School Choice

Mark J Van Ryzin. 21st Century Education: A Reference Handbook. Editor: Thomas L Good. Sage Publications. 2008.

This chapter reviews research on school choice programs, whose core objective is to provide high-quality educational options to students and their families. Advocates believe school choice programs will bring market forces to bear on underperforming public schools (Chubb & Moe, 1990). According to economic theory, the “inefficiency of public school districts might be reduced by increased private incentives to monitor by consumer-voters” (Kang & Greene, 2002, p. 5). In other words, competition should result in greater efficiency (i.e., higher test scores) as families behave more like educational consumers. Such consumer monitoring, as in other markets, should favor more efficient providers at the expense of those less efficient.

Many choice programs also attempt to address the issue of educational equity. In general, wealthier families have more educational options than low-income families, either through access to neighborhoods containing good schools or the financial resources to afford private schooling. Before choice programs, disadvantaged families had no recourse when presented with an underperforming local school. The establishment of formal choice programs can provide disadvantaged families with more options, creating more equity in educational opportunity.

It is clear that the goals of school choice programs (i.e., higher test scores, greater equity) are unassailable. But have these goals been achieved? To investigate this question, I review the research on school choice. Choice programs generally come in two forms: voucher programs, which enable students to attend private schools; and charter schools, which represent educational alternatives within the realm of public education. I first discuss the background and past performance of voucher programs and charter schooling; following this, I consider the types of families that are availing themselves of school choice and how these choices are made.

Voucher Programs

Vouchers represent cash payments or other incentives offered by state governments, school districts, or foundations that enable disadvantaged students to attend private schools. Private schools accepting students under voucher programs often must agree to abide by certain terms and conditions (e.g., elimination of admissions criteria for voucher students, increased accountability for voucher funds, etc.). When students take advantage of voucher programs to enroll in a private school, the public schools that lose these students generally also lose the public funding associated with them.


School choice programs have been in existence since the turn of the 19th century in states such as Vermont, which implemented a voucher program in 1869. Under this program, students in towns too small to support a public school could attend the private school of their choice using public funds. The goal of this program was educational equity, although the inequalities at that time fell upon rural students rather than the dis-advantaged inner-city students targeted by today’s voucher programs.

The first modern voucher program, the Milwaukee Parental Choice Program, was established in 1990 in Milwaukee, Wisconsin. This program was made available to approximately 1,000 students in disadvantaged neighborhoods. In 1995, in an effort to expand the available pool of schools, the Milwaukee program was expanded to include religious schools. Voucher opponents sued to halt the program, believing that the use of public money for religious pursuits via Catholic or other religious schools was a breach of the constitutional separation of church and state. The ensuing court battle culminated in a ruling from the Wisconsin Supreme Court in 1998 confirming the legality of the program.

In 1995, the Pilot Project Scholarship Program was established in Cleveland, Ohio. Tuition scholarships were granted to disadvantaged students that could be used at any private school (including religious schools) within the Cleveland school district. This program was also challenged in court, and the dispute was finally resolved by the 2002 U.S. Supreme Court decision (Zelman v. Simmons-Harris) in favor of the program.

Other voucher programs have been established in Washington, D.C.; New York City; San Antonio, Texas; and Dayton, Ohio. The state of Florida established the first statewide program, the Opportunity Scholarship Program, in 1999, and since then, Ohio and Georgia have followed suit. The first federally funded voucher program was established in January 2004, when Congress allocated $14 million to establish a program for low-income students in the District of Columbia.

Most of these programs are small, with a few hundred or a thousand students in each. However, the Milwaukee program has grown considerably and now supports more than 17,000 students in 121 private schools from kindergarten through twelfth grade, with up to $6,500 available for each student (Milwaukee Parental Choice Program, 2007). Overall, about 50,000 students participated in voucher programs in the 2005–06 school year.

