Empirical Research
On Educational Choice
What Are the Implications for Policy-Makers?

     Choice is everywhere in American education. It is manifest in the residential choices made by families with school-age children; it is capitalized in the housing prices found in neighborhoods. Choice also oc curs when parents decide how to care for their preschool-age child and in the consequences of those choices for their youngster's readiness for elementary school. It occurs when parents use their knowledge, skill, and social connections to get their children assigned to one teacher or another, to one program or another within a given school, or to one school or another within a given district. Choice is present when families, sometimes at great financial sacrifice, decide to send their children to private schools instead of public schools. And choice occurs when parents jockey for places in selective public high schools or when stu dents are chosen by lottery for magnet schools with specialized academic programs. In these and many other ways parents and students make choices that influence their educational futures. And in all in stances, these choices--and the options from which to choose--are strongly shaped by the wealth, ethnicity, and social status of parents and their neighborhoods.
     Choice is everywhere, yet for some critics of American education there is not nearly enough, or not nearly enough of the right kind of choice for the right sort of people. These critics usually portray public schooling as a monolithic and unresponsive bureaucracy, driven more by the self-interest of politicians and bureaucrats than by the interests of parents and students. They see increased choice as driving a lethargic educational system toward greater responsiveness and effectiveness for example, Chubb & Moe, 1990; Hannaway, 1991).
     Other critics, however, see choice in American education as serving :he interests of the already privileged and increasing the gaps between hose who are already successful at manipulating the system and those who are not (for example, Moore & Davenport, 1990). They see choice as driving the privileged and less privileged further apart, exacerbating school inequalities. They believe that many people are denied choices when they lack information, money, or accessible options.
     The issue for policy-makers, then, is not whether Americans do have educational choices. Many do. Nor is the question whether they should have educational choices. Virtually everyone in our democratic society--increasingly skeptical of institutional authority--agrees that parents should exercise some control over their educational choices. Rather, the issue confronting policy-makers is what kind of choices pol icy should promote, with what constraints, and for what purposes. As we demonstrated in Chapter 1, the emergence of choice as an educational policy issue can be traced, in part, to attempts by the modern state to accommodate its actions to increasing social diversity and to rising political pressure for more responsive public services. The emergence of choice can also be traced to more specific attempts by urban school systems, since the 1960s, to preserve the participation of the white middle class in public schools and to provide a positive vision of what public schooling can become in the face of increasingly strident criticism. These forces are unlikely to go away in the foreseeable future. While choice may wax and wane as a front-burner political issue, the pressures for attention to choice in American education are deeply rooted in the culture, social structure, and political fabric of this nation.
     Political debates around educational choice involve more than the usual amount of ideological posturing and political polemics--far out-pacing empirical investigation of the actual effects that diverse forms of choice have on individuals and institutions. This book brings together scholars who are doing serious empirical work on school choice in an attempt to begin to build a broader understanding of the institutional and educational effects of this policy. We believe that, since choice is a persistent motif of educational policy, it is important to encourage the creation of a cumulative body of research on the subject and to under-
     stand what that research might mean for policy-makers--local, state, or federal--who continue to confront the issue in its multitude of forms.

FOUR PROPOSITIONS ABOUT EDUCATIONAL CHOICE

     In this concluding chapter, we focus on the policy implications of empirical research on educational choice, drawing mainly on the re search reported in earlier chapters. We have distilled these conclusions into four main propositions. As with all propositions drawn from empirical research, these are tentative and subject to local conditions, but we think it is important to state them succinctly.

Proposition 1: Increasing educational choice is likely to increase separation of students by race, social class, and cultural background.

