresses policy concerns about the implications and outcomes of charter schools policy.

This article is divided into four main sections. The first section uses the literature to set the context for the analysis. The second section presents de- scriptive statistics of the charter schools included in these analyses, and the third presents multivariate models of charter school enrollment. The final section explores the implications of the analyses presented.

Literature

Much has been written about what charter schools are, how diverse they are, and the theory that underlies the idea (Bulkley & Wohlstetter, 2004; Gill, Timpane, Ross, & Brewer 2001; Nathan, 1996; Vergari, 2002). Here I focus specifically on literature that addresses three key concepts of this article: (a) equity and segregation in charter schools, (b) the involve- ment of for-profit educational management organizations (EMOs) in char- ter schools, and (c) the importance of understanding the state policy con- text in which charter schools operate.

The proponents of choice suggest that school choice, including charter schools, might improve the integration of schools by race and class by overriding the mechanisms by which most students are assigned to schools (Gill et al., 2001; Peterson & Greene, 1998; Viteritti, 1999). Oppo- nents fear, however, that the increased power schools have to choose stu- dents, along with financial and academic pressures, will lead schools to “cream skim” students based on race, class, and achievement. To date, the evidence on the sorting behaviors or outcomes of charter schools is mixed (Gill et al., 2001; Miron & Nelson, 2002). Gill et al. found that for the

80

N. Lacireno-Paquet

1998–1999 school year, in 11 of 27 states with charter schools, the charter school population was significantly lower income than that of traditional schools. Also, in many states the charter school population had more mi- nority students than public schools, but some states had charter school populations that were overrepresented by White students. The evidence seems to point not to cream skimming but to the fact that many charter schools are highly segregated by race and income (Frankenberg & Lee, 2003). Much of the research has also focused on how charter schools com- pare to traditional schools in terms of segregation (Buckley, Schneider, & Shang, 2005; Frankenberg & Lee, 2003; Gill et al., 2001; RPP International, 2000) and does not look at patterns of sorting of students by race and class within the population of charter schools.

A new and growing area of research is the categorization of charter school types. This work suggests that it can be misleading to group all charter schools together because there are important differences in charter schools in terms of their orientation to the market, their missions, and their affiliation with various groups such as EMOs (Brown, Henig, Lacireno- Paquet, & Holyoke, 2004; Henig, Holyoke, Brown, & Lacireno-Paquet, 2005; Henig, Lacireno-Paquet, Holyoke, & Brown, in press; Wohlstetter, Smith, Hentschke, & Malloy, 2004). Emerging research has shown that the categorizations of charter schools can help us understand what is going on at the school level (Henig et al., 2005; Lacireno-Paquet, 2004; Lacireno- Paquet, Holyoke, Henig, & Moser, 2002; Miron & Nelson, 2002). Indeed, Henig et al. (2005) suggested that charter school type is related to impor- tant organizational characteristics, such as grade configurations, school size, and so on. One of the main distinctions they made is between charter schools that are profit and market oriented—namely those associated with EMOs—and those that are more oriented toward social missions—those associated with social service organizations. Schools that are more market oriented have been found to be larger than more independent schools and less likely to delegate decision making in key areas to the school level (Brown et al., 2004). Others have also found that identifying charter school as associated with EMOs useful (Lacireno-Paquet, 2004; Lacireno-Paquet et al., 2002; Miron & Nelson, 2002). And although these differences in char- ter school type are beginning to be explored, some have voiced concern that for schools associated with EMOs the profit motive might become the primary focus, with quality and equity lost by the wayside (Miron & Nel- son, 2002; Molnar, Wilson, & Allen, 2004; Nelson, Muir, & Drown, 2000; Plank, Arsen, & Sykes, 2000).

The question that arises is this: What is known about charter schools as- sociated with EMOs? EMOs are essentially contractors of public schools, in that they manage schools under contract with some government agency or

81

Charter School Enrollments

with a charter school board of directors, although in many places EMOs initiate charter schools. Although not all states allow charter schools to as- sociate with management firms, some allow EMOs to hold a charter di- rectly. The General Accounting Office (2002) estimated that there are about 20 companies that manage at least 3 schools in more than one state. A re- port tracking the growth of EMOs found that for the 2004–2005 school year there are 458 EMO operated charter schools in the United States (Molnar, Garcia, Sullivan, McEvoy, & Joanou, 2005). This report also noted that EMOs focus on primary or elementary education and that, although there are both large and small EMO firms, the large firms (those that manage 10 or more schools) seem to be dominating the market. EMO-operated char- ter schools are predominantly located in urban areas (General Accounting Office, 2002; Wells, 2002), which reflects the pattern for charter schools overall.

Charter schools can only be understood in terms of the legal and regula- tory framework in which they operate (Buckley & Kuscova, 2003; Fuller, Gawlik, Gonzales, & Park, 2004; Witte, Shober, & Manna, 2003) because states play an important role in shaping the policy and market environ- ment in which charter schools operate. Witte et al. and Buckley and Kuscova found evidence that specific characteristics of state policies influ- ence the size and strength of charter schools in a state. Charter schools poli- cies vary across the states in several dimensions, including authorization and legal status, funding, transportation, and facilities. Through their deci- sions in these areas, such as whether to have only local school districts grant charters and so on, states set the broad contours of the charter school market and regulatory environment. These likely affect the extent of sort- ing that takes place among charter schools.

Characteristics of Charter Schools in Sample

Data on the charter schools in this study come from the 1999–2000 Schools and Staffing Survey Public Charter School Questionnaire. To be in- cluded, charter schools had to be in states with at least 10 charter schools. Table 1 lists the states and number of charter schools included. The charter schools in the 19 study states are diverse in a number of ways. In this sec- tion, I provide details on the internal organizational and other characteris- tics of the schools. In terms of organizational features of the schools, I look at the association with EMOs, grade levels served, size, curricular focus, and special programs offered. With regard to their student populations, I examine the minority enrollment and the percentage of students who are eligible for free and reduced price lunch (FARL).