Although many voucher programs are on relatively safe footing, others are under attack. For example, a voucher program was enacted in Colorado in 2004, but the Colorado Supreme Court declared it unconstitutional soon thereafter. A voucher program was established in the state of Utah in 2007 and then almost immediately repealed in a public referendum, with more than 60% of voters rejecting the program. This referendum was filed by Utahans for Public Schools, who surmised that, despite the passage of the legislation, there was not sufficient public support for the program.

Academic Achievement

Proponents argue that voucher programs can create greater efficiency in education because private and/or Catholic schools are simply more effective at educating students, especially disadvantaged students. Research provides some limited support for this point, finding that Catholic schools have small positive effects on student achievement and somewhat more substantial effects on student retention and graduation (e.g., Bryk, Lee, & Holland, 1993; Evans & Schwab, 1995; Grogger & Neal, 2000; Neal, 1997). However, there is no evidence regarding why Catholic schools have positive effects. It may be that Catholic schools are innately superior, for example, or it may be that the district-run public schools in disadvantaged areas are so inferior that any alternative is an improvement. Or it may be because of peer effects, in which students perform better when surrounded by higher-achieving peers (Hanushek, Kain, Markman, & Rivkin, 2003). Bryk and colleagues (1993), for example, only point to rather imprecise factors such as “high expectations of student success” and “individual attention” as the reasons behind the success of the Catholic schools in their sample. One noteworthy aspect of the research cited above is that the benefits appear to accrue mainly to African American and Latino youth in urban settings. Youth of other races in other settings do not appear to benefit, and again, it is not clear why this is the case.

Similar results have been found in actual evaluations of existing voucher programs. For example, Howell, Wolf, Peterson, and Campbell (2000) found that the privately funded voucher programs in New York City; Dayton, Ohio; and Washington, D.C., generated positive results but only for African American students; other ethnicities did not show any benefits. As a whole, the African American children across the three cities scored 3.9, 6.3, and 6.5 percentage points higher than the control groups in the first 3 years of the program. However, the size of the effects varied greatly among grade levels and cities at a given point in time, and the effects also varied substantially within grades and cities across time. This may be due to the extensive attrition of the study sample during the course of the study, in which at-risk students were more likely to drop out of the voucher program (e.g., students with learning disabilities, students from families on welfare) while students from families predisposed to religion were more likely to remain (Ladd, 2002). This data raises the possibility that the achievement gains were due primarily to selection effects. In other words, the positive results of these voucher programs may only be valid for those students who represent a good fit for private or religious schools; for other students, these programs may be ineffective.

The Milwaukee voucher program has been evaluated several times with varying results. This program, like most voucher programs, is not based upon true random assignment, so different analytic strategies were applied with differing results. In 1995, the initial evaluation found that the program had no impact on student performance (Witte, Stern, & Thorn, 1995). In contrast, a subsequent reanalysis of the data using a different control group found significant gains in both reading and math as a result of the program (Greene, Peterson, & Du, 1998). Finally, a third analysis that built upon the previous two and took into account important issues such as the differential rate of attrition in the treatment and control groups concluded that the program generated small gains in math (between 1.5 and 2.3 percentage points per year) but no gains in reading (Rouse, 1998).

In sum, the research offers no clear-cut findings on the effects of voucher programs on academic achievement. Some students and families appear to fit well into religious schools and benefit from voucher programs, but this is clearly not the case for the majority. The evidence that higher-risk students (e.g., disabled, low income) tend to drop out of voucher programs suggests that these programs may not be as successful with those students who need the most help.

Effect of Competition

Proponents argue that voucher programs represent competition to district-run public schools; in response, district schools will be forced to become more efficient through either increased attention to their own operations, the introduction of new programs and practices, or both. In contrast to the charter school sector, where a great deal of research has investigated this question, very little research has examined whether competition from voucher programs can generate improvements in district-run public schools. In general, this is due to the small size of most voucher programs.

