LIBRARY Michigan State University This is to certify that the dissertation entitled PARENTS’ VIEWS OF SCHOOL CHOICE: AN UNEXAMINED PERSPECTIVE presented by COURTNEY ANN BELL has been accepted towards fulfillment of the requirements for the Doctoral degree in Curriculum, Teaching, & Educational Policy Major Piofessor’s Signature 3/15/10“! Date MSU is an Affinnative ActiorVEquaI Opportunity Institution _._.--n-o-n-.----n-n—.-.-.—.- PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE MAR 0 A. 200M 2 7 u 37 W}? a? 9% 122007 2/05 c-JcTfiCIDateominda-ms PARENTS’ VIEWS OF SCHOOL CHOICE: AN UNEXAMINED PERSPECTIVE By Courtney Ann Bell A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Teacher Education 2004 ABSTRACT PARENTS’ VIEWS OF SCHOOL CHOICE: AN UNEXAMINED PERSPECTIVE By Courtney Ann Bell Scholars, using rational choice theory, have made predictions about how parents will behave given a more open market for schooling. The prediction that parents will choose better schools has met with the reality that few parents actually move their children out of failing schools. I argue that the poor predictive power of current models is the result of methodological and theoretical problems. In order to address these theoretical and methodological problems, this study took an interpretive stance toward answering the question, “Why do parents choose particular schools for their children?” The study also sought to address the more specific question, “Why do parents choose failing schools for their children?” The study was conducted in a large, Midwestern city in which the majority of city residents are African-American. Over the course of eight months, 48 parents were interviewed twice before their children began school in the fall and once afterward. The group of parents was stratified by grade level (5'h or 8‘”), family income, and school status (failing or non-failing). Modified analytic induction and descriptive statistics were used to analyze different dimensions of the data. This study shows that traditional explanations - disinterest, lack of information, and constraints - don’t explain parents’ choice of failing schools. Having eliminated the traditional explanations for choosing failing schools, I argue that we must investigate parents’ preferences if we are to more accurately predict their behavior. Through a case-based analysis of parents’ preferences, the study demonstrates that preferences are bound by parents’ cultural understandings of their children, schools, and communities. Four dimensions of parents’ preferences are critical to the choice process: academic achievement, child development, school quality, and community. The analyses show that preferences are not exogenous to choice models and ought not be treated as such. Based on the preference analysis, I propose a model that predicts parental choice of schools. The model, which includes both preference and constraint variables, relies heavily on revealed, rather than stated, preferences. Because the sample size is too small to test, the model’s potential applicability is considered for different subsets of the sample. Based on these analyses, I conclude that the model is a reasonable first step toward more accurate predictions of parents’ choice behavior. The study concludes that parents’ choices are not the unbounded, free-will, any- school-you-desire, kind of choices that free market advocates suggest. Rather, parents’ choices are best understood within social and historical choice trajectories. These trajectories, which rely on preferences, vary systematically. The implications of these trajectories and their systematic differences are considered. C0pyright by COURTNEY ANN BELL 2004 This dissertation is dedicated to all of the children in this country who do not find challenge, inspiration, and love behind the classroom door. ACKNOWLEDGEMENTS When I read other graduate students’ acknowledgements I thought the students were just being nice, almost like giving the teacher an apple at the end of the school year. But having gone through this process myself, I realize acknowledgements have almost nothing to do with niceness and almost everything to do with gratitude. The process of learning is painful and uncertain. The people I thank here are the people who most directly helped me embrace that pain and uncertainty. They demanded I face the challenge and they stood by my side as I struggled. I am humbled by their faith and generosity. First and most importantly, I must thank the parents and children of Gaston. Their willingness to explain their thinking and share their concerns with me has made this work possible and will hopefully, lead to better choice policy. Their efforts to give their children better lives are truly inspiring. Gary Sykes, my advisor from the minute I set foot on campus, has taught me what a scholar looks like -- a devoted father and husband, an avid athlete, a generous reader, and a sharp critic. I hope that some of his Renaissance approach to life has rubbed off on me and is visible in the pages of this dissertation. I know I’ll be saying it for years to come, thank you Gary. Suzanne Wilson’s teaching knows no boundaries - fi'om cooking to understanding higher education to learning to write - Suzanne’s sharp mind and deliberate scaffolding have helped me learn about myself, my work, and the world around me. Her fingerprints are literally, all over the pages of this dissertation. I cannot vi thank Suzarme enough, for walking beside me on what can feel like a lonely road of scholarship. I hope that I treat others with the care and respect she has treated me. My committee, through their attention and support, has dramatically improved this work. David Plank has argued with me often, but always in the spirit of making me a better scholar. David Labaree’s feedback on my writing early in my graduate career both hmnbled me and pushed me to think more carefully. His encouragement and advice at critical junctures in my career buoyed me up and helped me believe I was capable. Susan Florio-Ruane’s questions and alternative interpretations of data have taught me to think more deeply and more carefully about parents’ meanings. And David Arsen’s demands for conceptual clarity, his careful reading of my writing, and his intellectual generosity have improved this dissertation immeasurably. I am lucky to have been supported by such intelligent scholars and giving people. Other Michigan State faculty deserve my thanks as well. Lynn Fendler’s expansive knowledge, commitment to the intellectual development of her students, and flat-out spirit has inspired me. Ed Wolfe was always willing to read my work, give me advice, and perhaps most importantly, figure out how to fix my SAS code. He is a model of scholarly generosity. Mike Lipsey happily shared his love of geography with me and made the mapping analyses of the dissertation possible. And Chris Wheeler has Heated me like his own, supporting both my work and life experiences. My graduate student colleagues have made this two year process sustainable. My writing group, Mark Olson, Andrew Shouse, Jodie Galosy, and Ninna Roth, provided critical supportive feedback from proposal to final draft. Raymond Mapuranga and Linda Chard’s detailed and prompt answers to my questions were both insightful and generous. vii And finally, Kevin, Emily, Andy, Jodie, Richard, Lisa, Debbi, Benita, Stacey, Bebe, and Matt supported me in ways too numerous to count. Thank you friends, for not letting me take myself too seriously. The research reported in this dissertation was made possible (in part) by a grant from the Spencer Foundation. Though the data presented, the statements made, and the views expressed are solely my responsibility, the Spencer Foundation’s intellectual and financial support improved the project. And finally, a special thank you to my husband, Mark Olson. There is no one who knows more about this dissertation, the project, or the passions that underlie both of those endeavors. Mark has read and reread. He has asked questions and suggested changes. But most importantly he has believed. When I didn’t believe in myself and the work, he did. He was there with enthusiasm, laughter, beer, and the occasional tube of cookie dough. There is really no way to thank such deep and loving actions. I can only say that I know how lucky I am. viii TABLE OF CONTENTS List of Tables ........................................................................................ xii List of Figures ........................................................................................ xiii Chapter One: Background and Theoretical Framework ..................................................... l Situating the Study ........................................................................................................... 1 Choice, Choice Everywhere ......................................................................................... 1 The Historical Background of School Choice ............................................................. 1 Current Work ................................................................................................................... 6 Theoretical Framework .................................................................................................. 10 An Overview of the Dissertation ................................................................................... 15 Chapter Two: Method ....................................................................................................... 16 Study Design .................................................................................................................. 16 The Context .................................................................................................................... 17 Sampling ........................................................................................................................ 18 The List of Potential Schools ..................................................................................... 22 Determination of School Status ................................................................................. 22 Contacting Schools .................................................................................................... 24 How Does the Sample Compare to the City? ............................................................ 26 Gaining Access .............................................................................................................. 27 The Pilot Study .............................................................................................................. 29 Data Collection .............................................................................................................. 31 Methods ...................................................................................................................... 31 Instruments ................................................................................................................. 31 Data Analysis ................................................................................................................. 39 Chapter Three ............................................................................................................. 40 Chapter Four .............................................................................................................. 40 Chapter Five ............................................................................................................... 40 Conclusion ..................................................................................................................... 41 Chapter Three: The Process of Choice: The Case for Preferences .................................. 45 The Conceptual Landscape of the Choice Process ........................................................ 45 The Results of Choice .................................................................................................... 49 School Status .............................................................................................................. 49 Satisfaction ................................................................................................................. 50 Satisfied Parents and Failing Schools ........................................................................ 52 The Process of Choice ................................................................................................... 52 Predetermination ........................................................................................................ 53 Search ......................................................................................................................... 55 Choice ........................................................................................................................ 70 Parents’ Actions: Windows Into Parents’ Engagement ................................................ 77 School Discovery ....................................................................................................... 78 School Visits .............................................................................................................. 80 ix Potential and Choice Sets ........................................................................................... 82 Conclusions .................................................................................................................... 84 Chapter Four: The Social Construction of Preferences .................................................... 88 Summary ........................................................................................................................ 88 Propositions and Preferences ......................................................................................... 88 Proposition One: Children Influence Their Own Academic Achievement .............. 93 Pr0position Two: My Child is Typical .................................................................... 103 Proposition Three: I Know What Quality is for My Child ..................................... 117 Proposition Four: This Is My Community .............................................................. 130 Propositions Overlap ........................................ 140 The Elements of Preferences ....................................................................................... 144 Preferences are Co-produced ................................................................................... I44 Preferences are Relative ........................................................................................... 146 Preferences are Social and Historical ...................................................................... 148 Conclusions .................................................................................................................. 150 Chapter Five: A Model of School Choice ....................................................................... 152 Previously Advanced Models ...................................................................................... 154 The Empirical and Conceptual Landscape of Models for School Choice ............... 155 A Model of Choice ....................................................................................................... 165 Constraints ............................................................................................................... 167 Preferences ............................................................................................................... 173 Model Predictions ........................................................................................................ 1 81 Robustness of the Model .............................................................................................. 183 Limitations of the Model ............................................................................................. 188 Chapter Six: Conclusions and Implications .................................................................... 190 Review of the Argument .............................................................................................. 190 Chapter 1 .................................................................................................................. 190 Chapter 2 .................................................................................................................. 191 Chapter 3 .................................................................................................................. 191 Chapter 4 .................................................................................................................. 192 Chapter 5 .................................................................................................................. 192 Conclusions .................................................................................................................. 193 Trajectories of Choice .............................................................................................. 194 Applications of Rational Choice Theory to Parental Choice ................................... 201 Implications .................................................................................................................. 203 Choice is Unlikely to Deliver Equity and Improvement ......................................... 203 How Do We Improve Schools? ............................................................................... 205 Limitations and Additional Analyses ........................................................................... 206 Future Work ................................................................................................................. 21 1 Appendix A: Correlation Matrix ..................................................................................... 214 Appendix B: Interview Instruments ................................................................................ 216 Appendix C: Example of Family Choice History ........................................................... 223 Appendix D: Data Analysis Categories for Interviews ................................................... 227 Appendix E: Example of First Interview Summary ....................................................... 230 Appendix F: The Excluded Variables ............................................................................. 233 Appendix G: Description of Model Variables ................................................................ 236 Appendix H: Study Participants ...................................................................................... 239 References ........................................................................................................................ 243 xi 2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.1 5.1 5.2 5.3 5.4 LIST OF TABLES Description of Parent Sample ............................................................ 21 Frequencies of Gaston Residents’ and Study Participants’ Income, Race, and Education .................................................................................... 27 Data Collection and Instrument Development Schedule .............................. 33 Mrs. Gunnison’s Choice Set .............................................................. 58 Frequencies of Schools in Sets ........................................................... 63 Mrs. Borden’s Choice Set ................................................................. 66 Frequencies of Search Characteristics ................................................... 69 Reasons for School Placement in Unsuitable, Suitable, Eligible, Unattainable, and Attainable Sets .............................................................................. 74 Frequencies of Discovery Methods in Choice, Suitable, Eligible, and Attainable Sets ............................................................................................ 78 Frequencies of Schools Visited in Choice, Suitable, Eligible, and Attainable Sets ............................................................................................ 81 Potential and Choice Set Size, Quality, Cost, and Selectivity ........................ 84 Relationship Between School Status and Constraints ................................. 86 City and Suburban Choice Sets and Choice Schools ................................. 138 A Model of School Choice .............................................................. I66 Predicted Relationship Between Choice of a Non-F ailing School and Model Variables ................................................................................... 182 Education, Income, Race, and Final School Status for Sample and Two Subsets ...................................................................................... 185 Point-biserial Correlations and Chi-square Relationships of Model Variables and Non-failing Schools for the Sample, City Residents, and Public Schools ........ 186 xii LIST OF FIGURES 3.1 The Open Search Procedure ............................................................... 6O xiii Chapter One Background and Theoretical Framework Situating the Study How we ever got the idea in this country of telling people where they had to go to school, I’m not sure I know. I think it’s an aberration, an alien thought, really un- American. Lamar Alexander Choice, Choice Everywhere Choice is the mantra of the West. We choose where we live, we choose which car we want to drive, we choose which fast food restaurant we will support. It is not surprising that choice is so fundamental to our capitalist democracy. We view it as both our responsibility and our right to choose. It is a part of our collective fiber. Perhaps in response to, or perhaps in concert with, the collective commitment to choice, scholars of all stripes have concerned themselves with choice for many years. Philosophers wrestle with questions about the nature of choice. Psychologists consider the cognitive demands and influences of choice. Economists investigate the results and determinants of choice. From Descartes to Adam Smith, inquiries, essays, and experiments about choice abound. The Historical Background of School Choice This study investigates a particularly contentious variety of choice - school choice. School choice is not a new idea. According to Coons and Sugarman (1978), it dates as far back as 1792 and Thomas Paine’s text, The Rights of Man. But the notion as we know it now was popularized in Milton Friedman’s 1962 treatise on choice, Capitalism and Freedom. Friedman & Friedman (1980) argued, “One way to achieve a major improvement [in public schools], to bring learning back into the classroom. . .is to give all parents greater control over their children’s schooling, similar to that which those of us in the upper-income classes now have.” (p. 150). Friedman’s reference to the choice upper-income parents have was a reference both to the choice enjoyed as a result of being able to afford private and religious schools and the more common form of choice, residential choice. Residential choice is the idea that parents choose schools through the purchase of a home in a particular neighborhood. F riedrnan argued for school choice on two grounds, both of which remain salient in the political arena today. First he argued that all parents have the right to choose and we should support that right. And second, he argued that choice would work as a reform strategy, improving the overall quality of public schools. The Argument for Choice. Building on Friedman’s assertions, contemporary advocates argue that many of our schools, particularly those serving children of color and poor children, are failing. Schools are not responsive to change for at least two reasons. They are large bureaucratic organizations made up of complex structures that are meant to frustrate concerted change (Chubb & Moe, 1991). They have a monopoly over education (Hill, 1997). Together, these features result in an unresponsive institution with a captive clientele. No matter how poorly a school performs, children appear at the front door every fall. The children have no other choice. There is nowhere else for them to go. Advocates argue that children should not be doomed to educational failure because of where they live or because of lethargic, self-satisfied institutions. They argue that parents need to be given the freedom to send their children to good schools, not simply the neighborhood school. Parents, not government, should decide where children go to school. The argument for school choice conceptualizes parents as consumers. Advocates argue that parents gather information about schools’ test scores, specialized programs, and teachers (Armor & Peiser, 1997). Parents weigh the costs and benefits of attending certain schools and “vote with their feet.” Advocates argue that, given the opportunity, parents will choose academically superior schools. “Voting” will send market signals to failing schools and those schools will be forced to close or fimdamentally change. This argument promises improvements for both parents and schools. Because parents will choose schools they prefer, there is more likely to be a match between families and schools. The empowerment that comes from choosing and the improved match is likely to produce more satisfied and involved parents (Hill, 1997). Schools are likely to improve because they will be forced, through competition, to be responsive to parents’ concerns. Schools that are unable or refuse to respond will disappear and both schools and parents will be better off. The Arguments Against Choice. The critics of school choice have been unable to unravel the intoxicating logic of choice, so it behooves us to parse that logic. It has three parts: a statement of the problem (i.e., urban schools are failing), an assertion of parents’ behavior (i.e., they will choose the best schools for the children), and a set of deductive hypothetical results (i.e., parents will be satisfied and bad schools will close or improve). Generally, critics accept (or at least do not reject) the statement of the problem. They acknowledge that there are some schools, particularly urban schools, which do not serve children well. And for the most part, critics do not argue with the hypothetical benefits of choice. It could produce more satisfied parents and improved schools. Instead, critics focus their attention on parents’ behavior; behavior they describe as messy, human, and constrained. In contrast to choice advocates, choice critics portray parents as social actors rather than maximizing consumers. Critics argue that parents will not behave in the ways advocates have argued. For example, critics often take issue with the idea of parents choosing schools based primarily on academic criteria. They argue that parents care about academics, but they also care about other characteristics of schools such as location, the presence of their children’s fi'iends, or special non-academic features such as a marching band or a football team. Based on their children’s interests and needs, parents may reasonably consider different dimensions of schools and some parents may not consider academic dimensions at all. Critics conclude that, if parents choose schools for non-academic reasons, the argument falls apart and choice will not improve the educational system or the lives of children. Critics argue that, even if parents acted like consumers, there are practical considerations that would not allow them to take advantage of choice. Transportation for example, may limit options for low-income parents who have fewer resources to devote to taking a child across town to a better school. Parents may lack the information necessary to make an informed choice about schools. Information is costly to obtain. It takes time and even when parents are able to locate information sources they often are not user friendly and do not help parents sort through the complex, detailed information provided by the source (e.g., the State of Michigan’s Standard and Poor’s school rating website is one such source). Finally, social concerns might cause parents to stay in a “less desirable” school. Wells (1993) found that parents are sensitive to sending their children to schools in which they are the racial minority. They worry about their physical and emotional safety in these unfamiliar environments. This is particularly important if urban parents are supposed to use choice to send their children to wealthier, typically more white, suburbs. Finally, critics argue that ultimately, all parents will not choose schools. The most involved, interested parents will use choice. Choice will select for the parents and students who are most likely to agitate for change, thus leaving behind the children least able to advocate for themselves (through their parents). Critics argue that the government has a responsibility to attend to all children, not only the ones whose parents can advocate for them. Choice, the critics argue, will further stratify an already unequal system by allowing the government to abdicate its role in providing education for all children. Parental Choice. School choice hinges on parental behavior. If parents don’t choose “better” schools, choice simply does not work. Yet reports suggest that the logic of advocates’ assertions is not without its paradoxes. Through choice, charters, magnet schools, and vouchers, parents now have many more schools to choose from. And yet, only a tiny percentage take advantage of these expanded choices. Many Florida parents who used vouchers in September to remove their children from “failing” schools, have now returned to those same schools (Grench, 2002). Schemo (2002) and Robelen (2002) report that, of the 3.5 million children who are eligible under the No Child Left Behind Act of 2001 (N CLB) to move out of failing schools, only two to three percent have done so (N CLB, 2001). The national picture varies by region but is consistent: few parents who are eligible to opt out of failing schools have done so (Asimov, 2003; Gupta, 2004; Moses, 2004). If we are generous and assume that NCLB is new and more parents will participate in the long term, we might predict that eventually 10-12% of parents will move their children (quadrupling the numbers who participate now). This would still mean that the vast majority of parents choose to leave their children in “failing” schools. What explains this puzzling phenomenon? Current Work This study addresses the puzzle by investigating the question, “Why do parents choose particular schools for their children?” Though research on parental choice has touched upon this question, it has failed to help us understand what parents say and do. The problems are both empirical and theoretical. Empirical work on choice relies primarily on survey research (Armor & Peiser, 1997; Schneider, Teske, & Marschall, 2000; Witte, 2000), resulting in a surfeit of “single shot case studies” (Campbell & Stanley, 1965). Although these data shed descriptive light on important issues, they do not allow researchers to explore and understand the dynamic, situated aspects of parental reasoning. Researchers are limited to asking parents to retrospectively rank factors that influenced their choice of school (e.g., reputation, teachers, resources). The problems with self-report data and memory have long been a topic of discussion among both social science researchers and biographers. As Campbell and Stanley (1965) note, the data “provide very limited help, since the rival sources of difference are so numerous”(p. 7). In addition to the problems of cross-sectional data and hindsight bias, there is the problem of process specification. Research on choice relies on a poorly specified choice process. Bast and Walberg’s (2004) explanation of the how parents choose is representative of the literature more generally. They say: “Parents choose schools for their children based on costs and benefits (incentives), the availability of information, and the presence of opportunities (choices)” (p.432). The authors then go on to specify the incentives and costs of choosing the “right” or “wrong” schools, proposing that four constructs (incentives, costs, availability of information, and choice opportunities) shape behavior. But how those constructs shape behavior remains unspecified. Because much of the choice literature relies on rational choice theory (RCT), the unstated assumption in Bast and Walberg’s work (as in other’ s) is that the choice process can be described as one of utility maximization. Parents assign weights to the various costs and incentives associated with a set of schools and then choose the school that is the best balance of those weighted costs and preferences. But there are many ways to assign weights. How do parents even come up with weights? And what does the human calculus that leads to a “best balance” look like? What is the role of time in all of this weighting and maximizing? Does it happen all at once? In stages? The straightforward statement that parents choose based on incentives and costs does little to explain the choice process that actually occurs. Thus, the surveys and interviews that form the backbone of the choice literature largely treat the choice process as a black box, the internal workings of which we know little. One final empirical problem in the choice literature is its analytic reliance on relativistic measures. The reliance on relativistic measures has provided little insight into what parents say and what they do. Two common examples will illustrate the problem. One of the main outcomes choice researchers measure is satisfaction. The idea is that more satisfaction is an important choice outcome in and of itself. But there is a problem. Rubinowitz and Rosenbaum (2000) explain, “Satisfaction is almost always relative. While surveys usually ask questions about satisfaction as if it were an absolute, people can only answer relative to what they know or expect” (p. 129). The same logic applies to survey questions that ask parents to rank factors influencing their desire to choose a particular school. When most parents report they desire a “high quality” education or an “academically challenging” school, this does little to help us understand parents’ lived preferences. Again, parents can only define “high” or “challenging” as they understand it, based on their experiences. Thus, given the relativistic nature of those labels, reporting that parents are more satisfied or prefer high quality education does little to help us empirically understand educational outcomes and parents’ preferences, respectively. In addition to the empirical problems in the literature, there are also problems of theory. Applying RCT to issues of school choice, researchers assume that parents value “good” schools -- schools that have better academics. This assumption resonates with research that suggests some parents select academically superior schools (Armor & Peiser, 1997; David, West, & Ribbens, 1994; Witte, 2000). However, this assumption is inconsistent with other research, including interview studies (Wells, 1993 ), evidence from the United Kingdom (David, et. a], 1994; Woods, Bagley, & Glatter, 1998), and accounts of parents’ behavior (Robelen, 2002; Schemo, 2002). Some parents do not prefer what others might call the “best” school. For example, David et al. (1994) found that parents value a range of dimensions that may or may not add up to the “best” academic school. Together with the academic features of schools, parents valued the child’s happiness, the feel of the school, as well as the school’s proximity (David et a1, 1994). Wells (1993) found that the racial make-up of suburban schools (and the attendant challenges that went along with those racial differences) greatly influenced urban students’ desire to continue attending those schools. Thus, the assumption that all parents value a certain type of school and that there is a single scale against which “good” schools can be judged, may be more complicated than has heretofore been assumed. How parents define “good” schools brings us to the issue of parental preferences and their treatment in the choice literature. In a rational choice model, three constructs are often specified: information, costs, and preferences. These constructs hold very different status in the model. Individual choices hinge on costs and preferences. The individual assigns costs to various outcomes. Preferences vary by individual and are exogenous to the model. Scholars assume people have preferences and make choices given those preferences. Scholars do not, however, seek to explain those preferences (a recent exception to this generalization is Bowles (1998)). Choice research has investigated information and costs but parents’ preferences have been left under-developed. Many potentially important questions remain unanswered. Why for example, do some parents prefer a school with a band and others prefer schools with high test scores? How do parents come to have the preferences they do? How malleable are those preferences? The answers to these questions might hold the analytic solution to our present puzzle. Economists’ reasonable treatment of preferences as exogenous may be the theoretical limitation underlying our empirical puzzle. In an attempt to build on the existing literature, this study seeks to develop a situated and dynamic understanding of parents’ preferences by going inside the black box of the choice process. In contrast to the literature, the study relies heavily on parents’ actions in order to build an understanding of preferences that is grounded in behavior, thus avoiding some of the pitfalls associated with strictly relativistic measures. Finally, in the description and explanation of parents’ preferences this work argues for an evolution of our existing theoretical framework. Theoretical Framework There are a number of assumptions upon which the study rests. I assume that parents’ actions result from reasoned decision-making, and that this process involves an explicit or implicit consideration of constraints and preferences. Furthermore, I presume, as Cusick (1992) suggests, that “the participants described are reasonable men and women, an honest, hard-working, and well-intentioned lot who do what they do for good reasons. On their own terms, their actions make sense. The goal is to understand the educational world as they understand it” (p. 13). Finally, I assert that researchers’ traditional distinction between “choosers” and “non-choosers” is false (Schneider et al., 2000). According to this line of thinking, choosers move their children from assigned schools, “non-choosers” do not. Consequently, research has focused almost exclusively on the “choosers” (with the notable exception of Wells (1993 )). Since a small percentage of parents move their children, this focus has resulted in a narrow understanding of a very small group of parents. I take a different tack and, consistent with Cusick (1992), assume that all parents choose, whether they choose a neighborhood school or another. Given these assumptions, the goal of this research is to understand parents’ actions as they understand them. As implied earlier, this study uses the constructs of RCT to investigate parents’ choices. I use two broad categories rather than the three (costs, information, and 10 preferences) outlined in the previous section. Treating information as a constraint, I focus my attention on the constraints and preferences that influence parents’ behavior. Prior work has articulated relevant constraints (David et. a1, 1994; Gauri, 1998; Schneider et al., 2000; Wells, 1993; Witte, 2000). They include but are not limited to considerations of race, information, income, supply of schools, transportation, parental involvement, education, and employment. Building on this work, I focus on four constraints: the supply of schools, transportation, information, and income. The study is. however, inductive, so while these four constraints are all present in the study’s design, there is space for them to expand and change — as well as for new ones to emerge -- as the data are examined. Understanding constraints is important, but constraints alone do not sufficiently explain parents’ choices. We must also understand how particular constraints interact with particular preferences over time to produce choice behavior. One of the goals of the study is to develop a more nuanced understanding of preferences. So I begin the study with broad and flexible framework for investigating parents’ preferences. Based on other scholars’ thinking (Coleman, 1990; Lareau, 1989), my pilot work, and the work of sociologists (MacLeod, 1995; Wills, 1977), I posit that parents’ preferences have at least two dimensions - society and school. Parents’ societal preferences center on issues like family and work. These preferences are likely culturally bound. Lareau (1989) found that working and middle class families interacted with schools based on different assumptions of appropriate behavior. The former thought that, out of respect for schools and teachers, a parent should not “meddle.” Middle class parents viewed “meddling” as appropriate - even necessary -- advocacy for their children. These cultural views of the appropriate family 11 role in schooling may influence parents’ preferences for particular schools. How a parent understands the economy might also shape preferences. In my own pilot work, one mother explained that school did not help her get a good job so it may not help her son. Her view of the relationship between schooling and jobs influenced her preferences for schools, as did her own history with schooling, a result echoed in Okey’s (1990) study of why students drop out. Parents’ views of how school can or cannot facilitate economic success, and their own histories with schools may shape their preferences. Parents’ school preferences may also be critical to their ultimate choice. There are features of schools -- a foreign language program, high test scores, or a particular teacher -- which can influence parents’ choices. Prior work documents parents’ attention to such school features (Schneider et. al, 2000; Witte, 2000). A parent may prefer a school because it has a desirable feature: it may be close to a parent’s workplace or have a fantastic dance program. Perhaps the parent likes the peers at a particular school. Characteristics such as the race and socioeconomic status of peers are also important (MacLeod, 1995). If a choice school is likely to associate one’s child with desirable peers (e. g., children with special talents or high aspirations), parents may prefer that school. As children grow older, the issue of potential friendships at the choice school also shapes preferences about peers and therefore schools (David et al., 1994; Wells, 1993) All of the features described here can be associated with both schools and society. Schools are, after all, embodiments of society. This distinction between school preferences and societal preferences is an analytic one intended divide preferences into categories that might help us understand parents’ thinking. I do not take the distinction to 12 exist in parents’ minds, nor do I presume that ideas about schools are independent of parents’ ideas about society. Rather, I propose the distinction to be a useful and initial way of separating the dense complexity of parental thinking. A further complication is that parents’ constraints and preferences are contextualized. They are neither monolithic nor static; they shift and change over time. Parents may experience one dimension of their preferences more intensely than another at a given time. Acute events might cause parents’ preferences to shift or change altogether. The intensity with which any one constraint or preference is held alone, or with respect to the others, impacts choice. Though this study includes the somewhat static categories of constraints and preferences, it seeks to operationalize them as dynamic and situated, acknowledging that parents reason about these constraints and preferences in yet unexplored ways. In order to understand how this study is different from others, there is one point about the use of RCT that requires emphasis. Current applications of RCT to school choice have a normative dimension that was mentioned previously: scholars predict that given the opportunity, parents will choose academically superior schools. There is nothing about RCT that requires this prediction. RCT has the capacity to predict any kind of choice, not just the choice of academically superior schools. This normative stance has produced predictions (i.e., parents will choose academically superior schools) that do not match parents’ actions. As a result of the mismatch between predictions and actions, scholars have offered explanations for why parents do not act in the predicted fashion. Most of these explanations posit some kind of deficit in the parent or the system 13 more generally. Parents don’t have enough information about the schools that are available. They don’t understand how the application process works. They are too busy to investigate options as well as they might like to. Parents can’t afford to send their children to the schools they want to send them to. Parents don’t have transportation. There aren’t enough of the kinds of schools parents really want. The list goes on. The explanations researchers have offered thus far are explanations of what’s missing. Parents don’t have this or that. They don’t know how to this or that. The system doesn’t allow for this or that. This kind of theorizing and explaining leaves us with a long list of things to fix. We need to expand choice, give out vouchers, and have parent information centers. This view of the world gives policy makers “answers” to why parents don’t act the way we expect them to act. But this way of thinking leaves us explaining reasonable people as problems that thwart policy’s good intentions. Social scientists have encountered this problem before: Women were seen as problematic because their emotional states and moral reasoning differed from those of men, for example (Gilligan, 1982). Rather than rejecting RCT altogether, I employ the framework but do not assume parents will choose academically better schools. By utilizing a broad view of constraints and preferences that does not impose any a priori restrictions on the nature of parents’ academic preferences, this work uses RCT in order understand parents’ actions thereby allowing us to better predict them. Having explained the problem and theoretical framework that guides my investigation, I now turn to an overview of the study. 14 An Overview of the Dissertation The study is divided into six chapters that describe both the process and product of a project that began over two years ago in 2002. A brief description of each of the chapters follows: Chapter 2. Chapter 2 details the methodology used in the study, including the context, design, data collection, instrument development, and data analysis. Chapter 3. Chapter 3 details the choice process and selected choice behavior that demonstrate the parents in this study made interested, informed, reasoned choices for their children. This chapter’s view is broad and considers the processes and actions of all the parents in the study. Chapter 4. Chapter 4 focuses more narrowly on eight parents’ thinking in order to investigate parents’ preferences. Using the cases, I develop four dimensions of preferences that are crucial to the construction of parents’ choice sets and final decisions. Chapter 5. Chapter 5 builds on the work of the previous two chapters and articulates a model of school choice. The model relies heavily on parents’ revealed preferences and is likely to result in drastically different, more accurate, predictions of parents’ choice behavior. Thus, giving us a way to understand how it makes sense to choose a failing school. Chapter 6. Finally, looking across the other chapters, Chapter 6 outlines the larger argument of the dissertation. It considers the conclusions, implications, and limitations of the study. Finally, the chapter describes additional analyses that could be conducted on the existing data as well as future work that arises from the study’s conclusions. 15 Chapter Two Method Early in my graduate school career, Tyack (1976) and Allison’s (1999) work compelled me. Both considered a single problem from multiple perspectives. Tyack considered the history of compulsory schooling and Allison, the Cuban missile crisis. Through their work I became fascinated by the ways in which our conclusions are shaped by the questions we ask, the lenses we employ and the analyses we conduct. Before I knew what the dissertation was to be about, I spent time learning how to different apply lenses (i.e., critical, institutional, Foucaultian, feminist) to the same problem. The methodology of this dissertation reflects that interest in seeing the world in different ways. Study Design Since 1994, parents living in Gaston and the adjacent ring of suburbs have enjoyed choice options which include some 98 charter, 393 private, and 389 traditional public schools. In order to understand what influences parents’ reasoning in this rich environment, I designed a longitudinal case study of 48 parents’ thinking prior to their children attending 6’“ or 9th grade. The transitions fiom 5m to 6th and 8th to 9m grades are ones in which many children move from elementary to middle or middle to high school. At this juncture, parents might be particularly willing to consider alternative schools. The inquiry focuses on the eight months prior to the child starting their new school, a time believed to be formative to the ultimate choice (David et al., 1994). As I am investigating the question, “Why do parents choose particular schools for their children?” the design includes parents who previously sent their children to both failing and non- l6 failing schools across six types of schools (neighborhood public, magnet public, charter, non-religious private, religious private, and homeschool). This design maximizes diverse prior choices as a proxy for diverse parental thinking. The Context Like many cities in the Midwest, Gaston, a city of over 200,000 people is an old city that is in the midst of trying to reinvent itself. In the last 15 years, there have been numerous urban empowerment zones as well as significant tax breaks for large companies willing to bring their employees and their profits back downtown. While these initiatives have met with success, the divide - both racial and economic -- between the city and its suburbs remains large. It is both a racial and economic divide. The city is over 88% non-white, has a median family income of approximately $33,000 with just 11% of the population having earned a bachelor’s degree or higher. In contrast, the adjacent suburbs have on average a 18% non-white population that has a median family income of $62,000, with 28% of the population having earned a bachelor’s degree or higher.’ There is a long history of choice in Gaston. However, until recently, choice was available only through residential movement or attendance at tuition-based private schools. This changed dramatically ten years ago when the state passed its charter law. Since then, Gaston has been the site of much charter activity. There are almost 100 charter schools in the city and the surrounding suburbs, which serve as additional tuition- free altematives for parents who already enjoy a long history of magnet schools in the ' The state’s median family income is $53,457, 19.9% of its population is non-white, and 21.7% of the state’s citizen’s have earned a bachelor’s degree or higher. 17 city. Prior to the explosion in the number of charter schools and the increase in magnet schools that occurred almost simultaneously, Gaston was a district with strong neighborhood boundaries that determined which schools children attended. Now children are assigned to a school but parents can request their child attend any school in the district, in short, an option-demand system (Schneider et al., 2000). In neighborhood schools, priority is given to children that live in the catchment area, but if there is space, children from other neighborhoods are welcome. Because of the recent outpouring of students (more than 20,000 since 2000), Gaston is eager to keep its families in its public schools, and therefore, is much more open to intra-district choice than ever before. I chose Gaston for several reasons. First, Gaston was selected as the study site because it was a choice rich environment that had a history of choice. Ten years of charter schools provided the historical context in which parents would be likely to know they had options. Second, Gaston has many lower income families of color who demonstrate some amount of unhappiness with the present school system. Choice reforms have been aimed at just such a group, so Gaston provided a reasonable test case. Given the choice rich context and a group of parents who are likely to be open to choice, if parents were going to think about choice in the way reformers hypothesized, they would do so in Gaston. Sampling The sampling strategy for the study was initially straightforward. It rapidly became complex as access became difficult. I will first explain the simple version of the sampling strategy and then layer on the complexities and adjustments made in the process of gaining access. 18 Three intellectual concerns guided the sampling: the level of choice available to parents, family income, and diversity of parental thinking. The first two concerns were based on the literature. Scholars argue that parents need to have a variety of schools available to them and they need to be able to afford those schools (e.g., Chubb & Moe, 1990). These two concerns were dealt with by Situating the study in a section of Gaston that was rich with schooling options and by stratifying the sample by income. The final concern -- that I speak to parents who think about choice in different ways -- was more difficult to deal with. I used two variables as proxies for diverse parental thinking: the status of the 5th or 80' grade school (failing or non-failing) and the school type (neighborhood, private, etc.) By gathering a diverse sample with respect to school status and school type, I hoped to have access to a range of different ways parents thought about school choice. I will now explain the sampling frame in more detail. All elementary and middle schools five miles on either side of the border between Gaston and the adjacent suburbs were classified as failing and non-failing. This particular location was selected because there are many choice options available to parents - a high density of charter and private schools as well as nearby districts that can be accessed through inter-district choice. A school was considered failing if it was on the state’s list of failing schools or (in the case of private and religious schools) was unaccredited by any accrediting agency. To maximize access to diverse parental reasoning, a purposive sample of one failing and one non-failing school was selected for each of five school types (neighborhood public, magnet public, charter, secular private, and non-secular private). Principals at these schools were asked to supply a list of 5th (or 8th) graders’ addresses. By using census data, the addresses were assigned a median 19 family income based on their face block. Face blocks are the census’ smallest unit of analysis and are the houses that face each other between two city blocks. Addresses were then stratified into low, middle, and high income categories. Potential participants’ were randomly selected from each of the three income groups and contacted by telephone. This resulted in a sample of 48 parents stratified by their current status at a failing or non- failing school and median family income. Parents with similar characteristics replaced those parents who declined participation. Parents were compensated with a $30 gift- certificate to a store of their choice for their participation. Table 2.1 summarizes the actual sample. The 48 parents include an oversampling of 10% (four parents) to prepare for attrition. Happily, there was no attrition, so the study began and ended with 48 parents. 20 Table 2.1 Description of Parent Sample School status School type Reported family Total income Low Med High Failing Neighborhood public (1 school) 1 3 0 4 Magnet public (1 school) 1 3 0 4 Charter (2 schools) 2 4 2 8 Religious private (1 school) 2 0 2 4 Non-religious private (1 school) 0 2 0 2 Non-failing Neighborhood public (1 school) 2 1 1 4 Magnet public (1 school) 0 4 0 4 Charter (1 school) 0 4 O 4 Religious private (1 school) 0 2 2 4 Non-religious private (2 0 2 4 6 schools) Homeschooling 0 0 4 4 22 failing, 22 non-failing 8 25 15 48 I will now go back and explain in more detail the specifics of the sampling strategy. They are included here in the text, rather than in an appendix because they are important to understanding who ended up in the study and how they ended up there. These complications do not make for light reading, so they are separated from the rest of the text. A reasonable reader might choose to refer to the following three sections if she 21 has questions about the sample; however, she may skip next sections if the goal is to get a general sense of how the study was conducted. The List of Potential Schools In order to obtain a comprehensive, diverse list of schools located within five miles either side of the border between Gaston and the adjacent suburbs, a number of separate lists needed to be combined for no such accurate master list exists. The main list came from the state and included public, private, religious, alternative, and technical schools. Additional lists were used to make sure the state list was complete. Those lists of schools were found on websites including the Gaston Public Schools (GPS) website. the regional school district website, the ISACS website, www.yahoo.com, and the Lutheran and Catholic Schools websites. The schools which fell in this geographic area were then classified as failing and non-failing. There were no schools on the additional lists that were not on the state’s list; however, there were schools that were on the state’s list that were no longer open. Determination of School Status In metro-Gaston, a school is labeled failing if it does not make adequate yearly progress (AYP). The state produces a list every year of these schools which is printed in the Gaston Metro Times. Though many things are taken into consideration in the calculation for AYP (e.g., attendance, dropout rates, standardized test scores), the formula relies heavily (67%) on the school’s standardized test scores. The test scores are based on the state’s criterion-referenced exam, .which has been given in various forms since 1969. The test is the only common exam that all schools in the state administer to their 4th, 5’“, 7’“, 8m, and 10m graders. The test covers five areas: mathematics, reading, 22 science, social studies, and writing. According to the state, the exam is designed to measure what students know and are able to do in each of the content areas. Districts are encouraged to use exam results to determine students’ weaknesses and evaluate the match between the district curriculum and state standards. Because the state’s failing list is a political product with serious consequences for failing schools, it changed during the study. Some schools that were on the list got taken off, others added.2 The list available at the time of school selection (Summer 2002) was the most inclusive list the state produced, including over 1000 schools. Because it was clear that the list would change, I used two criteria to judge a school’s status: placement on the failing list and the summary test scores available to parents on the state’s website. In order for a traditional public, magnet, or charter school to be considered failing, it had to either be on the state’s failing list for two of the four subject matter areas (mathematics, science, reading, or writing) or have less than 45% of its children scoring at grade level. For private and religious schools, test scores and AYP data are not 2 In order to impose some order on the seemingly constant additions and deletions to the list, a school’s status remained constant over the course of the study. If I labeled a school failing in August 2002, that was its label in 2003 and 2004. This is a reasonable imposition if parents’ thinking is not sensitive to these lists and schools’ reputations are relatively stable across time. There is some recent evidence fi'om Florida which suggests that home prices are highly sensitive to the local schools grade (A, B, C, etc.) (Figlio & Lucas, 2000). Based on the data collected in this study, however, parental thinking is not highly sensitive to a school’s inclusion on the failing list or the school’s test scores. Not a single parent described a school as failing nor did any indicate that test scores are the most important factor in their thinking. Consistent with Holme (2002) and Pleasants (2000), parents used prior knowledge, reputation, and information gathered through social networks in order to make their decisions. 23 available so accreditation was used as a proxy for status. Unaccredited schools were considered failing while accredited schools were considered non-failing. Accreditation status was determined by a phone call to the school. Homeschoolers were not able to be classified in this way and therefore are left out of all analyses involving school status. Contacting Schools In order to investigate whether or not parents who sent their children to failing schools thought about their choice in systematically different ways than those who sent their children to non-failing schools, a purposive sample of one failing and one non- failing school was selected for each of five school types (neighborhood public, magnet public, charter, secular private, and religious). Principals were contacted by phone and asked to participate in the study. All of the principals with whom I spoke agreed to participate. There were two principals who did not return my calls and one who felt the study asked too much of the parents. These schools were replaced by schools of the same type and status. Principals in the charter, religious and private schools supplied a list of 5th or 8th graders’ addresses. Using block group census data, addresses were divided into three income groups (low, middle, high). Less than $39,000 was considered low, between $39,000 and $88,000 was considered middle income, and over $88,000 was considered high income. These cut-off points were determined by dividing the income range into thirds. Categorization with respect to income was performed to allow analysis of variance along that dimension. In order to reach the GPS magnet and traditional public school parents, 306 letters of invitation were sent to 5’Eh and 8”‘ grade parents at the four schools of interest. There 24 were a total of 29 envelopes that were never delivered and were returned to me. Parents then returned a self-addressed, stamped postcard with their name, address, and school. Of the 275 letters I believe were delivered, 37 post cards were retumed, a 13% return rate.3 Postcard addresses were then categorized into income groups. Within each income group, two addresses were randomly selected for each school. In order to locate homeschoolers, a search of the intemet using Google led to a list of homeschooling associations in the state. The associations were emailed with a description of the study and asked if they would pass along the message to their email lists. Four homeschooling associations agreed to do this. It is unclear how many families these messages reached. I received eight emails back from parents indicating they would be willing to participate. Of the eight, I asked if the parents resided in Gaston or one of the neighboring suburbs. One did not. Seven of the eight were eligible for the study, so the first four with which I was able to set up an interview were part of the study. I over-sampled so that up to 10% of the families could drop out and I could look at the impact of attrition. Families were contacted, resulting in a sample of 48 parents stratified by their current school status and family income. The positive response rate was 60%, the non-response rate 32%, and the negative response rate 8%.’ Parents with similar characteristics replaced those parents who declined participation. 3 Though this is a very low response rate, it is consistent with other one-shot mass mail response rates reported for urban areas (Dilhnan, 2000). ’ These response rates do not include the homeschooling parents who contacted me by email if they were interested in participating in the study. Further, if we remove the 16 parents who sent in their postcards 25 How Does the Sample Compare to the City? There is an important characteristic of the parents in this study that is a result of sampling but may go unnoticed. Though addresses were randomly selected, parents who I reached on the phone are not particularly random. As Fowler (1995) notes, when the telephone is involved, the sample tends to be biased toward relatively more wealthy families (those who can afford to have a phone or whose phone has not been turned off), and families that might be predisposed to participation in research. Further, 40 of 48 parents in the study previously opted out of their assigned neighborhood schools. Research on charters, vouchers, and magnets suggests that parents who opt out of their neighborhood schools tend to be better educated and more wealthy than their neighbors who remain (NCES, 2003). Table 2.2 compares the percentage of people in the sample and in Gaston on the basis of their income, racial background, and educational attainment. Though this is not an entirely fair comparison because there are suburban parents in the sample, it gives the reader an idea of how the sample is different from Gaston residents more generally. The last line in the table removes the suburban parents from the sample so that a comparison can be made between the city residents in the sample with the whole of Gaston. before the first attempt to reach them, the response rates change to 49% positive, 1 1% negative, and 40% non-response. 26 Table 2.2 Frequencies of Gaston Residents’ and Study Participants ’ Income, Race, and Education Group Income Race Education Low Middle High People of color White College graduate Gaston 49 39 12 88 12 11 Sample 17 52 3 l 73 27 44 Sample-city only 22 58 19 94 6 39 Note. All Gaston frequencies were calculated from 2000 census data As might be expected, the sample is made up of parents who have higher incomes, are more educated, have greater numbers of white parents, and have previously opted out of their neighborhood school. Though it is tempting to label these parents “marginal consumers,” parents who “search for more information then the average consumer. . .are more interested in and more “involved” with the product” (Buckley & Schneider, 2003, p .126), based on demographic data alone, that is a shaky assertion. But we can hypothesize that choice is likely to Operate as advocates have suggested among this group. The group may actually be more informed and “involved” with the product by virtue of the fact they have more education and have previously opted out of their assigned schools. But while we do not have good comparison group data, the best we can say is that these parents are as likely as other parents to enact the choice logic as it has been described by advocates. Gaining Access There was no problem gaining access to the charter, private, religious, and homeschools. As indicated earlier, principals (or parents in the case of homeschoolers) were contacted and, for the most part, if I was able to get the principal on the phone, they 27 agreed to have parents participate. The neighborhood and magnet access was another matter. Though I had secured written agreement from the principals of the schools I wanted to participate, the initial application to conduct research in Gaston was rejected. The Office of Research found it offensive and not in the district’s best interests. Over time, it became clear that neighborhood and magnet parents would not be in the study unless I was able to convince the Office of Research that I was not out to write an expose' of the district and that I would not offend or insult Gaston parents. This became possible through a Michigan State University professor who was trusted by the district and had personal connections to high level administrators in the district. He re-explained the nature of the study and his confidence in my intentions and abilities. After months of negotiation, the Office of Research agreed to allow me to do the study, but required that I change the method of gaining access to parents. I was not allowed to call parents. The district was unwilling to do two mailings so I was asked to mail the parents a letter and include a return postcard. All of the labeling and mailing of envelopes was performed by me at the Office of Research. I wondered if anyone would return the postcards at all. Luckily they did, and in the end, I had more than enough postcards from which to stratify the sample and make phone calls. Though neighborhood and magnet parents were recruited to the study using a method different fiom other parents, the selection effects neighborhood and magnet parents presented were similar to those who chose to participate from charter, religious, and private schools. The experience of gaining access to the neighborhood and magnet parents brought home to me the importance of access. The nature of the claims I could make were decided by the Gaston Office of Research. This seemed so strange. I had read about the 28 importance of access in my theory courses but I didn’t realize how deeply personal and idiosyncratic it really was. Throughout the process, I understood the district’s reluctance to allow me to do work there. It had, after all, been burned many times by white researchers ready to write, in painstaking detail, all of the things that were wrong with the city. I was surprised, however, at how personally I took the rejection and suspicion. My frustration came from an inability to adequately express my intentions and from my sense that I had little to no control over how I was interpreted. I knew I was not doing a study about the district, but I was at a loss for how to convince the district of that. If Robert Green, the professor who vouched for me, had chosen not to do so, this study would have been very different. A single person’s generosity and willingness to use his social capital made this research possible. I learned not to underestimate the importance of access again. The Pilot Study In the spring of 2002, I conducted four interviews with parents. The parents were recruited to participate through personal contacts. The interviews followed a semi- structured format and lasted between one and two hours. The four parents had either recently selected a school for their child or would be choosing a school in the next year. Three of the four children were entering middle school. One was about to enter high school. At the time of the interview, the children were enrolled in public schools in one of three districts, Gaston, an inner- ring suburb of Gaston, and an urban district over an hour away from Gaston. The pilot interviews were used primarily for two purposes. In my first three years as a graduate student, I conducted almost 100 interviews. But all of them were with 29 teachers and superintendents. I had never interviewed a parent. The first goal of the pilot was to familiarize myself with interviewing parents about their schooling decisions. The second goal was more pragmatic. I needed to figure out how to ask generative questions and then ask good follow-up questions. The pilot interviews met both of these purposes, albeit to varying degrees. I felt more comfortable speaking to parents after the pilot interviews, but I only made a modest amount of progress in leaming to ask good follow-up questions extemporaneously. The process of learning to ask good follow-up questions is one that requires experience (Walker, 1999), and the pilot work only launched the process. It did not complete it. The pilot study was generative. The four parents’ stories nominated a number of issues that had not been included in my prior conception of parental choice. These included the role of siblings, the parent’s views of schools as situated within a larger cultural environment, the parents’ beliefs about the difference between schools, and parents’ views of appropriate involvement in school. I concluded from the pilot work that parents’ preferences are more complex and situated than current empirical work demonstrates. Overall, the pilot work expanded my thinking about choice, so much so that the first interview instrument of the larger study comprised only five questions. I had learned enough to know that I had a lot to learn. Seven months after the pilot study, I began the data collection for the larger study. It is to this process I now turn. 30 Data Collection Methods Two face-to-face interviews were used to capture parents’ situated and dynamic reasoning with each of the 48 parents; one shorter telephone interview was used to follow-up. The first two interviews occurred at approximately eight months (t1) and two months (t2) prior to the beginning of 6th or 9th grade. The final interview occurred approximately two months (t3) after school began in the fall. The first interview averaged 60 minutes, the second, 49 minutes, and the final, 34 minutes. All told, 112 hours of interviews were conducted with the 48 parents. The first two interviews were conducted in person at a location chosen by the parent. Most interviews took place at parents’ homes, but in a few cases interviews were held at the parent’s place of employment, a restaurant, or their child’s school. The third interview was conducted by phone. All interviews were conducted at times that were convenient for and chosen by the parent. With permission, all interviews were audio recorded and transcribed. Instruments Validity. Before I explain how the interviews were conducted and analyzed, I will comment on the validity of the data collected. As Walker (1999) notes, there are two main concerns when one is interviewing cross-culturally: access and hesitancy. Hesitancy poses a particularly thorny validity problem. Even if a parent agrees to be interviewed, she might not be willing to discuss her thoughts and opinions openly with the interviewer. My social class, race, and outsider status all worked against me. I am a white, middle-class, non-parent who does not live in Gaston. I had reason to be concerned about the validity of that data I was about to collect. 31 I dealt with these concerns in two ways. First, as mentioned earlier, I had a sponsor (the principal of their child’s school). When I called a parent, I emphasized that the principal thought she would be a good person to speak to about how parents think about picking a middle or high school. I then explained the study and asked if it would be something she might be interested in participating in. The relatively high response rate (approximately 60%) indicates that the sponsor helped provide access and may have helped reduce parents’ hesitancy to speak with me. The second way I dealt with the validity concern that parents’ hesitancy might impose was designing the instruments and procedures so that a maximum level of trust could develop between me and the participant. There was not a single procedure or interview question that facilitated the development of trust. Rather, a lot of little procedures added up to what I hoped was a trust- promoting environment. For example, the parent chose the time and place of the interview. I explained the study in plain non- acadernic English. I made the first interview largely unstructured so that the parent, rather than the instrument, could be the center of the conversation. I sent thank you notes after each interview. And finally, I shared my summary of the first interview with each parents so that the she might make changes to it. I finished the study feeling like the trust necessary to gather valid information had developed. I did not, however, ask the parents if they trusted me, nor did I ask if they would be willing to participate in another study with me. There were many reasons I felt trust had developed. Parents told me many stories that were deeply personal. Some mothers cried as they talked with me. Others discussed issues that are not considered socially acceptable (such as talking about race, or their lack of knowledge about how to change 32 their child, or what they feel is wrong with their community). A number of parents asked for my home address so that they could continue to tell me how things were going for their child. No one dropped out of the study. Interviews lasted as long (if not longer) than I anticipated. I developed good relationships with the parents and some seemed genuinely sad to have the study end. Although I do not have concrete data about the level of trust that developed, I am confident that the data are as valid as interview data can be. To give the reader a better idea of the timeline of data collection, I now turn to the schedule by which data was collected and instruments were developed. Schedule. Interview data was collected in three waves, giving the opportunity to use the technique of grounded theory (Glaser & Strauss, 1979) to develop the second and third interview instruments. Preliminary analyses were conducted between rounds of data collection. Table 2.3 describes the data collection and instrument development schedule. Table 2.3 Data Collection and Instrument Development Schedule Month Data Collection Instrument Development January Conduct 1” interviews February Conduct 1St interviews March Conduct 1St interviews April Transcription and analysis of 1st interviews May Transcription and analysis of 1St interviews Develop 2"” instrument June Conduct 2"" interviews July Conduct 2'“I interviews Transcription of 2"‘1 interviews August Transcription and analysis of 2"d interviews September Develop 3rd interview Conduct 3rd interviews October Conduct 3"I interviews Note: All data was collected in 2003 33 The following section details the initial analyses that were conducted on the interview data as well as the development of the interview instruments. The final section explains the secondary analyses that resulted in the arguments presented in Chapters three, four, and five. The First Interview. The first interview developed from the pilot data. It was semi-structured and primarily inductive. It focused on ascertaining the child’s educational history and the parents’ thoughts about where the child would go to school in the fall (see Appendix B). The goal of the first interview was two-fold: to learn parents’ categories for thinking about choice and to develop enough trust that more directed questions in the second interview would yield valid data. By listening to the topics parents nominated, I was able to guide the conversation around participant-nominated categories of importance. This was critical to my study: I chose the use of qualitative methodology (interviews) because my intent was interpretivist (Erickson, 1986): I wanted to capture the experience of the parents, to understand how and why they were making the choices they were. The open-ended approach also allowed parents to talk or not talk about sensitive issues. Written demographic information was collected at the end of the interview to minimize any participants’ potential concerns with writing socially acceptable things and the inherent power dynamic between the researcher and participant. The first round of interviews, conducted in February and March of 2003, were transcribed with help from a transcriber. The first challenge of data analyses requires the display of data, often using innovative formats. I scanned Tufte’s (1983, 1990, 1997) volumes on the display of qualitative data; I examined displays that other researchers had used; I spoke with colleagues about how to represent what I was reading. After re- 34 reading each transcript, three types of data displays were made: a family choice history, a summary of the interview, and a categorization of the interview. Once completed, all three were put into a qualitative software program, N6 (the latest version of Nud*ist). I will describe each in turn. Examples of all three are available in Appendices C, D and E. The family choice history is a figure that summarizes where all of the children in the family have gone to school and why the parents chose those schools. This information, which I did not anticipate collecting, was volunteered by most parents during the first two interviews. It was verified with them over the course of the study and modified to reflect their corrections. The summary of the interview -- which was shared with parents at their second interview -- was an attempt to say back to the parent my understanding of their story. The summaries were necessarily uneven since I followed the parents’ lead in the first interview. Some, for instance, included parents’ thoughts about safety, others did not. Some included the parent’s entire choice set, others did not. Depending on what the parent talked about, the summary did or did not include the categories that proved, in later analyses, to be important. The summaries were used as a method of consolidating data as well as a reliability check. After the second interview the summaries were adjusted to reflect parents’ additions and deletions. The final data display was a categorization of the interview. The interviews were very chatty and many topics -- some unrelated to school choice -- were discussed. As I listened to and summarized the interviews, patterns began to emerge and I needed a way to keep track of them. For example, parents talked in broad terms about what they would and would not be willing to do. They talked about some schools that seemed reasonable 35 and others that were not. They mentioned their thinking about one child with reference to the others. In order to keep track of potential patterns, I began a list of categories. I worked back and forth between the interviews and the categories, adding new ones as they became necessary. If I was unable to provide a response for a certain category for a particular parent, I simply left it blank. The categorization was treated as a sorting mechanism, rather than a coding scheme. Dissimilar and vastly unequal text was allowed to be placed into the same category. The categories, like the summaries, were a first pass. They were meant to help me reduce the data but not so much as to lose potentially interesting, yet undiscovered facets of the choice process. Together these three data representations, each focused on a slightly different dimension of choice, acted as the basis for the development of hypotheses that could be tested in the second wave of interviews. This method of iterative conceptual development allowed me to take advantage of the highly nuanced, descriptive nature of the interview data (Miles & Huberman, 1994). During the time between the first and second interviews, I was able to perform two analytic tasks that helped me refine the emerging hypotheses and develop the second instrument. First, I wrote a number of analytic memos that outlined relevant patterns. These memos focused on topics like the role of siblings, parents’ understandings of how the next school fit with the previous one, and why parents’ chose the schools they did. Second, I wrote three detailed cases of individual parents as a way of getting clearer about a more holistic view of each parent’s decision making, as well as what in the data was interesting, and why it was interesting 36 given the choice literature.s From these preliminary analyses, I developed the second interview protocol (see Appendix B). After the first interview, I also realized that the 8’” grade parents spoke at length about what their children wanted. Other scholars have commented on the role children may play in choosing high schools (e.g., David et al., 1994; Wells, 1993; Woods, Bagley, & Glatter, 1998) -- and I had heard some of this in the pilot interviews -- but I was overcome by how pervasive it was in the interviews. I subsequently developed and gained approval to interview the 8’” grade children. This is a study of parents’ thinking, so I treated the interviews as an opportunity to better understand parents’ thinking. That is, the parents remained the focus of the study and my unit of analysis. The interviews with the 8’” graders allowed me to place parents’ thoughts into their family context. The interview focused on the child’s ideas and preferences about school (see Appendix B). Of the nineteen 8th graders in the study, 16 were interviewed. The remaining three were too difficult to schedule. The second interview. The second interview, conducted two to three months after the first, served five purposes. Before the first interview, there were certain dimensions of choice that I did not realize would be important to parents’ thinking. Only through conducting and writing up the first interviews did I realize that I needed to ask certain 5 I also kept a field journal throughout the entire nine months of data collection. This journal provided space for emerging hypotheses, notes to myself, questions left unanswered, and connections to the literature. The journal proved helpful during instrument development, data collection, analysis, and writing up the research. 37 questions (about the 8th graders for example). The first purpose of this interview was to ask questions of all the participants around issues that emerged from the first interview. The second purpose was to allow me to check on facts I thought I had gleaned from the first interview. Parents, after reading their summaries and family choice histories, corrected factual errors and added missing information to my account of their thinking. Third, the interview allowed me to document the most recent developments in the parents’ choice processes. In the winter, many parents did not know where their child would be accepted to school. By June and July, most parents had heard back from selective schools and knew where their children would attend school in the fall. Fourth, the second interview allowed me to test hypotheses that were generated in the first set of interviews. And finally, the second interview allowed me to interview the 8th graders in order to understand the family context of choice. All of the second interviews were transcribed and summarized by the researcher. Information from the second interview was categorized into the scheme that developed from the first interviews. Again, the summaries and categorizations were read into N6 and coded with the emerging coding scheme. The scheme focused on parents’ reasons for their behavior with respect to particular schools (i.e., reasons for vetoing a school, reasons for looking into a school). Potential hypotheses were revisited given the second wave of data. Many hypotheses dropped away because of a lack of evidence; what I once thought was important to parents’ thinking no longer seemed so, given the new data. Some hypotheses, however, remained. There was additional evidence for the significance of these hypotheses, though the evidence did not suggest particular 38 conclusions at this point. Given the new srunmaries and review of potential hypotheses, the third set of interview questions was developed. The third interview. The third interview, conducted by phone in late September and October, focused on parents’ satisfaction with their choice, parents’ educational history, the connections between parents’ educational experiences and the current choice. and parents’ views of standardized test scores. At the third interview, the children in the study had been in school between one and two months, depending on the child and date of the interview. The third interviews were transcribed by the researcher, summarized in the categories listed in Appendix D, and read into N6. The analyses conducted on the third wave of data are described in the following section. Data Analysis The final section of this chapter describes the analyses that were conducted for Chapters three, four, and five. The analyses that support these three chapters were not conducted sequentially; rather they were conducted in response to ideas and puzzles as they arose. Data analysis was neither a neat nor linear process. In the end, the three chapters take a decision making (or psychological), social, and economic approach to parents’ choices. Although I would not have predicted these distinctions at the study’s beginning, in retrospect the scheme makes sense given my interest in considering problems from multiple perspectives. We now turn to a description of the general analytic approach to each chapter. More specific details about particular analyses are found in the chapters themselves. 39 Chapter Three As described earlier, all of the interviews were put into N6. Using the modified coding scheme developed after the second interview, data were coded and tabulated, entered into an Excel database, checked for accuracy, and then imported into SAS (a statistical software package). Once in SAS, descriptive statistics were generated around issues of parental reasoning. As discrepancies emerged and additional categories were needed, the process was repeated. This process of coding and analyzing data allowed hypotheses to be tested, accepted or rejected, and reconceptualized as necessary. Chapter Four I began the analysis of parents’ preferences by reading the cases that had been developed after the first wave of interviews in addition to approximately 30 other parents’ summaries. I then brainstormed a list of cultural ideas that seemed to be common across parents’ stories. Using various cases, I wrote memos describing each of these cultural ideas or propositions. This served to narrow and clarify the meaning of the propositions. After I rewrote the memos into descriptions of the propositions and cases that exemplified those propositions, I performed the descriptive statistics that considered the propositions across the sample. This two-step process allowed me to utilize the richness and depth of the data collected. I used the individual cases to develop the propositions and the whole sample to check the prevalence of those propositions. Chapter Five If not for an article found late in the writing of this dissertation, which described the problems with stated and revealed preferences, the model of choice articulated in Chapter 5 would not have been. As I began writing up the study, I had a hunch that I 40 wanted to develop a better model of choice than had been articulated in previous choice research. But as I investigated the relationships between preference variables and the likelihood of choosing a non-failing school, I was left stunned. Few of the preferences (such as parental volunteerism and involvement) that I anticipated would be related to choosing a non-failing school were related. They were so unrelated, I began to doubt my judgment. I went back to the cases I was writing and realized my problem might be a measurement problem. Soon thereafter I began reading about the difference between stated and revealed preferences and began to understand the problem.6 I needed to investigate revealed preferences; in addition to the stated preferences I had already I investigated. I went back, coded more data and was able to develop the model I had hoped. The model relies on both quantitative and qualitative data; correlations and chi- squared analyses are the core of the former, and the propositions (in Chapter 4) form the core of the later. Conclusion I have already alluded to several important lessons I learned along the way. Access - how personal it is, how fragile it is, how little control the researcher has -- was one of the first lessons I learned. But there were several others. And while I will not delineate them all here, I close the chapter by noting two. 6 The difference between stated and revealed preferences is carefully explained in Chapter 5. For now, the reader can think of the difference as the difference between what people say and what they do. Researchers investigate stated preferences by asking people what they desire. They investigate revealed preferences by looking at people’s actions and inferring their desires. There are problems with each but in theory, they both measure the same construct—peoples’ desires (e.g., Hensher, Louviere, & Swait, 1998). 41 The first concerns the nature of this work. As graduate students we read books, write essays, and have class conversations about research quality. When we begin our dissertations, we think we know something about high quality work and the researchers that conduct that work. We have a little plan and then we go forth boldly to carry out that plan. I began this study thinking I had a handle on all the potholes that might be on the road ahead of me. I thought carefully about the relationship between the study design and the kinds of claims I could make. I practiced and got feedback on my interviewing techniques. I learned how to analyze categorical data. I thought I had those potholes filled in, wetted down, and smoothed out. How wrong I was! What I learned in the doing of this dissertation, is that becoming a researcher is not about making a travel plan and sticking to it. It’s about falling headlong into potholes and figuring out how to get out of them. It is about judgment. It is about learning how to learn. It’s not about automaticity. This is not to say that all my preparation was for naught or that studies should just be allowed to follow whatever path seems to feel right at the moment. On the contrary, that is not judgment. What I mean to say is that you cannot fully anticipate what the study will bring. And that is the nature of the work. You still brace yourself for those potholes, but being “ready” for the potholes is not the mark of a quality researcher. Being a good researcher is about knowing how to use your judgment. And though I can pronounce that with confidence, I am painfully aware that I am still a learner on this journey. The potholes remain. Which brings me to my final lesson. The fifth chapter in this study is still a pothole. And it demonstrates, with striking clarity, how ugly leaming to use one’s judgment can be. The chapter attempts an analysis that is uncommon in educational 42 research - the building of a causal model. I will not detail all the problems of the model here, you will read them for yourself. But using the fifth chapter as an example, I would like to share a final lesson about data analysis and judgment. I have learned that data analysis is more complex than I previously thought, and not simply because it yields surprising results and show-stopping problems. Data analysis is complex because it challenges your understanding of the problem and that challenge then requires the judgment of the re-made novice. You begin to see differently and then in the midst of seeing differently you are required to make wise choices. Choices about how to represent the data, what analysis to do next, and perhaps most visibly, how to report your “findings.” Data analysis remains what you do to “find” something, to reach a conclusion, to understand the relationships between variables. But it is also what you do to understand the problem and your data in a different way, in a way you could not have, had you not done the analysis. So in the case of Chapter 5, I present a model. A model that is arguably unparsimonious, difficult to read, even uninspired. But it is the analysis of the model -- which continues even after the numbers have been crunched and the argument laid ou -- that has led me to a deeper understanding of parents’ choices and preferences. The analyses I did to create the model are the very actions which make me question the model and see it in a new way -- a way that demands further attention, additional analyses and more judgment. To continue the earlier metaphor, the chapter (and data analysis more generally) creates a pothole -- a new set of understandings that must be analyzed and approached differently. And while this is quite distressing (especially when you are trying to “finish” a dissertation), it is the creation of these potholes that lies at the heart of research and inquiry. It is the potholes 43 that lead to new ways of thinking about old problems. And it is the potholes that lead to learning. Chapter Three The Process of Choice: The Case for Preferences In this chapter I describe a three stage process of choice that is both social and historical in nature. The process, which is made up of the predetermination, search, and choice stages, results in parents who are overwhelmingly satisfied with their decisions to send their children to both failing and non-failing schools. In order to understand why parents selected failing schools, I investigate the role that search strategies, parental engagement, and constraints played in the choice of failing schools. I argue that irrespective of school status (failing or non-failing), parents used information to make interested reasoned decisions. Further, explanations which focus on constraints or a lack of parental engagement do not explain the choice of failing schools. I conclude that parents’ preferences play a critical role throughout the choice process and it is those preferences we must understand if we are to make sense of reasonable, informed, caring parents who choose failing schools. The Conceptual Landscape of the Choice Process Choice is not new. Parents have been choosing their children’s schools for hundreds of years. Wealthy parents have selected private boarding schools, parents with strong religious convictions have chosen church and synagogue schools and perhaps most invisibly, parents have purchased their suburban homes as a form of school choice. Our current system of choice, which includes a range of types of choice (e.g., universal, controlled, option-demand choice), is a relatively new phenomenon (Schneider et al., 2000). The changes have been dramatic. The dissolution of attendance boundaries for assigned schools, the addition of vouchers, and the introduction of charter and magnet 45 schools has created a new dynamic of choice in many districts. These changes have made choice much more visible to the public. Further, these new dynamics have likely impacted the process parents use to choose schools. There are many literatures which could inform our theorizing about choice. The literature around the choice of college is particularly appealing. The phenomena are similar - parents and children with varying income and racial backgrounds choosing an educational provider from a range of institutions (e.g., private, public, selective, technical). The scope of the literature, which focuses both locally and nationally, is also attractive. Finally, the developmental stage of the literature is appealing — the descriptive, theoretical, and quasi-experimental work that has been conducted dates back almost 35 years. Though there is much to be learned fi'om this literature, I focus on the theoretical aspects of that literature, although the descriptive and quasi-experimental work also informs this analysis. In particular, this theorizing serves as the basis for my own conceptualization of the K-12 choice process.7 Hossler and Gallagher (1987) laid out a three-phase model of college choice. The model, which takes an inforrnation-processing stance toward cognition, built on the work of Jackson (1982) and Litten (1982) and delineated, in order, the three-phases as the predisposition, the search, and the choice phase (Hossler & Gallagher, 1987). The predisposition phase “is the developmental phase in which students determine whether or not they would like to continue their education beyond high school” (p.209). If they so 7 There has been what McDonough (1997) describes as three approaches to college choice research, social psychological, economic studies, and sociological status attainment studies. For a fuller description of these perspectives’ contributions, see McDonough (1997). 