82

N. Lacireno-Paquet

Organizational Characteristics of Charter Schools1

Of the 980 schools in the weighted sample, 64 schools (6.6%) can be characterized as managed by a large EMO.2 Another 45 schools (4.6%) re- ported management by other, smaller EMOs.3,4 Six of the study states do not have any EMO-managed charter schools. Most charter schools, includ- ing 67% of the large EMOs and 44% of the small EMOs, are located in large

83

Charter School Enrollments

Table 1

Number of Charter Schools, by State

State

No. of Charter Schools (% Charter Schools Sample)

Unweighted

No. of Charter Schools (% Charter Schools in Population) Weighted

Alaska 10 (1.19) 12 (1.19) Arizona 173 (20.40) 206 (21.02) California 112 (13.28) 132 (13.46) Colorado 52 (6.17) 58 (5.93) Connecticut 13 (1.54) 15 (1.56) District of Columbia 16 (1.90) 18 (1.8) Florida 59 (7.00) 64 (6.54) Georgia 22 (2.61) 25 (2.55) Illinois 10 (1.19) 11 (1.17) Kansas 12 (1.42) 14 (1.4) Massachusetts 27 (3.20) 33 (3.37) Michigan 119 (14.12) 135 (13.79) Minnesota 29 (3.44) 33 (3.37) North Carolina 42 (4.98) 49 (5.01) New Jersey 24 (2.85) 29 (2.96) Ohio 10 (1.19) 11 (1.13) Pennsylvania 22 (2.61) 28 (2.83) Texas 71 (8.42) 81 (8.28) Wisconsin 21 (2.49) 26 (2.63) Total 843 (100) 980 (100)

1Table 2 presents descriptive statistics on these and other characteristics. 2A large EMO, for the purpose of this study, includes the following for-profit management

firms in existence at the time of the survey: Advantage, Beacon, Edison, Leona, Mosaica, Na- tional Heritage Academies, and SABIS.

3Small EMOs are firms such as Smart Schools, Inc.; Designs for Learning; and other similar private, for-profit firms that operate fewer than 10 schools, usually in only one state.

4Another eight schools (unweighted) responded that they were managed by an organiza- tion other than a school district that also managed other schools but failed to list the name of the organization. These schools are not counted as being managed by an EMO because they could be managed by another type of organization.

or midsize central cities compared to 53% of all non-EMO or independent charter schools. Approximately 33% of large-EMO and 36% of small- EMO-operated charter schools are located in the urban fringe of a large or midsize city, compared to an almost equivalent percentage of non-EMO- operated schools (32.4%). Only 8% of the small-EMO-operated schools were located in a rural area or small unincorporated town, compared to 15% of non-EMO charter schools. No large-EMO charter schools were lo- cated in rural areas or unincorporated towns.

84

N. Lacireno-Paquet

Table 2

Summary Statistics

N Large EMO

Operated Small EMO

Operated Independently

Operated

No. of schools 980 64 45 871 Center of large or midsize city 520 (53.1%) 43 (66.9%) 20 (44.1%) 457 (52.6%) Urban fringe of large or

midsize city 319 (32.6%) 21 (33.0%) 16 (35.7%) 281.7 (32.4%)

Rural area or small, unincorporated town

141 (14.3%) 0 (0%) 9 (8.2%) 130 (15%)

No. granted by school district 369 (37.7%) 10 (16.1%) 8 (17.7%) 350 (40.3%) Newly created schools 725 (74.0%) 55 (84.4%) 40 (89.4%) 629 (72.4%) Years of operation (M, SE) 2.59 (0.02) 1.99 (0.05) 2.55 (0.08) 2.64 (0.02) Grade level: Elementary only 489 (50.0%) 49 (76.2%) 16 (35.5%) 424 (48.8%) School size (M, SE) 265.8 (3.59) 518.6 (12.92) 264.1 (32.49) 247.0 (3.40) No. with magnets for racial/

ethnic balance 66.8 (6.8%) 3.4a (6.9%) —a 63.3 (7.3%)

No. regular elementary/ secondary

525 (53.6%) 53 (82.3%) 24 (54.4%) 447 (51.5%)

No. with parental service requirements

388 (40.9%) 23 (34.8%) 10 (22.4%) 356 (40.9%)

Any admissions requirement 255 (26.1%) 11 (16.3%) 8 (18.2%) 237 (27.2%) Admissions test 33 (3.3%) 4a (6.7%) — 29 (3.3%) Standardized Achievement

Test 28 (2.9%) 4a (6.4%) — 25 (2.8%)

Student academic record 79 (8.1%) 3 (5.4%) 3 (7.9%) 72 (8.3%) Special student needs 102 (10.5%) 3 (5.4%) 3 (7.9%) 94 (11.0%) Special aptitude, skills, or

talents 28 (2.9%) 0 — 27 (3.1%)

Personal interview 188 (19.3%) 7 (11.1%) 8 (18.3%) 173 (19.9%) Recommendations 106 (10.8%) 5 (7.3%) 3 (7.6%) 98 (11.3%)

Note. Values are number and percentage unless indicated. EMO = educational manage- ment organization.

aCells for large- and small-EMO-operated schools are combined because National Center for Education Statistics regulations prohibit reporting cell sizes of less than 3.

Charter schools are most likely to have their charter granted by a school district (37.7%) or a state board of education (37.2%). Less frequently cited as the charter-granting authority are postsecondary institutions and state charter-granting agencies.5 Independent charter schools are more likely to be chartered by a local school district: 40% of non-EMO-managed schools compared to 16% of large- and 18% of small-EMO-managed schools. About 49% of large-EMO-operated schools and 48% of small-EMO-managed schools had their charter granted by a postsecondary institution. With many of the EMO-operated schools located in Michigan and the university authorizers therebeingparticularlyactive, thepercentageofEMO-operated schools authorized by postsecondary institutions is not surprising.

The vast majority of all charter schools (74%) are newly created or start-up charter schools. Another 16% are preexisting public schools that converted to charter status. The remaining 10% of schools reported being converted or preexisting private schools.6 A higher percentage of EMO-op- erated schools are newly created (84.4% of large-EMO and 89.4% of small- EMO-operated schools) compared to independent charter schools (72.4%).