However, some programs, such as the Florida voucher program, are quite large and represent a good case study for the effects of competition. The Florida program is based upon a system of letter grades assigned to each school; if a school receives two F’s within 4 years, the students in that school are eligible for the statewide voucher program. Greene (2001) found that schools receiving a grade of F were able to raise their rate of achievement more significantly that those schools receiving other grades. However, this study has been criticized on the grounds that it failed to adequately control for statistical confounds such as selection effects and regression to the mean (Camilli & Bulkley, 2001). Ladd (2002) also suggests that it is far more likely that the “increased scrutiny, shame, and additional assistance associated with being labeled a low-performing or ‘failing’ school is a more likely explanation” (p. 15) for the increase in test scores than a perceived threat from the voucher program. Similar results were found in a review of the Milwaukee program (Hoxby, 2001). Clearly, more research is needed before any conclusions can be drawn.


Given these results, it is difficult to determine whether large-scale expansion of voucher programs would have positive effects on student achievement. Uncertainties surrounding the reasons behind the success of Catholic schools, combined with the variations in effects found in research on existing voucher programs, call for a cautious approach in the promotion of voucher programs as a more widespread mechanism of school choice.

At the same time, the failure rate of students from dis-advantaged families is such that there does not seem to be grounds to eliminate or restrict these programs, especially given the decision by the Supreme Court confirming their legality. However, to ensure that voucher programs are successful and cost effective, these programs should specifically target the most disadvantaged families, particularly African American families for whom Catholic schools may have some positive effects. To ensure that disadvantaged students are given equal access to private schools under voucher programs, participating schools should be required to accept all applicants and to deal with excess demand through a lottery process, and the voucher should represent all of the costs associated with attending the school. Participating schools should also be required to provide some sort of evidence of accountability for public funds.

Even with these controls in place, there is no certainty that large-scale implementations of voucher programs would succeed. For example, if the success of Catholic schools is primarily due to peer effects, then the positive effects of moving to a Catholic school would be reduced or eliminated if large numbers of disadvantaged students participated in voucher programs. Thus, while voucher programs represent a significant component of school choice, there does not appear to be an opportunity for these programs to play a significantly larger role in the near future. More likely, voucher programs will continue to play a small, albeit important, role for some families within the school choice movement until more definitive knowledge is accumulated regarding the processes and effects of such programs.

Charter Schools

Unlike private schools that take part in voucher programs, charter schools are new public schools created by a school district or other school authorizer. As public schools, charter schools are free to all students and, in general, cannot restrict enrollment based upon race, gender, socioeconomic status (SES), special education status, or previous educational achievement. Charter schools must abide by state testing standards and laws related to teacher licen-sure, but are often free from district mandates regarding organization, administration, curriculum, and pedagogy. As part of the public school system, charter schools receive funding from the state based upon the number of students in attendance. In most cases, this funding is provided directly to the school, in contrast to traditional public school funding that passes through the local school district. Levels of funding on a per-student basis are generally quite similar between district-run and charter schools, although this is not always the case. Some states, such as Minnesota, have augmented the per-student funding to charter schools with “lease aid” funding, recognizing that charter schools, unlike district-run schools, often must obtain their own facilities.


In the 1980s, Albert Shanker, former president of the American Federation for Teachers (AFT), was among the first to advocate for greater choice within the realm of public education rather than through private school vouchers. His approach involved reforming the public education system through changes to school governance. Individual public schools would be given contracts or “charters” that would grant them more freedom to explore alternative methods of teaching and learning while at the same time demanding full responsibility for student success or failure according to state learning standards.

Minnesota was the first state to enact formal charter legislation in 1991, followed by California in 1992. Under this legislation, various organizations (i.e., school districts, universities, and nonprofit organizations) take on the role of “authorizer” or “sponsor” for new charter schools. The authorizer oversees the development of new charter applications, which are then forwarded to the state department of education for final evaluation. If the application is approved, the charter authorizer is then tasked with overseeing the operation of the new school and ensuring that state laws are upheld and state learning standards are met.