46 choose, students move on to the search stage in which they “gather information about institutions of higher education” (p.209). During the search phase some students also develop a sense of the criteria on which schools should be judged. The third and final stage is when the student actually decides which college or university to attend. Hossler and Gallagher (1987) emphasize the interactions embedded in all phases. They explain that throughout the phases, “individual and organizational factors interact to produce outcomes. These outcomes influence the student college choice process” (p.208). While I use this fiamework as an analytic device here, this study is not an empirical test of the applicability of this particular conceptualization of the choice processes. There are too many differences between K-l2 choice and postsecondary choice to simply lift the conceptualization of choice from context into the other. Three differences are worth noting. First, K-12 school attendance is compulsory, which means that parents m to make a choice. College choice is optional. Because there is no decision about whether or not your child will go to middle school, there is essentially no predisposition phase in the way Hossler and Gallagher (1987) describe it. Second, there is no history of assigned colleges. Through assignment, no one goes to the “neighborhood college,” although in some communities there are historical trends for students from various income levels to attend certain local and state schools. In K-12 education, there is a 150 year history of children being assigned to their neighborhood schools.“ This presents two problems - one practical and one conceptual - that result in a necessarily different search phase. From a practical perspective, many K-12 parents ’ Assignment to neighborhood schools began during the common school movement in the 1840’s (M.W. Sedlak, personal communication, July 16, 2004). 47 already have at least one school that their child is allowed to attend. This may mean they don’t need to “search” at all. From a conceptual perspective, parents may or may not continue to think of their neighborhood school as the school their child is “supposed” to attend. This could result in no search or a search in which every other school is compared to the assigned school. The final difference between college and K-12 choice is the age of the children. In the case of college choice, the student, whether or not she makes the choice alone or with her parents, is an adult. She can drive, vote, and maintain employment. The choice of 3 K-12 school is one made on behalf of the child. Depending on the family and the age of the child, children may participate in the choice to a greater or lesser degree (David, 1994; Wells, 1996), but the parent is legally responsible for the child’s welfare. For all of these reasons, the conceptualization of the process of college choice is helpful, but not directly applicable. I make use of work that has been done on the choice of college, but tinker with it to reflect the differences described above. In the following sections, I will make use of the predisposition, search, and choice stages, to describe the decisions Gaston parents made. I will not however, use them in the way Hossler and Gallagher (1987) intended. I do not view parents as information processors (which the model assumes). Thus, my use of the three terms Hossler and Gallagher (1987) suggest is a not consistent with the terms’ original intellectual foundations. I use the terms to describe and clarify parents’ thinking, which I take to be, at its core, a social and cultural non-linear phenomenon.9 9 This adaptation is in keeping with general shifts in the larger intellectual community. Information processing was a dominant framework within psychology in the 19703 when scholars believed that we 48 Before I turn to the three stages of choice, I feel obliged to tell the reader the punchline. My analysis leads to rather counter-intuitive findings, and by sharing the punchline upfront, the reader may more carefully and critically consider the data presented. The Results of Choice School Status As you will recall, the study began with 48 parents. There were four homeschoolers who could not be classified as sending their children to failing or non- failing schools. Half of the remaining 44 parents, or 22 parents, began the study with children enrolled in failing schools, the other half began with children enrolled in non- failing schools. In the fall, 25 parents enrolled their children in failing schools and 18 enrolled their children in non-failing schools. All four homeschool parents chose homeschooling in the fall. Though three more parents chose failing schools in the fall, this is not the most interesting part of the story. The more interesting part of the story is the lack of movement between schools of differing quality. Of the 22 parents whose children began in failing schools, 18 (82%) attended failing schools in the fall. Of the 22 parents whose children began in non-failing schools, 15 (68%) attended non-failing could model human reasoning in a linear fashion. This was influenced both by the desire to model human thinking on computers, and the move to include thought in psychology as well as the until-then dominant behavior. Since then, psychology has experienced several waves of intellectual shift, including the cognitive revolution and the more recent interest in social psychology, situated cognition, and cultural anthropology. My analysis reflects my own immersion in those more recent literatures that hold cognition to be culturally, socially, and historically situated. See Brunet for the history of psychology, etc. 49 schools in the fall. As a point of comparison, if students were equally likely to attend a failing or non-failing school, the percentage would be 50%. Unfortunately for the students who began in failing schools, we do not see evidence of equal likelihood. Rather, we see a striking proportion of students who continue to attend failing schools. There was, however, some movement between schools of differing quality. Four parents (18%) moved their children from a failing school into a non-failing school. A higher percentage (although not statistically significant) -- thirty-two percent (7 parents) -- moved their children the other direction, from a non-failing school to a failing one. Overall, there is a strong and statistically significant relationship between original school status (failing or non-failing) and final school status, 12(1) = 8.86, p = .0019, (b = .50. A child who previously attended a non-failing elementary or middle school had 9.64 greater odds of attending a non-failing middle or high school than a child who attended a failing elementary school. Because this study is based on a ptuposeful sample of parents who tend to be quite active in finding a school that matches their child, it is likely that the relationship between previous and final school status is underestimated. The underlying reasons parents reported for choosing particular schools will be taken up later in this chapter and in more depth in the following chapter. Satisfaction If a little more than half of parents chose failing schools, we might expect at least some of those parents to be dissatisfied with their choices. This is not the case. Parents were, for the most part, satisfied with their choices. Twenty-five parents (52%) were highly satisfied, 20 (42%) were satisfied, and three (6%) were not satisfied with their decisions. Most parents (94%) were satisfied or highly satisfied with their decisions. 50 Further, and perhaps more importantly, most parents (73%) felt that they had choices about where to send their children to school. Though there are only three parents who were dissatisfied and 13 who did not feel they had a choice about where to send their child to school, it is important if those parents all ended up in failing schools. We might then argue that although there is widespread satisfaction and sense of choice, those who are dissatisfied and do not feel they have choices are systematically forced into failing schools. This is a reasonable argument but it does not hold among families in this study. There was neither a statistically significant nor substantively important relationship between final school status and satisfaction. x2(2) = 4.49, p = .12, ¢ = .32, or final school status and sense of choice, 360) = 3.04, p = .10, of = .26. Parents who were dissatisfied and felt they did not have a choice had similar odds of sending their children to failing schools as parents who were satisfied and felt they had choices. Using less rigorous statistical standards, however, we might conclude there that there is a relationship between final school status and satisfaction and sense of choice. Because of the purposeful nature of the sample, it is quite possible that although there is not a statistically significant relationship in this sample, there is in the larger community. It seems likely that there exist parents who are legitimately dissatisfied and “trapped” into failing schools by circumstances such as income or school availability; certainly, this is the image portrayed in some literature (e.g., Chubb & Moe, 1991; Friedman & Friedman, 1980). Those parents present policy makers with relatively straightforward interventions: offer more choices and make the choices possible for parents who want them. This study is principally concerned with the parents who feel they have choices, select failing 51 schools and are satisfied with those schools. Those parents’ actions and thoughts are deeply puzzling and present a difficult policy problem for choice advocates and critics alike. Satisfied Parents and Failing Schools There are two types of explanations scholars offer for parents who choose failing schools. The first focuses on the parents and the ways in which their decisions are not optimal. Scholars argue that given the number of single mothers in urban areas, mothers don’t have time to adequately research choice options (e.g., Henig, 1994); they are essentially under-informed. Others point to parents who are less interested in their children’s educations or make decisions that are somewhat less than “rational” (e.g., Asimov, 2004; Holme, 2002). The second set of explanations focus on constraints, problems in the system that make for an imperfect market. For example, there is not enough high quality information (e.g., Arsen, Plank, & Sykes, 2000; Fisk & Ladd, 2001; Schneider et. al, 2001), there are not enough schools to choose from (e. g., Glatter, Woods, & Bagley, 1997), parents can’t afford the schools they’d like to send their children to (e.g., Moe, 1995), or parents don’t have transportation to the schools they might wish to select (e.g., Gewirtz, Ball, & Bowe, 1995). The remainder of this chapter considers the first set of critiques, those concerned with the nature of parents’ choice processes. This first set of critiques make assumptions about parents’ engagement and the choice process. It is those assumptions this work challenges. The Process of Choice In the following section I make use of Hossler and Gallagher’s (1987) model to describe the process of choice in three stages, predetermination, search, and choice. I 52 outline two search procedures that are used by a wide range of parents and result in similar outcomes with respect to satisfaction and sense of choice. I conclude by arguing that choice processes alone do not help us understand why parents choose failing schools. We begin with the first stage in the choice process - predeterrnination. Predetermination Gaston parents are in an option-demand choice environment. Option-demand choice is a system of choice in which children are still assigned to neighborhood schools but parents can opt out of those schools, most often into charter or magnet schools (Schneider et al., 2000). Of the 48 parents in the study, forty have opted out of their neighborhood schools. But many fewer have opted out of customary attendance patterns. Customary attendance patterns are the patterns made when children attend an expected sequence of schools (elementary, middle, and high). They are made up of what Fiske and Ladd (2001) call “feeder schools” and they exist in every sector: private, public, religious and homeschool. In Gaston, for example, if you go to St. Mary’s for middle school, it is expected that you will either go to St. Patrick’s High School or Atonement High School. Some children choose other schools, but most enroll in St. Patrick’s or Atonement. The moves from St. Mary’s to St. Patrick’s or from St. Mary’s to Atonement define the customary attendance pattern for St. Mary’s students and their parents. Customary attendance patterns are often unofficial and not written down (though in some cases they are). The customary attendance pattern provides a ready-made “next school” for most parents. This means that in the predeterrnination phase of choice, parents have to answer a question: Do I need to conduct a search? Among the 48 Gaston parents, thirty-three 53 conducted a search and 15 did not. There were a range of reasons parents did not conduct searches. Nine of the 15 parents reported that no other schools offered what they wanted. These parents knew the local schooling market from prior searches, felt confident that the school in the customary attendance pattern was exactly what they wanted, and were sure that other schools would not be able to offer what the customary school did. Three of the 15 said that the customary school was a good school with a good reputation so they were willing to try it. The remaining three parents reported that they were comfortable with the customary school and would reassess their decision after the next school year. The parents who chose the school in the customary attendance pattern reported that they felt confident about how their child was doing and they wanted their child’s success to continue. The implication was that they did not need to perform a search to “find” a school, they already had one. Mrs. Erhardt, a religious school parent, reasoned “They’re doing so well, why would I change? They’re thriving.” There is a lot to be unpacked in Mrs. Erhardt’s statement: her definition of “doing so well,” the implicit notion of why one would “change,” and the assumption that one can judge a school by a child’s progress. These issues will be taken up in the next chapter when we consider the social construction of parents’ preferences. For now, suffice to say that many parents, like Mrs. Erhardt, selected a school in the customary attendance pattern and their lack of a search did not preclude them from feeling confident and satisfied with their decision. Across the 33 parents who did conduct a search, two themes emerged as reasons for considering schools outside the customary attendance pattern. First, parents considered schools other than the customary school(s) because they were trying to find a particular feature the customary school did not have. One mother, Mrs. Smothers, 54 decided that the customary middle school did not have an appropriate autistic program, so she looked for other schools that did. Another mother, Mrs. Webb, was eager for her daughter to attend a more racially diverse high school; the customary school did not allow for this. A second theme in the reasons parents gave for opting out of the customary school attendance pattern was avoidance. These parents were trying to avoid a particular feature of the customary school. For example, two parents sent their 5m graders to the neighborhood elementary school and were very satisfied. The next school, the neighborhood middle school, worried them. The parents were concerned about the school’s leadership and felt “the kids are running the school.” One of the mothers, Mrs. Dorsey, worried that “anything can happen to her [Spirit, her 6"1 grade daughter] in that kind of a school.” Both mothers considered schools outside the customary attendance pattern because they wanted to avoid a school that was being “rtm” by the kids. Having explained the predisposition stage as the stage in which the parent decides whether or not she will conduct a search, we now turn to the search stage. Search There are two search procedures by which parents come to select a school for their child. The first search procedure, the open search, begins with a larger set of schools and overtime, in interaction with the schooling market, narrows to some smaller set of schools, final which the parent ultimately selects. The second search procedure, a closed search, is a procedure by which the parent has a single or perhaps two schools in mind. Over time, the parent gathers information about the school(s) and makes the necessary 55 preparations for the child to attend the school. We begin with a consideration of the open search, a procedure used by almost half of the parents. Mrs. Gunnison: An open search. Mrs. Gunnison has experience making assertive schooling decisions for her children. At age 16, her oldest son, Randy, was failing his classes and wanted to join the military. Mrs. Gunnison wanted Randy to be able to go back to school someday so she signed him out of high school. He is now married with two children and has a very successful career in the Navy. Doug, the middle child, presents a different set of concerns. He has a learning disability and never liked school. Mrs. Gunnison has been “up at the school” for Doug since he was in pre-school. Mrs. Gunnison removed Doug from the neighborhood middle school, Clark, because she felt it was too dangerous. In his new charter school, South Mill, Doug is working up to his potential and enjoys school much more than before. Jewel is a very different student than her brothers. She is a “talented and gifted” child, very talkative and curious. When Jewel was in first and second grade at her magnet elementary school, Park Slope, she would often finish her work early and begin “messing around” with her neighbors. This meant that Mrs. Gunnison was always getting phone calls from the school. At the end of second grade, Mrs. Gunnison grew tired of the discipline problems, concluding that Jewel wasn’t going to get what she needed at Park Slope. Jewel took an IQ test and was accepted at a private school for the gifted, which she attended in 3rd and 4m grade. Although she flourished in this environment, it became too expensive and Jewel returned to her magnet school for 5th grade. A year later, Mrs. Gunnison is still paying off those two years of Jewel’s education. 56 Because of her bad experiences with Doug’s schooling, Mrs. Gunnison would like Jewel’s middle school to be first and most importantly, safe. She explained, “To go to Clark is survival. And kids deserve more than that.” Mrs. Gunnison prefers a middle school with a clean environment, courteous staff, and energetic teachers who are able to control their classrooms enough to teach. Because Jewel is a smart, capable student, Mrs. Gunnison does not want a school where Jewel is forced to do busywork all the time, which will bore her and lead to discipline problems again. She wants Jewel’s natural curiosity to be nourished, encouraged, and channeled in productive directions: “If I can keep her focused and in the right atmosphere, she’ll want to learn, she’ll want to go on [to college]. When I asked Mrs. Gunnison, in the abstract, how she would rate different school choices, Mrs. Gunnison reported that homeschooling offers the highest quality education followed in order by, magnet, private, charter, religious, and neighborhood public schools. In choosing a middle school for her daughter, however, Mrs. Gunnison felt that only magnets and charters were “real” options. Private and religious schools are too expensive and Mrs. Gunnison has to work, so she cannot homeschool Jewel. The neighborhood school was not an option because of her experience there with Doug. Given these constraints, in February, Mrs. Gunnison mentioned eight schools (see Table 3.1). By June she had applied to South Mill, the charter school Doug attends, but was not completely satisfied that it was a good match for Jewel. Even through South Mill had “done wonders” for Doug and Mrs. Gunnison knew the teachers at the school very well, she was still worried about the degree to which Jewel would be challenged. 57 Table 3.1 Mrs. Gunnison ’s Choice Set Schools Status Type Applied Month mentioned & F ollow-up South Mill Technical Failing Charter Yes Feb-Brother attends there, likes Academy the teachers. Might not be challenging enough. Forward School Failing Magnet Yes June-Social worker encouraged her to look into it. Liked it immediately. Clark MS Failing Neighborhood No Feb-Never an option because of experience with Doug Bedrock Magnet MS Non- Magnet No Feb-Not an option, too much failing busywork F oskett MS Non- Neighborhood No F eb-Friend’s children attend there. failing (another district) heard only good things, Did not look into. Midwest Technology Non- Charter No F eb—Cousins attend there, too Academy failing much busywork DeLeone Academy Non- Charter No Feb-Got postcard in mail, never failing checked into it, Seemed like a “decent” school Madison Academy Non- Charter No Feb-Visited when considering 5" failing grades, was full then but liked it. Didn’t check into for 6". On graduation day in mid-June, the school social worker approached Mrs. Gunnison and asked where Jewel would be going to middle school. Mrs. Gunnison explains that Jewel had been accepted at South Mill but was still open to suggestions. The social worker insisted that Jewel attend either the magnet middle school, Bedrock, or Forward Academy. The social worker thought these were the only two schools “good enough” for Jewel. Mrs. Gunnison felt that Bedrock had too much busywork, so she wasn’t willing to send Jewel there. She had never heard of Forward, but within days she went to the school, visited classrooms when school was in session, and met with the parent/teacher liaison. Two things the parent liaison told Mrs. Gunnison made her believe Forward was the right place for her daughter. First, the parent liaison sends her 58 own children there. Second, the parent liaison indicated that, of the three children who graduated from Forward this year, two got into the best magnet high school in the city and one got into a very well respected private school in the suburbs. Before she even left the building, Mrs. Gunnison decided Jewel would go to Forward. There are many firings Mrs. Gunnison likes about the school. It is small and filled with children of different ethnic backgrounds. It has a looser structure based on students’ interests and creativity. With smaller classes, Mrs. Gunnison was told the teachers are able to really challenge the students. She feels this is exactly what Jewel needs. Finally, Mrs. Gunnison also likes the 12 hour/year volunteer requirement for parents. Though Jewel was less excited than her mom about going to Forward, she began the year with a positive attitude. Jewel likes Forward and has made new friends but Mrs. Gunnison is concerned that the curriculum is not challenging her daughter: “They’re reviewing a lot. I’m not sure if this turned out to be the right school for her. She’s not bored. She says she likes it. It’s so easy for her. That’s the part I’m worried about. They made it seem like they really challenge the kids when I went up there.” But Mrs. Gunnison doesn’t want to jtunp to conclusions so she is going to “give it until January and see how things go.” In the meantime, she is reconsidering Bedrock Magnet Middle in the hopes that Jewel could go to school there in January if things at Forward do not improve. Though Mrs. Gunnison rejected Bedrock earlier because of the amount of busywork, she said she was going to go up to the school to look around, explaining, “I know [Bedrock] is one of the better schools.” The open search. Stepping back from the details of Mrs. Gunnison and Jewel, we see that different schools held different status in Mrs. Gunnison’s mind. Some schools 59 were worth looking into, others were not. The set of schools she considered “real options” grew smaller as time went by. In January, there were certain schools that were already dismissed as options. Between January and July, Mrs. Gunnison became aware of particular schools, chose to investigate some, and apply to others. Friends recommended schools they knew, a professional suggested two magnet schools, Mrs. Gunnison got a postcard about a charter school. The sources of information varied, as did Mrs. Gunnison’s reasons for particular actions. In all this complexity, however, there is a structure. Figure 3.1 depicts the open search procedure over time. Un- Potential Set __ 1 mentioned Un- Choice Set _ 2 suitable Unapplied Suitable Set | . b) Un- Eligible Set _ 4 attainable Attainable Set Figure 3.1. The Open Search Procedure The schools Mrs. Gunnison considered can be sorted into five sets which become smaller and smaller as one moves down and to the right in the figure. I will briefly explain them here. 60 1: First, there is the set of schools that actually exist, the mtential set. These are the grade appropriate schools (of all types—private, religious, charter, etc.) that are in a two mile radius of Mrs. Gunnison’s home. Two miles would be approximately 10 minutes driving time, with traffic - a reasonable commute for many parents. In Mrs. Gunnison’s case, there were 13 schools like this, seven of which were tuition free. Of the 13 potential schools, Mrs. Gunnison only mentioned three. Ten schools that were in the two mile radius were not mentioned. We do not have data concerning why Mrs. Gunnison never considered these nearby schools “real options.” We do know that she mentioned eight schools, three within a two miles radius, five outside that radius. 2: The second set of schools is the choice set. These are all the schools the parent mentioned. Some schools in this set are deemed unsuitable by the parent and never looked into. We can think of the choice set as the schools that are on the parents’ radar screen. In our conversations, Mrs. Gunnison mentioned eight schools; three of which she chose not to pursue. The neighborhood middle school was unsafe and a local charter and a nearby magnet school emphasized too much busywork. Mrs. Gunnison felt all three were unsuitable. 3: The third, and still smaller, set of schools is the suitable set. These are the schools that the parent discussed as “real” options. Some schools in the potential set are investigated and applied to; others are eventually ruled out. Still others are never looked into and effectively remain in the back of parents’ minds. In Mrs. Gunnison’s case, there were five schools in her suitable set, three charter, one magnet, and one neighborhood school in an adjacent district. For a variety of reasons, parents do not apply to all the schools in the suitable set. Sometimes, the parent falls in love with one school and stops 61 he] 61 .L’) Illa; her search there. Sometimes, the child identifies one school she wants to go, and the parent investigates that school and not another one. And sometimes, the parent simply runs out of time and can’t apply to all suitable schools. Mrs. Gunnison found two schools that seemed reasonable; she never checked into the other three because she felt one of the two would be acceptable. 4: Schools in the eligible set are the schools a child could actually attend if everything worked out well. In other words, parents have completed whatever steps are necessary to be considered eligible by the school. Mrs. Gunnison applied to South Mill and Forward Academy, these schools make up her eligible set. Inevitably there will be schools the parent cannot get the child into (for reasons of low grades, test scores, not being selected in the lottery, the school not having space, etc.). These schools, though desirable, are effectively unattainable. 5: The final schools, the schools in the attainable set, are those that the child has been accepted to and is able to attend if the parent so chooses. In the case of Mrs. Gunnison, her eligible set and attainable set are the same since Jewel was accepted to both Forward and South Mill. This is not the case with all students. Many are not admitted to all of the schools in the eligible set, so their attainable set is smaller than their eligible set. From the attainable set, the parent selects the final school. A natural question one might ask is: How many schools were in each set? Table 3.2 shows that across the sample, parents had some 453 schools within a two mile radius of their homes, an average of 9.06 schools for each parent. Fewer than half (191 of 453) that number were mentioned in the choice sets. On average, for a given child, parents 62 considered 3.82 schools. They viewed 3.04 of those schools as suitable. They applied to and were eligible for 1.86 schools, and eventually were able to select from 1.38 schools.’O Table 3.2 Frequencies of Schools in Sets Potential set Choice set Suitable set Eligible set Attainable set Total 453 191 152 93 69 Mean 9.06 3.82 3.04 1.86 1.38 (SD) (4.67) (2.71) (1.94) (1.09) (0.57) Median 10 4 3 1 1 Mode 11 1 1 1 1 As the last line of the table makes clear, averages do not tell the whole story. The most common (modal) number of schools considered (choice set) is one; the most common number of schools applied to (eligible set) is one; and the modal number of schools parents select from (feasible set) is one. But there are parents whose process is not modal, and so we have two types of searches, one that looks much more like the modal process and, in fact, involves only one school which I will discuss as a “closed” search in the next section; the other, which looks like the mean, entails selecting the final school from a slightly larger pool of schools, the open search process that Mrs. Gunnison’s case illustrates. '0 The numbers reported here are summed over all the 48 parents so there are duplicates. The totals do not represent unique schools. This convention is used because the unit of analysis is the parent, not the school. 63 You will remember that 15 of the 48 parents did not search for a school because they felt satisfied with the school in the customary attendance pattern. An additional 22 parents conducted open searches similar to Mrs. Gunnison. This leaves 11 parents. These 11 parents conducted a search but it was quite different from the open search process. Rather than considering almost any school as a potential school, the remaining 11 parents had small choice sets, often just 2 or 3 schools, and they applied to only one school. In order to better understand a closed search, we will consider the case of Mrs. Borden. Mrs. Borden: A closed search. In contrast to Mrs. Gumrison (whose choice set included eight schools), Mrs. Borden did not consider a wide variety of schools and the two schools she did consider were outside the two mile radius that defines our potential set. This is similar to other parents who apply to a single school. Their search does not involve “finding” the right school. Instead, prior to making an active selection for their children, the parents have already narrowed down to a small number of schools. They gather information and compare schools, but in the end, they apply to just one school. Mrs. Borden’s thinking is representative of the group of parents who arrive at a decision through what I call a “closed” search. Mrs. Borden is a married, white mother of two children, Johnny and Clyde. While she discusses schooling issues with her husband, Mrs. Borden is the primary decision maker in the family. “He trusts me on these things,” Mrs. Borden explained. Though she is not college educated, Mrs. Borden has taken college classes and has worked in a variety of educational and business settings. She volunteers at the boys’ school and has been involved in their educations since they were young. From making 64 library trips to participating in the PTA to helping with homework and making sure Johnny and Clyde go to church every week, Mrs. Borden is engaged in boys’ day-to-day educational activities. Johnny (8* grade) and Clyde (6th grade) attend the same school, St. Mimi, which is a Lutheran private school that serves families through the 8th grade. Though the family is originally Catholic and Mrs. Borden attended Catholic school, both children have attended the Lutheran school since Clyde was in daycare and Johnny was in 2"" grade. Since the children began Lutheran school, the family has migrated to St. Mimi’s Lutheran congregation. God is the focus at St. Mimi and this was Mrs. Borden’s main concern when she originally chose a religious school for her sons: “1 want them to know that God is the biggest part of their life. . .when God is taken away from the equation, you end up with a lot of problems because they don’t understand why they’re here or what they’re here for and that somebody always loves them.” Over the course of the last school year, however, Mrs. Borden has had to consider other issues in her sons’ educations. Until the 8th grade, Johnny was doing well in school, getting As and Bs, cleaning his room, active in Boy Scouts. Things have changed dramatically this year. Johnny has started to earn mostly Cs and Ds. He doesn’t “take pride in his room.” Two badges away from being an eagle scout, he has quit Boy Scouts altogether. Mrs. Borden is worried about these developments but believes that, in time, Johnny will come back around. She is nervous but has faith that this is just a phase that he will eventually snap out of. She explained however, that Johnny does not want to go to St. Juliet, the high school that Mr. and Mrs. Borden planned on him attending. Many graduates of St. Mimi attend St Juliet, and Mrs. Borden has “always known” that Johnny would go there. Though Mr. and Mrs. 65 Borden have concerns about St. J uliet’s declining enrollment, lack of counselors, and potential financial problems, they prefer it to any other school. Mrs. Borden feels it offers a great education, small classes in which her son will get one-on-one attention, and a focus on God Johnny feels like a lot of his friends do not want to go to St. Juliet and he wants a “real” high school experience. St. Juliet feels too small and does not have the football, band, and social activities he associates with “real” school. His girlfriend and good friend will be attending Harry Smith, a large comprehensive high school in a neighboring district, and Johnny wants to be with his friends. The Bordens are committed to education and want both of the boys to go to college, saving for that eventuality through a state subsidized program for almost 1 1 years. Johnny’s education is paid for and Clyde’s is almost paid for. Mrs. Borden explained that college or not, “My job is to let them learn about life in a supervised environment and let them go out on their own and lead productive adult lives. You know, I want them to move out. You know? But I don’t want them to not be able to take care of themselves either. I want them to make good decisions.” Mrs. Borden’s choice set is below in Table 3.3. Table 3.3 Mrs. Borden ’s Choice Set Schools Status Type Applied Month mentioned & Follow-up St. Juliet HS Non- Religious No F eb-“Next” school for St. Mimi Failing students. Small, high quality, but Johnny doesn’t want to go there. Harry Smith HS Failing Neighborhood Yes Feb-Johnny has friends going (adjacent there. Large, comprehensive district) high school. A St. Mimi teacher is sending her child there. 66 Mrs. Borden thought a lot about whether or not to force Johnny to go to St. Juliet. Growing up, she was forced to change schools in high school and it had a dramatic effect on her behavior and grades. She got into a lot of trouble and did many things she now regrets. Mrs. Borden’s first choice is St. Juliet but she tried to balance her preference against Johnny’s, “If you force them to do things they don’t want to do, they’re either going to end up hating the situation or they’re gonna end up hating you.” Over the course of the spring and summer, Mrs. Borden decided to allow Johnny to go to Harry Smith. It was a difficult decision but, for a number of reasons, she made it nonetheless. First, a teacher at St. Mimi’s decided to send her daughter to Harry Smith. Mrs. Borden had extended conversations with the teacher and came to believe, through information the teacher had gathered, that the school was academically acceptable and provided a “socially safe” environment. Second, Mrs. Borden visited the school a number of times and many of her concerns were alleviated. She met a few counselors, saw some of the students, and walked through the building: It’s clean. The children seem really well behaved. There’s no like. . .low-life looking kids around. I don’t know (pause) somebody who is just obviously on drugs. I’ve been there at all different times of the day. I’ve never smelled smoke walking past the bathrooms. So obviously the hall monitor is doing his job. I mean I’m sure it goes on but. . .. (2mI interview) In addition to her own school visits, in the late spring Johnny began driver’s education classes at the school. He got to know some of the students in his classes and Mrs. Borden felt they were good kids. Together, these three developments reassured Mrs. Borden that Harry Smith was a reasonable school. 67 In October, Johnny was doing well. His grades were back up and he was happy. He seems to be adjusting well and hangs around with a “nice” group of friends who are mostly Boy Scouts and graduates of St. Mimi. Though there are no signs of problems. Mrs. Borden still views the current arrangement as a trial. “We’re gonna do it on a year to year basis. If we start having problems, having the wrong type of fiiends or grade problems or anything else, then I will put him in St. Juliet. I don’t want to use it as a punishment but it is a little more controlled there.” Overall, Mrs. Borden is highly satisfied with her decision and feels it has been good for Johnny. The closed search. Like Mrs. Borden, many parents have an idea about which school their child will attend. They do not consider every possible grade appropriate school, nor do they need to. They have their eye on two, maybe three, schools and, after some investigation, they make a decision and apply to the preferred school. Some parents, like Mrs. Borden, know where they want their child to go to school but then something happens; their child refuses to go, is not accepted to the school, or the parent’s preference changes suddenly (e.g., finding out one’s child is ADHD and preferring a school that can accommodate his needs). It is only then that the parents consider alternatives and begin a closed search. This much narrower, closed form of the search procedure differs fiom the open search in terms of the number of schools considered, but is similar in terms of the satisfaction, sense of choice, and final school status it can produce. Table 3.4 summarizes the similarities and differences between the types of searches. 68 Table 3.4 Frequencies of Search Characteristics Characteristic Search type None Open Closed Number of parents 15 22 l 1 Mean schools in choice set 1 7.5 3.4 Satisfaction Dissatisfied 0.00 (0) 4.55 (1) 18.18 (2) Satisfied or highly satisfied 100.00 (15) 95.46 (21) 81.82 (9) Sense of choice Few choices 13.33 (2) 31.82 (7) 36.36 (4) Have choices 86.67 (13) 68.18 (15) 63.64 (7) Final school status Failing 13.33 (2) 63.64 (14) 81.82 (9) Non-failing 60.00 (9) 36.36 (8) 18.18 (2) Homeschool 26.67 (4) 0.00 (0) 0.00 (0) Note. Numbers in parentheses are counts. These results contradict the notion that more choice is better. After all, parents in the group that did not conduct a search and had a choice set of one, were the most satisfied, had the greatest sense of choice, and were most efficient at choosing a non- 69 failing school.” As the sample is small, we must, however, interpret these frequencies with caution. If we want to determine whether or not a larger choice set is associated with the decision to attend a non-failing school we must evaluate the relationship between choice set size (not search type) and final school status. This evaluation was conducted as part of the model presented in Chapter Five (see Appendix A for results). To summarize, choice set size is not associated with higher levels of satisfaction or a greater sense of choice. It also is not associated with a particular group of parents. All of the correlations between choice set size and income, education, and race are not statistically significant or substantively important. In short, there is not a prototypical parent who considers many schools and then narrows down the potential schools as time goes by. All different parents -- low-income, white, black, urban, suburban -- engaged in the open search procedure. And all different parents engaged in no search or a closed search. We can conclude then, that the type of search does not alone explain why parents choose the schools they do. In particular, it does not explain why some parents choose failing schools. We now turn the third and final phase of the choice process, the choice. Choice As mentioned earlier, I use the Hossler and Gallagher (1987) process liberally. Having spoken for many hours with parents, I am convinced that “choice” is happening at almost every turn, not after the search process as Hossler and Gallagher’s (1987) model supposes. Throughout the search, parents are making choices: about schools to “ Only the relationship between choice set size and final school status is statistically significant and of substantive importance, r = -.47, p < .001. 70 consider, schools to investigate further, schools to remove from the running. Though there certainly is a moment when g choice has to be made, there are many choices throughout the process that provide us evidence of parents’ thinking. Therefore, my analysis of the choice phase spans the entire choice process, fi'om predetermination to choice. All three groups of parents — those who did not search, those who conducted a closed search and those who conducted an open search -- are accounted for in this analysis of choice. Parents’ gave many different reasons for their behavior. Their reasons were not only diverse, they were voluminous. Parents gave 74 different reasons for viewing a school as unsuitable. They gave 166 different reasons for choosing the school their child ultimately attended. To summarize parents’ reasoning, interviews were coded twice using two different coding schemes. The first coding scheme, generated inductively, used the categories of “holistic,” “academic,” “social,” “logistic,” “administrative,” and “other”. The second coding scheme was binary and used the categories of “interpretive” and “measurable”. I will first briefly describe the two coding schemes and then I will present the data on parental reasoning. The first scheme was focused on the substance in parents’ reasoning. Holistic reasons included, “they are thriving where they are” and the child “isn’t ready for that kind of school.” Academic reasons were those that focused on concerns around 99 66 classroom teaching and learning. The curriculum being “too basic, good teachers,” and “many leaming resources” were all coded as academic reasons for parental behavior. Parents mentioned many social reasons as well: “fiiends are going there” or the child “knows people there” or the students are “too rough” at that school, these reasons were 71 coded as social reasons. Logistical reasons were those that pertained to the location, transportation, and cost issues associated with a school. For example, reasons such as “moving to the area,” “too expensive”, and “close enough to sibling’s school” were all coded as logistical reasons. In contrast, reasons such as “couldn’t meet some of the school’s paperwork requirements” or “missed the application deadline” were coded as administrative reasons. Finally, there were a few reasons which didn’t fit well into any of the other categories. “That school has good lunch” and other equally puzzling reasons were coded into the other category. '2 The second coding scheme was binary and coded reasons as interpretive or measurable. Interpretive reasons were those which required an assessment of some school feature. They were reasons which might commonly be described as opinions: “matched my philosophical approach,” “Sarah can be successfirl there,” and “students are too rowdy.” These types of reasons relied on parents’ personal understandings of a school’s tangible features. In contrast, measurable reasons were reasons given by parents to which a number or third party measurement could be attached, including “small classes,” “less expensive,” and “close to home.” “Higher state scores” and “offers Spanish class” were also coded as measurable. '2 Parents discussed safety in different ways. On a survey “safety” is a comprehensive category that obscures the subtlety of the ways in which parents use the term. For example, some parents talk about unsafe schools as places that threaten their child’s ability to develop, socially and intellectually. Other parents discuss unsafe schools as socially chaotic places in which their child might get “mixed-up” in the wrong crowd. Depending on the context in which the parent discussed their concerns, “safety” was coded in the social, holistic or administrative category. 72 Both analyses are summarized below in Table 3.5. The headings used in the table are taken from Figure 3.1. “Unsuitable” refers to the reasons parents offered for not considering a school at all. “Suitable” and “eligible” refer to the reasons parents offered for pursuing some schools and not pursuing other schools, respectively. Reasons in the “unattainable” category are those reasons parents offered for their child not being accepted to a particular school. In many cases, the child didn’t get into a school the parent applied to or was waitlisted by that school. The “attainable” category includes all the reasons parents gave for selecting the school they eventually did. This category is the one most research on choice has historically focused on. 73 .8982 085 vacate 9.3 3:33 .«o 398:: 05 8 Bacon 8 ”finance =u 3 3235 288.. we c385: .58 05 9 macho.— .. ”ocoz as. 639862 2 35935 6.5. 33985 4.2. 3.2925 5:. 63.5262 3m 63296:: 8 sagas: «Na asaaaaz can sagas: 3n 332935 c .650 o 38m o EEO am see o 650 4 3 season 0 seaom 3. 38m arm 0235583. 2 ”sausages? 7 _.2 6333583. 2 .550 na Engage? 2: 38m 3: 38m 2: 38m 2 2582 cam Samoa 3: Samoa «.2 6933 Nam 22%: on gagaaaea. EN 28683. 