The average number of years of operation for all the charter schools is 2.59, but therearevariationsbytypeofschool.Theindependentcharterschoolsare older, with an average of 2.64 years of operation, compared to 2.55 for the small EMO schools, and only 1.99 for the large-EMO-operated schools.

In terms of the grade levels served by the charter schools, most of the schools (50.0%) were elementary schools, about 9% were middle schools, 22% were high schools, and 19% were schools that served combined grades. The pattern of grades offered varies by EMO status. Large-EMO- operated schools are much more likely to serve elementary grades com- pared to small-EMO-operated or non-EMO schools, with 76% of large EMOs serving elementary grades exclusively compared to about 35% of small-EMO-operated schools and 49% of non-EMO schools. EMO schools have been criticized for predominantly serving the elementary and lower cost grades (Miron & Nelson, 2002; Molnar et al., 2005). The EMO-operated schools were equally as likely as other charter schools to serve high school grades (22.2% each) but somewhat less likely to offer middle school only (3.4% vs. 9.3%) or a combination of grade levels (14.8% vs. 19.8%). In terms of offering combined grades, small-EMO-operated schools are the most likely to do so (24%), compared with only 9% of the large-EMO-operated schools and 20% of non-EMO-affiliated charter schools.

85

Charter School Enrollments

5Keep in mind, however, that not all types of authorities are able to grant charters in all states.

6Note that not all states allow public or private schools to convert.

The mean school size is 266 students, but disaggregating by EMO type unmasks some differences.7 Large-EMO schools have the largest aver- age enrollment, with 519 students. Considering that large-EMO-operated schools are mostly elementary schools, their large average enrollment is notable. Small-EMO-operated schools have a mean enrollment of 264. Other research has also documented that EMO charter schools, especially large-EMO schools, tend to be larger than other charter schools, perhaps as a way of achieving economies of scale that are important for profitability (Miron & Nelson, 2002; Plank et al., 2000).

Disaggregating by school origin, such as new start-up versus conversion schools also reveals different patterns of size. Whether managed by an EMO or not, charter schools that converted from public schools are the largest, with anaverageenrollmentof516students.Thiscomparestosmaller,newstart-up schools, which have on average 224 students. Private schools that converted to charter status are, on average, the smallest, with about 169 students.

Some charter schools are designed to serve at-risk students. A total of 6% are “specifically for students who have been suspended or expelled, who have dropped out, or who have been referred for behavioral or adjust- ment problems” (U.S. Department of Education, 2000, p. J11). The majority of these schools are located in Michigan.

A number of charter schools also have magnet programs. About 12% of all the charter schools have magnet programs, with 67 of them being pro- grams “intended to encourage students of different racial or ethnic back- grounds to enroll … for the purpose of creating racial balance or reducing racial isolation” (U.S. Department of Education, 2000, p. J12). The vast ma- jority of charter schools with magnet programs are not associated with EMOs. Only 3% of EMO-operated schools had racial-balance magnet pro- grams compared to 7% of the independent charter schools. Most of the magnet programs (72.8%) are in new start-up schools as opposed to public school conversion charters (24.3%).

Charter schools vary tremendously in their curricular focus. The survey asked schools whether they offered a regular elementary–secondary cur- riculum, had a special program emphasis (e.g., Montessori, language im- mersion, science focus, etc.), were designed primarily for students with disabilities, were designed to train students for vocational and technical occupations, or offered some other alternative program (e.g., serving at- risk students, etc.). Overall, about one half of all schools reported having a regular elementary–secondary curriculum (53.6%), with alternative pro- grams being the next largest category (22.9%), followed by schools with special program emphases (20.6%), special education schools (1.6%), and

86

N. Lacireno-Paquet

7Enrollment is for students in K–12 plus any ungraded students.

vocational–technical schools (1.2%). The pattern of school type varies greatly by management status, however. No EMO-operated schools are designed especially for students with disabilities, and no large-EMO-oper- ated schools are vocational–technical schools. The vast majority of large- EMO-operated schools (82.3%) are regular elementary–secondary schools.

Admission requirements are a potential tool for schools to carefully structure their student body, although not all states allow the use of such criteria. Many states do not specify whether admissions criteria are permit- ted, or more commonly, they permit preferences such sibling preference or geographic preferences. About one fourth of the schools have require- ments for admissions, even in those states where ostensibly these types of admissions criteria are not legally permitted, such as in Washington, DC. Some schools may have more than one admission requirement. Of those schools with admission requirements, 13% have an admission test and 11% use standardized achievement tests. Almost one third of schools with ad- missions requirements base admissions on the student’s academic record. A total of 40% of those with requirements base admission on students’ spe- cial needs, and 11% base admission on students’ special talents. Many of these schools (41.5%) also request recommendations as their requirement. The most frequently used admission requirement, however, was a per- sonal interview. Almost three fourths of the schools with admissions re- quirements use a personal interview. Independent charter schools are more likely to have admissions requirements, with about 27% of these schools having such criteria. Less than one fifth of large- and small- EMO-operated schools use admissions criteria.

Aquestion on the survey asks if the schools have a service requirement for parentsofenrolledstudents.Overall, about40%ofschoolshavesuchrequire- ments. Some charter school researchers (e.g., Lopez, Wells, & Holme, 2002) have worried that schools use parental service or volunteer requirements as a means of attracting certain types of students and families—those who are most involved and motivated, and those with the luxury of being able to de- vote volunteer time to a school. Non-EMO-operated charter schools were the most likely to have such requirements, with 41% having them compared with 35% of large-EMO-operated and 22% of small-EMO-operated charters.

Student Characteristics: Minority Enrollment and Low-Income Students

In this section I examine the characteristics of the students in charter schools. Specifically, I look at the enrollment of minority and low-income students in charter schools.