     Enhanced educational choice is often justified by its advocates as a strategy for improving educational opportunity. In the abstract, this seems a plausible argument: Low-income, minority parents and their children are seen, in this view, as being trapped by circumstances beyond their control in inferior schools. Providing choice to these par ents--in the form of open access to other public schools or vouchers redeemable in private schools--is seen as providing a way for these parents and their children to break free of the constraints of poor schools. Low-income, minority parents, the argument continues, probably have the same aspirations for their children as any other parents. Therefore, the argument concludes, providing choices to these parents will lead the children to better schools and all schools will improve as a result.
     As plausible as this argument sounds, it does not seem to be borne out in the actual effects of choice programs. What does seem to be the case is that both low-income parents and white middle-class parents seem to be favorably disposed toward increased educational choice, and that support for choice is higher among parents who view their children's schools negatively (see Chapters 4 and 6). But this generally optimistic picture of the effects of choice quickly disappears as we exam ine the more specific details of parental attitudes and responses to choice programs.
     Choice appears to have a stratifying effect, by social class and ethnicity, even when it is explicitly designed to remedy inequality. Lee, Croninger, and Smith (Chapter 4) find, for example, that while many parents express strong preferences for increased choice, nearly one-
     third of the parents they surveyed in the Detroit metropolitan area ex pressed no opinion about choice and these parents had significantly less education and family income than those who favored choice. They also find that parents who express a preference for high educational standards are less inclined to support choice and that there is considerably less support for cross-district choice plans that would bring together students from different social backgrounds than for within-district choice plans that would provide constrained choice for students of similar backgrounds.
     Likewise, Wells (Chapter 2) finds that in the St. Louis metropolitan desegregation plan--aimed at moving poor and working-class students from the inner city to the suburbs--youths who chose to leave the inner city and persisted in that choice differ markedly from those who left and returned or those who stayed behind. The highly committed choosers stand out in the degree of parental support for their decisions, in their attitudes toward educational achievement, and in their racial attitudes. Witte (Chapter 6) finds that even in a program heavily targeted toward low-income parents, those who choose the option of publicly financed enrollment in private schools are better educated and more likely to be involved in their children's schooling, both before and after their decision, than parents who did not choose. In their study of magnet school programs, Martinez, Godwin, and Kemerer (Chapter 3) demonstrate that students and parents who choose magnet schools differ systematically from those who do not in parents' education, educational expectations, and involvement. And Henig's study (Chapter 5) reveals quite different preferences between minority and white parents as represented by the characteristics of the magnet schools they choose.
     These findings reinforce an emerging pattern of evidence from other sources (see, e.g., Wells, 1991; Willms & Echols, 1993) showing that, regardless of the design of choice programs, they tend markedly to differentiate choosers from nonchoosers in ways that increase the social stratification of schools rather than reducing inequality. For this reason, authors in this volume urge caution and skepticism about the claims of advocates that choice will equalize educational opportunity. Certainly, minority and low-income parents need better schools, and simple equity would suggest that they ought at least to have the same choices as everyone else, but existing approaches do not yet provide these benefits.
     These findings on the effects of choice should be seen in the context of other research showing that (1) family background is a stronger predictor of children's success in school than school qualities and (2) over