By 2007, 40 states and the District of Columbia had enacted charter laws and over 3,900 chartered school were operating across the United States serving more than 1.1 million children (Center for Education Reform, 2007). California has the largest number of charter schools with over 700, while Florida, Ohio, Texas, Wisconsin, Michigan, Colorado, Minnesota, Pennsylvania, and Arizona have over 100. For the 2007–08 school year, 347 new charter schools opened across the United States, an increase of 8% over the previous year. In spite of these numbers, charter schools still only educate just over 2% of the public school students in the United States.

Research on charter schools has established several general patterns. One finding is that most charter schools tend to be smaller than district-run schools. The average enrollment of charter schools in 2002 was 242 students, while the average size of traditional public schools was 539 students (National Center for Educational Statistics, 2002). In addition, a sizable majority of charter schools have enrollments less than 200, while the majority of traditional district schools are larger than 200 students (RPP International, 2000). Charter schools are also quite popular, with 70% of charter schools having a waiting list (RPP International, 2000).

Research also suggests that charter schools often attract more disadvantaged and at-risk students. For example, a report by the U.S. Department of Education found that White students made up about 48% of charter school enrollment in 1998 compared to about 59% of public school enrollment in 1997–1998, and that charter schools enroll a higher percentage of students eligible for free or reduced-price lunch than do public schools in general (RPP International, 2000). A more recent report by the Department of Education confirms that charter schools contain a higher percentage of students eligible for free or reduced-price lunches (National Center for Educational Statistics, 2002).

However, this is by no means the rule nationwide. In some instances, the racial, socioeconomic, and/or “at-risk” composition of charter schools has more to do with local demographics than with any nationwide trends. For example, charter schools are more commonly found in central city areas versus suburban and small town areas (National Center for Educational Statistics, 2002), making them more accessible to disadvantaged families. In addition, many individual charter schools and charter school networks (e.g., Green Dot, Big Picture Company, Knowledge Is Power Program [KIPP]) specifically target at-risk populations. Beyond this, local factors such as the availability of public transportation, the cost and availability of appropriate facilities, and the political mobility of the local citizenry may dictate the locational decisions of charter schools and thus influence the composition of the student population (Henig & MacDonald, 2002).

Academic Achievement

As with voucher programs, proponents argue that charter schools are simply more efficient than district-run schools, mainly because they are free from district mandates. Research exists that both supports and contradicts this point. For example, several large-scale studies have found no significant difference in student achievement when comparing charter to district-run schools (Crew & Anderson, 2003; Hanushek, Kain, Rivkin, & Branch, 2007). In contrast, other large-scale research finds positive effects for charter schools of anywhere from 3 to 8 percentage points per year (Florida Department of Education, 2004; Greene, Foster, & Winters, 2003; Miron & Horn, 2002), while still other research finds positive effects for district-run schools of about the same size (Bifulco & Ladd, 2006; Miron & Nelson, 2002).

In each study of charter school achievement, the author or authors explain their findings by speculating as to the reasons for the differences (or lack of same) between charter and district-run schools. The result is virtual laundry list of potential confounds that could be influencing charter school achievement research. For example, Bettinger (2005) found that charter schools do not improve test scores or passage rates as rapidly as district-run public schools and speculated that this difference could be due to discrepancies in financial resources, teacher experience, and institutional immaturity. The author also found that charter schools in Michigan attracted students with lower “pre-charter” test scores than neighboring district-run public schools, meaning that differences in achievement could also be due to selection effects.

In another example, a portion of the 2003 Brown Center Report on American Education focused on “conversion” charter schools, which are schools that were formerly run by a school district but are now chartered (Loveless, 2003). Controlling for the fact that conversion charter schools serve higher-risk populations, the study found that achievement at conversion charter schools was higher than for both district-run schools and start-up charter schools. Other studies have reported similar results (e.g., Krop & Zimmer, 2005). Thus, the method used to create a new charter school may be an additional confound. If this is true, then it may also be important to consider “reconstituted” charter schools, in which an underperforming district-run school is closed by the school district or the state department of education and then reopened as a charter with a new administration and/or staff.