87

Charter School Enrollments

In terms of the characteristics of the students in the charter schools, there is great diversity both among the states and between schools. Minor- ity enrollment in charter schools is defined as the percentage of all students who are Asian or Pacific Islander, Black or African American, Hispanic, Native American or Alaska Native, or the percentage of non-White stu- dents. Overall, about 50% of all students in charter schools are minori- ty, but the pattern varies by EMO association. The large-EMO-operated schools have the highest percentage of minority students, with a mean of 64%; the small-EMO-operated schools have a much lower percentage (41.1%), and independent charter schools fall in between, with an average of about 49%. Of course, average minority enrollment varies by the urban location of the school. Table 3 presents minority enrollment by manage- ment status and urban location.

The 1999–2000 Schools and Staffing Survey asked several questions about the FARL program, including whether a school had any eligible stu- dents, whether the school participated in the program, and how many stu- dents are approved for the program. Most schools (95.7%) reported having eligible students. However, although all large-EMO-operated schools re- ported having eligible students, only 90% of small-EMO-operated schools did. Not all schools with eligible students actually participate in the pro- gram. Overall, about 70% of responding schools report actually participat-

88

N. Lacireno-Paquet

Table 3

Minority Enrollment, by Management Status and Urban Location

Overall Central Cities Urban Fringe Rural

Full sample (estimate) 50.1 64.4 35.9 29.2 SE 0.57 0.68 0.87 1.31 CI 48.9, 51.2 63.0, 65.7 34.1, 37.6 26.6, 31.8

Large-EMO operated schools (estimate)

64.4 76.3 40.2 0.0

SE 1.94 2.21 3.32 CI 60.5, 68.2 71.9, 80.7 33.6, 46.8

Small-EMO operated schools (estimate)

41.1 57.0 28.4 29.0

SE 2.26 3.10 3.47 5.7 CI 36.7, 45.6 50.9, 63.1 21.5, 35.3 17.6, 40.4

Independent charter schools (estimate)

49.5 63.5 35.9 29.2

SE 0.61 0.74 0.95 1.35 CI 48.2, 50.7 62.1, 65.0 34.0, 37.8 26.5, 31.9

Note. CI = confidence interval; EMO = educational management organization.

ing in the program.8 Large-EMO-operated schools were the most likely to participate, with 85% doing so, followed by 69% of independent charter schools and 53% of small-EMO-operated charter schools.

The average percentage of FARL-eligible students in each type of school also varies (see Table 4). Large-EMO-operated schools have the highest percentage of FARL-eligible students, with an average FARL-eligible en- rollment of 51%. This is not significantly different from the average for in- dependent charter schools (47.9%). However, small-EMO-operated char- ter schools, in addition to participating less frequently in the program, also have a significantly lower average FARL-eligible enrollment of 42%.

Further Questions

The data presented so far point to the importance of EMO association and internal school organizational characteristics in describing the popula- tion of charter schools. There appear to be some marked differences be- tween large-EMO-operated schools, small-EMO-operated schools, and in-

89

Charter School Enrollments

8Note that this question has a somewhat lower response rate than many on the survey. Weighted, only 872 schools responded out of a possible 979 respondents. Indeed, FARL data are known to have some measurement problems. In this case the problem may be exacerbated by the fact that the data are reported by the principal or director and do not come from admin- istrative records, and although some charter schools are more aggressive in identifying eligi- ble students, those schools that do not even participate in the program may underestimate the number of FARL eligible.

Table 4

Enrollment of Free and Reduced Price Lunch (FARL) Eligible Students, by Management Status

Mean % of FARL-Eligible Students

Full sample (estimate) 47.8 SE 0.50 CI 46.9, 48.8

Large-EMO-operated schools (estimate) 51.0 SE 1.67 CI 47.7, 54.3

Small-EMO-operated schools (estimate) 42.4 SE 2.1 CI 38.1, 46.7

All other charter schools (estimate) 47.9 SE 0.58 CI 46.7, 49.0

Note. CI = confidence interval; EMO = educational management organization.

dependent schools. Given the diversity of state charter school policies that apply to these schools, the multivariate analyses of enrollment patterns presented in the next section illuminate whether the differences between school types hold and begin to explain the differences found here.

Multivariate Analyses

In the next section I present multivariate analyses of two dependent variables: the enrollment of low-income and minority students in charter schools.9 The models include variables indicating school organizational characteristics and state charter school policy variables that are expected to have some influence over schools’ willingness and ability to serve tradi- tionally disadvantaged students. Table 5 contains the definitions for the variables used in both models.

The models use Ordinary Least Squares regression with the balanced re- peated replication (BRR) method of calculating standard errors. To sum- marize, the BRR method of calculating standard errors includes correc- tions for nonresponse and any nonequal probability of a school being selected for the sample.10,11 It also weights each responding school to pro- duce a population estimate; for example, in the FARL-eligible model, al- though the actual number of observations is 751, the estimated population size is 872.

FARL

The results presented here take the percentage of FARL-eligible stu- dents in a school, as reported by the principal or director, as the dependent

90

N. Lacireno-Paquet

9These dependent variables might be jointly determined and thus may not be completely independent. Here, however, I treat them as independent outcomes. I have conducted a sensi- tivity analysis, including estimating models using Zellner’s seemingly unrelated regression. These are available upon request from the author.

10Note that the charter schools were selected with certainty as the population was sur- veyed. Nonetheless, the replicate weights reflect the variability due to the nonresponse ad- justment. The nonresponse adjustment has both a variance and bias component (Kaufman, personal communication, February 2, 2004).

11Although the BRR method of calculating standard errors is the recommended method of the National Center for Education Statistics, others have suggested that cluster analysis (where schools in the same district or state are grouped together and the standard errors ad- justed appropriately) is a better method. I ran the models both ways. The results of the BRR and cluster analyses were robust with overwhelming consistency in significance, sign, and magnitude of the coefficients. The results of the cluster analyses are available upon request from the author.