6.3 2868:. ca 38232. a? 28082 8 season 33 65:6: mam 03:5 is Seam . a 3415 cans a. cane .mmuav a cane $15 a anus 6215 a amue ins eases? sausages sauce 638% snags fish Segafiw has SSEBESS 633% 538.3% $383.35 5 Susana 323m, aorxmaouesx Wm 2an Perhaps the most striking pattern in parents’ reasoning is the similarity across time. Parents offer the same reasons for determining a school is unsuitable as they do for vieWing a school as suitable. In all the school sets except one, parents offer holistic, academic, and logistic reasons most frequently. As the choice process evolves, parental reasoning relies on the same criteria. Holistic and academic reasons are either the most or second most frequently mentioned reasons in all the sets except one. We see the same pattern of similarity across time if we consider the measurable/interpretive coding scheme. Roughly half of the reasons offered for each set of schools are measurable, half are interpretive. Again, these similarities hold except for the case of the “unattainable” category, which we will return to in a moment. These analyses suggest that when we aggregate parents’ thinking, there is stability in reasoning across time. Throughout the choice process, parents tend to choose schools for holistic, academic, and logistic reasons. In their final decisions, academic reasons are the most frequently cited. This is consistent with much of the school choice literature (e.g., Gill et al., 2001; Schneider et al., 2000). The longitudinal perspective this study affords us suggests that parental reasoning is consistent over the eight months preceding the ultimate choice. It is important to consider that the reasons represented in Table 3.5 are parents’ reasons. If a parent said they choose a school because the “teachers are good,” that is how it was recorded. The teacher may or may not however, be “good” from another, perhaps more objective point of view. If we are to understand parents’ thinking we must understand how they see the world, but we must also understand how parents’ views match up with policy makers’ views. Even though we see academic criteria -- the criteria policy makers would most like parents to pay attention to -- playing a large role in 75 parents’ choices, this data does not support the interpretation that policy makers and parents are on the same page around issues of academics. They may be, but because there was no comparison to an external set of criteria, we do not know. Therefore, these data cannot be used to assess the soundness of parental reasoning from a policy perspective. ’3 It can, however, be used to establish a description of parents’ understandings of schools as well as a description of the reasons that are most prevalent in the choice process. In both analyses, the only set which does not fit the pattern described is the unattainable set. In the first analysis, the two most frequently mentioned reasons are logistical (40%) and administrative (30%). None of the other sets of schools have these types of reasons as the most or second more prevalent reason. In the second analysis, ninety percent of the reasons were measurable while only 10% were interpretive. Again, this is quite different from the 50/50 pattern we see at other points in the choice process. This difference is not, however, particularly surprising. After applying to a school, for the most part, parents — at least not the parents in this sample -- do not have control over whether or not their child is accepted to that school. Parents’ responses in this category hint at the bi-directional reasoning underlying parents’ choices. The reasons articulated by parents suggest that parents are not only acting on schools through their consideration, investigation, and application to particular schools, schools are also acting on parents’ thinking and choices through measurable actions such as never calling back, rejecting a student, or placing a student on the waiting ’3 Very few studies have tried to match parents’ stated reasons with the relevant school features. A notable expectation is Schneider et al.’s work (2000). Chapter 5 considers which this type of analysis is so rare. 76 list. Parents co-construct school choice: Parents choose schools and schools choose parents, not just at some final stage but throughout the process. Parents know this and they take account of it in their decision making. Up to this point I have argued that, although more than half of the parents selected failing schools, they were satisfied with their choices, felt they bad choices, and selected schools that were similar to their child’s previous school (as measured by school status). I have laid out the three phases of the choice process, predetermination, search, and choice. I have suggested that some parents choose the school in their customary attendance pattern, thus never moving out of the predetermination phase. Others use one of two search procedures -- closed and Open -- to identify a potential school for their child. I have also argued that across the choice process, parents use consistent reasoning that privileges academic and holistic concerns. I have not yet, however, taken up the critique that portrays parents as under-informed, disinterested, overwhelmed, and somewhat less than rational. It is to this critique I now turn. In the next section, I investigate parents’ actions across the search and choice phases of the choice process. By looking at parents’ actions, I propose that we can learn how parents’ are thinking about choice and to what degree their actions are interested and informed. Parents ’Actions: Windows Into Parents ’ Engagement The last section of this chapter makes inferences about parental engagement based on parents’ actions in the predetermination, search, and choice phases. We consider three actions - the methods parents use to find out about potential schools, the visits parents make to potential schools, and the choice sets parents construct -- in chronological order. 77 School Discovery Parents find out about or “discover” schools in many different ways: a friend recommends a school, the parent is reading the newspaper and there is an article about a school, or perhaps a parent drives past a school and notices it for the first time. By considering how parents come to nominate potential schools, we can better understand the reasoning underlying their choices. The 48 parents in this study were aware of 191 schools. The methods of discovery varied both within and across parents. Table 3.6 summarizes the methods into seven categories. Table 3.6 Frequencies of Discovery Methods in Choice, Suitable, Eligible, and Attainable Sets Discovery method Choice set Suitable set Eligible set Attainable set (n=l9l) (n=152) (n=93) (n=69) Social 39.27 42.1 1 41.94 28.99 Family 18.32 20.39 26.88 36.23 Assigned 18.85 11.84 13.98 15.94 Chance 8.38 7.89 2.15 2.90 Tools 5.76 6.58 9.68 8.70 Extra-curricular 5.76 7.24 4.30 5.80 Others 3.66 3.95 1.08 1.45 Note. N refers to the total number of schools in that particular category. Clearly, social and family connections are the primary method of discovering schools. Thirty-nine percent of the 191 schools were discovered by parents through some type of social connection (a friend’s recommendation, another parent’s suggestion, etc.). In the suitable and eligible sets, forty two percent of schools were discovered through 78 social connections. If we consider family members to be a “social connection,” approximately 58% of all schools considered in the study were discovered through social connections. The second most common method of discovering schools is by being assigned to them, accounting for 19 % of the total schools mentioned. Tools, such as the intemet, parent magazines, and the newspaper only accounted for 6% (or 12) of the 191 schools mentioned. These data suggest that parents find out about schools primarily through their networks of family and fiiends, acquaintances at church or the YMCA, school counselors or social workers. Other methods such as the intemet, TV, and direct mail are utilized by parents, but to a much smaller extent. The common sense notion that parents find out about schools through word of mouth is largely true for this group of parents. Discovery through connections is likely to carry with it both effort and information. There is more effort, for example, associated with having a conversation with the neighbor than there is with receiving a post card in the mail. The prior requires some willingness, the later that you have a mailbox. Though a social conversation may require less effort than for example, intemet research, it requires more effort than many other methods of discovery. This suggests that the primary method of school discovery used by Gaston parents demonstrates some amount of effort and interest. It also suggests that parents are basing their decisions on information. Conversations convey information. A friend is not likely to say, “You should check out Washington Middle School” without also telling you why she thinks you should consider the school. Numerous Gaston mothers relayed the information they had learned through their conversations. They retold stories like Mrs. Wise: 79 There’s another one I’m thinking of . . .a girlfriend of mine, the one that has her three [children] at Park Slope, well some school in Westville where it only cost a dollar. She said it only costs a dollar. But the only time I see her is when she’s dropping off her kids and I haven’t seen her in a while. You know so. ...she’s thinking of that school. And I don’t know, she’s been up there and said it was great. Yea, and it only costs a dollar. (1“ interview) Like the information gathered from school visits, we do not know the accuracy of the information gathered through social connections. But we do know parents value it highly. Almost to a person, parents felt that other parents are the best source of information about schools, perhaps because other parents’ concerns had a validity that resonated - even while concerns might have differed — with them. So again, these data suggest that parents’ decisions are interested and based on information they value. Parents exert effort in the form of conversations and they gather relevant information, judgments, and opinions from those conversations. School Visits A second set of actions that can help us make inferences about parents’ engagement is the behavior surrounding school visits. Parents gathered information about schools by going to visit. In fact, they became increasingly aggressive in their visits the further into the search process they got. If we hypothesize that a school visit is likely to result in information which will assist parents in decision making, Table 3.7 shows that most of the schools to which parents applied were also visited (76%). As we move to the right on the table, we see an increase in the proportion of schools parents visit. This suggests that as parents become more serious about a school as a potential 80 option, they are more inclined to visit the school. In the attainable set, the set of schools the child has been accepted to, parents visited 88% of the schools. Prior to the start of school, 44 of 48 parents stood in the building they ultimately selected for their child. Table 3.7 Frequency of Schools Visited in Choice, Suitable, Eligible, and Attainable Sets Choice set Suitable set Eligible set Attainable set (n=l91) (n=152) (n=93) (n=69) 87 85 71 61 (45.5%) (55.9%) (76.3%) (88.4%) Though a visit to a school does not guarantee what experts might describe as the essential knowledge necessary to make a decision about that school (e.g., standardized test scores, comparative analysis of similar schools, teachers’ certification status), a visit does provide parents with images of the physical plant, some of the resources the school has, the students who attend the school, and the adults who work there. In fact, across the families who participated in this study, school visits constitute an important mechanism for gathering information. Parents reported that visits were essential to their assessment of a particular school’s appropriateness for their child. These data support two conclusions. First, busy parents (working mothers, for example) cared enough to make time to visit the schools they were seriously considering. And second, most parents in this study made decisions based on information they believed was useful. It is also important to note that those visits involved both interpretive -- “I liked the feel of the school” - and measurable -- “There is a gym, a stage, and a cafeteria -- information. 81 Though experts may disagree about the accuracy or relevancy of the information, parents are nonetheless making decisions based on information that mattered to them. Potential and Choice Sets Parents’ actions regarding school discovery and school visits allow us to examine the engagement inherent in their choices. As I have already mentioned, however, the data do not allow us to judge the validity of the information collected or the decisions made, for we do not have independent measures of school or teacher quality. Even if we had independent measures of quality, there would still be space for questions. Parents’ and policy makers’ definition of a good teacher for example, may differ because one or the other is wrong (e.g., makes a invalid assessment or does not consider all the dimensions of “good”). Their definitions may vary because they have different values. They may disagree about which traits are most important in a good teacher. So while we cannot directly evaluate the accuracy of the information parents gathered or the “soundness” of their reasoning, we can indirectly evaluate the results of that information and reasoning. We can evaluate the end product of parents’ search procedures — the choice set. If we compare the choice sets parents construct and the potential sets presented by geographic location, we can evaluate the results of parents’ reasoning using external criteria. In comparing the potential and choice sets, at a minimum, size, quality, cost, and selectivity should be investigated. Though there are certainly other relevant school characteristics, these four criteria provide a snapshot comparison between parents“ constructed choice sets and the schools that are available to them. In this study, size refers to the average number of schools. Quality refers to the average proportion of 82 schools that are non-failing. Cost refers to the average proportion of schools that are tuition free. Selectivity refers to the average proportion of schools that have selective admissions procedures. As Table 3.8 describes, parents construct choice sets that are on average, smaller and of higher quality than the schools that geographically surround them. Parents’ choice sets contain 3.77 schools in comparison to the nine potential schools which surround parents’ homes. In other words, parents consider less than half the number of schools that are within two miles of their homes." Parents’ choice sets also contain a greater proportion of schools that are non-failing. On average, choice sets have 51% non- failing schools in comparison to 36% of the average potential set. Choice sets also contained a slightly higher proportion of schools that are free and selective as compared to the potential set. Neither of these differences was statistically significant and the associated d-values suggest they were not of substantive importance. " This is slightly misleading. By saying that parents considered less dran half of the number of schools in the potential set, I mean they considered that number of schools. They may or may not have considered the specific schools that make up their potential set. 83 Table 3.8 Potential and Choice Set Size, Quality, C 0st, and Selectivity Mean (1 Potential set Choice set Size 9.00 3.77"' 1.01 Quality 0.36 0.51" 0.38 Cost 0.68 0.72 0.09 Selectivity 0.51 0.55 0.08 Note. Means were compared on a matched sample t-test. " p=0.05, "p<0.0001 While these data do not support the conclusion that parents are making “good” decisions as judged by proxies that are currently being used by policymakers, together with other data, they do support the assertion that parents are making interested, informed, reasoned decisions. We have considered three actions which, taken together, suggest that Gaston parents made interested, informed, reasoned decisions. Their actions stand in contrast to the deficit characterizations inherent in the assumptions and policy recommendations of some researchers and the media. The prevalence of social networks, school visits, and relative high quality choice sets across the sample, suggests that we cannot explain the choice of failing schools by some sort of parent failure. We must consider other explanations. Conclusions I have described the choice process as a process that is roughly separable into three phases, predetermination, search, and choice. This process results in three groups of parents, parents who do not conduct a search but chose a school in the customary 84 attendance pattern, parents who conduct a closed search, and parents who conduct an open search. All three sets of parents are similarly satisfied and empowered with a sense of choice. Though the groups differ greatly in the number of schools in their choice sets, they make statistically and qualitatively similar choices. Process alone does not explain parents’ choice of failing schools. Parents’ actions, visible throughout the choice process, suggest that parents made interested, informed choices that were based on stable criteria and resulted in smaller, higher quality choice sets. We see evidence of parents’ interest in their effort to visit schools and carry on conversations with others about potential schools. We see evidence of the informed nature of parents’ choices in their school visits and conversations with other parents. We see evidence of parents’ reasoning and skill in their consistent emphasis on academic and holistic criteria and ability to create smaller choice sets with more non-failing schools. From all of this evidence, we see a picture of informed caring parents who are trying to find good academic schools in which their children can mature. Again, we cannot pin parents’ selection of failing schools on disengagement or a lack of appropriate skills. One final explanation for parents’ choices (that is embedded in the critique outlined at the beginning of in this chapter) is this: parents are simply under tremendous constraints. They don’t have access to transportation. They can’t afford tuition. They don’t have access to accurate information. These are plausible explanations for many parents, but as I will show, they do not apply to this particular group of parents. Not only is there little qualitative data to support this explanation, there is no statistically significant or substantively important relationship between choosing a failing school and 85 parents’ income, education, the accuracy of information used, or the number of schools the child was eligible to attend (see Table 3.9). Table 3.9 Relationship Between School Status and Constraints Constraints 12 DP p (11 Income 0.31 1 0.58 0.083 Parents’ education 1.15 1 0.28 0.17 Accuracy of information 0.67 l 0.41 0.12 #schoolsinattainable set 0.03 l 0.86 -0.03 This evidence suggests that constraints are not playing a decisive role in determining whether or not children end up in failing schools. Certainly parents are under constraints, even wealthy parents who can afford cars, tuition, and pay someone to gather accurate information. But again, for this group of parents, constraints alone do not explain why sensible parents choose failing schools. Thus far I have suggested that the choice process, parents’ engagement, and constraints do not explain parents’ choice of failing schools. As mentioned in the conceptual framework, RCT considers preferences to be exogenous to any model. Based on the fact that we have eliminated the leading explanations for choosing a failing school. we must turn to an alternative explanation, the possibility that preferences are endogenous. In the next chapter, I demonstrate the ways in which parents’ preferences (the independent variable) are not independent of the odds of choosing a failing school (the dependent variable). I argue that the social construction of parents’ preferences 86 warrants a careful consideration of the distribution of those preferences over racial. educational, and income groups. 87 Chapter Four The Social Construction of Preferences Summary Rational choice theory is used by many economists to elegantly and simply portray parental choice. Parents gather information, have preferences, are constrained by circumstances, and make choices. The messiness of human existence drops away -- as it necessarily does in much social science. The method of this study, however, was interpretive, and aimed to understand choice from the parents’ perspective -- in all of its human, messy, contextualized detail. Using eight cases, I explore four propositions that anchor and bound the social construction of parents’ preferences. I argue that the four propositions describe dimensions of parents’ preferences — academic achievement, child development, school quality, and geography — crucial to our understanding of parents’ choice sets and school selections. The social and historical nature of these propositions suggests that certain preferences are unlikely to be distributed evenly across parents, and thus might be considered in choice policies aimed at more equitable provision of education. Propositions and Preferences Rational choice theory has two broad categories which inform models of human interactions: constraints and preferences. When applied to school choice, preferences can be thought of as parents’ desires for particular features, outcomes, or values related to the schools their children attend. Parents can prefer a school that is close by, or they can prefer to have their oldest and youngest child be in the same school, or they can prefer a positive peer group for their child. Though these are different types of preferences, we 88 can imagine that any one, or all three together, could inform parents’ decisions about where to send their children to school. If we think of preferences as parents’ desires, constraints act to limit or modify those desires. Parents’ income, access to transportation, and information about schools constrain parents’ expression of their preferences. Less obvious constraints -- such as a child not being accepted to the parent’s preferred school or missing the application deadline for a school -- also play a role in constraining parents’ preferences. Though both constraints and preferences are important in school choice decisions, constraints have been relatively well specified in the literature (e.g., Gauri, 1998; Moe, 1995; Peterson, 1998; Rasell & Rothstein, 1993) while preferences remain underspecified. This chapter focuses on the ongoing dynamic process by which parents’ preferences are created by human beings, that is, the social construction of parents’ preferences. Preferences are situated in and informed by culture. Many scholars from diverse disciplinary backgrounds have developed explanations of the role culture plays in shaping and influencing an individual’s ideas and actions (e.g., Bruner, 1990; Geertz, 1983; Quinn & Holland, 1987). Bruner (1990) describes the “folk theories” that inform our understandings of ours and others’ minds. He says, All cultures have as one of their most powerful constitutive instruments a folk psychology, a set of more of less normative descriptions about how human beings ‘tick,’ what are own and others minds are like, what one can expect situated action to be like, what are possible modes of life, how one commits oneself to them, and so on. (p.35) 89 Geertz (1983) describes a similar construct but calls it “common sense.” For Geertz, our common sense is “what the plain man thinks when sheltered from the vain sophistications of schoolrnen” (p.77). Fuller, Holloway, Rarnbaud, and Eggers-Pie'rola (1996) explain the role culture plays in our actions through yet a third concept, “cultural models.” Citing Quinn and Holland (1987), Fuller et al. defrne cultural models as presupposed, taken-for-granted models of the world which are widely shared (although not necessarily to the exclusion of other, alternative models) by members of a society and that play an enormous role in their understanding of that world and their behavior in it. (Quinn & Holland, 1987, p. 4). Whatever construct one chooses - folk theory, common sense, or cultural models -- there is a common thread: culture is the central framework through which people make sense of their own and other’s lives. But cultures differ and the same behavior can take on varied meanings depending upon the culture through which it is interpreted. An example may help illustrate this point. A student explains to a professor that he is struggling with the ideas he is trying to write in a paper. The professor tells the student to just go ahead and try to write the paper. The professor assures the student that he is likely to learn how to get the ideas out clearly by actually trying to write them down. The professor’s suggestion is consistent with the notion, pervasive in Western culture, which views experience as the best teacher. If one holds the belief that the best way to learn is to go ahead and do the thing you are trying to learn, the professor’s advice is perfectly reasonable. If however, one believes an alternative view, let us say that learning is the result of long apprenticeships and closely supervised practice, the professor’s words may be understood as reckless and unhelpful 90 advice. The point here is that the professor’s words are interpreted within a cultural understanding of learning, and that understanding may differ fi'om culture to culture. What is sensible and encouraging common sense in one culture may be unreasonable and discouraging in another. Parents, depending on the culture(s) to which they belong, have different theories about how the world works with respect to schools. These theories. which I will henceforth call propositions, shape parents’ preferences. Again, an example will illustrate the concept. Across income levels, one commonly held idea which informs 8th grade parents’ preferences in this study is the notion that it is not desirable to send a teenager to a school she doesn’t want to go to. Parents worry a child who currently does not like school will want to drop out of school if the parent sends her to a school she doesn’t prefer. They worry that the child will have a hard time making friends and will be socially isolated, thus making academic success highly unlikely. They worry that their child will rebel, stop applying herself, and get involved in the “wrong group.” Though parents differ in their explanations as to why this is a bad idea, most of the 8th grade parents prefer to find a school their child wants to attend, or at least does not adamantly refuse to attend. Because most teenagers want to stay with their friends, depending on where the child’s fiiends are going to school, parents may be confronted with a set of more or less desirable schools. From a policy maker’s perspective, parents may therefore prefer schools that appear to be less than stellar academic institutions. If, however, one looks at the propositions upon which parents’ decisions are made, their preferences and ultimate choice become understandable. 91 As the example demonstrates, culturally constructed propositions about how the world of school operates can influence parents’ preferences. Parents simply presume that this reasoning is “common sense,” yet in another culture, considerations of the adolescent’s desires would not even enter the picture. There are many such propositions that are germane to school choice, some of which are common across cultures and others that are unique to particular cultures. Based on the interview data of this study, four central propositions are identified. These propositions help us understand why parents have the preferences they do. They also nominate four categories of preferences that are crucial to our understanding of parents’ choices. Using a navigation metaphor, if we think of parents’ preferences as buoys, propositions are the anchors which hold those buoys within a particular space. The buoy does not always float directly above the anchor, it moves with the wind and current, but the buoy does stay within a limited proxirrrity of the anchor. Parents use these buoys and preferences to navigate school choice. Propositions are beneath the surface of the water and are therefore, out of sight. But propositions are tied to the features of parents’ thinking we can see, their “buoys” or preferences. As already noted, four propositions were identified. For each proposition, the format of the argument is the same. First, there is a description of the proposition. Then, two cases are used to illustrate the ways in which the proposition influences parents’ preferences. The cases are compared to one another and finally, we consider the proposition across all 48 parents in the study. Unless otherwise noted, all the parents described in this chapter are African-American parents who live in Gaston. 92 Two of the four propositions cut across income, education, and racial categories, two are specific to the large and diverse group of African-American parents who live inside the Gaston city limits. All four of the propositions demonstrate the ways in which parents’ preferences are shaped and bound by their cultural context. Those preferences. in turn, bound the schools parents consider as well as their final choices. Though not inclusive, the four dimensions of preferences that are described by the propositions -- academic achievement, child development, school quality, and geography -- help us understand much of the variation across parents’ choices. The propositions illuminate the logic underlying what policy makers might call ‘bad’ decisions. Proposition One: Children Influence Their Own Academic Achievement All of the parents in the study recognize that schools matter. They believe that there are some schools that are better than other schools and they want better schools for their children. But parents do not think that the school is the sole determinant of academic achievement. They do not attribute academic achievement only to schools. In fact, seventy-one percent (34 of 48) of parents said that the child and/or the home play an equal or more important role than the school in a child’s academic success. Parents gave numerous examples of the ways in which children can either frustrate the best intentions of schools or overcome the highest barriers schools present. One mother, Mrs. Martinez explained that her daughter, Ashley, was so motivated and intelligent that she was enrolled in third grade but was doing all the homework and class work for the fifth grade in addition to her 3" grade work. The school would not let Ashley move up a grade nor would it allow her to cease doing 3rd grade work; so Ashley just did twice the amount of 93 work she needed to be doing. The school presented obstacles; Ashley overcame them through hard work and intelligence. This notion that the individual plays the decisive role in her own success is the basis of a meritocratic society (Boudon, 1973; Willis, 1977). MacLeod (1995) says, ‘Any child can grow up to be president.’ So says the achievement ideology, the reigning social perspective that sees American society as open and fair and full of opportunity. In this view, success is based on merit, and economic inequality is due to differences in ambition and ability. Individuals do not inherit their social status they attain it on their own. (p. 3) Though most parents do not use MacLeod’s language, they generally believe a child’s ambition and ability play an important role in her school success. This brings us to our first proposition: Children influence their own academic achievement. Mrs. Carol and Mrs. Hawill believe that their children are active collaborators in their own achievement. Through their cases, we will investigate how the attribution of achievement to their children influences the mothers’ preferences and choice sets. We will then consider how those preferences and choice sets shape their final choices. Mrs. Carol. Denzel is in the 8th grade at the neighborhood middle school. He is a good looking young man who has a passion for boxing and making people laugh. Mrs. Carol explained that Denzel is artistically talented but hesitant to show off his rapping and singing skills in fi'ont of other people. This is Denzel’s second time through the 8th grade. The year prior to the study, Mrs. Carol explained that Denzel did everything in school but his work. His teachers said he was an enjoyable child to have in class; he didn’t give them any trouble but he just wouldn’t do his work. Though Mrs. Carol was 94 frustrated that the school waited until the end of the year to tell her Denzel was going to be retained, she ultimately agreed that he needed to get himself together. Part of doing that was repeating 8th grade. Mrs. Carol explained, however, that Denzel wasn’t held back because he couldn’t do the work, he was held back because he chose not to do the work. In her words, Denzel is the type who “does the minimum” to get by. Mrs. Carol and Denzel talk on a regular basis about the importance of school and trying to do your best in school: I also try to. . .I don’t want to use the word ‘encourage’ because it’s kind of more than that. I try to coach him, try to push him in the right direction, telling him don’t just go and be there. Go and be the best that you can be while you’re there. ...I tell him don’t just pass the test, ace the test. And that’s the kind of mentality that I want him to have. (2“ interview) Though Denzel did much better in 8th grade the second time around, he still did not get good grades. Mrs. Carol sees Denzel’s lack of motivation as a problem. Consequently, Mrs. Carol doesn’t feel she has a lot of choices about where she can send Denzel to school. Like many other families, the Carols live paycheck to paycheck, so a private school is not an option. Even if she had the money, Mrs. Carol feels that Denzel’s lack of interest in school would simply be a more expensive problem if she were to send him to private school: [If] I’m sending him to public school for fi'ee and he’s going there and he’s not putting forth no effort, then I would really be fighting a losing battle to spend all my money and send him over there [pointing to the private school down the street] and he’s still... ‘Cause you’ve got to be motivated to learn, you know, and 95 that starts fiom within. And. . .that’s what I . . .I’m working on getting him motivated. And that’s hard. (1St interview) Mrs. Carol does not associate Denzel’s lack of motivation with the school environment. Though she concedes that schools can do things to make a child more motivated -- encouraging children or having a clean, welcoming physical environment -- Mrs. Carol believes the problem of under-achievement is Denzel’s. Until he chooses to be more motivated, his academic performance is unlikely to change. Mrs. Carol worries that in schools that expect more effort and work from her son, he will simply “fail out” more quickly and be in trouble more often. Mrs. Carol’s preferences are influenced by her understanding of the cause of Denzel’s poor school record. She cannot afford nor does she believe it would be a good use of money to send Denzel to an even more challenging private or religious school. She prefers a magnet, charter, or neighborhood school, for she considers those types of schools “real options” because she can afford them. Yet Mrs. Carol’s choice set did not contain any charter or magnet schools. It contained two neighborhood schools, one that Denzel talked about attending and another that is his assigned neighborhood school. Mrs. Carol explained that -- given his grades and previous school history -- Denzel cannot be accepted at a magnet or charter school so she did not include them in her choice set. She worried that, even if he were accepted, he would quickly get kicked out due to his refusal to do work. Denzel’s grades, which are, in Mrs. Carol’s estimation, caused by a lack of motivation, shape both her preferences and her choice set. Indeed, they act as a filter for what Denzel’s mother thinks is possible for her son. 96 In the fall, Mrs. Carol enrolled Denzel in his assigned neighborhood high school, Pinecrest. When I talked with her in October, she was satisfied with his performance. Mrs. Carol went to parent-teacher conferences, met his teachers, and found that Denzel is making satisfactory progress in all of his classes. “Right now, he isn’t failing any classes and he seems to like it alright. But we’ll see.” He has continued to box and is talking about trying out for the football team next year. Mrs. Hawill. Alecia, who has a twin brother, Brandon, is also in the 8th grade. Both Alecia and Brandon attended a charter middle school the Hawills “loved.” By her mother’s description, Alecia is a bright young lady who gets “very good grades.” She keeps to herself, has a few close friends, and loves to read. She also has her future mapped out, right down to where she wants to go to college and medical school. Alecia has researched the schools and expenses associated with going. She knows what kind of grades she needs to get in and she is planning ahead. Mrs. Hawill believes Alecia will be a doctor someday and wants to provide her every opportunity to reach that goal. We were raised that education is important. You know? People say it’s the key that unlocks the door. You can do whatever you want to do. If you’re educated, you can do. So I don’t want them to be denied anything because they didn’t know. They need to have an education. That’s something that nobody can take away. (2“ interview) Mrs. Hawill explained that Alecia has some personal characteristics that enable her achievement: “A lot of people, I’m not saying just children, a lot of people are stuck in culture. She’s not. She doesn’t see color at all. At all. And that alone will take her far.” Alecia is the type of child for whom color doesn’t matter. Mrs. Hawill feels Alecia 97 is accepting of all people, doesn’t judge, and is able to get along with everyone. She feels these traits will help Alecia succeed in school and life more generally. Mrs. Hawill described Alecia as a very independent thinker and yet a trusting person. This combination means that Mrs. Hawill has to “keep her challenged.” There was no way Mrs. Hawill could send Alecia to a neighborhood public school: “My daughter is very trusting and I feel if she went to a public school where she wasn’t challenged she would get. . .I would lose her. She needs a school where she will be pushed.” Mrs. Hawill realizes how motivated and driven Alecia is and wants to provide a supportive environment that will help her realize her dreams. Mrs. Hawill’s beliefs about Alecia influence her preferences for private, religious, magnet, and charter schools, all of which Mrs. Hawill feels can challenge Alecia. Mrs. Hawill believes that neighborhood schools have an “unchallenging environment,” are therefore, are not an option for Alecia. Homeschooling also is not an option because Mrs. Hawill works and she does not feel that homeschooling is a healthy social situation for a young person. Mrs. Hawill’s choice set contained one private, one religious, three magnet, and one charter school. Alecia began school in the fall at a magnet school, Xavier. Through Alecia has liked Xavier and done well academically, she has been admitted to Jefferson (another one of the magnet schools in Mrs. Hawill’s choice set), the most academically prestigious high school in the city. Alecia will be transferring to Jefferson in the 10m grade. Satisfied with Xavier, Mrs. Hawill is nonetheless is eager for Alecia to attend Jefferson because she believes it will be more challenging for her. Moreover, Alecia wants to go. 98 Mrs. Carol and Mrs. Hawill. Both Mrs. Carol and Mrs. Hawill attribute their children’s school success to intrinsic motivation. Denzel is unsuccessful because he chooses not to do his work even through he is perfectly capable of doing it. Alecia is successful because she is motivated, not “stuck in culture,” and has clear goals. Children’s characteristics are central to both mothers’ choices of schools. Though each mother acknowledges that schools play a role in academic success, they focus on motivation being critical to success. And that motivation is attributed to the child, not the school. Mrs. Carol acknowledges that schools can provide the right environment (by “being encouraging” or “welcoming”) for success, but ultimately academic success comes down to Denzel “stepping up.” Mrs. Hawill also acknowledges that some schools are able to challenge students more than other schools but, for her, the school features are secondary to the fact that Alecia is smart and motivated and will “do well wherever she goes.” From the mothers’ perspectives, their children’s motivation comes from within and academic success (or lack thereof) is a result of that motivation. They attribute academic success to their children, not their children’s schools. Mrs. Carol’s and Mrs. Hawill’s preferences are shaped by their beliefs about their children - who they are and what they need. These preferences shape their choice sets, and ultimately, school choice. Mrs. Carol doesn’t think that Denzel would do his schoolwork in a private or religious school, so those schools are not options. She also doesn’t think he can get into a charter or magnet school, so those are not options. This leaves her with one type of school to choose from: neighborhood schools. Attributing low achievement to Denzel focuses Mrs. Carol’s attention on motivating Denzel rather than on finding a school that can help Denzel be successful. Because she does not blame 99 the middle school for his low achievement, the next school is unlikely to be viewed as the solution to the problem of that low achievement. There are similarities and differences between Mrs. Carol and Mrs. Hawill. Though Mrs. Hawill cares very much where Alecia goes to school, she does not attribute Alecia’s school success to the school. She attributes it to Alecia. This is similar to Mrs. Carol. Because Alecia does well in school, however, Mrs. Hawill’s choice set includes a wider range of potential school types than Mrs. Carol’s choice set. Mrs. Hawill feels Alecia’s academic record makes it possible to consider many different schools. Further, Alecia’s motivation and intelligence give Mrs. Hawill confidence that Alecia will be able to keep up wherever she goes to school. This is quite different from Mrs. Carol." Perhaps the most important difference between the two mother’s choice sets is their quality. Mrs. Carol considered two failing schools, one that has 17% of its 1 1m graders at grade level and the other (Pinecrest) which has 14%. Mrs. Hawill considered six “non-failing” schools. Of the three magnet schools she considered, the percentage of students proficient on the “111 grade state tests were 48%, 57%, and 71%. These mothers have qualitatively different choice sets. Both children ended up attending public schools, so the striking difference in choice set quality cannot be attributed to being able to afford a good school. Further, the Carols and Hawills live less than a mile from one another. They have different preferences; those preferences are visible in their choice sets. Both ’5 The mothers’ preferences are based both on their children’s current motivation and their prior academic record. Alecia does well in school and has done well in school. Denzel is not doing well in schools and has not done well in the past. The children’s records and their cunent motivation are related but distinct influences on their mothers’ preferences. 100 mothers attribute their child’s achievement in part, to their child. But because Alecia is successful in school and Denzel is not, their mothers construct very different choice sets. Across the Sample. Across the sample, children who had better grades had greater odds of ending up in selective schools" but they had similar odds of ending up in “non-failing” schools. There was not a statistically significant nor important relationship between the child’s grades and whether or not they ended up in a failing school ()6 (2) = 0.57, p = .75, (p = 0.11). There was however, a moderate and statistically significant relationship between children’s grades and whether or not they ended up in a selective school (x2 (2) = 6.89, p = .03, (p = 0.40). In other words, if a child got good grades (As and Bs) in school, she was more likely to attend a selective school than a child who did not get good grades. If parents are attributing their child’s success to their child (more than the school) the way Mrs. Hawill does, this pattern is reasonable. This finding contradicts however, a finding in the voucher literature. The evidence from some voucher experimentsl7 suggests that students who participate in voucher programs have lower test scores than other children who are eligible to participate but choose not to (Gill etal., 2001; Howell, 2000; Moe, 1995). This contradicts the finding above in which children that earn higher grades have similar odds of attending a selective school (which a voucher school is). Based on the Gaston findings, we would predict that voucher students would be more likely to have higher, not lower, test scores as compared with their peers. There are three potential explanations for this apparent contradiction. ’6 Magnet, private, religious, and homeschools are considered selective schools. '7 Voucher programs give parents access to selective schools in the form of private and religious schools. 101 First, it is possible that voucher children’s parents attribute academic success more to the school than the child. This explanation is plausible because voucher parents tend to be more dissatisfied with their child’s former school than similar parents who do not use vouchers (Gill et al., 2001; Witte, 2000), thus making them more likely to attribute low performance to the school, not the child. Alternatively, if voucher parents are similar to Gaston parents and do attribute academic success to the child; perhaps they are willing to try vouchers because they have simply run out of explanations for their children’s school failure. Vouchers may be a last ditch effort to change their child’s academic performance. This is plausible because many parents who participate in voucher programs have more formal education than comparable parents who elect not to participate in such programs (Gill et al., 2001). Higher levels of education might be associated with a stronger belief in the efficacy of schools, thus making parents more willing to change their child’s school in order to improve the child’s academic achievement. Finally, the contextual features of Gaston (in particular the inclusion of magnet schools which require a test for admission) may influence who attends selective schools in ways that are different from the contexts in which voucher experiments have been conducted. In sum, the data suggest that parents in this sample believe that, together with other factors, their children influence their own academic achievement; this belief shapes parents’ preferences. In particular, parents have views of which schools their children can gain admission to. If it appears that a child will not be accepted at a school, a parent is not likely to consider that school a real option. Parents also have views about which types of schools (private, charter, etc.) are appropriate for their child. Some parents for 102 example, may never consider a private school or a magnet school. Because many parents attribute academic achievement to their child’s motivation and hard work. the schools they consider “real options” (i.e., their preferences) vary with their assessment of their children’s academic characteristics. This assessment hinges on the parents’ attribution of their child’s academic achievement to the child or the school. In general, Gaston parents do not attribute academic success or failure to schools. They see it largely as a function of the child. This finding is consistent with the Phi Delta Kappan/Gallup poll which reports that only 16% of citizens feel the achievement gap between White students and Black and Hispanic students is “mostly related to quality of schooling” (Rose & Gallup, 2003). Eighty percent of respondents believe that other factors cause the achievement gap. In those citizens’ views, family life and poverty cause poor academic achievement, not bad teachers and inappropriate curriculum. This is a problem for schooling markets. If parents don’t attribute poor academic achievement to schools, policy makers must either intervene in parents’ beliefs about what influences academic achievement or they must convince parents to attribute academic achievement to schools. Both of these options are unprorrrising as they require policy makers to change what people think, a task notoriously difficult to do (e.g., Cohen, 1988). Proposition Two: My Child is Typical The parents in this study are attentive to their children. They know what their children are good at, what they need help with, and what hurts their feelings. At length, parents described both the good and the bad about their children. When asked to describe their children, they said things like “he’s a good boy,” “she’s a very sensitive, sweet child,” “she is helpfirl, respectful, and perceptive,” and “he makes us laugh all the time.” 103 They also described more frustrating aspects of their children: “He doesn’t want to do what you want him to do” and “she needs to be a leader, not a follower” and “she is just lazy sometimes.” Though there were a couple of parents who seemed to believe their children could do no wrong and a few parents who were at wits end about how to change their child’s behavior, most parents described their children as good kids, young people learning how to be in the world with all the joys, temptations, and challenges of growing up. Because parents pay attention to their children, they are often asking themselves, either implicitly or explicitly: “How is my child doing?” When their children are acting stubborn or starting to notice the opposite sex or generally behaving in a novel way, parents wonder, “Is this normal?” In order to answer this question, parents refer to their sense of what is normal or average. This usually means that parents compare their child to his or her siblings, cousins, neighbors, or friends that are the same age. If their child seems to be in the same ballpark as other comparable children, the parent is reassured. As an 8"h grade mother said, “They don’t come with instruction manuals!” Parents have to rely on their own sense of what is happening with their child in order to ascertain whether or not a particular action is worthy of concern. Almost three-quarters of parents felt their children were “typical” children and most parents (83%) were hoping that the next school would support and continue their child’s current success. Of course, what constitutes “typical” is relative. With respect to schools, what is typical in one community may not be typical in another community. Demographic features, ways of speaking, and institutional features differ across contexts, for schools and the children in them vary (e.g., Kozol, 1991; Rubinowitz & Rosenbaum, 104 2000). One version of typical is not better or worse than another, but some versions are privileged in particular outside contexts. For example, it rrright be typical in one community for children get good grades, play an instrument, and be on a sports team. In another community, it might be typical for children to get good grades but spend most of their fiee time helping their parents on the family farm or at the family business. These versions of typical may be privileged in different ways in outside contexts such as college admissions. A child fiom the first community may be viewed as being “well-rounded” and therefore desirable; a child from the second community might be viewed as having been “sheltered” and is therefore less desirable. Schools influence parents’ perceptions of typical, which brings us to our second proposition: “My child is typical.” With Mr. Dish’s and Mrs. Grisson’s cases, we consider how parents use reference groups to assess their child’s typicality, for those reference groups give meaning to “doing well” and enjoying “success.” We will also consider how those reference groups shape parents’ preferences for particular schools, which in turn shapes their choice sets and ultimate school selections. Mr. Dish and Mrs. Grisson, like most of the parents in the study, view their children as typical. Though each child has particular characteristics an outsider might not consider typical, if we are to understand parents’ preferences, choice sets, and final choices, we must understand their views of their children. 