Gronberg and Jansen (2001) found that Texas charter schools focusing on at-risk children performed better than comparable district-run schools, while charters focusing on non-at-risk children performed slightly worse. Using almost the same data, Hanushek and colleagues (2002) found that new charters performed significant worse than district-run schools, but this effect disappeared once the charter schools were at least 2 years old. The conclusion that charter schools become more effective as they age was echoed by Sass (2006) using data from Florida. In Arizona, charter elementary schools were found to be superior to district-run elementary schools, but this situation was reversed for high schools (Solmon & Goldschmidt, 2004). In California, however, charter high schools were found to be superior (Raymond, 2003). Research has also found that students experience slowed growth in test scores during the first year in a charter school; thereafter, students recover and potentially accelerate to a rate beyond that of district-run schools (Booker, Gilpatric, Gronberg, & Jansen, 2007).

These examples provide valuable insight into the fundamental weakness of the extant research on charter schools. Contradictory findings are to be expected when there is no general consensus on the most appropriate methods for comparing the performance of charter and district-run schools. For example, many studies still rely on year-to-year comparisons that do not take into account the changing nature of the school population over time. Research has found that year-to-year patterns of test scores are very erratic and unstable, with a near-zero correlation from one year to the next (Linn & Haug, 2002). Many studies also make use of small samples, and findings from such samples can be highly unstable (Kane & Staiger, 2002).

Betts and Hill (2006) have issued guidelines that should contribute to better research on charter schools:

  • To control for selection effects, studies should be based upon random assignment where possible; where this is not possible, student characteristics and educational background must be measured and controlled.
  • Large samples of students should be used to enhance the stability of the results.
  • Individual student progress should be measured over time by tracking student data as they move between schools.
  • Family factors, such as income and involvement, should also be taken into account, and the traditional measure of family income (i.e., student eligibility for free or reduced-price lunch) should not be used.
  • Test characteristics must also be examined to ensure that, for example, an overly easy test is not creating a “ceiling effect” on test scores. The authors also hint at, but do not directly endorse, the inclusion of a wider range of achievement-related data beyond standardized test scores.
  • Rather than treating charter schools as a “black box,” important school-level factors must also be taken into account; these include curricula, pedagogical practices, years in operation, funding levels, staff qualifications, staff turnover, and school goals and philosophy. The authors even advocate the consideration of the environment in which a school operates (i.e., benign vs. hostile).

The challenge of opening the “black box” was taken up by Buddin and Zimmer (2005), who disaggregated the charter sector according to the origin of the school (conversion vs. start-up) and the school’s learning model (classroom based vs. nonclassroom based). They subsequently found significant differences in performance among the categories. They concluded that the freedom to innovate, a core component of the charter concept, has created a heterogeneous charter sector; as a result, the impact of school-level factors must be better understood before a more accurate picture of student achievement in the charter sector can emerge. Other researchers have proposed alternative starting points from which to disaggregate charter schools. Some examples include:

  • Founder type, such as for-profit versus nonprofit (Henig, Holyoke, Brown, & Lacireno-Paquet, 2005) or Education management organizations (EMO) versus non-EMO (Loveless, 2003)
  • Orientation, such as market oriented versus mission oriented (Brown, Henig, Lacireno-Paquet, & Holyoke, 2004)
  • Pedagogy, such as traditional, progressive, or vocational (Carpenter, 2006)

Perhaps the most advanced of these classification systems was created by Sass (2006), who utilized several dimensions at once: school age, start-up status (i.e., new vs. conversion), founder type (i.e., for-profit vs. nonprofit) and school mission (i.e., whether certain types of students were targeted for enrollment). However, given the sharp contrast between the simplicity of most of these classification systems and the complexity of the guidelines issued by Betts and Hill (2006), it is clear that much work remains to be done.