     the past 15 years there has been an increase in the racial, ethnic, and economic isolation of students in American public schools (Fuller & Clarke, 1994; Hanushek, 1994; Orfield, 1993). If the propensity to choose and children's performance in school are heavily influenced by parents' social class and educational background, then it seems plausible to expect that, other things being equal, increasing parental choice will accelerate both the social stratification of schools and the gap in student performance between schools enrolling high concentrations of poor and working-class students versus those with predominantly white, middle-class students.
     These findings should give policy-makers pause. It seems likely, for example, that interdistrict or metropolitan choice plans would provide enhanced opportunities for inner-city parents and students who have a strong achievement orientation but would further isolate parents and students whose expectations are less well formed and whose knowledge of how to take advantage of complex choice options is limited. It also seems likely that within-district choice programs focused on inner-city schools would further separate parents and students based on their educational background, their prior involvement in school, and their knowledge of how to engage complicated choice schemes. Hence, even choice programs that are designed to increase educational opportunity could have the effect of further stratifying parents and students within groups that are already at a disadvantage in the existing structure.
     This sort of conclusion should come as no surprise to those who are knowledgeable about the operation of markets, even the constrained markets represented by the programs described in this book. Markets create product differentiation and segmentation of consumers by pro viding for the free play of preferences around alternatives. Among the distinctions that markets make are those based simply on consumers propensity or ability to choose. It should hardly be surprising, then, that some parents are at a relative disadvantage in understanding whether they have choices, or what those choices might be if they should choose to exercise them. As Henig reveals in Chapter 5, even simple awareness of the term magnet school differed systematically based on parents' ethnicity and social class.
     Nor should it be surprising that sometimes choices are conditioned on the family's economic resources, such as the ability to provide transportation to another school or to miss work and wait in line to sign up for a popular choice. This feature of markets becomes problematic only if there is some overriding public interest in helping people to make "good" choices, as there is when educational choice programs progresively
     separate certain parents and students from access to higher- quality educational programs or to other parents who value such pro grams.
     This connection between educational choice and social stratification poses a serious challenge to policy-makers. Policies that intentionally create or aggravate social stratification based on income, race, and social class should provoke special scrutiny. It is unclear at this point whether the stratifying effects of choice programs are a consequence of the de sign of the programs or simply the inevitable collective result of the individualistic exercise of choice. Some things are relatively clear, how ever. One is that the current design of choice programs, even those designed to enhance equal opportunity, is not adequate to deal with the stratification issue. Another is that if public funds are used in ways that foreseeably increase racial segregation and inequality, they may violate the Constitution. Rectifying the tendency of choice programs to increase social stratification will likely require more governmental intervention rather than less. Insofar as choice advocates see enhanced choice as a way of taking the government out of private educational decisions, attention to the stratifying effects of choice will likely provoke political conflict.
     The evidence also suggests that the designers of choice programs should look hard at the problem of nonchoosing parents. A large part of the stratification problem seems to result from parents and students who simply do not choose, rather than from differing preferences among those who do choose. That is, once parents and students make the decision to choose and actively exploit the opportunities that decision presents, they seem to have preferences that are remarkably similar across race and social class (see Chapter 5). This finding suggests that the design of choice programs should focus more on getting large pro portions of families to make choices, rather than simply catering to the preferences of active choosers. The stratification problem probably requires a careful rethinking of both demand-side and supply-side features of choice programs:

· Would more carefully designed parent information programs yield a higher proportion of active choosers?

· Would requiring all parents to choose, rather than passively making choices available only to active choosers, result in a decrease of social stratification in parent choices?

· Would a gradual introduction of the idea of choice to parents and students, by stressing initial choices of teachers and alternative instructional approaches within schools, for example, yield a higher proportion of active choosers when parents are presented with larger choices, such as the opportunity to choose a school outside their attendance area?

     These questions cannot be answered in the context of existing choice programs. They would require more carefully designed programs than presently exist. Nor are they questions that are amenable to simple, ideologically determined answers. They would require considerable attention to details in the design of choice programs (see proposition 3, below). In the meantime, our advice to policy-makers is to treat with considerable skepticism the claim that educational choice enhances equality of opportunity.

Proposition 2: Greater choice in public education is unlikely, by itself, to increase either the variety of programs available to students or the overall performance of schools. Coupled with strong educational improvement measures, however, choice may increase variety and performance.

     Another common argument for enhancing school choice is that it will increase the quality and diversity of educational offerings and, consequently, boost student achievement. One version of this argument is that choice increases competition among schools, leading to a better fit between the preferences of parents and educators, a greater focus on learning in new school organizations, and hence more learning on the part of students. Another version of the argument is that bureaucratic administration distracts teachers and principals from their central mission and shifts their energy to the narrow demands of political constituencies. Substituting market incentives for bureaucratic controls reduces the role and costs of urban school bureaucracy, and refocuses the energy of educators on classroom innovation and student learning.
     The evidence on this argument is mixed. Lockheed and Jimenez (Chapter 7) find in their cross-national study that achievement differences between public and private schools, after carefully controlling for the social background of students, favor private schools. They also find that while public and private schools are similar in resource levels, they differ both in the locus of management control--private schools focus more decisions at the school site--and in their degree of attention to academic learning--private schools focus more time on academic subjects. This evidence is consistent with the findings of Coleman, Hoffer, and Kilgore (1982), Chubb and Moe (1990), and Hannaway (1991), who highlight differences in achievement results and organizational characteristics between public and private schools.
     Other analysts have suggested, however, that aggregate achievement differences between public and private schools are relatively small, that they may be explained by unmeasured motivational and other background differences, and that the variation among schools within the public and private sectors in both achievement results and organizational characteristics is wider than that found between the two sectors (e.g., Willms & Echols, 1993). This suggests that it is more important to understand the organizational characteristics of effective schools, whether public or private, than to focus on public-private differences (Murnane, 1985).
     Witte's study (Chapter 6) of Milwaukee's public voucher program for poor parents choosing private schools suggests that achievement effects are weak and variable and favor neither public nor private schools. He also finds a relatively high turnover rate--about one-third-- among students whose parents have chosen private schools; this suggests that, at least in the early stages of the program, many parents see little advantage of private over public schools once they have experienced private schools.
     The evidence from public school choice programs is equally ambiguous. Henig (Chapter 5) finds, in his study of Montgomery County magnet schools, that the curricular themes of magnet programs seem to have less influence on parental choice than other characteristics, such as the school's ethnic composition and characteristics of teachers. Henig concludes, "parents may not be looking for particular instructional themes and styles so much as for the kind of energy, creativeness, and extra resources that some schools build around their magnet programs." The major premise of magnet school programs is that parents of all races, income levels, and social backgrounds can be drawn to schools by developing schools around distinctive curricular and instructional themes. Henig's study suggests that active choosers may be basing their decisions on other grounds and that attempts by public school systems to engineer choices by creating alternatives may not work under permissive conditions like those found in Montgomery County. It is also possible that, despite the differences in labels among magnet schools, there may be very little difference in the actual curriculum and instruction within the schools. Hence, parents may be making reasonable choices by focusing on characteristics other than curricular themes. One of the earliest studies of public school choice, in the Alum Rock voucher experiment, found little evidence of systematic variation in instruction among schools that had nominally different themes; it also found that parents tended to choose schools based on characteristics, such as location, that had little to do with schools' curricular themes (Bridge & Blackman, 1978; Elmore, 1990).
     These ambiguous findings about the relationship among choice, school innovation, and student performance lead us to the conclusion that introducing choice will not, by itself, result in large changes in educational programs or student performance. It does make sense, however, to think about choice policies operating in tandem with other educational improvement initiatives to foster variation in educational programs and to focus school leaders on student performance. Knowledgeable designers of public school choice programs have, for a long time, argued that choice plans need to be combined with policies that reinforce high expectations that all students will achieve and that pro mote the systematic development of alternative instructional strategies, rather than simply relabeling existing strategies (Fliegel, 1990). The evidence on differences between the bureaucratic environments of public and private schools also suggests that dramatically streamlining and focusing central school bureaucracies on supporting high-quality instruction in schools could result in more attention to student learning.