105 Mr. Dish. Mr. and Mrs. Dish have four children they are responsible for, three are their grandchildren and one is their daughter. All are between the ages of 14 and 8." Mr. Dish, a retiree, is the primary person responsible for schooling decisions. Mr. Dish describes Tasha, a 5th grader at a science magnet school, as a typical child who makes fiiends easily, is self-confident, outgoing, and bright. She is a hard worker and can “be stubborn when she wants to be.” Like her siblings and fiiends, Tasha has many learning opportunities outside of school. For many years, she has attended an overnight summer science camp in Minnesota. Recently, she was selected to participate in a youth visit to Australia with People to People, an organization committed to developing leaders who want to work toward global peace. She, like her older siblings, will spend this summer traveling to another country and learning about the issues associated with global peace. Mr. Dish explains that his goal is to keep Tasha on the right track. “Part of my agenda is to keep the kids busy, to keep them on the same track as the kids around them.” In Tasha’s case, “the kids around them” are the kids in her magnet school, at her summer science camp, and in programs like People to People. Mr. Dish believes the kids around Tasha are on the right track and he wants Tasha there with them. Mr. Dish explained, In any school you’ve got a group of kids who are focused on really moving ahead and learning and then you’ve got another grouping of folks who basically. . .don’t care and. . .try to make failing the standard. And you’ve got both of those ’8 Mr. Dish is a European-American who has been married for over 20 years to Mrs. Dish, an African- American woman. In describing his children’s physical appearance, he said, “If you saw them on the street you’d think they were Black.” 106 groupings mixing. And we’re always concerned about which direction we keep em kids focused, going in the right direction. (1St interview) Mr. and Mrs. Dish focus their efforts on trying to keep Tasha engaged and focused in the right direction. They believe it is possible to get in with the group of kids who try to “make failing the standard.” They are determined to have their children stay with the other group of children and stay focused on learning. Mr. Dish has been successful thus far in keeping his two oldest children moving “in the right direction.” For both, be selected Wilson Middle School, the science magnet school into which Tasha’s elementary magnet school feeds. Mr. Dish feels “They [Janet and Terrell] have thrived at Wilson. I know all the teachers and they know me. We work together to keep the kids going.” He explained that both Janet and Terrell have excelled at Wilson. Janet, the oldest, recently won a statewide science competition. He commented that the Wilson students “beat out Upland Hills and some of the very high end private schools to win that competition.” Mr. Dish has watched Janet thrive and excel for three years at Wilson. He believes Wilson provides a generative learning environment for his children. Tasha, like her older brother Terrell, is dyslexic. She has received special education services throughout elementary school and although she struggles in reading, she excels in other subject areas. Terrell has made a lot of progress on his dyslexia at Wilson. Terrell’s special education teacher is trained in the Johnson method, a particular approach to dealing with dyslexia. This method has proven to be an excellent match for Terrell’s needs. Terrell has made so much progress that he recently made the honor roll and his reading is markedly improved. Because Terrell has been so successful with his 107 special education teacher and Tasha has a similar form of dyslexia, Mr. Dish is eager to have Tasha begin working with Terrell’s teacher. Primarily because of his experiences with his two older children, Mr. Dish did not consider sending Tasha anywhere but Wilson. Wilson was the only school in his choice set. He said, “I’m real pleased [with Wilson], that’s why that’s where she’s going.” He also explained that he is “real into the science. Lots of others parents could take it or leave it but I want that. I want it to continue in their education. I think it will be a real asset when they are older.” Based on his experience with Janet and Terrell, Mr. Dish believes Wilson will be a school in which Tasha’s dyslexia needs will be met, she will continue on the science path she currently is on, and she will be with other students who will keep her on the right track. In late October, after Tasha had been attending Wilson for over two months, Mr. Dish observed, One thing I’ve noticed is that she’s working very hard. She was a good worker at Applebee [her elementary school] but she is working even harder at Wilson. I think the resource room is really important to her. They get a lot of emotional support. I haven’t had reason to doubt my reasons to put her there. It might be working even better than I anticipated. But I’m reading the tea leaves there. We’ll see. (3’6 interview) The decision to send Tasha to Wilson seems to have been a positive one, but —- ever watchful of his children’s progress -- Mr. Dish is reserving judgment. Meanwhile, the two oldest children continue to do well. 108 Mrs. Grisson. Mrs. Grisson is the mother of six children, three boys and three girls. Shawn, the third oldest is in the 8th grade at the neighborhood public school, Evergreen. Mrs. Grisson describes Shawn as a “good kid” who has a big heart and wants to help people. He likes to read (especially Harry Potter), talk on the phone, and go to the movies or the mall with his fiiends. Like Mrs. Grisson, Shawn has had some “learning problems.” Early on in school, Shawn had a hard time learning how to read, but “he’s mostly grown out of that now.” Mrs. Grisson explained that Shawn used to have some type of disability but she could not remember the label. He did have an IEP and she used to attend meetings at the school about his plan. A few of his elementary school teachers really reached out to their family and “tried to teach my kids.” From making sure their family had a turkey on Thanksgiving to making sure the children understood their homework, Mrs. Grisson feels lucky that her children have had some good teachers along the way. Shawn’s 6th grade teacher changed everything for Shawn by taking him under her wing and making him believe he could learn. Mrs. Grisson feels that Shawn has overcome many of his reading problems and she credits this both to his teachers and his “growing out of it.” Mrs. Grisson sees Shawn as similar to his peers. He likes what his friends like. He has lots of friends and is self-confident. Recently the state and national standardized test scores were mailed home. After talking with her neighbors, Mrs. Grisson discovered that a lot of the students in Shawn’s class scored “average” on the tests. “He scored average and below average in a few areas.” Though Mrs. Grisson wasn’t sure she understood everything the test report said, she felt that Shawn’s scores seemed to say that he was about average, “along with everybody else.” From Mrs. Grisson’s perspective, 109 Shawn is a typical teenager who likes the same things his friends like. He has made good educational strides. He is reading well, no longer needs and IEP, and gets along with his teachers. Mrs. Grisson also sees a lot of herself in Shawn. I did like going to school but I did have problems learning. I didn’t get things like I should. I barely made it out of high school anyway. I did not graduate when I was supposed to. But I went on back. I could have dropped out, but I didn’t. I went back and finished. You know I had to pick up one of those. ....transcripts, the other day and whewl I didn’t do very good. I barely made it through but I’m glad I did. And 1 sort of see that in him [Shawn]. (3rd interview) Like, Mrs. Grisson, Shawn has some learning problems but he perseveres. He may only get average test scores and grades but he continues to go to school and put forth good effort. All of this makes her think that her son is typical. In thinking about where to go to high school, Mrs. Grisson considered three schools. She focused mainly on the neighborhood high school, Pinecrest, though early on in the process she did entertain the idea of sending Shawn to a nearby magnet school. Ultimately, Mrs. Grisson did not apply to the school because she felt Shawn would not be able to get in because of his grades. She worried that it would really “disappoint him” if he applied and wasn’t admitted. “That wouldn’t be good for him right now.” The third and final school Mrs. Grisson briefly considered was their old neighborhood’s high school. Because the Grissons recently moved to their current neighborhood, Shawn’s older sister, Tracy, will be graduating this year from their previous neighborhood’s high school, Roanoke. Mrs. Grisson has been very pleased with Tracy’s progress at Roanoke. 110 “She’s done so well. It’s been just a great school for her.” Mrs. Grisson describes Tracy as a very responsible young lady who gets good grades and is currently applying to college. Mrs. Grisson feels that Roanoke is a good school and she would have liked Shawn to attend so that he could be with his sister. But Shawn has made new friends, adjusted to the new neighborhood, and doesn’t want to go to Roanoke, which he sees as another big change. Mrs. Grisson understands this and feels it is in Shawn’s best interest to keep him in a situation they know is working, especially when it is the situation he prefers. In the fall, Shawn was doing well in all of his classes except math. Mrs. Grisson said none of the family is very good at math. Shawn liked his history teacher and guidance counselor the most and often spent his free time during the day with them in their classroom and office. Mrs. Grisson is satisfied with her choice but is worried about Shawn’s math class and all the social changes high school brings more generally. For now, she feels that firings are going well, but she intends to keep her “eyes open”. Mr. Dish and Mrs. Grisson. Both Mr. Dish and Mrs. Grisson see their children in a positive light. Tasha and Shawn have things going for them. They both like school and earn respectable grades. Tasha has been selected to be a part of People to People and Shawn has largely overcome his battle with learning to read. Both children are at positive points in their development as students and people. Mr. Dish and Mrs. Grisson recognize where their children are and they would like the next school to support their positive trajectories. But the features of each child’s positive trajectory are quite different. Mr. Dish and Mrs. Grisson are comparing their children to qualitatively different peers. Their decision making relies on different reference groups. Mr. Dish 111 compares Tasha’s progress to her siblings, her magnet school peers, and the prototypical child who travels around the world in the 6th grade; in short, achievement- oriented, college-bound children who attend selective schools and summer science camps. Mrs. Grisson compares Shawn’s progress to his siblings, children in the neighborhood, and herself; people who like school, work hard, struggle, graduate from high school, and attend some college. Compared to their respective reference groups, Tasha and Shawn are on track. They are typical. The parents’ assessment of progress, which relies both on a reference group of other children and on the parents’ own histories of schooling, shapes what Mr. Dish and Mrs. Grisson prefer in a middle and high school, respectively. Mr. Dish prefers a school that will work with Tasha’s dyslexia and continue the science focus she began in 1’1 grade. Based on his older children’s positive growth and development, Mr. Dish knows that Wilson can support Tasha’s science development as well as continue working on her dyslexia. He does not need to consider any other schools, for Wilson offers what he wants. Mrs. Grisson prefers a school that Shawn can be accepted to and where Shawn is likely to continue his forward progress. She also prefers a school where Shawn may be lucky enough to have more teachers like his 6th grade teacher. These preferences are as likely to be met at Pinecrest as the previous neighborhood high school, Roanoke, so Mrs. Grisson prefers the school Shawn wants to attend and the one that will provide the most stability given their recent move. Both parents selected schools which allowed their children to follow the customary attendance pattern and stay with their friends, in part because they view their children as typical and Pinecrest and Wilson is where typical children go to school next. 112 Mr. Dish and Mrs. Grisson’s choice sets were shaped by their preferences. Those preferences hinged on an assessment of their child as typical, in comparison to some reference group. If preferences are shaped by parents’ reference groups, it is easy to imagine that, depending on who a parent compares her child to, the child’s progress and status may look quite different. Thus choice sets may look quite different depending on the characteristics of the reference group. Had Mrs. Grisson compared Shawn to Mr. Dish’s reference group, she may have concluded that Shawn was lacking certain kinds of knowledge and experiences. She may have concluded that she needed to search for different schools. The same might be true of Mr. Dish. A different reference group may have prompted a different choice set and ultimate selection. We can only speculate. Both Mr. Dish and Mrs. Grisson ultimately selected schools that are failing. Mr. Dish chose a selective school (magnet) and Mrs. Grisson chose a non-selective school (neighborhood), though on standardized test scores, the schools look quite similar. Both parents were satisfied with their choices. In October, they reported that Tasha and Shawn were doing well, getting good grades and developing supportive close relationships with a couple of their teachers. Their children had friends and were happy. Across the Sample. In these two cases, both parents viewed their children as typical and both selected schools that fall within the customary school attendance pattern. This was not the case with all the parents. In fact, viewing a child as typical had no statistically significant relationship with choosing a school in the customary school attendance pattern ()6 (1) = 0.0061, p = .94, (p = .011). There was also no evidence of a relationship between viewing your child as typical and the child ending up in a non- failing school ()52 (l) = 2.53, p = .11, q) = -.24). Nor was there evidence of a relationship 113 between children being viewed as typical and ending up in a selective school (x2 (1) = .014, p = .91, (p = .017). Parents who viewed their children as typical were no more or less satisfied with their choices as compared with those who did not view their children as typical (x2 (2) = 1.33, p = .51, (p = .17). Overall, there were many different outcomes for children who were considered typical by their parents. They attended schools that are selective and non-selective, failing and non-failing, inside and outside customary school attendance patterns. The cases of Mr. Dish and Mrs. Grisson demonstrate the larger pattern across parents: a parent’s preferences, choice set, and final school choice decision are influenced by their understanding of their child as typical, but that understanding does not play out in predictable ways with respect to the variables policy makers might be interested in. The variable policy makers and choice scholars are interested in is opportunity. One goal of choice is to create more opportunities for urban children to get out of failing schools. On this goal, parents’ actions are not encouraging. Parents tended to choose schools that were, along status and selectivity dimensions, similar to the schools they just left. There was a very strong and statistically significant relationship between the child’s original school status (failing v. not-failing) and their final school status (x2 (1) = 11.21, p <. 01, (p = .50, OR = 9.64). Of the 22 children who began the study in a failing school, 18 ended up in a failing school (82%). Of the 22 children who began the study in a non- failing school, 15 ended up in a non-failing school (68%). This relationship tells us that children who attended 5111 or 8’“ grade in a failing school had 9.6 times the odds of ending up in a failing school as compared to children who attended a 5th or 8th grade in a non- failing school. This is discouraging. If choice is supposed to create the opportunity for 114 different educational outcomes for children, one part of that is getting children enrolled in different (and better) schools." Scholars have long argued that we are most comfortable with the familiar (Dewey, 1990). The evidence in this study suggests how this works psychologically: The status quo is “normal,” it is what parents and children come to expect. And so parents judge their children against the typical, in this case schools that are “working” for their children, even though they are failing. A similar pattern exists for school selectivity. There is a very strong, statistically significant relationship between the child’s original school selectivity (selective v. non- selective) and their final school selectivity (x2 (1) = 14.89, p < .001, qr = .57, OR = 14.17). Of the 20 children who began the study in non-selective schools, seventeen ended up in non-selective schools (85%). Of the 28 children who began the study in a selective school, 20 ended up in a selective school (71%). Children who attended selective schools had 14 times greater odds of ending up in a selective middle or high school as compared with their peers who spent 5th or 8th grade in a non-selective school. Again, this pattern is caused in part, by parents coming to see their children’s progress as typical. So being in a school that does not accept just anyone is how the world works, it is taken-for-granted. Together, these findings suggest that, although there is movement between failing and non-failing, selective and non-selective schools, the vast majority of children attend middle and high schools that are similar to their elementary and middle schools. ’9 Because the measure of failing is blunt, we cannot, using that criterion, ascertain whether or not Mr. Dish and Mrs. Grisson enrolled their children in better schools. The issue of determining whether or not children are attending better schools is taken up in the final chapter of the dissertation. 115 Parents preferred and selected schools that were similar to their child’s previous school for many reasons. The widely held proposition that “my child is typical” offers us one way to understand parents’ preference for schools that are similar to the child’s previous school. It is possible that parents’ ideas about what constitutes an appropriate school implicitly rely on comparisons to other children (reference group(s)). Parents consider choice sets that are consistent with those ideas. They then choose schools from those choice sets. For example, parents with children in selective or non-failing schools have a specific set of experiences and understandings. They might be surrounded by children who are interested in school, speak two languages, and do enrichment activities in the summer. Parents then may come to prefer schools that will hold their children’s interest and are challenging. They may prefer a middle school with a foreign language program because, after all, their child had language instruction in elementary school; why stop now? They may prefer schools that offer interesting summer activities to engage their children. Parents of children in non-selective or failing schools have similar influences on their preferences. They might be surrounded by children who struggle through school, have one or two teachers who show a strong interest in them, and have close supportive friendships with other children. Those parents then may come to prefer schools that respect their child as an individual, have teachers who know how to teach and love their children, and allow their children to remain in close supportive fiiendships. Both sets of parents then choose schools that are consistent with their idea of typical, ultimately selecting schools that are similar to the ones they just left. Mr. Dish and Mrs. Grisson show us that there are multiple definitions of typical. These definitions are developed through comparisons made with the reference group. 116 Similar to the chain of influence we saw in the first proposition (which went from a parent’s assessment of their child’s academic motivation, to preferences for certain schools and school types, to the selection of an appropriate school), we see parents’ assessment of their children’s progress influencing the schools they are willing to consider and ultimately select. Thus far, the first two propositions demonstrate the ways that folk theories shape and bound parents’ preferences along two dimensions: academic achievement and child development. We now turn to the third proposition. Proposition Three: I Know What Quality is for My Child The parents in this study are from Gaston. Though many of them live in different neighborhoods than the ones they grew up in, they have, for the most part, always lived in Gaston. Not only have they lived in Gaston all of their lives, ninety-four percent of them were also educated in the Gaston Public Schools. Except for three of the 36 Gaston parents, as a group, the parents liked school and have positive memories of growmg up in Gaston and its schools. Raised and educated in the city, Gaston parents have opinions about their city’s schools -- both public and private. They have ideas, which are informed by tremendous local knowledge, about which schools are right for their children. Many of the 36 city residents interviewed for this study send their children to schools inside the Gaston city limits. Of those 36 parents, twenty-three chose failing schools. The preferences of those 23 parents appear to be in opposition to the state’s assessment of those schools’ quality. From an outsider’s perspective, the parents appear to prefer low quality schools. But quality can be judged in many ways and the state’s 117 assessment of school quality relies almost exclusively on standardized test scores. Parents’ interpreted quality in much broader terms and they made decisions based on their own interpretations. As we will see in the next two cases, parents believe they know what a “quality” school is for their child and make decisions accordingly. Through the cases of Mrs. Moore and Mrs. Robinson, we will see how parents use their own -- rather than the official -- assessment of quality to select schools which match their preferences. We will see how parents’ definitions of school quality shape their choice sets and final school selections. I limit the explanation of this proposition to the parents who live in Gaston for one reason. Suburban parents’ preferences were almost never in conflict with the “official” assessment of school quality. Suburban parents were not positioned between conflicting accounts of their communities and schools. So while they too believed that they know what is best for their child, those judgments were rarely in conflict with the state’s “failing” list. Suburban parents’ preferences might well be influenced by this proposition, but there was no discussion of it in the interviews and the proposition was most clearly evident in Gaston parents’ preferences. Mrs. Moore. Mrs. Moore is a Gaston resident who grew up and was educated from elementary school through college in the city. She is married and the mother of four children, one of whom is a CPA and lives in Maryland with her husband and brand new baby. Mrs. Moore’s other three children range in age from 18 to 13 and live at home. The youngest, Clifton, is in the 8th grade at a neighborhood middle school, Evergreen. Mrs. Moore has been very pleased with the school. She feels Evergreen has excellent programs, cares about the children, and tries to reach out to families. The 118 Moores always attend basketball games and family fun nights and Mrs. Moore volunteers for many of the parent group events. Mrs. Moore feels Clifton has had some wonderful teachers at Evergreen. A few teachers in particular have been nurturing to Clifton by encouraging him to participate in the many programs the school offers. One program in particular -- Mentoring with Pride -- has had a big impact on Clifton. Mentoring with Pride requires children to stay after school to meet with a college mentor and work on the computer twice a week. Mentors and children work on the child’s academic weaknesses. Mrs. Moore explained that once Clifton started the mentoring program he was “just a changed kid, much more outgoing and self-confident.” Mrs. Moore is aware, however, that Evergreen has been on the state’s list of failing schools. She explained that she disagrees with the state’s designation. “My kids have had wonderful teachers. No matter what they say about the school didn’t make the test, these are wonderful teachers.” Mrs. Moore feels that you “can’t judge a school only by test scores anyway. Scores aren’t everything.” Mrs. Moore explained that, at times, she and her husband have strong disagreements about where the kids should go to school. Though they both want the same things for their children, they don’t always see the same school as the right choice. For example, when deciding on a high school for Clifton, Mr. Moore felt it would be good for Clifton to go to Truman High School to be with his older sister, Rose. Though Rose has excelled at Truman, Mrs. Moore didn’t feel Clifton would do well there. Mrs. Moore felt the school was too big and didn’t provide the kind of social environment Clifton needed to feel comfortable and be successful. Clifton, in comparison to his sister Rose, struggles in school and has to work very hard for his grades. Mrs. Moore preferred 119 a school which would give Clifton a lot of personalized attention. Mrs. Moore felt the classes must be small because it seems “almost impossible to get what you need if there are 39 other children in the classroom.” She also preferred a school filled with teachers who care about and take an interest in their students. Both Mr. Moore and Clifton had ideas about where Clifton should go to high school. Mr. Moore felt Clifton should attend Truman; Clifton was interested in Carnegie. Both are public high schools outside their neighborhood, but inside Gaston. Mrs. Moore considered those schools, but ultimately selected South Mill Academy, a charter school that she discovered through a girlfriend. The girlfriend’s own children have attended the school for almost two years and she has been very pleased with her children’s progress. Mrs. Moore visited the school, gathered information about it on the web and decided it was the right school for Clifton. She felt the school had the right mix of small classes, personable, caring teachers, a good curriculum, and a number of after-school programs that Clifton is interested in. South Mill was neither Mr. Moore’s nor Clifton’s first choice. Though both of them approved of the school, they preferred other schools. By October, however, everyone was won over. Clifton was getting good grades, had made new friends, was participating in the sports programs, and was very happy at South Mill. Mrs. Robinson. Like Mrs. Moore, Mrs. Robinson is a married, college graduate. She too was born and raised in Gaston and attended Gaston Public Schools. Currently, she teaches kindergarten in the district. Mrs. Robinson has two children, Raven and Robert. Robert is a kindergartener in his mother’s classroom and Raven is a 5m grader at Park Slope, a small magnet school. Mrs. Robinson has been very pleased with Park Slope because she feels the teachers have been very supportive and attentive toward 120 Raven. Raven has earned excellent grades throughout elementary school and always scores very high on the state’s standardized tests. Mrs. Robinson has been very involved in Park Slope since Raven was a 1” grader. Over the years, she has spent a great deal of time at the school, volunteering for various projects and fundraisers. Most recently, Mrs. Robinson was the keynote speaker at the fifth grade graduation. When Mrs. Robinson began to look for schools, she was principally concerned with finding a school for Raven that had serious academics. “I just want her to achieve. Because she’s going to college.” Though Mrs. Robinson did not take test scores into consideration during her search process, she did look into each potential school’s curriculum. She was particularly interested that Raven would be challenged and exposed to the level of classes that would prepare her for high school. In addition, Mrs. Robinson also felt strongly that Raven needed a place where she would be socially comfortable. Raven is a “quiet, obedient child,” who loves school and is motivated to do well. In thinking about where Raven would fit in, Mrs. Robinson considered the way the children at the school carried themselves and acted toward one another. Mrs. Robinson’s final concern centered on the sense of community at the school. Raven’s elementary school was very small and “everyone knew everyone.” Mrs. Robinson preferred a middle school that would be similar to Park Slope in terms of size and community feel. Overall, Mrs. Robinson preferred a school that was small, academically focused, and had a strong sense of community. When Mrs. Robinson began to check into schools, she was advised by other parents and teachers that the best place for Raven was the magnet school which most Park Slope children go on to attend, Bedrock Middle School. Most parents at Park Slope 121 felt that Bedrock was both the logical progression after Park Slope and the premier middle school on the south side of the city. Though Mrs. Robinson applied to Bedrock, she was unimpressed with what appeared to be disorganization in the main office. Staff did not provide clear answers to her questions about the application deadline or timeline for decisions on submitted applications. If the disorganization in the office was any indicator of how the school functioned, Mrs. Robinson wasn’t sure Bedrock was the best place for Raven. She also applied to and was accepted at two well respected charter schools. Though she liked one of the charter schools very much and it came highly recommended from a colleague, Mrs. Robinson was drawn to a third school, a magnet school called Lincoln Middle School. She described her impressions of this school: The first time I went up to Lincoln. . .I thought, ‘this is so nice.’ Because I felt it was a regular school. They had um, uniforms. I didn’t see children running around. They were very well behaved. I was quite impressed. It was clean. . .I got to meet some of the teachers and they explained the types of things they do in their classrooms. . ..they said that the parents automatically become a part of their family. They had sign up sheets there and all that. You’re expected to participate. I thought that was good because I like to be known. You know, it’s just better if something goes wrong or whatever, they know who I am. (3rd interview) Though Mrs. Robinson liked Lincoln, she felt the choice of which middle school was a difficult one. Many of Raven’s friends were going to Bedrock and it had an excellent reputation among Park Slope parents. Mrs. Robinson felt there must be some good reasons for parents’ enthusiasm for the school. But Mrs. Robinson was really drawn to 122 Lincoln. Her visit to Lincoln convinced her that Lincoln was more of what she wanted in a school for Raven. It was smaller, more fiiendly, appeared to be more organized, and had the academics Mrs. Robinson wanted. In the end, Mrs. Robinson chose Lincoln. Like Mrs. Moore, Mrs. Robinson has been very pleased with her choice. Lincoln has been a supportive, welcoming environment for Raven. Raven has continued to earn excellent grades and enjoy school. Mrs. Moore and Mrs. Robinson. Mrs. Moore and Mrs. Robinson had their own definitions of school quality. They used those definitions to construct choice sets and choose schools. In fact, both mothers selected schools which, as of late October, were highly satisfactory to themselves and their children. But their selections stood in opposition to their farme and friends and the state’s official assessment of quality schools. Both Mrs. Moore and Mrs. Robinson selected schools that are, from the state’s perspective, failing. But from the mothers’ perspectives, they are not.20 Mrs. Moore and Mrs. Robinson had their own ideas about high quality schools. Their views were based, in part, on hours of involvement in their children’s middle and elementary school educations. Mrs. Moore wanted to find a school with small class sizes that had caring teachers and after-school activities. Mrs. Robinson wanted to find a smaller school with a good curriculum and a strong sense of community. They both found schools that matched those preferences and sent their children to them. From the 2” I do not know if Mrs. Moore and Mrs. Robinson know that the schools they choose are failing. 1 did not ask any of the parents if they knew the status of the school because I did not want to convey the message that the state’s list should be consulted or that their parenting skills were going to be judged based on whether or not they chose a non-failing school. 123 mothers’ perspectives, they did not choose failing schools, they choose schools which they thought would be good places for their children to continue their educations. Neither Mrs. Moore nor Mrs. Robinson used state test scores as a measure of school quality. Both mothers did, however, use other measures of school quality. Mrs. Moore took account of the types of teachers at the school, she investigated the school’s curriculum, she gathered information about the school’s activities and building space, and she spoke to her friend who knew the school first-hand. Mrs. Robinson investigated how the children behaved and dressed at the school, she checked into the curriculum as well as how some classes were structured, and she looked into the ways parents were involved in the school community. Above all, the mothers wanted their children to be academically successfirl at their new school. But their assessment of whether or not their child could be academically successful did not rely on standardized test scores. In their assessment of school quality, they considered academic (teacher quality, curriculum), holistic (small school size, parental involvement) and social (children’s behavior, after-school activities) characteristics, but not the official criteria -- state standardized test scores.” Mrs. Moore and Mrs. Robinson selected South Mill and Lincoln despite a set of alternatives their family and fiiends thought were better options. Mrs. Moore considered three schools in total: South Mill, Truman, and Carnegie. The percentage of 11th graders in each school that was proficient on the state tests were 26, 17, and 13, respectively. Even through Mrs. Moore’s husband and child preferred Truman and Carnegie, she felt 2’ I did not tell parents the state test scores of the schools in their choice sets and not one parent asked me. I decided not to tell them because of the implicit judgment they might feel if they were to choose a school with lower test scores. 124 that the small class sizes and caring teachers at South Mill were in Clifton’s best interest (which also has the highest scores of the three schools but is on the state’s failing list). She trusted her own judgment, which stood in opposition to her family’s and the state’s assessments. Sinrilarly, Mrs. Robinson chose Lincoln despite the state’s assessment and fact that most of Raven friends’ parents felt Bedrock was the best school around. Mrs. Robinson also considered four other schools that had 35, 39, 19, and 22 percent of their 7”I graders proficient on the state exams. Lincoln had 30% of its 7th graders at the same level. Mrs. Robinson chose Lincoln -- not for its test scores, which were not the highest or the lowest of her choice set schools -- but for its characteristics as a high quality school. Lincoln offered the curriculum, small school size, and sense of community she preferred. Both mothers chose schools that matched their (but not the state’s) definition of a high quality school. They reported that they felt they had a choice about where to send their children to school and were highly satisfied with their decisions. Mrs. Moore and Mrs. Robinson had different ways of thinking about the state’s official assessments of schools. Mrs. Moore feels test scores do not tell the whole story about a school. With respect to Clifton’s middle school, the scores did not reflect her experience with the teachers. Mrs. Moore dismisses test scores as a way of evaluating schools. She trusts her own judgment of school quality and if Clifton’s early success and happiness at South Mill is any indicator, that trust is well placed. Mrs. Robinson deals with test scores differently. She does not dismiss them; in fact, she feels they are valuable. But rather than using them to judge a school, she uses them to understand how Raven is doing. She explained, “They [test scores] help you know what a child’s weaknesses are.” For Mrs. Robinson, a school’s scores are less important than a child’s 125 scores. “Too much emphasis is placed on those tests for schools. What matters to me is [that] Raven is achieving.” Though Mrs. Robinson does not reject the validity of test scores as a measure of school quality out-of-hand, she does not use them to help her decide where to send Raven to school, for she does not think the school is the important unit of analysis when considering those scores. Both mothers are familiar with test scores, but they do not rely on them to judge schools. Rather they use their own, rather than the state’s, definition of school quality to select schools for their children. Across the Sample. Though all but a few of the 48 parents feel test scores are important indicators of school quality, not one parent chose a school based on test scores alone. In fact, over nine months and an average of two hours and 25 minutes of interviews with each of 48 parents, only five parents ever mentioned test scores. And none of the five rejected or selected a school because of test scores. They considered test scores to be a piece of information that describes a school, but not t_h_e_ piece of information. Like Mrs. Moore and Mrs. Robinson, the rest of the 48 parents had their own definitions of school quality and they used those to choose schools. This pattern is consistent with and supported by an analysis of parents’ choice explanations. Parents gave many reasons for checking into a school or deciding a school was not a reasonable option. Statements of parents’ reasons were tabulated for all of the schools they mentioned throughout the interviews. Together, parents offered 415 reasons for their actions with respect to particular schools. However, only 6 of the 415 reasons (1.4 %) were related to test scores (e. g., “It had much better scores than the other,” “I was worried about the [state test] scores”). In other words, roughly 98% of parents’ reasons were unrelated to the primary market signal intended to measure school quality 126 (test scores). It is clear that these parents decided what school quality looked like for themselves; they did not rely on the official designations from the state. This finding is not to be misinterpreted with the notion that parents dismiss test scores and do not feel they are important. Many parents (64%) said test scores tell you something about a school, such as how the teachers teach the curricultun, how well the children are prepared for the test, and how well the community and school are working together to help students achieve. Nor is it to be confused with the idea that parents’ notions of school quality are largely disconnected from the rest of the community’s notions of school quality. Every parent had an opinion about which schools in their community were good and which ones were bad. When asked to rank types of schools in the abstract, parents with different amounts of income and formal education, ranked schools in surprisingly similar ways. Ranking the six types of schools in the study by the quality of education they provided, parents ranked private and religious schools first and second followed in order by magnet, neighborhood, charter, and homeschools. Parents who selected failing schools were as likely as those who selected non-failing schools to rank private or religious schools as the best or second best educational opportunity, 38(1) = .037, p = .847, 0 = .24. In an abstract sense, there appears to be consensus around quality among parents who chose both failing and non-failing schools.22 So how then does it make sense for parents to have an abstract agreement on school quality and believe test scores are important, and at the same time choose schools that are low on their own rankings and do not do well on standardized tests? There are at 22 This finding is consistent with F iske and Ladd’s (2000) findings in New Zealand and Woods et al.’s (1998) findings in England. 127 least two kinds of answers: The first concerns parents’ ability to mediate the official and their own assessment of school quality; the second concerns parents’ differential assessment of quality in schools close to their experience versus schools further away. If a parent is faced with media accounts of all the things that are wrong with their community’s schools and yet their child is coming home from school happy and apparently having learned something, the parent must mediate the conflicting evidence. Mrs. Moore and Mrs. Robinson rely on their own experiences in their children’s schools to help them sort things out. This was common among other parents. Some parents, like Mrs. Moore, got involved in their child’s education and came to disregard the state’s assessment of their child’s school. Those parents had many first-hand experiences which told them their assessment, not the state’s, was correct. Other parents, like Mrs. Robinson, paid attention to tests but shifted the unit of analysis, focusing on how her child was doing on the standardized tests rather than how the school was doing. These parents reasoned that, as long as their child was performing well and coming home happy, the school (and the parent’s assessment of the school) must be doing its job. These two methods of mediating one’s own assessment of school quality and the official assessment were common among parents in failing schools. Across the sample, Gaston parents relied on their own assessment of school quality and were not, for the most part, motivated by the official account.23 A second explanation to the apparent conflict between abstract notions of school quality and more tangible indicators of school quality has to do with the halo around local 23 Parents’ hesitancy to attribute poor academic achievement to the school is consistent with the Gallup poll results mentioned in the first proposition. 128 schools. For many years, the Annual Phi Delta Kappa/Gallup poll on school quality has turned up the same result. Every year since the poll has been asking citizens to grade schools, “the public gives the schools high marks, and the grades improve the closer people are to the schools.” (Rose & Gallup, 2003, p.44). This year, sixty eight percent of parents gave the public school their oldest child attends an A or a B while 48% gave their community schools the same grade. Twenty-six percent of parents gave the nation’s schools an A or B. It appears that giving the nation’s schools lower marks does not preclude a citizen fi'om giving their oldest child’s school higher marks. In a parallel argument, assessing private, religious and magnet schools as offering the highest quality education does not preclude a neighborhood or charter school from also being a high quality school. Though test scores are a seemingly objective set of criteria that could potentially span the distance from the abstract (the nation’s schools or private schools) to the concrete (my child’s school or the neighborhood school), parents choose not to rely on them. Instead they place their trust in their own assessments of school quality; they know what’s best. Trusting one’s judgment has implications for one’s preferences. Across the 48 parents, we see that trusting one’s judgment means that the preference for a high quality school includes, but is not limited to, academic concerns. Contrary to normative assrunptions, the bulk of those academic concerns do not rely on standardized test scores. Gaston parents believe test scores are important, but they judge academic quality on broader terms that include criteria such as class size, curriculum, and teachers. To the extent parents pay attention to test scores, they mediate any disconnects between their 129 assessment of school quality and the state’s assessment in ways which foreground their judgment over the state’s. Having considered the ways in which parents’ preferences around school quality are shaped and bound by culturally held beliefs, we now turn to a consideration of the relationship between parents’ notions of community and their preferences around geography. Proposition Four: This Is My Community Gaston, like many cities across the country, has a long history of racial and social class stratification and discrimination (O'Connor, Tilly, & Bobo, 2001). As Massey and Denton (1993) document, the flight of middle class people and their jobs to the suburbs have left many American cities “hyper segregated.” From a schooling perspective, this white flight has concentrated African-American, Hispanic, and poor children in city schools, schools that are financially disadvantaged compared to their suburban counterparts (Kozol, 1991). Schools are more segregated now than during the years preceding Brown v. Board of Education (Orfield & Yunn, 1999). It is easy to understand, then, that the relationship between the city and the suburbs is marked, on both sides, by a sense of “us” and “them.” The suburban-urban relationship in Gaston -- skeptical at best, hostile at worst -- is fireled at least in part by the television, newspapers, and radio. Many city residents and activists believe the media actively seeks to report all the bad news in Gaston or at least tells the most distressing version of any story having to do with the city. The “spin” as one mother said, is “always for the worst.” Many suburbanites are unwilling to stop for gas in Gaston and yet, thousands of people continue to choose to live in Gaston. Some Gastonians have no choice; they 130 simply cannot afford to move to the suburbs. But some enjoy living in the city. They grew up in Gaston. Their family and friends still live in the city. They participate in the cultural life the city provides: A lively music scene, a museum with one of the largest collections in the country, a waterfront and city market. They are comfortable and happy. Gaston is home. So what is a self-respecting, intelligent person to do when the media is constantly reminding you that others view your home as inferior, dysfturctional, and corrupt? How does someone make sense of driving to the suburbs and watching the roads change from pothole covered hazard zones to smooth asphalt autobahns in the space of a few blocks? How does someone explain to themselves why the bus system between the city and the suburbs frustrates any attempts to move between the two in a reasonable amount of time? One answer is in the fourth proposition. Gaston parents are often aware that they live in a place others would chose not to. To paraphrase Hirschman (1970), in choosing to stay, they choose voice rather than exit. They choose to move their children out of bad schools and into better schools. But they do not leave the city. Their loyalty may not be to the school district, but it is to the city. These parents do not want their children to grow up in a different community with different kids; they just want to control the enviromnent in which the same kids grow up together. Through Mrs. Wise and Mrs. Brown we will see how parents’ desire to be a part of the Gaston community shapes their preferences for schools. We will also see the ways in which those preferences shape their choice sets and ultimate school selections. Mrs. Wise. Mrs. Wise is a single working mother with two children, Seymour, a 10111 grader at Pinecrest High School, and Sarah, a 5th grader at Park Slope, a nearby 13] magnet school. Mrs. Wise grew up in Gaston and went to Gaston Public Schools through high school graduation. A few years ago she moved from her old home to the one she currently owns. She explained, I love it. I love it here. I’ve lived in this area since 1977 and I had no clue that this little community even existed. Until my real estate agent told me about it. Of course, it was the last house and I was like “Hhh, that’s just too small.” And it just surprised me. We came over here and I was like “Hum.” I told my mom, “I have a garage, there is a porch”. . .And then, I like the area, you know, it’s on a dead end. The golf course is right there. The kids are within walking distance of my parents. The kids didn’t have to change schools. I was closer to work. I’m 12 minutes away from work. It was just perfect. The Lord blessed me. (2“d interview) Mrs. Wise loves her home but she also loves what her home provides: a safe, dead end street, in a pleasing setting that is within walking distance of her children’s grandparents. When asked if she ever considered moving to the suburbs, she said immediately, “No. No, I’d never move there.” Mrs. Wise has grown up, gone to school, and lives in Gaston. It is her community, her home. Mrs. Wise acknowledges however, that it is difficult to find the kind of school she is looking for. It’s just so hard, trying to find caring teachers who really care about their students and who care about their learning. You know, what they should be learning. And I just wish too that the students here in the city had the opportunity to learn as much as students in the suburbs. You know. (1St interview) 132 Over the course of our three interviews, Mrs. Wise explained that the suburban education is better than a city education but, “I do feel that there are some good schools in Gaston. There just aren’t enough of them on the south side. So many of those acaderrries are on the north side. But I’m not going to give up our Gaston Public schools yet. I’m not.” Mrs. Wise prefers to send Sarah, a good student, to a middle school in Gaston. She does not want to send her to a suburban school. Mrs. Wise searched a great deal in order to find Sarah a school she felt good about. There were seven schools in Mrs. Wise’s choice set. Three schools were Gaston magnet schools, two were charter schools, and two were neighborhood schools. Of the seven schools, Mrs. Wise preferred the three magnet schools to the others. Sarah eventually ended up attending what Mrs. Wise considered to be “the best middle school in this area,” Bedrock Magnet Middle School. Sarah has adjusted to Bedrock and is “challenged” at school. She continues to be picked up from school by her grandfather with whom she does her homework and spends time until Mrs. Wise gets home from work. Mrs. Wise has met Sarah’s teachers and feels that she has found a school that pays attention to her daughter and wants what is best for her. “They care at Bedrock. And they make sure the kids are leaming what they should be. I’m very satisfied.” Mrs. Brown. Like Mrs. Wise, Mrs. Brown was born, raised, and schooled in Gaston. Though they grew up on opposite sides of the city, they share a sense of place. Mrs. Brown went away to college where she became an engineer. As soon as she was able, she moved back to Gaston. She has one child, Sebastian, a fifth grader at Massey Christian School. When discussing the purchase of her current home, she explained that her husband works in a town about 80 miles west of Gaston: 133 Oh yeah, he has a haul and a half. When we were looking at homes, we were looking in Roanoke, Durham, and Franklin. But I don’t want to live out there. All my family’s over here on the east side of Gaston and so I have a very, very good husband so he just drives. (1St interview) Mr. Brown is committed to his family so he drives over an hour to work, making that long drive for both social and family reasons. Mrs. Brown is committed to her family and comfortable in her community. Like many African-American parents in Gaston, they know that the suburbs are not always welcoming places for them and their children. Mrs. Brown worried about moving to a district with good public schools because most are predominantly white: I’ll tell you, when I was looking into moving. . .it would have been into Bayfront [an adjacent district]. Okay? And I was very concerned. I didn’t really want to do it because I don’t need Sebastian to be the first. He doesn’t need to be the pioneer. He doesn’t need to be the first little boy to be, you know, some little kid to call him a nigger. And unfortunately, I don’t know. . .I don’t care how many people want to believe that it is not true. That doesn’t happen. It happens. It happens. (2“’1 interview) In the end, the Browns were able to find a home they love in a neighborhood just minutes from Mrs. Brown’s mother and extended family. They chose not to move to Bayfront, though they could afford it. But their choice of residence had implications for Sebastian’s schooling. Mrs. Brown “talks to everyone about schools” and spends time each year looking on the intemet for the list of things a child his age should know. She enrolls Sebastian in 134 football, golf, baseball, soccer, basketball, Cub Scouts, and a university run science and mathematics enrichment program. In her words, she is “crazy” about his education. She does not, however, feel that she has a lot of choices about where to send Sebastian to school. She feels that she cannot send Sebastian up the street to the neighborhood middle school, both because of the low test scores and because Sebastian would struggle socially. He’s not a typical urban kid because we’ve sheltered him. I hate to say this too. . .um. ..Sebastian’s never had a fight. He wouldn’t know what to do. You know, the first year [at a neighborhood school] he’d probably get beat up because he’ll be so astounded that someone hit someone and. . .and meant it in a harmful way. (1St interview) Mrs. Brown focused her search efforts on private, religious, and magnet schools. She applied to two schools: Massey Christian” and Ashford, a very prestigious Gaston magnet school. Mrs. Brown did mention two other schools, but they were dismissed because one became logistically impossible (her husband’s job changed so he no longer drove that direction to work) and the other was too expensive. Over six months, Mrs. Brown got the run-around from Ashford. The principal would not return her calls and she could not get a straight answer about when the entrance exam would be offered. In J une, she exclaimed, “I am a parent who cares. A parent who plans. So I said, ‘forget it.’ If I have such problems now and no way to resolve them, what if he’s there and there are problems? What will I do? I’m gonna lose 2" The middle school associated with Massey is in a different building than the elementary school Sebastian currently attends. 135 my mind.” Mrs. Brown ultimately decided to continue Sebastian’s education at Massey Christian and has been very pleased. He has adjusted to changing classes and having six different teachers. He is still getting all As and Bs. Mrs. Brown plans for him to stay at Massey through the eighth grade. Mrs. Wise and Mrs. Brown. Mrs. Wise and Mrs. Brown chose schools that are failing. Massey Christian is unaccredited and Bedrock Magnet Middle has 39% of its students reading and doing math at the proficient level on the state’s standardized exam. Like the two mothers in the previous proposition, Mrs. Wise and Mrs. Brown are satisfied with their choices and feel the schools they have chosen are good matches for their children. The mothers do not see Massey and Bedrock as failing. They are schools that, given their constraints, Mrs. Wise and Mrs. Brown prefer. If we can set aside the fact that both mothers chose failing schools, we can see an important relationship between the mothers’ preferences and their choice sets. Mrs. Wise and Mrs. Brown’s preferences are influenced by their connection to Gaston, their commturity. Those preferences in turn, influence the schools that are a part of their choice sets. And those choice sets include the final school the mother selects. A brief look at Mrs. Wise’s thinking might illuminate the relationship. Mrs. Wise wants to continue living in Gaston. Though she had the opportunity to move to the suburbs, she chose not to. Despite her belief that suburban schools may be better, Mrs. Wise does not want to move. Gaston is her home. Her family lives there. Her friends live there. She is comfortable and happy. Her children encounter children like themselves there. Her desire to remain in her community, however, influences her schooling preferences. The seven schools in her choice set are all schools in Gaston that 136 are tuition-free. She does not consider any tuition-based schools or schools across district lines. From that set of Gaston-based schools, Mrs. Wise then chose the magnet school that she believes has the best reputation and offers her daughter the most educational opportunities. In Mrs. Wise’s and Mrs. Brown’s cases, their membership in and commitment to their communities shape their preferences. They both prefer selective schools that are in Gaston. They do not look across city lines. They construct choice sets that reflect their preferences and ultimately choose schools that are consistent with those preferences. Policy makers may view their choices as poor or ill-informed, but they are not. The three schools Mrs. Wise applied to are all failing. Stevenson Magnet Middle School has 30% of its students performing at the proficient level. Forward Academy has 29% of its students at the same level. Bedrock, the school Mrs. Wise chose, has 39% of its students performing at the proficient level. Mrs. Wise did not use test scores to chose Bedrock and she does not believe test scores tell you very much about a child’s progress. But she did choose the school that happened to have the highest test scores. If we equate test scores with quality, as so many accountability schemes do, Mrs. Wise chose the highest quality school. But very few state departments of education would consider a school in which less than 40% of children are proficient, a high quality school. Mrs. Wise chose the best school of the ones she considered. But all the schools she considered are failing. Across the Sample. If the nature of one’s choice set is influenced by where one lives, we should see differences across suburban and city parents’ choice sets and choice schools. And we do. The choice sets of suburban parents contained more than double the proportion of non-failing schools as compared with city parents’ choice sets. F urther, 137 suburban and city parents Opted out of their own communities at similar rates but with quite different results. City parents considered a slightly higher proportion of external schools (schools outside their own community) but selected a lower proportion of non- failing schools as compared with suburban parents. Across these patterns, there is an important finding: depending on the communities in which parents’ reside, parents’ commitments to educating their children in their own community result in unequal rates of choosing non-failing schools. Table 4.1 summarizes the patterns which lead me to this conclusion. Table 4.1 City and Suburban Choice Sets and Choice Schools City parents Suburban parents Choice set % non-failing 43 91 % community schools only 58 67 % community and external schools 42 0 % external schools only 0 33 Choice school % non-failing 33 88 % of external schools non-failing 25 75 % external 25 33 Depending upon the community in which parents’ reside, their commitment to educating their children in their own community produces differential outcomes. This happens as a 138 result of both the choice sets parents construct as well as the final decisions they make. Because many parents’ choice sets contained private schools, it is not possible to compare parents’ choice sets in terms of standardized test scores. We can however, compare the prevalence of non-failing schools in the choice sets as a way of determining the degree to which parents in different communities had access to non-failing schools. The comparison is as you might expect. Despite similar rates of considering external schools (42% v. 33%), suburban parents’ choice sets contained a much greater percentage of non-failing schools than Gaston parents’ choice sets. The mean suburban choice set contained 91% non-failing schools in comparison to the mean Gaston choice set which contained 43% non-failing schools. The modal response is even more disturbing. Save one mother, all the suburban parents’ choice sets had 100% non-failing schools. It did not matter whether or not suburban parents stayed in their own communities because almost all of their options were non-failing ones. In contrast, more than half of the Gaston parents’ choice sets contained less than 30% non-failing schools. For a parent whose choice set contained three schools, this means that just one of those schools was non-failing, the other two were failing. Of the 36 Gaston parents, twenty-five (or 67%) had choice sets that contained zero or one non-failing school. These choice sets produced unsurprising results for children, tlrirty-three percent of Gaston children attended non-failing schools in the fall, while 88% of suburban children attended non-failing schools. Parents whose choice sets contained higher percentages of non-failing schools were more likely to choose non-failing schools. But we cannot attribute this result to Gaston parents’ lack of trying. Gaston parents opted out of their community at almost double the rates than the national or state averages. 139 Twenty-five percent of Gaston parents selected schools outside of Gaston. The state estimates that less than three percent of children are involved in inter-district choice and less than ten percent attended private school. Nationally, fourteen percent of parents send their children to public schools of choice (this includes magnet schools and, therefore, overestimates the number of parents sending their children to schools outside their own communities) and nine percent of children attend private or religious school (N CES, 2003). Gaston parents are not more hesitant than most to leave their community in order to find a school that better matches there preferences. They consider schools outside the community but those schools are not overwhelmingly non-failing schools. Looking across this data, it is clear that these two groups of parents, who looked outside of their communities at similar rates, sent their children to schools of different status. Perhaps again, this is a restatement of the inequalities that exist across urban suburban lines. Depending on where you live, it may not be advantageous to want your children to attend city schools. If you live in Gaston, having a commitment to the city decreases the odds of your child attending a non-failing school. It does not make it impossible (33% of city parents chose non-failing schools), just less likely. Suburban parents do not face the same circumstances. Irrespective of suburban parents’ commitment to their communities, their children are still likely to end up in non-failing schools. As the choice sets demonstrate, there are simply a greater percentage of non- failing schools to choose from; they ahnost can’t go wrong. Propositions Overlap I have portrayed each of the propositions as separate and distinct from one another. This has been done in order to simplify and emphasize each. This 140 simplification does not, however, do justice to the complexity of parents’ preferences. Multiple propositions influence any single parent. In order to demonstrate the overlapping influences of multiple propositions, we will briefly reconsider the case of Mrs. Hawill. As demonstrated in the first proposition, Mrs. Hawill believes that children influence their own academic achievement. Alecia is a very academically-driven young lady and Mrs. Hawill prefers a school which can support and challenge Alecia. Mrs. Hawill’s understanding of children as participants in their own success influences her preferences, but the third proposition, the commonly accepted notion that as a parent, you have to decide for yourself what quality it, also influences her preferences. The following description will illuminate the ways in which this proposition influences Mrs. Hawill’s preferences for Alecia’s high school. Alecia did not score high enough on Gaston’s magnet school entrance exam to be accepted to her first choice schools -- Washington and Jefferson. In speaking with other students and their parents, Mrs. Hawill learned that Alecia scored higher on the exam than some of her friends who were admitted to the two schools. Upset, she filed an appeal which was denied. Alecia was very disappointed for “about a week,” but eventually began to accept the rejection. Mrs. Hawill said that Alecia believes that Washington and Jefferson are the ones who lost out. Alecia was at the point where she “wasn’t going to let it bother her.” But Mrs. Hawill wanted answers. Either there was a mistake in the central office or Alecia was not being given fair treatment for reasons Mrs. Hawill couldn’t imagine. She wanted to know what happened. 141 Over the last half of the summer and the beginning of the fall, Mrs. Hawill continued to try and get an answer for Alecia’s rejection. She refused to accept the explanations Gaston officials gave her. In June, she said, “I just want an answer. I want to know why. And I want an answer that makes sense to me.” It took many months, but Mrs. Hawill eventually met with the superintendent of the Gaston schools and he explained that there had been a mistake. Alecia was free to chose between Washington and Jefferson and transfer mid-year. Though Mrs. Hawill appreciated the superintendent’s admission of an error, she was not willing to move Alecia two months into the school year. Together, Mrs. Hawill and Alecia had selected her current high school and they felt it was best for Alecia to finish the school year there. In the spring, once Alecia has adjusted to her current high school, they would decide if she wanted to switch to Washington or Jefferson. Throughout the appeals process, Mrs. Hawill decided for herself. She did not agree with the district’s rejection of Alecia and she sought an explanation. Even when Alecia was given the green light to attend either Washington or Jefferson, Mrs. Hawill thought carefully about the pros and cons associated with changing schools mid-year. Because of the prestige associated with Washington and Jefferson, other Gaston parents suggested Alecia immediately leave her current schools. But again, Mrs. Hawill decided for herself. She had carefirlly chosen Alecia’s school and was not going to change course simply because others thought she should. Mrs. Hawill’s preferences are influenced by more than one proposition. She is not alone. Mrs. Carol views Denzel as a typical 9111 grader (Proposition 2) in addition to attributing his poor academic achievement to his lack of motivation and effort 142 (Proposition 1). Mrs. Robinson, who has her own definition of high quality (Proposition 3), is not worried about sending Raven to a particular school because she believes Raven will be success at almost any school due to her work ethic and intelligence (Proposition 1). Mrs. Robinson also did not seriously consider any schools outside of Gaston. She, like the other mothers who want to send their children to school in Gaston, is committed to the city. She works for the district, takes her children to the public library, and believes that there are good schools out there for Raven (Proposition 4). All of the parents’ preferences develop within a cultural context. Those contexts do not neatly or uniformly shape preferences. Even if they did, interviews with 48 purposefully selected parents cannot be used to establish the prevalence of a given proposition. Rather, I have used the interviews to locate and describe four of the cultural propositions that are relevant to the shaping of parents’ preferences. In doing so, I have nominated four areas of parents’ preferences -- academic achievement, child development, school quality, and geography -- that are crucial to understanding parents’ actions. I have argued that parents’ preferences, choice sets, and selections are shaped by propositions, which are based upon culturally specific common sense or taken-for- granted assumptions. 1 have articulated those propositions and the ways in which they influenced Gaston parents’ preferences. But as it stands now, parents’ preferences appear diverse and idiosyncratic. There are, however, three unifying elements of preferences. It is to those elements we now turn. 143 The Elements of Preferences Three elements of preferences cut across the apparent idiosyncrasies of individual parents and allow us to focus in on the specific dimensions of parental preferences that might have implications for choice policy: (a) parents’ preferences are co-produced; (b) parents’ preferences are relative; and (c) parents’ preferences are social and historical. I will briefly explain each element, draw connections between the element and the eight parents in this chapter, and then suggest some policy implications for each element. Preferences are Co-produced As any teacher, principal, or parent will tell you, education is not only about teaching. It is also about learning. Scholars have commented at length on the co- production of education (Cohen, 1988; Jackson, 1968; Schneider, et. al, 2000). C0- production implies that the production of education requires students, parents, and school officials to be working toward the same goal. To simplify, ambitious learning requires ambitious teachers and ambitious teaching requires ambitious learners. One does not exist without the other. There are, of course, exceptions. Bright students can get an education despite terrible teachers and great teachers can motivate academically unaccomplished, apathetic students to scholarly excellence. But these are not the norm. Likewise, some schools are successful without the support of parents, but the task of educating children is greatly facilitated by parents and schools working together in respectful, supportive ways. In the same way that education requires the compliance and even enthusiasm of multiple actors, parents’ preferences also require the compliance of other actors in the choice system. In other words, preferences reflect the interactions parents have with 144 schools and the choice system more generally. For example, Mrs. Hawill preferred Washington and Jefferson over the school Alecia attended, Xavier. But when Alecia was not accepted to Washington and Jefferson, Mrs. Hawill was forced to investigate other schools. Mrs. Hawill and Alecia came to prefer Xavier to three of the other schools in her choice set. The preference for Xavier was not produced in isolation. Rather it was co-produced within the choice system that rejected Alecia’s application to Washington and Jefferson. Mr. Dish preferred a school that would continue the advanced science work that Tasha had been doing in her elementary school. This preference was co-produced by the educational system available to him when he enrolled Tasha in first grade. He had heard through other parents that Tasha’s elementary school was a good school. Though he approved of the science focus of the school, he did not choose the elementary school for that focus. Five years and three children later, Mr. Dish was adamant that he find a school that would support Tasha’s science studies. Had the district not had a magnet elementary school that focused on science, Mr. Dish might never have preferred a middle school that had a science focus. His preference for a science magnet school was co- produced by the educational system in which he was embedded. Both examples illustrate the role that opportunity plays in the co-production of preferences. There is a more subtle dimension of co-production that focuses on the interaction inherent in co-production. As before, we will examine it with a case. Mrs. Carol attributes her son’s repetition of the 8th grade to his lack of effort. She does not prefer a private or religious school for Denzel because she believes it would simply be a waste of money. He would just, in her words, “flunk out more quickly.” Her preference 145 for a non-tuition based school is, in part, based on her assessment of the cause of Denzel’s achievement. Imagine however, if Denzel attended a school that used different teaching methods and engaged him, thus making his academic performance better. Mrs. Carol’s preferences might be different. She might prefer a religious or private school. These interactions, which occur between Denzel and his middle school, influence Mrs. Carol’s preferences. The notion that preferences are co-produced complicates the portrayal of preferences in the choice literature. It suggests that we must consider the interactional nature of preferences. We can look at parents’ school selections and make inferences about them, but we must be careful because those selections embody parents’ preferences as they have been co-constructed with the current choice environment. Choice environments vary and it is likely that some parents’ preferences will be co-constructed in ways that are not the same as other parents with different demographic characteristics (i.e., social class and race). Overall, the co-construction of preferences argues for a more nuanced and multi-dimensional account of parents and the choice contexts in which they choose. Preferences are Relative As demonstrated by the parents in this study, school characteristics such as “good teachers” or “good test scores” do not mean the same thing across parents. Parents all want good teachers and good test scores but what counts as good is variable. What satisfies one parent may not satisfy another and vice versa. Rubinowitz and Rosenbaum (2000) said, “Satisfaction is almost always relative. While surveys usually ask questions about satisfaction as if it were an absolute, people can only answer relative to what they 146 know or expect” (p.129). In an urban district, parents may feel certified teachers who are teaching in their subject area are good teachers. In the suburbs, teachers who have master’s degrees in their teaching area may be considered good. Good is not the same from context to context. Further, parents’ definitions of good do not always match up with policy makers’ definitions of good. Seven of the eight parents described in this chapter chose schools they felt were good for their children, schools with which they felt satisfied and comfortable. The state considers those schools failing. Clearly, the parents in this study and the state have different expectations that are not consonant with one another. From a policy perspective, when parents report high quality teaching as the most important school characteristic, we must be careful not to assume either that parents mean the same thing as one another or that their definition of high quality teaching matches that of researchers and policy makers. The relative nature of preferences presents a particularly difficult measurement problem for policy makers because there is not, for better or worse, a single, agreed upon scale of quality. While this may frustrate policy makers who are working vigorously to improve all schools, it is the essence of a market for any good. Variable preferences create niches to which suppliers then cater their products. The danger here is that variable preferences turn into unequal outcomes through the power of the market. In support of others’ warnings, the nature of preferences opens the possibility that choice may exacerbate an already stratified educational system (e.g., Gauri, 1998; Schneider et al., 2000). 147 Preferences are Social and Historical Parents are social and historical actors. Their preferences are also social and historical. We will first consider the social dimensions of preference and then the historical dimensions. Looking back across the eight cases, we see that parents have preferences that are social, informed by social norms, and information gathered through social connections that informs and shapes those preferences. For example, Mrs. Moore prefers to find a school in which Clifton, who is a relatively shy 14-year-old, can be socially comfortable. This is a social preference. She does not see him as the kind of child who can do well in a large comprehensive high school like his sister. This assessment is based on her understanding of the kind of child who can do well in that type of school as well as what a comprehensive high school feels like for the children going to school there. And finally, Mrs. Moore discovers the school she ultimately selects for Clifton from a girlfriend, a social connection. This same analysis can be done for the historical nature of parents’ preferences. For clarity, I will stay with Mrs. Moore, though a similar analysis can be done for any of the other seven parents. Mrs. Moore has recently witnessed a dramatic change in Clifton’s school engagement and performance. She credits this to the Mentoring with Pride program. She prefers a school with small class sizes that will provide the kinds of individualized attention likely to continue Clifton’s recent change. Mrs. Moore’s preference for small classes is historical, meted in her prior understanding of Clifton’s dramatic change. Mrs. Clifton’s preference to find a charter school for Clifton is based on her historical knowledge of the Gaston Public Schools. She knows that the high schools in the district are large comprehensive high schools. Based on her experience 148 with her older daughter who was accepted to a number of magnet schools, she also knows that Clifton does not have the grades or disposition to “survive in one of those schools.” Hence her preference to look for a charter school that will help Clifton succeed. Mrs. Moore’s preferences have historical dimensions that help explain why she prefers what she prefers. If we step back for a moment from the cases and consider a more general example, the social and historical nature of preferences becomes clearer. Take the example of a parent having graduated from college. Parents who are first generation college graduates and have benefited financially from that experience, are likely to have a strong belief in the power of education. Having benefited from their college attendance, they are also likely to have learned particular lessons they will pass on to their children, such as the need to take Advanced Placement classes in high school or prepare for the SAT through preparation courses. These experiences are likely to result in preferences for college preparatory high schools, extra-curricular activities, schools with high graduation rates and good counseling departments. Though these are hypothetical preferences, it is easy to imagine how they might come to exist for a first generation college graduate. We could generate these hypotheticals for many different historical circumstances, but I think the point is clear. History matters. I end with the words of Mrs. Springfield, a mother of three girls who attend a religious school in Gaston. Mrs. Springfield explained that she had experiences that really changed her ideas about what she wanted for her daughters. Her family doesn’t understand her ideas. Mrs. Springfield is the only person in her family to attend college and she explained how different histories give her and her family different preferences 149 for schools. Her words explain better than I can the role history plays in shaping parents’ preferences: Everyone thinks I’m crazy about school. They don’t understand why I’m so stressed out about their educations. But I’ve had experiences that make me this way. For them [her family] the mentality is “The school’s right down the street. Why not go here?” And if you have not been exposed to anything different, if all you know is a classroom with 30 students then why would you even think about what it would mean to have your child in a classroom with 15 to 16 kids? If your whole exposure is where the entire student population is of the same race, why would you consider, why would race or diversity be an issue for you? (2“’1 interview) Conclusions Mrs. Springfield’s words capture the essential argument of this chapter: all parents’ preferences are shaped and bound by culture. Those preferences influence all aspects of the choice process, most notably, the construction of the choice set. So the assertion that parents will simply match schools with their preferences is somewhat misleading. Parents will (and do) match schools with their preferences, but those preferences are themselves already restricted, bounded, shaped. The schools parents perceive to be “real options” are not determined by the actual set of schools that exist, the set that we might see if we drove through the neighborhoods of Gaston. Rather, the schools parents perceive as “real options” and the ones that they ultimately choose, are a subset of that larger set, a subset that is in large part determined by parents’ preferences, which is, in turn, shaped by their culture, history, social networks, educational beliefs. In 150 the next chapter, I propose a model of choice that takes account of the co-constructed, relativistic, and socio-historical nature of parents’ preferences. The model includes variables that measure the four dimensions of preferences this chapter nominates as crucial to parents’ choices: academic achievement, child development, school quality, and geography. The model, grounded in parents’ actions, offers us a new conceptualization of the factors influencing parents’ choices. 151 Chapter Five A Model of School Choice For economists, an explanation is, at its core, a prediction. As mentioned earlier, almost 97% of parents who are eligible to move from failing schools have elected to stay in those schools (Asimov, 2003; Moses, 2004; Schemo, 2002). These 97% of parents do not fit our current “model” of parental choice; we have not predicted their behavior. Previous chapters have offered partial explanations for the mismatch between current predictions and parents’ behavior. In Chapter 3, I suggested that the choice process, parents’ dispositions, and constraints alone, do not explain parents’ school selections. In Chapter 4, I argued that preferences, which are delimited and bound by cultural beliefs, dramatically shape parents’ choice sets and ultimate school selections. Across both chapters I described the historical and social nature of choice as well as the ways in which choice behaviors are inextricably embedded in the particular choice context in which they occurred. While illuminating and interesting, the story I have told thus far, is partial. In this chapter, I synthesize and build on the insights of previous chapters. I propose a model of parents’ choice of schools for their children, in which a non- failing school is one that is accredited and/or is not on the state’s list of schools in need of improvement.25 The proposed model aims to help researchers understand why some 25 Though I distinguish a non-failing school based on state compliance with NCLB, there are alternative ways other researchers might define non-failing. One can imagine for example, defining non-failing using a value-added measure (e.g., Sanders & Horn, 1994) or a combination of measures which account for other 152 parents choose failing schools. The model I propose grows out of my data analyses. Future studies will need to test the model. Before I describe the layout of the chapter, I owe the reader an explanation. From the beginning, this study has been interpretive. It has sought to understand the world from parents’ perspectives. One might reasonably question whether or not the proposal! of a model is consistent with such goals. For me, it is. Far too much work in the social sciences does post hoc theorizing. Data is collected using large-scale instruments and researchers busy themselves with developing models (i.e., equations) that describe the relationships between the measured variables. Through this procedure it is possible to develop models that describe reality, but it is at least as likely that one will develop a model that misses critical features of the phenomenon, thus mismodeling the object of study. Further, in areas of inquiry that are theoretically nascent (e.g., school choice) this type of backwards theorizing can be downright counter-productive, producing “results” that obscure more than they enlighten. A more appropriate tack to take is to use interpretive data to develop models (a proposed set of variables that cause an outcome) which can then be tested. This method allows the researcher to begin with a broad, nuanced understanding of the phenomenon and work her way down to a parsimonious model that essentializes the phenomenon but does not misrepresent it. Of course no modeling procedure is full-proof. Predictive and unpredictive models have been produced using both of the procedures I have just described. The devil is most certainly in the details. This chapter steps back from parents and their stories in order to inform features believed to be relevant to achievement such as teacher quality, instructional leadership, trust, or graduation rates. 153 the larger body of school choice research. This deliberate step is taken so that the wider community’s conceptions of choice might be informed by parents’ thinking specifically, and the insights of interpretive work more generally. I begin the chapter with an overview of the tacit and explicit models other researchers have employed in conceptualizing choice. I then propose a model of choice, explaining its variables, and assessing it against the sample. Finally, I consider the predictions of the model and suggest some of the model’s limitations. Previously Advanced Models For readers curious about choice, it is easy to get distracted. There are shelves full of research (and rhetoric) on almost every dimension of choice imaginable. If one believes that literature on choice outside of education can also inform this discussion, one can peruse scholarship on choice in shopping, residential selection, romantic partner, parenting practices, and career choice. In education, the research has focused on both the supply and demand sides of choice. On the supply side, scholars have studied existing schools’ responses to the introduction of charters and vouchers (Cohn, 1997; Gill et al., 2001; Moe, 1995; Witte, 2000; Woods et al., 1998). On the demand side, they have documented the effects of choice on important outcomes such as parental satisfaction, parental involvement, and economic and racial segregation (Armor & Peiser, 1997; David et al., 1994; Fuller et al., 1996; Gauri, 1998; Gill et al., 2001; Rasell & Rothstein, 1993; Schneider etal., 2000; Wilhns & Echols, 1993). Although research has documented aspects of the demand side of choice, no coherent explanation of parental choice of schools has been advanced. I focus here on the demand side of choice, describing a causal model that predicts parents’ selection of 154 schools. I will briefly review the literature on both constraints and preferences so that the model’s connections to and differences from the existing literature will be clear. My aim here is to generate a sense of the conceptual landscape for the reader. I do not attempt to mitigate the serious disagreements in the literature around issues of measurement, control groups, or effect sizes (to name a few of the more contentious disagreements). Rather, I aim to lay out the concepts that can roughly be considered constraints and preferences in others’ work. I should clarify, however, that many scholars do not use the language of constraints and preferences, so the categorization of variables as constraints or preferences is my doing, not theirs. The Empirical and Conceptual Landscape of Models for School Choice Constraints. First let us consider constraints. A constraint is a condition which limits the expression of preferences. Parents do not simply select any school they desire - - their actions are constrained by education, income, information, and, in some circumstances, their child’s academic achievement. While educational researchers do not use that language, they describe many concepts that might be considered constraints. Most often constraints are conceptualized as structural features of a given situation that limit parents’ ability to select the school they most desire. There are many categories of “constraint” in the literature: parents’ education, income, information, the child’s prior academic record, and time are some of the more common constraints. Of these, education, income, information, and the child’s prior academic record are most frequently investigated. There is strong consensus in the literature that the education and income of parents who use charters, magnet schools, and vouchers are important objects of study. 155 The logic goes something like this: The more education a parent has, the more likely she is to read newspapers and learn of choice programs, and the more likely she is to have the skills to navigate the educational bureaucracy. And as income increases, parents are more likely to have enough disposable income to actually choose a private or religious school for their child. Despite this logical portrait of constraints, the profile of parents participating in each of these forms of choice varies greatly, in part, based on the policy details of the particular programs. Vouchers and charters have, with a few notable exceptions, been used by lower-income, parents of color. However, those parents have been relatively more educated and of higher-income than their peers who elect not to participate in these programs, for “All programs of school choice are more likely to be used by better-informed families” (Gill et al., 2001, p.156). Witte (2000), Gauri (1998), Wells (1996), and David et al. (1994) found that parents who are more educated are not only more likely to participate in choice, they are more likely to be successfirl.26 Specifically, Gauri (1998) found that if a mother had gone to college, her child was 4.3 times more likely to attend a top school. In Dayton, Washington, DC, New York, San Antonio, Cleveland, and Milwaukee, mothers who participated in voucher programs had 2’ In Witte’s study, successful choosers were those whose children remained in their voucher schools. Successful choosers in Gauri’s study were parents whose children attended a school whose test scores were in the top third of the available schools. Wells’ parents were successful if their child transferred to and remained enrolled at the transfer school in the suburbs. 156 more formal education than comparison groups (Beales, 1995; Howell, 2000; Metcalf, 1999; Meyers, 2000; Peterson, 1998; Peterson, Myers, & Howell, 1999).27 There has been a great deal of discussion about the role of information in choice. Information can be almost anything, but is most often thought of as facts about access to choice programs or school quality. For example, the enrollment procedure and deadline for a charter school would be considered information, as would school level test score data published by the district or state. Across these conceptualizations, a lack of information or inaccurate information hinders parents’ abilities to make informed decisions. After all, if a lower-income parent believes that charter schools charge tuition, she is very unlikely to consider those schools for her child.28 A number of scholars have documented that a lack of information or access to poor quality information hinders the potential positive effects of choice (Holme, 2002; Schneider, et al., 2000; Schneider, Teske, Roch, & Marschall, 1997; Wells, 1993). Poor information might, for example, include inaccurate information about a charter school’s admissions procedure or misinformation about the grade point necessary to attend a magnet school. Gauri (1998) found that parents who use specific information about the quality of schools in their decision-making are more likely to choose a top school. Many 27 Following Bourdieu (2000), one could argue that parents’ socioeconomic status influences their aesthetic sense or taste, and therefore income and education should be considered preferences rather than constraints. This is a reasonable assertion. It is not however, the argument that has been made in the literature. Income and education are treated as circumstances which limit a parent’s ability to exercise her preferences. 2’ It is important to note there that parents make choices about how much information to collect. The opportunity costs (the cost of gathering information in terms of an opportunity foregone) associated with gathering information vary across families. 157 scholars have stressed that a properly functioning market must have access to accurate information (e.g., Arsen et al., 1999). Recognizing the importance of information, NCLB requires districts to alert parents to the fact that their children are attending failing schools and their options which result from their child’s attendance at that failing school (N CLB, 2001). More and better information assists parents in making knowledgeable choices. The final constraint that has frequently been investigated is students’ prior academic record (Armor, 1997; F iske & Ladd, 2000; Howell, 2000; Metcalf, 1999; Moe. 1995; Witte, 2000). Even if a parent chooses a particular school, her child’s grades may not be high enough to get in. Thus the child’s prior academic record may limit the parent’s expression of her preferences. Despite charges that they accept only the best students, voucher, charter, and magnet school advocates have been eager to point out that students who participate in these programs are not different from the population of students from which they came. In an extended literature review, Gill et. al. (2001) conclude, Most of the existing targeted voucher programs also serve relatively low- achievement students; there is no evidence that voucher schools are “creaming” high-achieving students from the public schools. The same can be said of charter schools, at least on average. In three prominent charter states where data are available, charter students have test scores below statewide averages. (p.156) Blank, Levine, and Steel (1996) reach a similar conclusion about magnet schools. “These statistics reveal that while a portion of magnet schools and programs do serve higher-achieving students, primarily the gified-and-talented programs, most magnet 158 programs serve a broad distribution of students in big city school systems” (p.170). Though the evidence varies depending on the program design, overall, children who participate in choice programs have similar records of academic achievement compared to students who do not participate in choice programs. In addition to constraints, scholars also delineate relevant parental preferences. It is to those that I now turn. Preferences. Before we consider the literature, I would like to distinguish two types of preferences, stated and revealed. Stated preferences are preferences detached from a corresponding behavior. For example, if I report on a survey that I prefer a school with higher test scores, I have a stated preference for high test scores. I have done nothing which might demonstrate that preference. Stating that I prefer schools with high scores is different than actually choosing a school with high scores. In market research, stated preferences are often elicited through an experiment in which the researcher provides the participant with a hypothetical situation and the participant indicates his or her preferences with respect to that situation. In contrast, revealed preferences require an action. If I selected a school with test scores that are higher than the other schools in my choice set, ceteris paribus, a researcher can assert that I have revealed a preference for schools with higher test scores. There may, in fact, be other reasons I selected that school, but for the purposes of understanding my actions, I have revealed my preference for high test scores. There is a body of literature which documents the challenges associated with both stated and revealed preferences (e.g., Loureiro, 2003). Though in theory, both types of preferences measure the same underlying construct -- desire - there are validity concerns 159 with each. Simply because someone reports they prefer high test scores does not mean they select schools with high test scores. On the other hand, inferring that a parent prefers high test scores based on their selection of a school with high scores may be invalid. They may have actually chosen the school for its excellent marching band or theater program. Scholars who study behavior which can be described by both stated and revealed preferences increasingly argue that a combination of both stated and revealed preferences can and should be used to assess the validity of each (e.g., Cameron, 1992; Hensher et al., 1998; Herriges, Kling, & Azevedo, 1999). As I review choice scholars’ investigations of preferences, I focus first on the preferences which might be considered revealed preferences and then on stated preferences. One way to think about the difference between stated and revealed preferences is to think of the former as requiring forward projection and the latter as requiring backward projection. A good example of the forward projection can be found in David et al. (1994). These researchers lay out a careful analysis of parents’ preference for “discipline” at a school. They infer (or project forward) that parents who articulate a preference for discipline select schools that have better discipline than the other schools under consideration. Because the study does not match parents’ words with their actions. we have no way of knowing if, in fact, the preference for discipline is actualized in a choice. You can imagine however, that this stated preference for a school with discipline could be a revealed preference if we knew parents’ action (i.e., if they choose schools with strong discipline). “Discipline” is not the only characteristic for which this is true. With some notable exceptions (such as satisfaction), preferences can be either stated or revealed depending on the data available to the researcher. Because the status of 160 preferences depends so heavily on the data collected, most studies rely on either stated or revealed preferences, but rarely both. A brief example will demonstrate the conceptual and methodological challenges of tracking both kinds of preference in schooling choices. If one were to set about studying parents’ revealed preferences, she would need a good deal of highly specific information. Let us say a parent articulates a preference for schools that “have high test scores.” In order to determine that parent’s revealed preference for test scores, the researcher would have to know the test scores of the school the parent chose. Only then could the researcher determine whether or not that score was “high” and then infer that the parent had a preference for high test scores. This would still be a weak claim. A stronger claim would require the researcher to know the test scores of the other schools the parent considered. This would allow the researcher to compare the parent’s preferences in the available options, thus permitting inferences about both the nature and strength of the preference for high scores. Unfortunately, this type of data collection requires a great deal of highly specific, individualized information. which is time consuming and costly to obtain. lrnagine the task of discerning parents’ revealed preferences for characteristics such as “atmosphere” or “discipline,” concepts which are not routinely measured and documented for individual schools (though parents report they are factors in their final decisions). Though I can only hypothesize, it seems likely that the field’s heavy reliance on stated preferences stems from the relative ease (and lower cost) of gaining access to those kinds of data. Researchers have made use of revealed preferences for characteristics that are routinely collected as part of state level data (such as race and free and reduced lunch status, and test scores) (Armor, 1997; Fiske & Ladd, 2000; Schneider et al., 2000). Those 161 studies have taught us important lessons about what parents prefer. Overall, those studies suggest that most parents prefer schools with fewer poor children and children of color.29 Fiske and Ladd (2000) suggest that this is not a consequence of outright racism or classism, but rather, parents use student demographics as a proxy for school quality. In the United States, student demographics are, unfortunately, a reasonable proxy for both school resources and test scores (The Education Trust, 2004; Kozol, 1991; O'Connor, Tilly, & Bobo, 2001). Based on parents’ choices in Massachusetts, Armor and Peiser (1997) found that parents selected, and therefore prefer, schools that have a lower percentage of drop-outs, a higher percentage of students planning on attending four-year colleges, and higher 10th grade reading and math scores. Though Ni and Donahue (2004) did not estimate preferences, they found that in Michigan, the flow of students went toward schools that serve children who are better off financially, have fewer students of color, higher standardized test scores, and higher graduation rates. For several reasons, it is difficult, however, to make broad claims about parents’ preferences based on these studies. First, all of these studies use aggregated data that are available at the school or district level. They do not investigate individual parents’ preferences within the parents’ particular choice context, thus they do not eliminate other rival explanations. It could be, for example, that if you live in the central city, simply choosing an out-of-district school results in choosing a school that is better off financially, has higher academic indicators, 29 It is unclear whether or not Schneider et al. (2000) found the same pattern. They report that parents who cared about diversity ended up in schools that were “more than 1 standard deviation above the district average” (p. 175). It is unclear, though likely, given the racial make up of Community Districts 1 and 4, that those more diverse schools also had fewer minority students. 