Effect of Competition

A great deal of research has examined the effects of charter school competition on district-run schools, and some evidence can be found to support the assumption that competition would have positive effects; however, the literature is far from unanimous on this point. In a study of charter programs in Arizona and Michigan, traditional public schools responded to the competition of charter schools by raising their achievement and productivity, where productivity is defined as achievement per dollar spent, controlling for incoming achievement differences of the students (Hoxby, 2003). This effect was found to be significant even when accounting for preexisting trends in achievement and productivity before the introduction of competition. Interestingly, the improvements in traditional public schools occurred once charter competition reached a level where reductions in enrollment began to create significant consequences for school staffing levels. The author proposed a number of explanations for the positive relationship between competition and achievement/ productivity, including:

  • The financial pressures of competition may bid up the salaries of the best teachers, thus keeping people in teaching (or drawing new people into teaching) that would otherwise pursue other career paths;
  • The need to attract families may force schools to issue more information about their performance, resulting in more informed choices by parents;
  • In order to ensure high levels of satisfaction, schools may have to be more receptive to parent participation in school business; and
  • The pressure to produce results may force schools to abandon traditional pedagogical techniques and curricula that are found to be unsuccessful.

In North Carolina, nearly 100 new charter schools opened in a 3-year period in the late 1990s and researchers studying the impact of this new source of competition found significant test score gains in district-run schools as a result of the competition (Holmes, DeSimone, & Rupp, 2006). Their findings suggest that the closer a charter school is to a district-run school (i.e., the greater the competition), the greater the achievement gains.

At the same time, there is also research that reports mixed findings. For example, Sass (2006) linked charter school competition in Florida to a slight increase in math scores at nearby district-run schools, but found no change in reading scores. In addition, a review of research that examined more than 40 studies of the effects of the competition found that a number of these studies reported beneficial effects on student achievement while others reported no significant effects (Belfield & Levin, 2002).

One explanation for the disparate results is that, as with research on voucher programs, there are extraneous factors in play. For example, both Bettinger (2005) and Sass (2006) raise the possibility that improvements in district school test scores may result from the exodus of certain low-performing groups of students to charter schools. Complicating the picture, there may also be factors involved that are external to the schools themselves. For example, a study that included school districts in Massachusetts, New Jersey, and Washington, D.C., found that charter competition had not induced large changes in district operations, despite the fact that a significant number of students had left district schools for charter schools (Teske, Schneider, Buckley, & Clark, 2000). The authors concluded that public policies isolated the school districts from the financial effects of departing students by making up a portion of the lost funding. Another factor helping to mask the effects of charter schools was steadily increasing enrollments, so that total student numbers were rising in these districts even as their percentage of students was falling. As with the research on the effects of voucher programs on nearby district-run schools, it is clear that more work remains to be done.


Although not every charter school is successful, it is clear that some do provide a valuable alternative for students who struggle to succeed in district-run schools. Rather than eliminating charter schools as a whole, a more effective strategy may be to focus on those charter schools that are not succeeding. Toward this end, more attention could be given to the role of the charter autho-rizer, which is critical to the success of a charter school. In fact, researchers have pointed to quality oversight from authorizers as the biggest weakness in the charter sector today (Palmer, Terrell, Hassel, & Svahn, 2006), with one study going as far as labeling the authorization function the “Achilles’ heel” of the charter movement (Lake, 2006, p. 1). If charter schools mature into an effective component of school choice, the role of the authorizer must receive increased attention. This effort is already under way with the establishment of guidelines for effective oversight of charter schools (e.g., National Association of Charter School Authorizers, 2005).

To ensure that charter schools are targeting the neediest populations and thus helping to create more educational equity, some states are considering the establishment of dedicated sponsorship organizations. These organizations would sponsor new charter schools only in high-need areas, relieving nonprofits and postsecondary institutions of the responsibility of authorizing the majority of the new charter schools.

The Process of Choosing

Who is Taking Advantage of School Choice?