     School choice is often portrayed in policy debates as an either--or issue: Either parents and students will have choices or their choices will be subject to bureaucratic control. Choice advocates often argue-- incorrectly--that present public school systems do not permit choice and that providing choice requires eliminating public controls on enrollment decisions and providing money directly to parents and students. Opponents of choice often argue--also incorrectly, we have found--that most choice plans have the same invidious effects, regardless of how they are designed.
     The world of school choice portrayed in this book is rather different from this either--or picture. It is a world in which specific decisions in the design and implementation of choice programs have consequences for parents and students that are sometimes consistent and sometimes inconsistent with the intentions of policy-makers. Design details matter, and we are only beginning to understand how and why they matter.
     One major design issue that we previously mentioned is the degree to which choice programs are designed simply to provide an option for those who wish to choose--what might be called "option-demand" programs--or to require active choice of all parents and students--what
     might be called "universal-demand" programs. All of the programs studied in this volume are option-demand programs, and we have seen that such programs create, by their design, a large category of non- choosers who are disproportionately poor and minority and who tend to engage their children's schools less actively. So the design of the program itself disadvantages certain parents, by enhancing opportunities for parents who are already active choosers and disregarding those who are less inclined to be active choosers. While option-demand programs may seem reasonable to school administrators and policy- makers, because they appease vocal and active clients and they create highly popular schools, they accentuate differences in opportunity between choosers and nonchoosers. Hence choice programs should include not only strategies that offer opportunities for already active choosers but also strategies that attempt to increase the proportion of active choosers and decrease the proportion of nonchoosers.
     Another design issue has to do with the educational content of the options made available to parents in choice programs. From the studies reported in this volume it appears that the designers of choice programs fail to think either very systematically or very deeply about the class room-level content of the options they are offering parents and students. In the metropolitan St. Louis case described by Wells (Chapter 2), the parties settling the desegregation case assumed that, by definition, suburban schools would be preferable to inner-city schools for urban minority youth. In fact, judging from the reports of students, these schools seemed neither well prepared to accept students from the inner city nor particularly aware of the additional responsibility that the arrival of these students entailed. There was, in other words, no fully developed educational logic to the program, only a kind of logistical logic that entailed the movement of large numbers of students from one place to another.
     The studies by Henig (Chapter 5) and Martinez, Godwin, and Kemerer (Chapter 3) suggest another dimension of the educational content issue. There is evidence in both studies that when educators deliberately design programs around coherent educational approaches, parents are attracted to these novel and distinct options, and these programs seem to have positive educational effects. Henig, for example, notes the appeal of foreign-language immersion at Oak View School in Montgomery County to both white and minority parents, while parents chose along racial lines when it came to schools with blurry identities. The Martinez team's study shows that when school administrators develop options that fit cultural and educational preferences of parents, families choose them and they may experience positive effects on their children's learning,
     notwithstanding increased stratification by family social class and prior school involvement. This latter effect stems from school authorities' insistence that opportunity to enter a multicultural school depends upon the child's prior academic performance.
     For the most part, though, there is little evidence that educators focus very systematically on the supply side of educational choice. The studies by Henig (Chapter 5) and Blank, Levine, and Steel (Chapter 8) suggest that while magnet school programs carry different labels, there is not much evidence that would lead one to predict how they would differ in their curriculum, teaching, and achievement outcomes. In the absence of such genuine differences, it seems highly unlikely that parents would base their choices on program characteristics that hold direct achievement benefits for their children.
     The studies in this volume suggest that the design and implementation of choice programs is a fertile area for future empirical work and an area on which policy-makers should focus their scrutiny. The notion that introducing demand-side choice will lead more or less automatically to the creation of distinctive supply-side educational alternatives seems implausible. Parents on the whole choose programs based more on their social composition and convenience than their educational content, except where educators have invested great care in developing distinctive school options. Furthermore, since parents and students with the least social capital seem also to be the ones who are least likely to engage in active choice, there are few demand-side incentives in choice programs for educators to engage in the deliberate design of programs that appeal to, and work well for, the most disadvantaged students. So it seems unlikely that choice, by itself, will stimulate creativity and improvement in the development of new, more effective educational programs. The problem seems to lie in the fact that the designers of choice programs have focused most of their attention, in all but a few cases, on demand- side issues, such as who gets to choose and how choices will be coordinated, rather than on crucial supply-side details, such as how schools and classrooms actually differ.

Proposition 4: Context matters in the design and implementation of school choice policies.