162 and fewer minority students. Second, the school features noted are highly correlated, and it is therefore terribly difficult to disentangle which of the factors -- lower poverty, better test scores, or students of color -- is the preference, and which are simply correlates of those preferences. Stated preferences present their own set of methodological and conceptual challenges. Stated preference research investigates a wide variety of preferences, ranging from the preference for a good education to the preference for school uniforms. There has been little theorizing about why particular preferences might play a larger or smaller role in parental thinking so there are not neat categories of preferences across studies. There are patterns however. Researchers have generally investigated preferences that fall into the categories David et al. (1994) describe as “the three P’s -- the academic results or performance [italics in original]; the atmosphere/ethos or pleasant feel; and the school’s ' location or proximity to home” (p.136). Godwin, Kemerer, and Martinez (1998) add two additional dimension preferences: family values and parental involvement. Based on a priori or ex post stated preferences, these studies conclude that parents prefer schools that perform better, are welcoming and inviting to them and their children, are close to home, match their values (in the case of vouchers this often includes religious values) and have high levels of parental involvement both in school and out of school (Armor, 1997; David, 1994; Godwin et al., 1998; Lee, 1996; Wells, 1996). Parents prefer good schools that are close to their homes, in which their children can be comfortable and successful. But “good” and “close” and “comfortable” and “successfirl” are relative terms, defined both in relation to other choices and in relation to a parents’ own values and beliefs. In order to accurately predict parents’ actions, this 163 model relies almost entirely on revealed prefers. Though there are measurement challenges associated with revealed preferences, these preferences reflect parents’ actions and are therefore more reliable than their counterpart -- stated preferences. Before I turn to the presentation of the model, there is one caveat about preferences and constraints which deserves articulation. One person ’s constraint is another ’s preference. There are a few variables that are difficult to classify as constraints or preferences. Race, for example, is discussed in the literature as a characteristic which geographically locates families in areas in which there are systematic differences in satisfaction with the neighborhood schools (Charles, 2001; Massey & Denton, 1993; Stier, 2001; Wells, 1996). In part, due to the practice of red-lining, white parents tend to live in the suburbs, giving them access to what many believe are better schools. African-American and Latino parents, on the other hand, tend to live in the city, which gives them access to schools viewed as lower quality schools (Rubinowitz & Rosenstein, 2000). If we use this depiction of a racially segregated society, one views race as a constraint. If, however, we take a less structural view and consider the data about race as a social construct, we might consider it a preference. Fiske and Ladd (2000) found that parents select schools based on the ethnic mix of the children who attend those schools. Though the evidence in the school choice literature is thin, the evidence of racial preferences in other literatures such as the residential housing literature is well-established and robust (Henig, 1996; Massey & Denton, 1993; O'Connor et al., 2001; Stier, 2001). Parents feel more comfortable sending their children to schools in which their child will not be racially isolated. This is a preference. 164 Race is not the only variable which has dimensions related to both preferences and constraints. As mentioned earlier, Bourdieu’s (1984) pioneering work conceptualized social class as a collection of preferences rather than a set of constraints. For the purpose of a model, it does not matter a great deal if a variable is conceptualized as a constraint or preference, as long as it is included in the model and the relationship between it and the dependent variable is properly specified. But in the world of policy and practice, it does matter. It matters to policy whether or not income acts primarily as a constraint or a preference. If, as voucher advocates have argued, income functions primarily as a constraint, one solution to parents choosing failing schools involves giving them vouchers so that they can afford non- failing schools. If however, income frmctions primarily to shape the types of schools parents prefer for their children, a policy solution might be parental education classes or de-tracked neighborhood schools with higher standards. Unfortunately, this study only begins to sort out these distinctions. Our understanding of choice and the policy solutions appropriate for the problems of choice would be greatly enhanced by a line of research that investigates and theorizes about such matters. A Model of Choice I begin by describing the model variables. Each variable is then evaluated for its likelihood of contributing significant, independent explanatory power to the model. I consider the robustness of the theorizing implicit in the model, and finally, I take up the 165 1.30 Models, by their very nature, omit variables. For those limitations of the mode interested, Appendix F includes an analysis of the variables omitted fi'om the model. The proposed model, summarized in table 5.1, includes 10 variables, five of which are constraints, the other five of which are preferences. Taken together, the variables predict a parent’s school choice. Table 5.1 A Model of School Choice Construct Measure Constraints Income Parent’s educational attainment Potential set size Cohort consensus School visits Preferences Previous school status Choice set status Choice set size Sibling’s school selectivity Race 3° I build a model based on a purposeful sample of interview data. The sample is small. Statistical significance and measures of effect size must be taken as a piece of data but not the piece which determines inclusion in or exclusion from the model. Further, without regression analysis, one can only reason through the ways in which one variable may be confounded with another 166 Constraints Parents ’ income. Other studies have documented the relative success of parents with higher incomes (David et al., 1994; F iske & Ladd, 2000; Gauri, 1998; Schneider et al., 2000; Wells, 1996; Witte, 2000). Thus, despite the lack of a relationship between income and school status in this sample (36(1) = .67, p = .40, o = .13, OR = 1.67), parents’ income is included in the model for two reasons.3 ’ First, as evidenced by the Milwaukee choice program, even two or three thousand dollars toward an education can make the difference between choosing a tuition-based or tuition-free school (Gill et al., 2001; Witte, 2000). Second, the ability to pay for a private or religious education has been shown to produce differential outcomes in parents’ choice success. In Chile, Gauri (1998) observed that “‘successfirl’ choosers correlated with increased household income and assets” (p.109). There are many potential competing reasons why this might be the case. Parents with higher incomes might have the social capital which best matches more successful (middle-class) schools. Higher income families may have mothers who do not work outside the home and thus are able to spend more time searching for and evaluating schools. Or perhaps more wealthy parents are connected to social networks in which 3' The lack of a relationship is likely due to the purposeful nature of the sample, which is stratified by income. Further, the purposeful sample over-represents high-income parents who have selected failing religious and private schools, thus diminishing the relationship which likely exists between income and final school status. 167 more accurate school information is commonplace. Whatever the reason, parents with higher incomes are more likely to choose a non-failing school for their child. Parents ’ educational attainment. The more education a parent has, the more likely she is to select a non-failing school. She is more likely to view her own knowledge as appropriate for the task of assessing teachers, principals, and counselors at the school (Lareau, 1989). She is more likely to be aware of the controversies in schools (e.g., class size, teacher certification, funding for arts and music) and therefore able to ask questions about issues that are likely to influence her child’s education. She is more likely to have the cultural capital to advocate effectively for her child in school interactions such as parent teacher conferences, open houses, and application proceedings (Epstein, 2001; Lareau, 1989; MacLeod, 1995). A mother with more education is more likely to have the skills to navigate the system of school choice. She may know what kinds of questions to ask over the phone, she may know how and feel comfortable asking for teacher recommendations for magnet and charter schools, and she may speak the discourse schools expect parents to use -- a middle class discourse (Williams, 2004). In short, the more education a parent has, the more likely she will be to choose a non-failing school for her child. This hypothesis is supported by decades of research on the influence of parent’s education on children’s academic outcomes (e.g., Epstein, 2001; Schneider & Coleman, 1993).32 32 Again, the purposeful sampling strategy causes there is an inverse, non-statistically significant relationship between education and final school status (12(1) = 1.15, p = .28, 1b = -.l6, OR = .51). This is caused by the over-representation of more highly educated parents who have selected failing schools. 168 Potential set size. At its best, market competition results in higher quality products at a lower cost to the consumer. If we apply market thinking to choice we would expect that over time, schools would increase the quality of their product as they compete with one another to attract parents and children (Friedman & Friedman, 1980). There are many scholars trying to determine the degree to which choice actually produces this competition and improvement among schools (e.g., Woods et al., 1998). If we consider the schooling market from a parent’s perspective, the size of the market matters. More schools means more choices. Which may mean better choices. The variable “potential set size” measures the number of grade level appropriate schools within a specified range say, a two mile radius of the parent’s home.33 It is included in the model because it gives us insight into how the market functions from parents’ perspectives. Having a variable in the model that helps us understand how the market influences parents’ thinking is important because the market varies so dramatically from neighborhood to neighborhood and city to city. Over 73% of Gaston parents felt they had choices. The average potential set for those parents contained nine grade appropriate schools (six of which were free). We would expect this level of parent satisfaction and potential set size would be associated with attendance at a non-failing school. It was not. There is an inverse relationship between the number of schools within two miles of the parent’s home and the final school status (r = -.39, p < .01). In other words, as the number of potential schools goes up, the odds of attending a non-failing school go down. ’3 Two miles would be a 10-15 minute drive, depending on traffic. This is a conservative estimate of the distance parents might be willing to travel to take a child to school. 169 This makes sense if we consider two dimensions of the schooling market, quantity and quality. Charter, magnet, private, and religious schools tend to be located in urban areas (NCES, 2003). There are more schools in urban areas, which translates into more choices for parents. But urban areas tend to have a disproportionate number of schools that are not preparing their students for success (The Education Trust, 2004; Moe, 1995; Orfield, Losen, Wald, & Swanson, 2004). In Gaston (an urban area), the inverse relationship between potential set size and the odds of attending a non-failing school suggests that more does not necessarily mean better. Gaston parents feel they have choices but comparatively few of those choices are non-failing. Because school choice is in its early stages of development and it is uncertain whether or not more choices will lead to fewer failing schools, this model predicts that there will be an inverse relationship between the number of schools in an area and the odds of attending a non-failing school. Cohort consensus. Recent scholarship has focused on the role of social networks in the flow of information and resources (Bridge, 1995; Brown & Duguid, 2000; Edin & Lein, 1997). Previous chapters detail the important role other people play in shaping parents’ preferences. While there are many potential resources parents might access from their relationships with others, here I focus on the quality of information gathered through social connections.” In particular, the degree to which parents’ school-based social network gives them access to non-failing schools. 3" Of course, not all parents acquire information through such networks. However, in this study, 92% of parents reported learning about a school from friends, other parents, or family. 170 The cohort consensus measures the degree to which a particular context gives parents access to non-failing schools through social connections. An example from the study will clarify my meaning. Together, a group of five 8th grade parents nonrinated 29 potential high schools. Five of the 29 schools were mentioned by at least half of the parents. Those five schools seemed to be “in the air” for that group of parents. The schools are, to some degree, the cohort’s consensus. Four of the five were non-failing schools. Thus 80% of the cohort’s consensus was around non-failing schools. In short, this particular social network gave parents access to a good number of non-failing options. Other parents were not so lucky. Some parents’ school-based social networks gave them no access to non-failing schools. For the purposes of this model, each parent’s cohort consensus is defined as the percentage of the consensus schools that are non- failing. Because social networks are so important to parents’ choices, the strong relationship between the percentage of non-failing schools in a network and the child’s final school status is not surprising. The point-biserial correlation between these two variables equals .81 , which, according to Cohen (1988), suggests a strong relationship between network access and the final choice.35 The model predicts that the greater the cohort’s consensus, the greater the odds of attendance at a non-failing school. School visits. Though there are multiple ways of conceptualizing the role of school visits in constraining parents’ choices, this model focuses on the information ’5 All correlations have been squared and the r-square value interpreted using the following guidelines: small = .01, medium = .09, large = 25 (Cohen, 1988, p. 82-83). 171 gathering dimension of the school visit}6 The model focuses on that aspect of the visits because parents talked most often and most adamantly about that dimension. School visits is measured as the percentage of schools in the choice set the parent actually visits. Gaston parents felt strongly that school visits were an important way to learn about both the positive and negative aspects of a school. Mrs. Webb asked, “How you gonna know if you like it if you don’t see it and decide?” Mrs. Steward stressed the importance of visits. She said, “Well, you can look at them and that’s it. Look at them and say, I’m not sending my child there. Sometimes someplaces you can walk in and automatically know its worth being there.” The point-biserial correlation between school visits and final school status equals is 0.42 (p<.01), indicating a moderately strong relationship between final school status and school visits. The model predicts that the greater percentage of schools a parent visits, the higher the odds of her child attending a non-failing school.37 Together, the constraint variables account for many of the variables other scholars have investigated, albeit in a slightly different way. The constraint variables in the model take account of the money parents have available to spend on education, parents’ 36 Alternative conceptualizations of school visits include school visits as a measure of parental effort in the process of choice or a measure of parents’ comfort and skill in negotiating the schooling bureaucracy (e.g., Lareau, 1989). 37 It is likely that the percentage of schools visited by a parent measures more than just how much information a parent gathers about potential schools. It is likely confounded with (at least) the conceptualizations mentioned earlier: the level of interest in the choice decision and the parent’s comfort in navigating the bureaucracy. Without statistical controls, there is no way to sort out the influence of each. 172 potential skill and experience navigating the education system, the number of schools they have to choose from, the quality of schools available to them through school networks, and the information available to them in the form of school visits. These variables add to the literature by specifying school visits as a crucial source of information, including a variable which measures the extent of the market, and measuring the access parents’ have to non-failing schools through their school-based social networks. Obviously, there are constraints missing from the model. I revisit these at the end of the chapter when discussing the limitations of the model. Preferences The preferences included in the model are quite different from those other scholars have articulated. They are most different in that four of the five are revealed preferences, preferences that are inferred from parents’ actions. This is purposeful. As mentioned earlier, I use revealed preferences because they are more likely to lead to a predictive model that will be usefirl to policy makers. You may wonder where some of the familiar preferences -- safety, discipline, and school uniforms -- fit into the model. At the end of this chapter, I explore and explain the excluded variables. Previous school status. Except in the case of a child’s first school, when a parent chooses a school, she has done so before. A parent’s prior choices embody her weighting of constraints and preferences at a given point in time. In particular, the status of that previous school -- failing or non-failing -- implies something about the parent’s preference for a non-failing school. We can infer, though it may not be the case due to 173 other preferences and constraints, that a parent who previously enrolled her child in a non-failing school has a preference for non-failing schools.38 Previous school status is the status -- failing or non-failing -- of the child’s prior school. In the case of Gaston parents, this was the status of their child’s 5th or 8th grade school. Among Gaston parents, there is a very strong relationship between previous and final school status, 380) = 11.2, p = .001, 4) = .50, OR = 9.64. Children who previously attended non-failing schools had 9.6 greater odds of attending another non-failing school than did children who previously attended a failing school. This relationship does not describe all parents’ actions. There were some parents who moved their children out of failing schools and into non-failing schools (and vice versa), but these actions were not the norm. Parents who previously chose a failing elementary or middle school tended to chose a failing middle or high school and parents who previously chose a non-failing elementary or middle school tended to choose a non-failing middle or high school. The model predicts that a parent who preferred a non-failing school in the past has greater odds of choosing a non-failing school. This variable should not be interpreted as portraying parents as Marxist automatons. Just because parents chose a failing school in the past does not mean they are destined to choose one in the firture. As the cases in the previous chapter demonstrate, parents have agency, they interact in the social world, and they make choices. My point here does not take that agency away. Simply put, if a parent chose 3’ Parents do not prefer a failing or a non-failing school. They neither speak in those terms nor desire a failing school for their child. These are simply labels given to the schools, based on policy makers’ definitions. These labels are not part of the discourse or lived experience of the parents in this study. 174 one particular type of school (e.g., a non-failing, a private, or a selective school) in the past, they have greater odds of choosing it again in the future. Prior actions do not determine future choices, but they do provide us a sort of summary measurement of parents’ preferences. Choice set status. Parents’ choice sets are windows into their preferences. We can learn what kinds of schools they are aware of and what kinds of schools they prefer. If a parent only considers failing schools in her choice set, it is difficult, if not impossible to select a non-failing school. The constitution of the choice set is perhaps the most important dimension of parents’ thinking to understand. And it often goes overlooked. Of all the major studies on parents’ drinking, only a handful include substantive data about the other schools parents considered (e.g., Schneider et al., 2000). If we are to understand why parents are both satisfied and choosing failing schools, we must understand the schools they consider “real options” for their children. We must understand their construction of the choice set. Choice set status was defined as the proportion of choice set schools that were non-failing. As the choice sets bore little resemblance to the geographically available schools, choices sets are a measure of parents’ preferences. Some parents only considered failing schools and others considered only non-failing schools, with the average choice set containing approximately half (49.6%) non-failing schools. As you might anticipate, there was a very strong and statistically significant relationship between final school status and choice set status (rpb = 0.81, p < .0001). Because of the relatively small sample size, we cannot use statistical procedures which would allow for causal arguments. Based on the logical relationship between parents’ choice sets and their 175 ultimate choice as well as the pattern across Gaston parents, the model predicts that the higher the proportion of non-failing schools in the choice set, the greater odds a parent will choose a non-failing school. Choice set size. Our common sense tells us that there must be alternatives in order for parents to have a choice. If other schools exist but parents cannot, due to transportation or income, select them, we would hardly say the parent has choice. Taking the logic one step further we would want to include a variable that measures the number of schools a parent can choose from, their choice set size. We would assume that parents with larger choice sets would be more likely if only by chance, to chose a non-failing school. But predictions about how the number of choice set schools relates to the odds of choosing a non-failing school are not straightforward. Choice set size, a revealed preference, interacts with income.39 Parents who select particular suburbs in which to purchase homes do so because of schools (Holme, 2002). It is quite common however, in these suburbs, to have only one or two high schools. Suburban parents often do not have a choice about which high school they send their child to because children are assigned to schools based on their neighborhood location. But we would hardly say those parents do not have choice. They simply execute their choice at a different time -- when they buy a home. Another exception to this seemingly clear logic is religious and private school parents. We would hardly say they do not have choice, after all, they can afford tuition. But many parents who ultimately chose private schools in this study actually considered, on average, fewer ’9 It may interact with other variables but based on the Gaston parents’ stories, income is one of the variables with which it interacts. 176 schools than their counterparts who chose some form of public education (2.81 versus 3.77 schools, respectively). To complicate predictions further, it is costly to investigate many schools. With each additional school a parent investigates, the limited resources of time and energy are spent. It is possible that parents who have smaller choice sets gather better information about those schools because their time and energy are targeted narrowly rather than being spread thin across many schools. Further, the psychological literature that finds people can only keep five plus or minus two things in their short term memory, suggests that it is cognitively difficult to keep all of the details of eight or nine schools in mind. Parents who consider many schools must sort through a great deal more information in a fixed period of time in order to arrive at a similar decision. Though there is no literature which describes the optimal number of schools to consider, it is possible that there is a point of diminishing returns once a parent begins to consider a larger and larger number of schools. One additional piece of evidence bears on this variable’s relationship to the choice of a non-failing school. Among Gaston parents there is an inverse, modest relationship between the number of schools the parent is aware of (and mentions) and the possibility of ending up in a “non-failing” school. The point-biserial correlation between these two variables equals 037 (p<.01), indicating a moderately strong relationship between choice set size and school status. This inverse relationship suggests that more choices 177 may not necessarily result in better outcomes.“ Consistent with the latter arguments presented above, the model predicts that the larger the choice set, the smaller the odds of choosing a non-failing school. Sibling ’s school selectivity. As has been demonstrated in the previous chapter, parents’ notions of typicality shape their preferences, choice sets, and ultimate decisions. Parents’ notions of typicality are informed by their experiences with older siblings and since older siblings can be so varied, so too can parents’ conceptions of “typical.” Though there is a vast popular literature on the role of siblings in families (e.g., Leman, 1998; Newman, 2001), the role of siblings has gone unnoticed in the choice literature.“ Sibling’s school selectivity is a measurement of whether or not the older sibling’s school had selective admissions criteria or costs money (private, religious, magnet, and homeschools are considered selective). It is a revealed preference for a selective education. Among Gaston parents this measure is strongly related to ending up in a non- failing school (78 = 7.03, p = .021 , o = .48, OR = 9.75). Children whose siblings attended selective schools had 9.8 greater odds of attending a non-failing middle or high school than children whose siblings did not attend selective schools. 4° Given fact that the parents in this study are arguably “active” consumers, the pattern may be exactly reversed in the general population. Larger choice sets may be associated with higher odds of attendance at a non-failing school. This will have to be empirically determined. 4' There have been a number of scholars who have included the number of siblings in regressions, however, they do not investigate the role siblings might play in the development of preferences. Instead, they focus on the amount of time parents have to devote to a particular child, given the total number of children in the family (e.g., Gauri, 1998; Plank, Schiller, Schneider, & Coleman, 1993; Willms & Echols, 1993). This is an important but different dimension of siblings. 178 Two components of school selectivity -- cost and competition - can help us understand why sibling’s school selectivity may predict final school status. If a child is to attend a selective school, his or her parent has to be willing to pay for that education and/or take the necessary steps to apply to that school. This oflen includes soliciting reference letters, signing up for the admissions exams, and transporting the child to school each day. We can think of phrents’ choosing a selective school for a sibling as 3 revealed preference for two intertwined preferences, parents’ willingness to gain an “exclusive” education for their child and their skills in obtaining such an education. Parents who have chosen a selective school for a sibling might be both more skilled and more comfortable performing these tasks than parents who have not made such a choice. The parent may or may not select such a school for the target child, but if they preferred such a school, they would likely have the skills and resources to do so. The skills and preferences associated with choosing a selective school for a sibling are probably similar to (if not the same as) those associated with choosing a non- failing school. Those skills and preferences that are measured by sibling’s school selectivity, likely contribute to the strong relationship between sibling’s school selectivity and final school status. The model predicts that parents who have chosen a selective school for an older sibling are more likely to choose a non-failing school for the target child. Race. Race matters. It matters to people of all colors. We only need look as far as our segregated cities and suburbs to understand that people have strong racial preferences (Rubinowitz & Rosenbaum, 2000). And we need only look a bit further to 179 the busing efforts of the last 30 years to see evidence of strong racial preferences in the schooling (Orfield, Eaton, & The Harvard Project on School Desegregation, 1996). The Gaston parents’ racial preferences spanned a wide spectrum. A white father explained “I wouldn’t send my son to Washington cause it’s 99% black. Why would I want to send him to a school where he is the minority?” But another white father said, “I want them to be exposed. I want them to be able to get along with all people, not just white, middle class people.” Black parents articulated a similar range of preferences: some parents felt very strongly that they would not send their child to an all-white school and others searched out schools with that very demographic profile. Most parents, however, had preferences somewhere in the middle. Though many parents (33 of 48) preferred schools that were, in their words, “diverse”, they felt they had to settle for schools in which their children were in the racial majority. They said that diverse schools were simply “too hard to find.” Across the sample, the preference to have your child go to school with children who look like him or her, was related to choice of a non-failing school (x2 = 5.52, p < .027, (p = .35, OR = 6.71). White children had 6.7 greater odds of ending up in a non- failing school than did children of color. Perhaps again this is simply a restatement of the circumstances that led to cries for expanded choice. One explanation for this relationship is that there are not enough non-failing schools that serve African-American and Latino students. Whatever the reason -- a desire for homogenous schools, a need for higher quality schools serving children of color, or some other reason -- the model predicts that in the current educational climate, parents of color who prefer their children to be in a socially 180 safe environment (one with a reasonable number of children of color) have lower odds of choosing a non-failing school. White parents have greater odds of choosing such a school. Together, the preference variables offer a new and different account of parents’ preferences. The preferences rely heavily on revealed preferences, although there is one stated preference as well. They also take explicit account of the co-constructed, relativistic, and social and historical nature of preferences. The preference variables measure parents’ preferences for non-failing schools; the number of non-failing schools they considered as part of their choice set; their proclivity toward and skill at considering multiple schooling alternatives; their preference for and skill in obtaining a selective education for their child; and their preference for a student body of a particular racial make-up. Like the constraints, there are preferences missing from the model. They will be considered at the end of the chapter. We now turn to a brief summary of the model’s predictions. Model Predictions Given what has been explained about the literature on each of the variables and the data this study brings to bear on the variables; Table 5.2 summarizes the model predictions for each of the variables. 181 Table 5.2 Predicted Relationship Between Choice of a Non-F ailing School and Model Variables Variable Relationship to final school status Constraints Parents’ income + Parents’ educational attainment + Potential set size - Cohort consensus + School visits + Preferences Previous school’s status + Choice set size - Choice set status + Sibling’s school selectivity + Race - The constraint predictions are straightforward. As parents’ income and educational attainment increases, the odds of attending a non-failing school increases. As the number of schools in a two mile radius of the parent’s home increases, the odds of attending a non-failing school decreases. As the proportion of non-failing schools available to a parent through her social network increases, the odds of attending a non- failing school also increases. And finally, as the proportion of visited schools in the choice set increases, the odds of attending a non-failing school increases. 182 The preference predictions are also straightforward. Children whose previous school status is non-failing have greater odds of attending a non-failing middle or high school. As the number of schools in the choice set increases, the odds of attending a non- failing school decrease. As the proportion of non-failing schools in a choice set increases, the odds of attending a non-failing school also increase. Children whose sibling attends a selective school have greater odds of attending a non-failing school. African-American, Latino, and other racial minorities have lower odds of attending a non-failing school compared with their European-American counterparts. Robustness of the Model Because this model is an inductive model, it is necessary for us to consider the robustness of theorizing implicit in the model. One way to do this is to consider the predicted relationships in subsets of the sample. We might wonder for example, if income has the same positive relationship with final school status for city dwelling, lower income parents as it does for suburban dwelling, middle income parents. In order to ascertain the likely limits of the logic undergirding the model, the statistical relationships (chi-square association and point-biserial correlation) for two subsets of parents were recalculated. I will briefly describe each of the subsets of parents and then review each variable’s relationship with final school status for the two subsets of parents. First, there are the Gaston residents. These are parents who send their children to all different types of schools but who live within the Gaston city limits. As a subset, they have lower incomes, send their children to more failing schools, and have more members 183 of “minority” groups.42 They are not, however, different from the whole sample in terms of educational attainment. This group is particularly relevant because choice policy is most clearly targeted toward them -- urban parents “trapped” in urban schools (Chubb & Moe, 1990; Friedman, 1962). The second subset, the public choosers, comprised parents who selected charter, magnet, and neighborhood schools for their children. These parents are another group toward whom choice policies are aimed, parents who are unable (or one might argue, unwilling) to send their children to tuition-based schools. These parents have slightly less education than the parents in the other groups. They are similar to city residents in their income levels and the numbers of them that choose failing schools for their children. Finally, they are have more white parents among their ranks than city residents but not as many as the whole sample. Table 5.3 describes the sample and each of the two subsets. ‘2 Minority is in quotation marks because these African-American and Hispanic parents are not and have not been in the numerical minority in Gaston for the better part of 30 years, as far back as most of the parents can remember. People who look like them are in positions of power in the city government, the schools, the business community, and the police force. The use of the term “minority” in this situation obscures the reality of Gaston’s look and feel for its residents and visitors. 184 Table 5.3 Education, Income, Race, and Final School Status for Sample and Two Subsets Subset of sample All parents City residents Public choosers Measure (n=44) (n=36) (n=32) Education . Did not graduate from college 61.36 61.1 1 65.63 College graduate 38.64 38.89 34.38 Income Lower 54.55 61.11 59.38 Higher 45.45 38.89 40.63 Race African-American or Hispanic 79.55 94.44 84.38 European-American 20.45 5.56 l 5 .63 Final school status . Failing 56.82 66.67 65.63 Non-failing 43.18 33.33 34.38 Note. All values are reported as percentages of the given subset. Public choosers are parents who selected charter, magnet, and neighborhood schools for their children. Having described the subsets, 1 now turn to the relationships between the odds of ending up in a non-failing school and the model variables in each of the sample subsets. Table 5.4 summarizes both the point-biserial correlations and chi-square associations for the variables. For chi-square statistics in which there was a cell less than five, exact statistics were calculated and are reported in the table. In most cases, this raises the p- value, making it a more difficult to reject the null hypothesis. For a complete account of how each model variable was calculated, see Appendix G. 185 Table 5.4 Point-biserial Correlations and C hi-square Relationships of Model Variables and Non- failing Schools for the Sample, City Residents, and Public Schools Final school status All parents City residents Public choosers Measure (n=44) (n=36) (n=32) Point-biserial correlation Potential set size -0.39** -0.29 -0.31 Choice set size -0.37* —0.33"' -0.46" Cohort consensus 0.50” 0.42“ 0.50" Percent visited 0.42" 0.40“ 0.50" Choice set status 0.81"" 0.76“" 083*" Chi-squared association Education x2 = 1.15 x2 = 1.46 x2 = .37 (p=.28,0R=.51 q)=-.20,0R=.39 qr=-.ll,OR=.6I Income x2 = 0.67 x2 = .058 12 = .16 tp=.l3,0R=l.67 tp=.04,0R=.I.19 cp=.07l,OR=l.35 Previous school status x2 = 11.2“ 12 = 5.84" x2 = 6.79" qr = .50, OR = 9.64 qr = .40, OR = 6.0 (p = .46, OR = 9.0 Race x2 = 5.52“ x2 = 0.27 x2 = 1.72 (p = .35, OR = 6.71 (p = .09, OR = 2.1 q) = .23, OR = 3.56 Sibling’s school selectivity 12 = 7.03“ x2 = 4.00“ 12 = 6.66" (p = .48, OR = 9.75 (p = .40, OR = 6.5 (p = .54, OR = 16.26 Note. In the analyses of sibling’s school selectivity and sibling’s school status, families with a single child or siblings that were not school age (K-12 grades) were dropped from the sample. This resulted in N=30, N=25, and N=23, for all parents, city residents, and charter, magnet, and neighborhood parents, respectively. *pS0.05, "pS0.0I, "*p<0.0001 (two-tailed) Education, income, choice set size, school visits, choice set status, previous school status, and sibling’s school selectivity have consistent relationships with final school status across the subsets of parents. This suggests the theorizing that produced the 186 model relies on relationships that are stable across subsets. It also suggests that a particular subset of parents is not exerting a disproportional influence on the sample. There are however, three variables that deserve a brief explanation. The potential set size variable is not significant among city residents or public choosers. Though the table does not show it, potential set size achieved borderline statistical significance (each had p values of 0.08). Because the city resident and public chooser subsets are smaller than the whole group, it is possible that weaker relationships lose their statistical significance in those groups. For the reasons articulated in the description of the variable it was important to include potential set size even through statistical significance was not achieved in the smaller subsets. The second variable which deserves explanation is choice set size. The relationship between choice set size and frnal school status was not statistically significant for the city residents. This is likely due to the lack of variance among city parents. On the whole, they considered many more schools than did the suburban parents. Suburban residents had small choice sets and chose almost all non-failing schools. This is an interesting finding and suggests that, depending on the distribution of non-failing schools across urban and suburban lines, choice set size may not be related to final school status. To complicate matters, the relationship between choice set size and final school status for public school choosers is even stronger than for the whole sample. It may be that choice sets are simply larger when they include public schools and as mentioned earlier, more difficult to manage. Whatever is causing the relationship between choice set size and final school status, it has yet to be pinned down empirically and deserves further exploration. 187 The final variable, race, also loses statistical significance in the city residents’ subset. This is because city residents are almost all African-American (94%) and the sample is not large enough to detect what variation there might be around race with so few cases of white parents in the city. Looking at the other two groups - which have small but still larger percentages of white parents - the relationship between race and final school status is stronger and in the case of the whole sample, statistically significant. Having made a case for the inclusion of particular variables in the model, I now turn to the limitations of the model. Limitations of the Model One critique of this model is its reliance on the measurement of failing and non- failing schools. Through the media writes about those distinctions as if their meaning were stable and unambiguous, students of politics know that these distinctions are human judgments invented by a small number of people with competing interests. One can argue that the distinctions are flimsy, shifting, even arbitrary; they should not be used to understand parents’ thinking. While this argument should be taken seriously, every measurement is imprecise. There is uncertainty and error in every measurement. The question is one of degree, not kind. While the reliance on such a political measurement troubles me, it is reasoned, logical and consistent. If a researcher has to pick a measurement on which to hinge an analysis, school status is as good as any other. What troubles me much more are the current limitations presented by the data we have on schools of all types. It is to this limitation I now turn. In presenting this model, there is a subtlety in parents’ choices that has been lost. Having listened and transcribed more than 100 hours of interviews with these 48 parents, 188 I am convinced that with a few notable exceptions, parents chose schools that were better than the rest of their choice set. The schools were better both in terms of test scores and in terms of the fit with their child’s and family’s needs. Parents made good choices and improved their children’s life chances by making decisions on their behalf. But because of the blunt instrument of failing and “non-failing,” this subtlety is lost. Twenty-five parents appear to have made poor choices; they chose failing schools. And some of those parents chose schools that truly are, from a policy perspective, worrisome. But yet, there was wisdom and forward motion in their choices. The puzzle then is how do we better represent the complexity of parents’ thinking and actions? My hunch is that if we had different data, data on ethos, school environment, family needs, parenting practices, and the match between a particular child and a particular school, we would have a much been chance at understanding the wisdom (and lack of wisdom) in parents’ thinking and behavior. If we had more sensitive measurements we would be able to see exactly how the chosen school compares to the other schools in the choice set. We would be better able to predict parents’ actions. 189 Chapter Six Conclusions and Implications Review of the Argument Before I offer some conclusions about the findings presented here, I would like to briefly review the arguments that have been made in the preceding chapters. As the chapters are long and take up different types of analyses, it is easy to lose the thread which weaves them together. I will present a summary of the arguments in each chapter and then offer some conclusions that build on those arguments. Chapter I Scholars, using rational choice theory have made predictions about how parents will behave given a more .open market for schooling. The prediction that parents will choose better schools has met with the reality that few parents actually move their children out of failing schools. I argued that the poor predictive power of current models is the result of methodological and theoretical problems. Scholars have studied parents’ thinking using deductive methodologies that afford little space to understand parents’ logic on its own terms. And scholars have used rational choice theory as it is classically defined, with preferences assumed to be exogenous to the model. In order to address these theoretical and methodological problems, this study took an interpretive stance toward answering the question, “Why do parents choose particular schools for their children?” The study also sought to inductively develop a case for and explanation of the potential endogeneity of preferences. The study builds on current literature by using rational choice theory, specifically the constructs of preferences and constraints, to offer a potentially more accurate model of the choice of non-failing schools. 