Many factors influence the decision to select either a private school or alternative public school as a substitute for a traditional public school. One example is parental educational attainment, with more educated parents being more likely to avail themselves of choice options (Armor & Peiser, 1998; Coleman, Schiller, & Schneider, 1994; Godwin, Kemerer, & Martinez, 1997; Witte, Sterr, & Thorn, 1995). However, it is not clear how this factor actually influences the decision to exercise choice. For example, parental educational attainment could reflect the value that families place on education, with the families that value education more highly being more likely to investigate the educational options and exercise choice (Coleman et al., 1994; Ogawa & Dutton, 1997; Schneider, Marschall, Teske, & Roch, 1998). Alternatively, parents with higher educational attainment generally have more widespread, diverse interpersonal information networks, which provide these parents with greater access to information about the educational choices that are available and thus potentially increase the likelihood that an actual choice would be made (Schneider, Teske, Roch, & Marschall, 1997).

Participation in school choice has also been correlated with parenting variables. For example, parents participating in school choice report more positive parenting practices and more time spent with children, and children in these families report higher levels of familial support (Godwin et al., 1997; Shumow, Vandell, & Kang, 1996). In this case, participation in school choice can be seen as an expression of greater parental involvement.

Other factors include demographic variables, such as family income and race, with some research finding that families exercising school choice tend to be White and/or have higher incomes (Armor & Peiser, 1998; Godwin et al., 1997; Lee, Croninger, & Smith, 1996; Martinez, Godwin, & Kemerer, 1996; Saporito, 2003). This research, combined with the findings cited earlier regarding parental educational attainment, has led to the charge that choice programs are “skimming” the best families from traditional schools, leaving the rest of the families behind (Levin, 1998; Moore & Davenport, 1990; National Working Commission on Choice in K-12 Education, 2003). Other research has refuted this charge, finding that parents exercising choice were more likely to have lower incomes, belong to minority groups, and live in higher risk neighborhoods (Shumow et al., 1996; Witte et al., 1995).

The inconsistency in these findings may be due to the differences in the overall goals of the various choice programs and the accompanying student demographics that are served. With regards to charter schools, there are usually no restrictions on who can apply for admission and generally no admissions requirements, such as grade point averages or test scores. Thus, as noted above, charter schools are more likely to attract disadvantaged families with students that have not been successful in district-run schools. Many voucher programs also tend to attract more disadvantaged families, especially those like the Milwaukee program that are specifically designed to provide an opportunity for these families to send their children to private schools (Witte et al., 1995). In contrast, some choice programs have slightly different goals and serve a different demographic. For example, the goal of the choice program in Philadelphia is to decrease racial segregation in the city’s public schools through the establishment of magnet schools, which are allowed to be restrictive in their admissions practices (Saporito, 2003). Although many students in Philadelphia apply to these magnet schools, only the higher-scoring students are admitted and these students are more likely to belong to wealthier nonminority families. As a result, the magnet schools have become very integrated while the other public schools in Philadelphia, having lost many of their White students to the magnets, contain an even higher percentage of poor minority families. Thus, the Philadelphia program did produce the “skimming” effect, but it can be seen as an inevitable byproduct of the program design.

In sum, it appears as though the design of the choice program will have an impact on the types of the families that exercise school choice. However, it also appears that the factors of parental education and parental involvement do apply across different kinds of programs. Even within the low-SES, minority population targeted by the Milwaukee voucher program, the more educated families were more likely to participate (Witte et al., 1995), indicating that some form of “skimming” was taking place even within a relatively disadvantaged population.

How are the Choices Made?

In addition to influencing who chooses, race also plays a role in the choices that are made. In many studies, distinct same-race preferences have been found. For example, in a study drawing from urban areas in the northeastern United States, White families displayed a marked tendency to avoid schools with high concentrations of minority students, even if the schools possessed good academic reputations; at the same time, racial preferences did not appear among African American families, who tended to view racial diversity as a strength (Saporito & Lareau, 1999). Other research has found that the implementation of school choice has led to increasing racial segregation in Massachusetts (Armor & Peiser, 1998) and Maryland (Henig, 1994), with both White and minority families showing a preference for schools dominated by their own race.