     The local cultural and institutional context of choice has an important impact on its consequences for different groups of parents and students. Market theories, for the most part, assume that consumers preferences can be described and aggregated in relatively simple ways and that all consumers are more or less engaged in the same process
     of rational search for value-maximizing choices, operating with similar information and few practical constraints. The studies in this volume cast considerable doubt on this simplistic view. Different groups of educational clients seem not only to have very different predispositions to choose, they also seem to bring very different cultural and social assumptions to the choices they are expected to make.
     Wells (Chapter 2) finds distinctive differences among individuals in the same group in the cultural capital they bring to educational choices. For example, minority parents and students who did not take advantage of the St. Louis metropolitan transfer option were more inclined to be distrustful and fearful of whites, less inclined to regard whites as their superiors, and less inclined to accept the view that schooling is the road to upward mobility. These attitudes are formed from a lifetime of experiences with individuals and institutions, and they are not likely to be easily changed by one novel intervention like school choice. Non- choosers might, in other words, be actively expressing cultural values, rather than simply failing to express a preference for something we would like them to choose. This applies to suburban whites--seeking to preserve culturally familiar school settings--as much as to working-class Hispanics who seek culturally consonant (nonmainstream) school set tings. Kane's evidence on Pell Grants (Chapter 9) suggests that despite billions of dollars spent on "college vouchers," the constrained expectations and habitual pathways of low-income high school graduates have proven stronger than economic incentives.
     Lee, Croninger, and Smith (Chapter 4) show how the social and demographic context of choice exerts considerable influence on the way people understand their options. The Detroit metropolitan area is a largely black inner city surrounded by largely white and considerably more affluent suburbs. Racial isolation by political jurisdiction is extreme due to the nation's most segregated housing market. In this con text, their data suggest that choice comes to be construed as a possible route of escape from the inner city by some minority parents and as a possible threat to homogeneity and predictability by white parents. There is, for example, a negative relationship in their survey data between a desire for high academic standards and support for choice. One would predict the opposite relationship if parents saw educational choice as a way of expressing a preference for high-quality education. This finding only makes sense if one associates choice with a movement away from homogeneous, racially identifiable "high-quality" schools. Whether choice is perceived as a device for improving education, or as a threat to established cultural and social boundaries, depends on the specific social context in which the issue arises.
     While it appears that context matters a great deal in the design and implementation of choice policies, it is also clear that we know very little about the interaction between context and policy in this area. The studies in this volume have only scratched the surface of this interaction. It is unlikely that we will be able to reliably predict the effects of choice programs on different local populations in the absence of evidence and deeper understanding of these contextual forces.