91

Table 5

Variable Definitions

Variable Name Variable Definition

School management variables Operated by a large-EMO firm Indicates a school is operated by a large-EMO firm Operated by a small-EMO firm Indicates a school is operated by a small-EMO firm Non-EMO charter school Excluded or reference group Interaction of urban and large EMO Indicates a school is in a large or midsize central city and is managed by a large EMO Interaction of urban and small EMO Indicates a school is in a large or midsize central city and is managed by a small EMO

School history variables New start-up school Excluded or reference group Converted public school Indicates a school is a converted public school Converted private school Indicates a school is a converted private school Charter granted by school district Indicates a school’s charter was granted by the school district

School characteristics Operating years Number of years in operation (since first offered classes) Elementary grades only Excluded, reference group Middle school grades only Indicates a school serves exclusively middle school grades High school grades only Indicates a school serves exclusively high school grades Combined grades Indicates a school serves combined grades Racial/ethnic balance magnet Indicates a school has a magnet program with racial/ethnic balancing provisions Total enrollment in school Total enrollment for the school in Kindergarten to Grade 12 and any ungraded students

(continued)

92

Admissions and other criteria Has any admissions criteria Indicates a school has special requirements for admissions other than proof of residency, age, or

immunization Admissions tests Requirement is an admissions test Standardized tests Requirement is a standardized achievement test Academic record Requirement is academic record Special student needs Requirement is special student needs (i.e., students at risk or with disabilities) Personal interviews Requirement is personal interview Recommendations Requirement is recommendations Parental service requirement Indicates whether parents are required to participate or volunteer at the school

Special programs Regular elementary/secondary

curriculum Excluded, reference group

Special program emphasis Indicates a school has a special program emphasis such as Montessori, foreign language immersion, and so on

Special education program Indicates a school is primarily for students with disabilities (reference group is a regular elementary or secondary school)

Vocational/technical program Indicates a school is primarily for students being trained for technical occupations Alternative program Indicates a school is designed to offer an alternative or nontraditional occupation and is not a

regular, vocational/technical, or special program, or special education school For at-risk students Indicates a school is specifically for students who have been suspended or expelled, dropped out,

or been referred for behavioral or adjustment problems

Table 5 (Continued)

Variable Name Variable Definition

93

Geographic location Urban Excluded, reference group Suburb Indicates a school is in the suburb or urban fringe of a large or midsize central city Rural Indicates a school is in a rural or unincorporated area

State policy variables Cap on no. of schools Indicates school is in a state that has a cap on the total of charter schools State-only authorizer Indicates school is in a state where the only chartering authority is a state level one (such as a state

charter board or state board of education) Local-only authorizer Indicates the only chartering authority in the state is local school districts Schools have independent legal

status Indicates schools in the state have an independent legal status (i.e., not dependent on school boards

or other entities) Private school conversions

permitted Indicates school is in a state where private conversion schools are permitted

Funding varies by grade level served Indicates funding for charter schools in the state varies by the grade levels served by the school State funding for at-risk students Indicates school is in a state that provides supplemental funds for serving at-risk or low-income

students State start-up funds available Indicates school is in a state that provides start-up funds for charter schools Transportation not required Indicates school is in a state where transportation of charter school students is not required State facilities funding Indicates school is in a state that provides additional funds to charter schools for facilities Districts must provide facilities Indicates school is in a state where local school district is required to provide facilities to charter

schools

Note. EMO = educational management organization.

variable. This variable serves as proxy for the poverty level of the student body.12 Not all schools that have eligible students participate in the pro- gram, and over 200 schools failed to report the figures for the actual num- ber of students who are approved for the FARL program. For these rea- sons, I use the percentage of FARL-eligible students as the variable to estimate the low-income population of schools. Although the variable per- centage of FARL-eligible students also has some missing data, 89% of schools reported eligibility statistics.13

The school-level variables in the model can be grouped into six catego- ries: (a) school management, (b) school history, (c) school characteristics, (d) admissions criteria, (e) special programs, and (f) geographic location. Two interaction variables are included that indicate whether a school is managed by a large EMO and is located in a central city and whether a school is operated by a small EMO and located in a central city. These two variables are included because EMOs tend to disproportionately locate in central cities and thus the interaction terms can identify the differential im- pacts of EMOs in central cities.

The state policy variables can be grouped into the following categories: (a) authorization and legal framework, (b) funding, (c) transportation, and (d) facilities. The data on state policy variables come from Nelson et al. (2000) and RPP International (1999). Table 5 provides all of the variable definitions.

The model estimates and goodness-of-fit statistics are presented in Ta- ble 6. The model overall is statistically significant and explains about 21.32% of the variation in the percentage of FARL-eligible students in char- ter schools. The small amount of explained variation in the dependent variable may be due to the FARL measurement problems noted previously.

Looking first at the management status of charter schools, the variables indicating charter schools are operated by large or small EMOs present some interesting results. Charter schools operated by a small EMO but not in a central city location, on average, enroll about 11 percentage points fewer low-income students than independent charter schools. The coeffi- cient on the variable indicating that a school is operated by a large EMO in

94

N. Lacireno-Paquet

12Although eligibility for FARL is not a perfect proxy for family poverty status, one usu- ally has to be of low-income status to qualify and given the lack of data on the true income of students in a school, it is the most commonly used measure of school-level poverty.

13I wanted to see if schools with missing data on FARL eligible appear different from those without missing data. The schools with missing data have a much lower minority enrollment, with an average of 23% of minority students compared to 53% for the schools without missing FARL data. In addition, the schools with missing data are less likely to be elementary schools. Five of the 92 schools are EMO-operated charter schools.