However, these racial preferences do not necessarily indicate the presence of active racism in school choice decisions. When high-quality information about schools is not available from the school districts themselves, parents often fall back on the opinions of friends or relatives, school guidance counselors, their child’s peers, or the media (Neild, 2005). These interpersonal information networks are often racially segregated, and the information passed through these networks is often limited according to the experiences of the families in the network (Schneider 2001; Schneider et al., 1997). As a result, the schools most often recommended by African American parents, for example, are those schools that African American students most often attend. In this way, racial segregation can be self-perpetuating, even in a choice environment. These information networks can also represent a powerful differentiating factor among families, in that parents with wider networks, or networks containing individuals who happen to be school teachers or administrators, often have more and better information regarding school choice options (Neild 2005). As such, the lack of high-quality, easily accessible information about choice options can be seen as contributing to inequity in educational opportunity.

Another potential explanation for the racial preferences cited above is the confounding influence of SES. Wamba and Ascher (2003) suggest that when White parents display a preference for White-dominated schools, they may simply be avoiding poorer schools in disadvantaged neighborhoods, which tend to be dominated by people of color. At the same time, minority families in these disadvantaged neighborhoods may place more value on their children going to school nearby, as opposed to across town, and this may represent a desire for safety and convenience rather than an expression of racial preferences.

The idea that low-SES, minority families have unique values related to schooling that are not shared with wealthier and/or nonminority families has some support in the research literature. For example, Schneider and colleagues (1998) found that minority and low-income parents were mostly concerned about test scores and good discipline, whereas parents with a college education and a higher income were more concerned about moral education and the teaching of values. Other research has argued that a concern for school safety differentiates minority, low-SES families from middle-class White families (Lee et al., 1996). The claim that low-SES and/or minority families may value different things in education has given rise to concern over “sorting,” where these innate preferences bring about greater segregation in schools (Schneider et al., 1998).

In contrast to the research on racial preferences, other studies of school choice have found that academic quality is more important than racial composition (Kleitz, Weiher, Tedin, & Mailand, 2000; Tedin & Weiher, 2004). However, this research relies on survey data from parental interviews, which are sometimes restricted to choosing families (e.g., Kleitz et al., 2000) and in some cases present hypothetical choices that were not actually available (e.g., Tedin & Weiher, 2004). Thus, it is debatable whether (1) the preferences expressed by the parents were actually their true preferences and not impacted by social desirability; (2) the results from choosing families can be generalized to nonchoosers; and (3) the preferences expressed by the parents would result in any active choice given the schooling options that were actually available to them in the real world.


While parental education and parental involvement appear to play a role in determining the families that exercise school choice, there does not seem to be widespread agreement on whether racial preferences impact the choices that are made. The concerns regarding “sorting” are merited, and in response some researchers have argued that choice programs should contain strict guidelines for admission (“controlled choice”) to ensure that schools do not become imbalanced in terms of race and/or SES and thus contribute to greater segregation (Alves & Willie, 1987). Segregation can also be addressed through the design of the choice program itself. For example, voucher programs specifically targeting disadvantaged minority families have succeeded at moving at least some of these students into racially homogeneous private schools, leading to greater overall diversity across the school district (Gill, Timpane, Ross, & Brewer, 2001).

To combat the impact of differentiated and segregated information networks, every effort should be made to educate the public regarding the school choices that are available. Some researchers even advocate for a “forced-choice” policy (e.g., Gill et al., 2001), in which parents are required to make a choice and thus incented to educate themselves on the options that are available.


The goals of the school choice movement are laudable, and some evidence exists that the various choice programs around the country are taking steps toward achieving these goals. However, school choice must be considered a work in progress, with most of the potential of this concept still unrealized. Charter and voucher opponents in state legislatures continue to drive this point home while calling for the elimination of voucher programs and/or moratoriums on new charter school creation. Although it would not be advantageous to move backward on school choice, there is clearly much more that can be done to provide high-quality educational options to families and achieve greater equity in education.