SCHOOL CHOICE AND EDUCATIONAL IMPROVEMENT

     The advocates of choice, like all reformers, are optimists regarding their own proposals and pessimists about competing reform proposals. Hence debates around choice tend to be couched in simplistic terms. Choice is alleged to remedy the defects of excessive bureaucracy and to make schools more responsive and effective, without incurring any significant social costs. We think the evidence suggests a more cautious and skeptical view of choice as an instrument of educational reform. Enhanced choice does seem to activate certain parents and students; it does not, however, activate certain others. It may produce useful innovations in previously unresponsive systems. And, other things being equal, it seems to increase the social disparities between those who choose and those who do not. There is strikingly little evidence that enhanced choice triggers the kind of educational improvement on the supply side that its advocates predict and little evidence that active choosers are looking for distinctive educational programs when they make their choices. There is some suggestive evidence that cultural and political contexts have an impact on the effects of choice programs, and we currently understand little about these effects.
     What do these overall findings suggest about the role that enhanced educational choice might play in the long-term improvement of American education? As we noted in the beginning, the issue presented to policy-makers is not whether there should be educational choice but what kind of choices public policy should promote, within what constraints, and for what purposes. In democratic societies the role of policy- makers is less one of deciding whether or not to grant choice than of how to guide, orchestrate, and augment the choices that are available to parents and students.
     We have suggested that using choice as an instrument of improvement requires considerably more governmental action and involvement than the rhetoric of choice advocates leads one to believe. If, for example, we want choice to produce more equitable access to better education
     for all students, then it is clear that we have to design policies that do a more effective job of engaging the large proportion of parents and students who currently appear to be nonchoosers. In the absence of serious progress on this front, it is unlikely that enhanced choice will produce anything resembling more equitable access. If, for example, we want choice to result in the improvement of school and classroom programs, then it seems clear that we have to dedicate much more effort than we have in the past to designing and implementing distinctive educational alternatives. In the absence of serious progress on this front, it is unlikely that choice will do anything other than simply move high achievers around from one school to another, mistaking the effect of concentrating strong and motivated students for an effect of the school or the choice system. If, for example, we want to develop choice systems that are sensitive to the cultural and social differences among groups of parents and students, then it seems clear that we have to understand a good deal more than we presently do about how individuals understand and construct their choices from their cultural backgrounds. In the absence of progress on this front, it seems likely that we will continue to produce choice systems that will work well for people who already have choices but will fail to engage those who do not.
     Using choice to improve education, then, is a serious and complex task that is not easily amenable to guidance by simple ideological principles. We have suggested that the idea that choice will produce better results with less public authority or bureaucracy is highly problematic. Policy -makers should take seriously both the distributional impacts of choice and the achievement effects for specific groups of students. Policy-makers are accountable not only for the beneficial effects of choice policies on those who choose, but also for the detrimental effects on those who, for whatever reason, fail to choose. And policy-makers are accountable not just for the enhanced consumer satisfaction of people who are already active choosers, but also for the overall improvement of opportunity and performance for all students. These democratic responsibilities--and the public interest in both fairness and school effective ness--require policy-makers to be skeptical and deliberate in their use of choice as an instrument of reform. They should ask tough questions about the effects of choice policies and develop hard evidence on which children benefit and which do not.

REFERENCES

Bridge, R. J., & Blackman, J. (1978). A study of alternatives in American education:

Vol. 4. Family choice in education. Santa Monica, CA: Rand Corporation.

Chubb, J., & Moe, T. (1990). Politics, markets and America's schools. Washington, DC: Brookings Institution.

Coleman, I., Hoffer, T., & Kilgore, 5. (1982). High school achievement: Public, Catholic, and private schools compared. New York: Basic Books.

Elmore, R. (1990). Choice as an instrument of public policy: Evidence from education and health care. In W. Clune & I. Witte (Eds.), Choice and control in American education: Vol. 1. The theory of choice and control in American education (pp. 285--318). New York: Falmer.

Fliegel, 5. (1990). Creative non-compliance. In W. Clune & I. Witte (Eds.), Choice and control in American education: Vol. 2. The practice of choice, decent ral ization, and school restructuring (pp. 199--216). New York: Falmer.

Fuller, B., & Clarke, P. (1994). Raising school effects while ignoring culture? Review of Educational Research, 64(1), 119--157.

Hannaway, J. (1991). The organization and management of public and Catholic schools: Looking inside the black box. International Journal of Educational Research, 15, 463--481.

Hanushek, E. A. (1994). Making schools work: Improving performance and controlling costs. Washington, DC: Brookings Institution.

Moore, D., & Davenport, 5. (1990). School choice: The new and improved sorting machine. In W. Boyd & H. Walberg (Eds.), Choice in education:
     Potential and problems (pp. 187--223). Berkeley, CA: McCutchan.

Murnane, R. (1985). Comparisons of private and public schools: Lessons from the uproar. Journal of Human Resources, 20, 263--267.

Orfield, G. (1993). The growth of segregation in American schools: Changing patterns of separation and poverty since 1968. Alexandria, VA: National School Boards Association, Council of Urban Boards of Education.

Wells, A. S. (1991). Choice in education: Examining the evidence on equity. Teachers College Record, 93, 156--173.

Willms, D., & Echols, F. (1993). The Scottish experience of parental choice in schools. In E. Rasell & R. Rothstein (Eds.), School choice: Examining the evidence (pp. 49-68). Washington, DC: Economic Policy Institute.