95

Table 6

Regression Results (Ordinary Least Squares Regression With Balanced Repeated Replication–Based Standard Errors)

% FARL Enrollment

% Minority Eligible

School management variables Operated by a large-EMO firm –5.34 (3.20) 0.62 (3.77) Operated by a small-EMO firm –10.82*** (2.42) –6.75 (3.41) Non-EMO charter school Reference group Reference group Interaction of urban and large EMO 20.03*** (3.90) 10.93* (4.34) Interaction of urban and small EMO 21.19*** (4.28) 1.58 (5.09)

School history variables New start-up school Reference group Reference group Converted public school 14.39*** (1.57) 3.80* (1.53) Converted private school –3.33 (1.99) –1.93 (1.67) Charter granted by school district 2.66* (1.26) 0.26 (1.35)

School characteristics Operating years –0.24 (0.31) –1.82*** (0.38) Elementary grades only Reference group Reference group Middle school grades only –0.71 (1.59) 4.78* (1.91) High school grades only 0.92 (1.62) 2.59 (1.44) Combined grades 4.65*** (1.23) –2.02 (1.29) Racial/ethnic balance magnet 8.51*** (2.22) 6.79** (2.19) Total enrollment in school –0.004* (0.002) 0.01*** (0.001)

Admissions and other criteria Has any admissions criteria –0.47 (2.38) 3.15 (2.57) Admissions test –13.92*** (2.64) –3.10 (3.17) Standardized test 13.46*** (2.69) 4.29 (3.18) Academic record –11.28*** (2.15) 6.94** (2.65) Special student needs 12.70*** (2.64) 2.90 (2.54) Personal interview –4.83* (2.36) –5.49* (2.39) Recommendations –0.17 (2.42) 1.42 (2.07) Parental service requirement 0.10 (1.08) 6.04*** (1.21)

Special programs Regular elementary/secondary curriculum Reference group Reference group Special program emphasis 1.59 (1.32) –1.49 (1.30) Special education program –6.24 (3.76) –14.17*** (4.15) Vocational/technical program –8.68** (2.97) 7.25 (4.70) Alternative program 0.36 (1.33) –2.91* (1.27) For at-risk students 8.41** (2.83) 4.81 (2.75)

Geographic location Urban Reference group Reference group Suburb –14.15*** (1.29) –23.54*** (1.17) Rural area 0.61 (1.74) –25.55*** (1.69)

(continued)

an urban location is large, positive, and significant (20.03). This suggests that these schools enroll higher percentages of FARL-eligible students. The variable indicating a charter school is operated by a small EMO in a central city location is also large, positive, and significant (21.19). With the coeffi- cient on small EMO, it suggests that urban, small-EMO-operated schools enroll about 10 percentage points more low-income students compared to similar independent schools.

The results on the variables indicating the use of admissions criteria are also noteworthy. The use of admissions criteria overall is not significant, but certain individual criteria are important for understanding the varia- tion in the enrollment of low-income students. Some criteria seem to de- press the percentages of low-income students enrolled, such as the use of admissions tests, academic records, and personal interviews. Others, such as the use of standardized tests and special students needs, have the oppo- site effect.

At the state level, a number of the regulatory and authorization vari- ables prove to be important in explaining the variation in the dependent

96

N. Lacireno-Paquet

State policy variables Cap on no. of schools –9.65*** (1.47) –1.05 (1.75) State-only authorizer 0.30 (2.49) 2.36 (2.64) Local school district-only authorizer –20.21*** (2.43) –17.64*** (2.59) Multiple authorizers Reference group Reference group Schools have independent legal status –2.34 (1.58) –10.92*** (1.63) Private school conversions allowed 4.10 (2.39) 0.41 (2.29) Funding varies by grade level served –1.48 (1.79) 5.91** (2.00) State funding for at-risk students 9.98*** (1.45) 11.50*** (1.34) State start-up funds available 6.77*** (1.17) 9.11*** (1.45) Transportation not required –27.65*** (1.42) –27.17*** (1.68) State facilities funding –11.56*** (1.84) –16.75*** (2.03) Districts must provide facilities –7.83*** (1.71) –19.27*** (1.79)

Constant 73.29 (3.57) 85.94 (3.54) Model properties

Unweighted N 751 843 Weighted N 872 979 F statistic F(39, 50) = 45.32,

p = .0000 F(39, 50) = 87.13,

p = .0000 R2 0.2132 0.3062

Note. Standard errors are in parentheses. EMO = educational management organization. *p ≤ .05. **p ≤ .01. ***p ≤ .001.

Table 6 (Continued)

% FARL Enrollment

% Minority Eligible

variable. Schools located in states with a cap on the total number of charter schools serve a significantly lower percentage of FARL-eligible students— about 9.65 percentage points fewer. Schools in states with local school dis- tricts as the only charter school authorizer (vs. having multiple author- izers) also serve much lower percentages of FARL-eligible students (–20.21). The coefficients on the regulatory and authorization policy variables sug- gest that in an open or less restricted market, charter schools will serve low-income populations perhaps in addition to serving other niche mar- kets. Although policymakers may try to manage the competition poten- tially introduced by charter schools (e.g., by limiting authorizers to local school districts or by restricting the total number of schools permitted), these types of policies may result in artificial constraints on the market, or these types of policies may give too much power to local school districts to control the competition by authorizing only the kinds of schools they want as alternatives to the schools and programs they offer, rather than the types of schools entrepreneurs, parents, community groups, and others believe are necessary.

A number of the funding variables were significant, which is not sur- prising when we consider that funding may be the most important factor in charter schools’ willingness and ability both to meet the needs of stu- dent and to generate a profit (or at least not a loss). The coefficient for the variable indicating that states provide additional funds for serving at-risk or low-income students is 9.98, suggesting that schools in these states en- roll higher percentages of low-income students. Schools in states with start-up funds available also enroll a higher percentage of FARL-eligible students, about 6.77 percentage points more.

The negative coefficient (–27.65) on the variable “transportation of charter school students not required” (where the law requires neither the charter school nor the school district to provide transportation to charter students) is expected, although its magnitude is startling. The large magni- tude indicates the extent to which transportation can be a barrier keeping low-income families from exercising their choice options, which has been a concern of some skeptics (Godwin & Kemerer, 2002; Henig, 1994). Low-in- come families may not have the resources (time and/or money) to trans- port their children to schools that may be located far from their homes. This result lends support for these concerns.

With regard to facilities, which charter school operators frequently cite as the major obstacle in getting their school up and running (RPP Interna- tional, 1998, 2000), the results were somewhat surprising. The negative and significant coefficient (–11.56) on the variable indicating that a school oper- ates in a state with additional facilities funding suggests these schools serve slightly lower percentages of FARL-eligible students. This is surpris-

97

Charter School Enrollments

ing because one might expect that having facilities funds would free up money from the general per-pupil amount that would otherwise go for fa- cilities but could now go toward instructional programs or other services.

The estimates for the policy variables suggest that state policies can play an important role in influencing the types of students who get served by charter schools. The regulatory and funding environment created by state policies affect the set of incentives and disincentives that charter schools face in determining the types of students to serve and target. The use of the state policy variables advances our understanding of what might encour- age or discourage charter schools from serving or enrolling low-income students.

Minority Enrollment

Recall that the average minority enrollment in the full group of charter schools is 50% but that the percentage varies significantly by charter school management status. Large-EMO-operated schools enrolled the highest percentage of minority students (64%) and small-EMO-operated schools enrolled the lowest (41%). In non-EMO schools about 50% of students were minorities. Here I examine whether this pattern holds in a multivariate en- vironment controlling for other school organizational and state policy characteristics.

The model is presented in Table 6. Again, Ordinary Least Squares re- gression with the BRR method of calculating standard errors is used. All variable definitions can be found in Table 5. Minority enrollment was mod- eled identically to the FARL-eligible model presented before; however, the sample size is slightly larger because minority enrollment has less missing data. The model is significant overall, and it explains about 30.62% of the variation in the percentage of minority students in charter schools.

Looking first at the school management variables, those indicating management by EMO firms are not as important in this model. The coeffi- cient on the variable indicating a small-EMO firm manages a school is –6.75 and is marginally significant (p = .051). Schools located in urban areas and managed by large-EMO firms enroll significantly higher percentages of minority students compared to other charter schools—about 10.93 per- centage points more.

Geographic location is important to control for given the racial segrega- tion of many of our towns and cities. Charter schools located in both the suburbs and rural areas enroll significantly lower percentages of minority students than do urban schools: –23.54 and –25.55, respectively.

The coefficients on the admissions criteria variable suggest that the ef- fects of particular criteria may differ between minority students and low-

98

N. Lacireno-Paquet

income students. Schools that use students’ academic records for admis- sions enroll about 6.94 percentage points more minority students, all else constant. This differs from the coefficient in the FARL-eligible model, in which the coefficient was large and negative (–11.28). The use of personal interviews for admissions was also significant, although negative (–5.49) and consistent with the coefficient in the FARL-eligible model. Schools that have a parental service requirement enroll significantly higher percentages of minority students—about 6.04 percentage points more than schools that do not, holding all else constant. The positive coefficient is noteworthy be- cause there is some concern that parental service requirements act as bar- rier to the participation of minority families.

The coefficients on the policy variables indicate that the regulatory, funding, transportation, and facilities policies exert influence on charter schools’ enrollment of minority students. The magnitude, sign, and signifi- cance of the coefficients are fairly consistent with the FARL model pre- sented earlier.

In terms of the authorization and legal framework, the effect of having only local school districts authorize charter schools, holding constant all other variables, is negative and significant. Schools in such states enroll about 17.64 percentage points fewer minority students. The magnitude and significance of the coefficient is similar to the FARL model.

Many of the funding variables were statistically significant and had signs in the expected direction. The coefficients on the variables funding varies by grade level served, state funding for at-risk students, and state start-up funds available are all significantly positive, as expected. Schools in states with these policy variables enroll significantly higher percentages of minority students, on average, compared to schools in states where these policy characteristics are not present.

Not requiring transportation of charter school students, by either char- ter schools or host school districts, significantly depresses the percentage of minority students in charter schools. This is a similar relationship as in the FARL-eligible enrollment. The coefficient suggests that schools in states with this policy enroll about 27% fewer minority students than those in other states.

The two facilities variables have large, negative, and significant coeffi- cients. The coefficient for whether the state provides funding for facilities is –16.75. The direction of the coefficient is somewhat of a surprise, al- though it is negative like the coefficient in the FARL-eligibility model. There are several possible explanations for this finding. Perhaps states with supplemental facilities funds have lower overall per-pupil reim- bursement amounts. Another consideration is that facilities funding was frequently not included in early legislation but rather was often added

99

Charter School Enrollments

later. It is possible that states were more likely to add facilities funding when their charter schools are serving a more White population. In states where school districts are required to provide facilities, schools are also likely to enroll lower percentages of minority students. Indeed, in these states, charter schools, holding other variables constant, serve 19.27 per- centage points fewer minority students than schools in other states.

Conclusions

The models presented here examine the influence school organizational and state policy characteristics on charter schools’ willingness and ability to serve low-income and minority students. These results demonstrate why the school organizational characteristics and the state policy context are important for understanding charter school outcomes.

School-level characteristics help explain a fair amount of the variation in the dependent variables. Organizational characteristics, such as associa- tion with EMOs, do matter, although not necessarily consistently across various student populations. Schools operated by large-EMO firms in ur- ban areas enroll higher percentages of low-income and minority students. These firms may be filling a demand for professional services in communi- ties that lack them but where there is the desire to offer educational choices to their communities. Schools operated by small-EMO firms in urban areas also enrolled higher percentages of FARL-eligible students than independ- ent charter schools. On the other hand, some school practices, such as the use of admissions criteria, may be tools to sort students or shape student populations in ways the critics of choice fear.

The results of the models suggest that the characteristics of state charter school policies are also important for encouraging or discouraging charter schools to serve disadvantaged students and the results are relatively con- sistent between the two student populations examined. Specifically, pol- icymakers need to pay attention to the issues of authorization, funding, and transportation. States need to carefully craft their policies to ensure the market and regulatory environment does not create incentives for charter schools to avoid traditionally disadvantaged student populations.

References

Brown, H., Henig, J. R., Lacireno-Paquet, N., & Holyoke, T. T. (2004). Scale of operations and locus of control in market- vs. mission-oriented charter schools. Social Science Quarterly, 85, 1035–1051.

100

N. Lacireno-Paquet

Buckley, J., & Kuscova, S. (2003). The effects of institutional variation on policy outcomes: The case of charter schools in the states (Occasional Paper No. 79). New York: Teacher’s College, Colum- bia University, National Center for the Study of Privatization in Education.

Buckley, J., Schneider, M., & Shang, Y. (2005). Are charter school students harder to educate? Evi- dence from Washington, DC (Occasional Paper No. 96). New York: Teacher’s College, Co- lumbia University, National Center for the Study of Privatization in Education.

Bulkley, K. E., & Wohlstetter, P. (Eds.). (2004). Taking account of charter schools: What’s happened and what’s next? New York: Teachers College Press.

Frankenberg, E., & Lee, C. (2003). Charter schools and race: A lost opportunity for integrated educa- tion. Cambridge, MA: The Civil Rights Project at Harvard University.

Fuller, B., Gawlik, M., Gonzales, E. K., & Park, S. (2004). Localized ideals of fairness: Inequal- ity among charter schools. In K. Bulkley & P. Wohlstetter (Eds.), Taking account of charter schools: What’s happened and what’s next? (pp. 93–120). New York: Teachers College Press.

General Accounting Office. (2002). Insufficient research to determine effectiveness of selected private education companies (GAO–03–11). Washington, DC: Author. Available from http://www.gao.gov/new.items/d0311.pdf

Gill, B. P., Timpane, P. M., Ross, K. E., & Brewer, D. J. (2001). Rhetoric versus reality: What we know and what we need to know about vouchers and charter schools. Santa Monica, CA: RAND.

Godwin, R. K., & Kemerer, F. J. (2002). School choice tradeoffs: Liberty, equity, and diversity. Aus- tin: University of Texas Press.

Henig, J. R. (1994). Rethinking school choice: Limits of the market metaphor. Princeton, NJ: Prince- ton University Press.

Henig, J. R., Holyoke, T. T., Brown, H., & Lacireno-Paquet, N. (2005). The influence of founder type on charter school structures and operations. American Journal of Education, 111, 487–588.

Henig, J. R., Lacireno-Paquet, N., Holyoke, T. T., & Brown, H. (in press). When does mission matter? Exploring a typology of charter school orientation. In M. Kayes & R. Maranto (Eds.), A guide to charter schools: Research and practical advice for educators. Lanham, MD: Scarecrow Press.

Lacireno-Paquet, N. (2004). Do EMO operated schools serve disadvantaged students? The in- fluence of policy. Education Policy Analysis Archives, 12(26). Retrieved July 7, 2005, from http://epaa.asu.edu/epaa/v12n26/

Lacireno-Paquet, N., Holyoke, T. T., Henig, J. R., & Moser, M. (2002). Creaming versus crop- ping: Charter school enrollment practices in response to market incentives. Educational Evaluation and Policy Analysis, 24, 145–158.

Lopez, A., Wells, A. S., & Holme, J. J. (2002). Creating charter school communities: Identity building, diversity, and selectivity. In A. S. Wells (Ed.), Where charter school policy fails: The problems of accountability and equity (pp. 129–177). New York: Teachers College Press.

Miron, G., & Nelson, C. (2002). What’s public about charter schools? Lessons learned about choice and accountability. Thousand Oaks, CA: Corwin.

Molnar, A., Garcia, D., Sullivan, C., McEvoy, B., & Joanou, J. (2005). Profiles of for-profit educa- tion management organizations, 7th annual report, 2004–2005 (Document No. 6103). Tempe: Commercialism in Education Research Unit, Arizona State University. Available from http://www.ecs.org/html/Document.asp?chouseid=6103

Molnar, A., Wilson, G., & Allen, D. (2004, February). Profiles of for-profit education management companies, 6th annual report, 2003–2004 (EPSL-0402-101-CERU). Tempe: Commercialism in Education Research Unit, Education Policy Studies Laboratory, College of Education, Di- vision of Educational Leadership and Policy Studies, Arizona State University. Available from http://www.asu.edu/educ/epsl/CERU/Documents/EPSL-0402-101-CERU.pdf

101

Charter School Enrollments

Nathan, J. (1996). Charter schools: Creating hope and opportunity for American education. San Fran- cisco: Jossey-Bass.

Nelson,F.H.,Muir,E.,&Drown,R. (2000).Venturesomecapital:Statecharterschool financesystems. Washington,DC:OfficeofEducationalResearchandImprovement,U.S.DepartmentofEd- ucation. Available from http://www.ed.gov/rschstat/eval/choice/charterfin.pdf

Peterson, P., & Greene, J. P. (1998). Race relations and central city schools: It’s time for an ex- periment with vouchers. Brookings Review, 16(2), 33–37.

Plank, D. N., Arsen, D., & Sykes, G. (2000, May). Charter schools and private profits [Elec- tronic version]. The School Administrator.

RPP International. (1998). A national study of charter schools: Executive summary. Washington, DC: Office of Educational Research and Improvement, U.S. Department of Education.

RPP International. (1999). A comparison of charter school legislation: Thirty-three states and the District of Columbia incorporating changes through October, 1998. Washington, DC: U.S. De- partment of Education.

RPP International. (2000). The state of charter schools 2000: 4th year report. Washington, DC: Of- fice of Educational Research and Improvement, U.S. Department of Education.

U.S. Department of Education. National Center for Education Statistics. (2000). SASS & PSS Questionnaires, 1999–2000 (NCES 20000–310). Washington, DC: Author.

Vergari, S. (Ed.). (2002). The charter school landscape. Pittsburgh, PA: University of Pittsburgh Press.

Viteritti, J. P. (1999). Choosing equality: School choice, the constitution, and civil society. Washing- ton, DC: Brookings Institute.

Wells, A. S. (Ed.). (2002). Where charter school policy fails: The problems of accountability and equity. New York: Teachers College Press.

Witte, J. F., Shober, A. F., & Manna, P. (2003, August). Analyzing state charter school laws and their influence on the formation of charter schools in the United States. Paper prepared for the American Political Science Association Annual Meeting, Philadelphia. Available from http://pmanna.people.wm.edu/research/WitteShoberManna2003_APSA.pdf

Wohlstetter, P., Smith, J., Hentschke, G., & Malloy, C. (2004, April). Understanding the playing field: Strategic alliances. Paper presented at the Annual Conference of the American Educa- tional Research Association, San Diego, CA.

102

N. Lacireno-Paquet