190 Chapter 2 The study was conducted in Gaston and its adjacent suburbs. Gaston is a large, Midwestern city in which the majority of city residents are people of color. Over the course of eight months, 48 parents were interviewed twice before their children began school in the fall and once afterward. The group of parents was stratified by grade level (5th or 86‘), family income, and school status (failing or non-failing). The study took an interpretive stance and sought to understand the choice process from the parents’ perspectives. A mixed methods approach was taken to data analysis. Modified analytic induction and descriptive statistics were used to analyze different dimensions of the data. Chapter 3 Traditional explanations -- disinterest, lack of information, and constraints -- don’t explain parents’ choice of failing schools. I draw on evidence from the choice process as well as the relationships between traditional constraints (e.g., inaccurate information, income, education) and the odds of choosing a failing school in order to make this argument. Based on the choice of college literature, I describe the choice process as having three phases: predisposition, search, and choice. I argue that this process divides parents into three groups, those who do not conduct a search for schools, those who conduct a closed search, and those who conduct an open search. After an analysis of the parents’ reasoning (which is visible through their actions of discovering schools, making school visits, and constructing choice sets), I conclude that -- irrespective of search procedure -- parents used information to make interested, reasoned decisions. These decisions, over half of which were the selection of a failing school, left satisfaction and a sense of choice. Having eliminated the traditional explanations for 191 choosing failing schools, I argue that we must investigate parents’ preferences if we are to more accurately predict their behavior. Chapter 4 Through a case-based analysis of parents’ preferences, I argue that preferences are bound by parents’ cultural understandings of their children, schools, and communities. Those preferences are visible in the choice sets parents construct as well as their ultimate school selections. I explore four versions of cultural understanding: “Children contribute to their own academic achievement,” “My child is typical,” “1 know what quality is for my child,” and “This is my community.” Across these propositions I demonstrate that parents’ preferences (and therefore, choices) are co-constructed, relative, and social and historical. Preferences are co-constructed because they are defined in interaction with the choice context. They are relative because there is not a unitary scale against which school quality is judged. And they are social and historical because they rely on other people for their sensibility and are influenced by present day ideas and beliefs that were developed in the past. Together, these cultural propositions illuminate four dimensions of preferences that are critical to the choice process: academic achievement, child development, school quality, and community. Finally, the analysis suggests that preferences are not exogenous to the choice models and ought not be treated as such. Chapter 5 Based on the analyses in Chapters 3 and 4, I step away fiom parents’ thoughts to propose a model that predicts parental choice of schools. The model, which includes ten variables -- parents’ income, parents’ educational attainment, potential set size, cohort consensus, school visits, previous school status, choice set size, choice set status, 192 siblings’ school selectivity, and race -- relies heavily on revealed, rather than stated, preferences. Because the sample size is too small to test, the model’s potential applicability is considered for different subsets of the sample. Based on these analyses, I conclude that the model is a reasonable first step toward more accurate predictions of parents’ choice behavior. Having reminded the reader of the content of each chapter, I will now look across the chapters and offer some conclusions that bear on both the conduct of choice research and our understanding of choice. Conclusions Data. Patterns. Theories. Data. Patterns. Theories. This beguilingly simple mantra -- data, patterns, theories -- rings in my mind as I share the conclusions I draw from this study. The mantra, taken from Anderson’s (2003) work with pre-service science teachers, simplifies the messy work of any science, social or otherwise. The scientist gathers data, finds patterns in those data, and theorizes about how those patterns make sense. Having summarized the patterns laid out in the chapters, we now turn to the theorizing. There are two kinds of theorizing relevant here: theorizing about the use of rational choice theory and theorizing about why parents choose failing schools for their children. I will eventually comment on both, but I begin with the latter. At a surface level, parents who chose failing schools did not see them as failing. They chose those schools because they preferred them. Given their other perceived options, the schools they chose were reasonable selections. The schools may not have 193 been perfect but the parents would not (and did not) describe them as failing. We could attribute parents’ choice of failing schools to ignorance; after all, they did not use test scores in making their decisions. But as Chapter 3 demonstrates, that is not a reasonable explanation, given the evidence on the information and higher proportion of non-failing schools in their choice sets as compared to schools in a two mile radius. I propose that the explanation for parents’ actions lies in a different conceptualization of parents’ choices. Parents’ choices are not the unbounded, free-will, any-school-you-desire, kind of choices that free market advocates suggest. Nor are they a predetermined, Marxian production of race or social class. Rather, they are the result of both structure and agency. They are both produced and made. Parents’ choices fall within trajectories of choice and, if we are able to understand the forces that shape choice trajectories, we will be able to understand why some parents choose failing schools and others choose non- failing schools. Trajectories of Choice The Oxford English Dictionary defines a trajectory as “The path of any body moving under the action of given forces” (Simpson & Weiner, 1991, p.1364). Based on the data presented in previous chapters, I conclude that parents’ selections of failing schools are explainable if we consider those choices to be a part of a choice trajectory. A trajectory of choice is a path made up of choices influenced by two forces: social and historical. I will first explain the two forces and the role they play in parents’ thinking, and then I will make the case that these forces produce unequal odds of choosing particular schools. 194 The social forces shaping choice. As we have seen over and over and over again in this study, choice is social. It does not happen inside a parent’s head, cut off from the rest of the world. At every stage of the process, there are people -- other parents, spouses, other family members, and the children themselves -- influencing the constraints and preferences that comprise a decision. Though there are many examples of the social nature of choice in each of the chapters, I will focus on three. Social networks shape parents’ preferences and choice sets by giving them access to information about particular schools. As the data on school discovery suggest, almost all parents rely on their social networks for information about schools. But not all networks are created equal. As we see in Chapter 5, some networks provide access to non-failing schools while others do not. Depending on who is in one’s network, parents have differential access to failing and non-failing schools. Another example of the social nature of choice is in the use of reference groups. Reference groups, or the people to whom the parent implicitly or explicitly compares her child, are by definition social. As demonstrated in Chapter 4, parents rely on their assessment of how their children are doing in order to determine what types (e. g., religious, private, etc.) of schools will be a part of the choice set. This assessment hinges on the reference group. As Mr. Dish’s case demonstrates, because Tasha is doing well and he wants her to stay on track with her peers, he only considers one school, the school in the customary attendance pattern. Mrs. Webb made an almost opposite decision based on her daughter’s reference group. Mrs. Webb deliberately moved Alexandra out of the customary attendance pattern because she wanted Alexandra to be different from the kids she had grown up with. Mrs. Webb searched for other schools because she reported that 195 it was time for Alexandra to be challenged in ways her peer group could not challenge her. In both cases and across the 48 parents, the reference group played a large role in defining both what the parent was looking for and which schools rrright be able to match those preferences. But sinrilar to networks, reference groups vary. Some parents are comparing their children to others who struggle in school and accomplish something important if they graduate fi'om high school and earn a higher degree. Other parents are comparing their children to peers who read chapter books in their spare time, score high on standardized tests, and attend four year universities across the country. A final example of the social nature of choice comes from geography. Almost any geographer will tell you geography is not about space, it is about place. And the difference between the two is people. Through interactions with one another, people make spaces into places. As the fourth proposition demonstrates, parents have certain loyalties to the places in which they live. Those loyalties influence which schools make it into the choice set. Parents like Mrs. Brown and Mrs. Wise are unwilling to consider schools outside of Gaston. They are committed to Gaston as a place, a community, a home, a culture. But again, different places have qualitatively different schools. So even if two parents have similar commitments to their communities, they may have vastly different choice sets because of the particular schools those communities provide. The historical forces shaping choice. The second force which shapes choice — history — is, of course, related to the first. History relies on people but is distinguishable fi'om the strictly social because of its implicit notion of time. It is a force which takes on lived experience and learning, processes that require time. It is to the historical force we now turn. 196 The historical force might not be as obvious to readers as the social force. It is however, just as pervasive. We will consider the historical forces that shape parents’ choices in three instances: the customary attendance pattern, family history, and the co- construction of the choice set. Choice as a reform is layered on top of a pre—existing system of schooling. In that system there are a range of school types (e.g., homeschool, private, charter, and neighborhood), each of which has its own set of customary attendance patterns. For neighborhood schools, the option-demand system creates customary attendance patterns. Children are assigned to specific schools and it is the parent’s choice to select other schools should they so desire. For private and religious schools, a common set of values and sense of community cause parents to select the same schools over and over again until there are customary attendance patterns built up for various schools. Customary attendance patterns influence preferences and choice sets in at least two ways. First, they can act as a ready-made option, causing some parents to forego conducting at search at all. Second, by acting as a point of comparison, the schools in the customary attendance pattern can act to shape the criteria upon which other schools are judged. A move away from the customary school might only be justified by a clear gain on criteria established by an analysis of the customary school’s strengths and weaknesses. For example, a parent whose neighborhood school does not have an all-day kindergarten might focus her search around finding a school with all-day kindergarten because that is the characteristic nrissing from the customary school. The point here is not that the customary attendance pattern can be negative. Rather, the point is that customary attendance patterns vary. Some include non-failing schools, others do not. Some include multiple schools, others 197 just a single school. The particular schools in the customary attendance pattern help to shape parents’ preferences, choice sets, and final decisions. And the customary attendance pattern is inherently historical, emerging over time through the choices of earlier generations of parents, organizations of school districts, and the like. A second instance in which we see history influencing parents’ choices is through families. Certain skills and beliefs are passed down through families and are born of family experiences. These skills and beliefs influence parents’ current choices. For example, if a parent chooses a charter or private school for the older sibling, she carries the skills and knowledge acquired in that process to the choice for the next child. This idea was supported in Chapter 5, where we saw that a sibling’s school selectivity is related to a parent’s choice. Additionally, parents also acquire beliefs about schooling from their own experiences in school as students. These beliefs and skills influence their current choices. Recall Chapter 4, in which Mrs. Grisson struggled in school but worked hard and graduated from high school. She said, “I barely made it through but I’m glad I did. And I sort of see that in him [her son Shawn].” Mrs. Grisson’s understanding of Shawn’s struggles is directly informed by her prior experience in school. Like Mrs. Grisson, Mrs. Hawill’s current actions are informed by ideas she learned in her childhood. She said, “We were raised that education is important. ...If you’re educated you can do. So I don’t want them to be denied anything because they didn’t know. They need to have an education.” Both mothers draw explicit connections between the past and the present. Their educational histories matter to their current choices. As MacLeod (1995) and Lareau (1989) demonstrate, however, not all educational histories are alike. Though many people have positive school experiences, those 198 experiences differentially convey the requisite skills and knowledge for interacting confidently and productively with schools. Some families, in particular middle class families, already possess the skills necessary to make school interactions simple and productive. Other parents have to learn those skills once they become parents. The skills and experiences necessary to interact with schools in positive ways are not distributed evenly across parents’ educational histories. This is painftrlly captured in Ted Okey’s (1990) study of dropouts: For the students and their families in his study, dropping out was not a problem, it was a solution to a problem: The alienation of generations of students, their parents, and their extended families from school. This has implications for choice. The final instance of historical force is found in the co-construction of parents’ preferences. The co-construction of preferences has an historical dimension that was not mentioned in the detailed description in Chapter 4. Recall Mr. Dish, Mrs. Hawill, and Mrs. Carol. Mr. Dish wanted Tasha to continue her exemplary science education so he considered only a science magnet school. Alecia, Mrs. Hawill’s daughter was unable to attend her preferred school because the magnet school did not accept her in the initial application process. And Mrs. Carol only considered certain schools because she understood Denzel’s poor grades as his, not the school’s problem. In each of the cases, an event or series of events in the past influenced the current choice. For many reasons (e.g., funding) academic opportunities are unevenly distributed across children and schools. Few children have the opportunity to attend a magnet science elementary school. Children who do not do well on standardized tests get rejected from academic magnet schools. Children who earn poor grades do not have the same academic 199 opportunities as children who earn higher grades. Though all parents’ preferences are co- constructed, some parents enjoy constructing their preferences within an environment rich with opportunity. The environments in which other parents co-construct their choices are much more limited. Like the other instances of historical force, the co- construction of preferences plays a role in choice that is likely to produce systematically different outcomes for parents and children. Di/ferent forces lead to difierent choices. In the classic text, Distinction, Bourdieu (1984) makes a compelling case that our tastes (i.e., preferences) are shaped by our location in society. He argues that education and social class shape all of our preferences: the books we want to read, the furniture we buy, and even the art we consider beautiful. The distribution of tastes across a society is not the deeply personal idiosyncratic phenomenon some market advocates assume. According to Bourdieu, it is not equally likely that an educated physician and a less educated factory line worker will consider the Mona Lisa beautiful. Based on Bourdieu and the evidence that social and historical forces are unequally distributed across parents’ choices, I assert that preferences are unequally distributed across parents. Thus the odds of choosing a particular school, be it non-failing, selective, private, or any other, are inherently unequal. Let us briefly consider two parents, Mrs. Smith and Mrs. Bear. Mrs. Smith is a high school graduate whose own parents did not participate in her education. Mrs. Smith works for the phone company and her social network gives her little access to information about non-failing schools. Mrs. Bear on the other hand, is a college graduate whose own mother was involved in every aspect of her education. Mrs. Bear works for a hospital and her social network provides her with 200 access to information about many non-failing schools. It is not surprising then that Mrs. Smith chose a non-selective failing school for her son while Mrs. Bear chose a selective non-failing school for her daughter. By “not surprising” I do not mean to imply that the mothers’ choices were inevitable. I am not of the mind that social and historical forces predetermine parents’ school choices , but this study suggests that they do shape parents preferences and those preferences, together with constraints explain parents’ choices.43 Applications of Rational Choice Theory to Parental Choice We now consider the remaining conclusions surrounding the use of RCT in investigations of parental choice. The first conclusion has already been specified in Chapter 5: better predictions of parents’ actions require a greater use of revealed preferences. Stated preferences may still be useful, perhaps essential, to accurate predictions of parents’ actions; however, current work relies too heavily on them. Though this may pose challenges, such as increasing the cost of conducting studies of parents’ preferences, the use of revealed preferences is likely to move choice research forward by specifying (and debating) the relevant variables and their measurement. The second conclusion regarding RCT concerns endogeneity. Endogeneity is one of the most pervasive problems in social science and parental choice is not immune. As explained earlier, preferences are usually assumed to be exogenous to models of choice. This simply means that the preference variables are independent of the dependent ‘3 Because these trajectories of choice emerged from the study, of course, I can only assert that trajectories of choice exist and that they shape parents choices in the ways I have sketched here. Another study would need to test the notion of choice trajectories. 201 variable (in this case, the choice of a non-failing school). The direction of the causal arrow is clear. Some examples of endogeneity problems include Trade is a cause of international peace OR states trade if they anticipate peace. Democracy causes economic growth OR economic growth causes democracy. Campaign money causes candidates to win elections OR donations go to anticipated winners. (emphasis in original) (Gelpi, 2004, p. 3) If we apply this to thinking to parental choice, we might ask, “Does a preference for non- failing schools cause attendance at a non—failing school?” or “Does attendance at a non- failing school cause a preference for non-failing schools?” Given the results presented in Chapter 4, it is clear that the answer to both of these questions is yes. Having a child enrolled in a school in which children are reading at grade level, teachers call home, and there is no violence causes a preference for schools with those traits. At the same time, having a preference for high test scores, high-quality teachers, and safe schools causes a parent to choose such a school. Both are true. Thus, we conclude that preferences are endogenous to the choice of school and future models of choice must take account of this complexity. I have painted a complex picture in this dissertation. I have argued that parents’ choices make sense if we understand them to be embedded in trajectories of choice. I have also argued that those trajectories of choice rest on preferences that are culturally bound and therefore endogenous. What are the implications of all of this complexity? I will explore two. 202 Implications First, I will consider the implications for choice as an equity enhancing, school improving reform strategy. And second, the implications for our current system of schooling. In the final two sections of this chapter I will consider the limitations of the study, as well as the questions this work leaves unanswered. Choice is Unlikely to Deliver Equity and Improvement Of the 48 parents in the study, 41 chose non-assigned schools. They are more educated and wealthier than the average Gaston parent. They were overwhelmingly satisfied with their decisions and almost three quarters of them felt they bad choices about where to send their children to school. All of the 48 used information to make interested, reasoned decisions. And yet 25 chose failing schools. If the parents in this study reasonably chose failing schools, we can realistically expect that other parents will do the same. My study raises some doubts about the efficacy of choice in responding to certain longstanding problems in US education. But not for the reasons others tend to nominate. Choice will not fail because parents will make bad choices (though they may); and it will not fail because parents will not have choices (which they may not). Choice may fail to deliver the dramatic changes promised to individuals and the system for three reasons: attribution, test scores, and red flags. I will explain. Parents do not attribute academic success to schools. The achievement ideology, the notion that anyone can succeed if they just work hard enough, is entrenched in our collective psyche. To acknowledge the critical role schools play in advantaging some and disadvantaging others would undermine this belief and our faith in it. It would undo 203 the very logic of our educational system. And so, year after year in the Gallup poll, parents attribute the achievement gap, a measure of academic achievement, to families. to parents, to communities. If children are not learning what they should at school, it is because their parents are not involved or they come from broken homes. It is not because their schools are under-funded and they are not given the same opportunities as other children to learn ambitious curriculum with supportive and well educated teachers and principals. It is agency, not structure. If parents do not attribute academic achievement to schools, why would we expect them to en masse, “vote with their feet?” Parents do not use test scores to decide between schools. They have other measures of quality that are valid and reliable for them and they match their preferences with those other measures. For the most part, their measures work well; they construct choice sets that, on average, contain more non-failing schools than the set of schools within two miles of their homes. Parents generally believe test scores can tell you something about a school, but they do not use them in their search processes or comparisons of schools. But test scores, whether we like them or not, are external, somewhat objective, criteria on which schools can be compared. Though the scores do not summarize the complexities of teaching and learning, they do offer a relatively unambiguous unitary scale on which to compare schools. There is no other measurement that offers such benefits. If there is not a common academic measure that parents use to compare schools to one another, how do we expect comprehensible signals to be sent from parents’ movement of their children? And finally, the red flags are few and far between. Because of the racial and class segregation of our schools and workplaces, we rarely come into contact with people who 204 might have different skills and knowledge of schools. This means that the red flags, the little facts that might cause a parent to question her own thinking, are unlikely to happen. As Mrs. Springfield said, If all you know is a classroom with 30 students, then why would you even think about what it would mean to have your child in a classroom with 15 to 16 kids? If your whole exposure is where the entire student population is of the same race, why would you consider, why would race or diversity be an issue for you? The opportunities to discover that education can be different for your child are limited. Our closed social networks keep the same people with the same experiences talking to one another. This isolation means that information and the action it can spur, are unlikely to travel across racial and class boundaries. Based on these three reasons — attribution, test scores, and red flags — choice is unlikely to produce the kind of wholesale improvement in equality and quality choice advocates hope for. Thus, we are left with a dilemma. How Do We Improve Schools? If it is unlikely that choice will lead to a significantly improved, more equitable school system, we are still faced with the question choice advocates tried to answer with their reform: How are we going to improve schools? Do we go with the free market and make everyone choose, hoping that if everyone is forced to choose, the invisible hand of the market will work its magic? Or do we batten down the batches and develop even stronger accountability schemes that will force teachers and principals to do better in their classrooms and schools? How do we choose? 205 Sedlak (1995) writes about our inability to develop policies that sit between hegemonic imperialism and passive abdication of responsibility. He argues that we tend to swing back and forth in our policy making, unable to find a happy medium. This holds for the history of educational reforms concerning choice. The former system of schooling, when all children were required to attend their neighborhood schools, was the hegemonic imperialism and the free-market vision of choice in which everyone has to choose a school for her child, is the passive abdication of responsibility. The puzzle then is, how do we find a middle ground? Quite frankly, I am still left thinking about how we find it; but I am quite convinced it is necessary. We cannot simply abandon choice as critics might have us do. The happiness, satisfaction, and academic improvement produced by parents’ choosing to send their children to different schools cannot be ignored. There is power in choice. But we cannot ignore the thousands of parents who reasonably choose failing schools for their children. The consequences for those children are simply too great. If we are able to find a middle ground, my hunch is that it will not only rely on the invisible hand of the market. It will also rely on our own hands. I doubt there is any other way. Limitations and Additional Analyses Unlike medicine and the physical sciences there are few, if any, immutable laws that govern the social sciences. Context matters. This work strives to investigate a particular group of parents in order to shed light on new ways of thinking about choice. It is not a study that “proves” that choice works one way or another for all people. It is not a study from which we can generalize to all Gaston parents or even to other interested parents in other cities across the country. These were a particular group of people who 206 had a particular set of stories to tell. My work here has tried to take those stories seriously and learn something new from them. The contribution of the study is in its ability to challenge our previous assumptions about how choice operates so that we can then test revised fiameworks and models empirically. Given those goals, I would like to consider three other limitations of the study which bear on its reliability and validity. The first issue which still concerns me is my role as the other. I am a white middle class woman who interviewed mostly parents of color. There was a power dynamic present in my conversations with parents that I think is almost impossible to overcome in three interviews. Whatever distrust may have existed probably caused a social desirability effect. Parents may have withheld information or exaggerated the reasonableness of their actions in order to convey the message that they are a good parent. For example, a parent who chose the neighborhood school because she simply was too busy to check into any other schools or she was in a fight with her child so she hadn’t wanted to go check into other schools, may have told me that she thought it was better to send her child to the neighborhood school because his friends were there or she felt it was a good school. The desire to appear to be a good parent may have caused parents to emphasize certain dimensions of their thinking and not others. This is not to say parents were lying to me, though they may have been, but rather, that distrust may have caused them to omit some details and include others in order to construct a particular picture of themselves as parents. With another researcher conducting the study, it is possible the data would have been slightly different. Again, other analyses (such as the comparison between the number of non-failing schools in the choice and geographic set and the consistency in status between the child’s previous and 207 new school) lead me to believe that the issue of distrust was not enough of a problem that we should consider the data fimdarnentally flawed. We should however, realize that the data were collected under a particular set of circumstances that may have influenced the data’s reliability. A second limitation of the study regards coding and reliability. Due to the limited funds available to pay another person to code data, I was the only person to code the data. Though this is common in dissertations, it limits the study’s reliability. It is possible there was miscoding and uneven application coding criteria. Had there been other people to code the data, inter-rater reliability checks would have been performed and we would know more about the coding reliability. Additionally, nuances within the coding scheme may have been worked out sooner and more clearly than they are now. Though there were safeguards to help me negotiate this limitation (double checking of the accuracy of the coding, iteratively revising the coding scheme, and writing multiple memos reporting on analyses over time), it remains a potential source of error. The final limitation of the study I will address here is that of the “observer effect”. 1 am sure that for a handful of parents, my presence changed their behavior. Two of the parents explicitly told me that they were interested in my “help” thinking about where to send their children to school. Another asked my opinion about some of the schools she was considering. In each of these conversations I tried to explain that I didn’t have answers about which schools were best but that they might want to look at some more information about the schools. I referred them to the county level school district’s website as well as the state’s website, both of which have school level descriptive and empirical data. For these three parents, my presence made it more likely that their child 208 would end up in a non-failing school but for all the parents there was a more subtle observer effect. I forced parents to verbalize their thinking when I asked them questions about what they are looking for in a school or how things were going at the new school. I made them impose order on their thoughts and then share them with me. This is an intervention. Through 1 did not treat them as “research subjects” (Bogdan & Biklen, 1992), I did, through my actions and words change the natural setting that exists when I am not there. Though I cannot know exactly how my presence changed the natural setting, I imagine parents were more inclined to ask questions when they went to visit schools, actually investigate schools they might otherwise would have dismissed or even just think more about their choice because I was asking them questions. Overall, these potential changes to the unaffected process of choice are more likely to result in a child attending a non-failing school because they engage parents in the choice process. Another way to think about limitations of the study is to think of the additional analyses that might have filled in or slightly altered my ideas. Because I had such a rich data set with which to work and a limited amount of time in which to do it, there are still analyses left to do. I will describe some of the additional analyses which might have been conducted so that the reader may better understand the both the scope and limits of the actual study. Perhaps the most important analysis left undone is the analysis of the history of family choice. I have data for each parent and all the children in the family concerning where family members went to school and why the parent chose those schools. Though the data suffer from all of the problems of hindsight bias, it is a rich source of revealed 209 preference data. There is much to be learned about nature and meaning of the current choice through its relationship to prior choices. Throughout data analysis, I focused quite deliberately on parents’ preferences. I did this for all the reasons outlined in the first chapter. This meant however, that information and constraints more generally, did not receive the same level of attention as preferences did. Had I more time, I would have conducted more thorough analyses of both constructs. What actually counts as information in choice decisions is vastly under- estimated in the choice literature. A careful analysis of information would help us understand why some parents value certain kinds of information over other kinds and what deterrrrines the status of information in a choice process. Further, it would help us understand the social life of information, so apparent in parents’ thinking. Like information, constraints have been portrayed in the literature quite narrowly. As the cases in Chapter 4 demonstrate, parents’ preferences and constraints are intertwined. Denzel’s poor academic performance is the basis for Mrs. Carol’s preferences and it is a constraint on the schools be can actually be accepted to. Thus, a category like child’s academic achievement, is not as analytically distinct as we might, for theoretical purposes, like it to be. A more careful treatment of constraints would allow us to better understand the construction of constraints as well as their relationship to other variables influencing parental choice. A final analysis that was not conducted with the data was an analysis of the geography of choice. It would be very helpful, both for schools, and for policy makers, to better understand how geography places a role in choice. I have school names and addresses for all of the schools in both the geographic and choice sets. It would be 210 helpful to quantify the distances between the parents’ home and potential schools so that stronger arguments about the nature of convenience could be made. Because I also have access to siblings’ schools and the parents’ account of how the siblings’ school played a role (or didn’t) in the choice process for the target child, a more fine-grained analysis could even look at the distances parents nominated as important and see how those distances compare to the distances associated with other potential schools. There is some disagreement in the choice literature about the degree to which parents’ choose schools out of convenience and a careful analysis of the geography of choice would add to that conversation. The analyses sketched here are not the only ones left to do but they have the greatest potential to contribute to the argument outlined here. I would like to turn now to the questions this study could not have answered. They are questions that arise from this work. Future Work There are many questions that arose from this study, however, not all are equally crucial to our immediate understanding of choice. Data on some questions would allow us to make changes to current choice schemes which might significantly alter their effectiveness. Data on other questions would deepen our understanding of choice but play a smaller role in informing choice policy. In order to get a sense of the scope of questions that have arisen from this study, I list ten of the questions and then describe my next steps. 1. Can a large scale model which utilizes revealed preferences better predict parents’ choices? 211 2. In what ways, if any, are choice markets segmented? 3. Do trajectories of choice exist across choice contexts? If so, how do they function? 4. To what degree do parents’ choices match their stated and revealed preferences? 5. How does a parent’s attribution of academic success influence choice? 6. How do family educational histories interact with the current choice environment to influence choice? 7. What role do 8th graders play in the choice of high school? 8. In what ways does the social life of information influence choice? 9. How does the presence or absence of a customary attendance pattern influence choice? 10. What fraction of marginal consumers must choose non-failing schools in order for choice to improve the system? I mentioned at the beginning of the chapter that the goal of any science is to develop theories that are based on patterns in the data. I suggested one such “theory,” the notion of trajectories of choice. From a broader perspective, trajectories of choice are part of a much larger issue, social inequality. But as inequality is one of the persistent social problems of our time, I am faced with a dilemma that all people interested in social problems face. Social problems are not disciplinary. They do not rest in sociology or psychology or economics. They do not belong to one discipline or another. And while this can be positive (scholars from different traditions can use different methodologies to more fully illuminate the problem), it is a challenge for a novice researcher who does not 212 herself belong to a discipline. Thus, I suggest that my next scholarly task is to better understand exactly how trajectories of choice might productively be investigated. As Loury (2002) found in his ground breaking treatise on racial inequality, a particularly stubborn social problem, new ways of thinking about old problems require insights from various disciplines. Analysis of social problems does not simply benefit from interdisciplinary perspectives; it demands them. And so, as I look out onto the scholarly horizon, I anticipate spending much of my time working across the social sciences in the hope of bringing a variety of frameworks and tools to bear on school choice and the trajectories upon which it lies. 213 < x€=onn< ~80 «N0 2.0 3.0 5.0- 2.0- 2.0 mmod 5.0 0 1.3.0 :30 sign? 08.0- 3.0 mmd 2300 3.0- v8.0- :36 ehhnmd- 2.9- and nnmd :Nvd ammd mNd reswvd- 2..de- eibmd VNd {ind 1:.de ~Nd 2.0 M20- .mmd. 33.0- “:0- 000.0 0 :0 :30- S .0- M20 2.3.0 2.0 .3390 2.3.0 2.0 .00 3.3.0 0N0 N0- 800 :28 Boa; s is: 9 0205 H8 885 mass. 8m meann— Bm .25: was; -3828 Em .«o E 28 com 86:0 .55 35:0 muzfim we egg :8 new no: :8. .35 85m flaw “2.x. 8mm >2; o\.. 28 a 5 £8 a E £8. a -329 2:85 35..— ..EE .Szum RIBS. 5253er 214 mafia com fl0umm_> Hum 85:0 5 ME .32 L95 Hum 865 3:5 N €03 .080: So. V no: Ac. V a: .8. v a. .22 215 Appendix B Interview Instruments Interview One 1. I don’t know very much about [name of current school]. Will you tell me about it? 1. [name of child] will be going to sixth (or ninth) grade next year. Can you tell me a little bit about what you’re thinking about for next year? 2. Why those school(s)? 3. Are you looking for any specific things in a school for _? 4. Is there stuff that I should have asked that I haven’t? Things that you are thinking about that we haven’t touched on? Interview T we The Update 1. Can you tell me a little bit about what has happened since we spoke in the spring? Be sure to ask for follow up information o What are they worried about with respect to certain schools? 0 What are the logistics associated with each choice? 0 Which school do they think they have the most information about? 0 What information about schools do they pay the most attention to? 2. To what extent do you feel like you have choices about where to send (child) to school? 216 Summary and Decision Map . I’d like to show you this decision map and have you check it for me. I’m sure I’ve made mistakes so I’d like to have you correct them for me. Are there any corrections I need to make? Things you would like to add? . Next, I’d just like to read you what I wrote for myself as a summary of what we discussed last time. Again, I’m sure I don’t have this just right so please tell me where I have not said the right things. Or maybe haven’t paid attention to the things you really want me to understand. Odds and Ends . As I was summarizing our first interview I realized that there were things I didn’t understand and questions I should have asked. So, let me just look at my notes to myself and see if I can have you fill in the missing pieces. Ask them to answer the questions. Questions for Everyone . Some parents explained to me that they think some types of schools are better than other types of schools. I have these little stickies with a bunch of different types of schools. I’ll give you examples of each kind so that we both are on the same page about what I mean by a certain type of school. [Do relevant examples] Could you rank them for me and put the best one at the top of the page and the others lower on the page wherever you think they should go? 217 6. 10. Ok, now, I’m just going to go down your list and will you tell me which types of schools you feel are real or legitimate options for your child? [Follow up with “why?” for ones that aren’t real options] You mentioned that you have gone up to the school for (conferences, _etc.). I’m trying to understand all the different ways parents are involved in their children’s lives. Could you describe your involvement in (your child’s) education? The last big thing I learned about from the other parents, and this may sound really dumb to you because you are a parent, is that some parents think about their children in terms of who they are as a person not only who they are as a student. That made me realize that I hadn’t really asked parents about who they think their child is. So I was hoping you could just describe (your child) to me as a person? 0 What are they good at? Bad at? o How have they changed over the years? 0 How do they compare to their fiiends? Do they like the same things? 0 Do you consider them typical? We’ve talked about what you would like for (your child) in terms of the future. Now I’m wondering what that might look like. Where do you think he’ll be when he’s 23 or 24 (20 or 21)? What will he be doing? What will have happened to him during those years? Pick pseudonym Now the last thing I need to have you do today is to pick a fake name for you, your spouse and the kids. It will be the name I use if I need to quote something you said in the dissertation. 218 l 1. Do you have any concerns or anything you are wondering about the study or me or what I’m up to? Eighth grader Interview 1. Your mom has been telling me about her thoughts about you going to high school. I was hoping I could get your perspective on school next year. Where do you want to go to school? Probes I Why do you want to go there? I Are your friends going there? I Was there anywhere else you thought about going? I Do you feel like you have a choice about where you go to school? Why or why not? I Is __ (school) your first choice? I If I gave you a million dollars today and I said it could only be used on school, what school would you choose to go to? Why? 2. I’m trying to understand how kids think about school and what they want to be when they get older. What jobs can you see yourself doing when you get older? Probes: I Do you think school will help you get ready to be a (occupation)? I How do you think it will help? 219 OR I Why don’t you think it will help you? 3. This is a little bit of a hard question but I’ll give it a try. In your opinion, what if anything, does school teach you? 4. Are you excited about anything related to school next year? Probe: I What things and why are you excited about them? 5. Are you worried about anything related to school next year? Probe: I What things and why are you worried about them? 6. Is there anything else I haven’t asked about that is important to you? Other things that are on your mind as you get ready to go to high school? Interview Three Say hello, ask them how everyone is doing, explain that you are calling to do our very last conversation. Is now a good time or should we set up another time? It will take about 15 minutes. ASK ABOUT RECORDING 220 [If they picked the school b/t June and now. . ..] a. how did they find out about it? b. What did they like about it? c. What were their concerns about it? (1. Why didn’t they end up choosing x, y, 2 schools we talked about? The Update 2. How are things going for [child]? a. Does s/he like it? Seem to be adjusting? b. I don’t know if this is too early in the year to be asking or not but. . .Have you had much contact with the school—like meeting the teacher? Seeing the classroom? Going to a school event? 3. Overall, do you feel like X school is a good place for [child]? How do you know? What would tell you if it wasn’t working out? 4. Last time we talked, you were concerned about x, y, 2. How have things been going with respect to those things? Old Questions 5. When I was typing up our interview from this summer I realized that I didn’t ask you a couple of questions I asked the other parents. If it’s ok, can I ask you those now? I Ask questions on the bottom of the second write-up 6. Ask everyone about the organized activities (read them what you have). What will they do this year? 221 New Questions Now I just have two more questions and then we are done. . .. 7. I am been reading a lot of books recently and they made me think about your experiences in school. I was hoping that you might tell me about what school was like for you—did you like it? did you feel like you were a good student? Good test taker? What were you like in class? a. Are there any parts of your experience you want repeated for [child]? b. Are there any parts of your experience that you hope [child] can avoid? 8. What do test scores tell you about a child? About an education? That’s it for me. Do you have any questions or concerns at this point? I’ll be sending them some things throughout the year-«Double check the address. Thank them again. 222 one» 505m ms 220 one» .2 2:: ms 220 3an fiwacfiflm 39.8... .3 BED 0292.5 8.295 amfieksxe 6353M u 58%.... 8968» 30:8 .3: eozaozvm .«o 733 235 8282332 .838 we 89¢. 05% 3900. 328 223 3305a 2: v.85 .mEx 08.3. .8320 Bow e :85 mess 888.on 8 cm ovum am «a £080 .8 $3 .805 :63 350.5 .520 8:3 33:38 3 Scan 28: 323% 33966.45 Soazox .3: 083.5 Haggai ...ow 2 25... as a as 03 wfitow 33 on? v.33 2C. 3% 09.0233 w§6- .6\ 336m. .2.an 3983‘ :3 no 93 .5: as. was? .605 2: a: 30:8. Bostonian 05 83 33.6% hmamaeoaQ gomwzom 033.5 £95 82 -cooawaz ska.“ Negev. _ _ _ _ _ soon om sou—emu con—.83 323 com “mom N70 MVWVOFM v5 «:55 224 00m $05 _ N— . : _ S 28:8 32930 E: Basses 0 800m 0 003 03 0003.5 05 sow 000m am 80.0 0.03 00003 0.005: _ 0 _ n _ 0 _ n 820:3. 38 .5 4 0335 008—0 : ~8an 0:03:05 comN m 2:8 505 0003.5 cm 205 230 00: 36:8 «009. 05 a: 0:05 a 005 _ N . _ :3 no 050 0:29—05 000m em .ow 2 08: 60:00. 0:95 cow—.80 x to: 02.. has; 39:0 as a .05 08.03.2022. 08.02.2362 0:0 000000 08 wSuow 05 003 05 a: .3 30m 003 030 $20 2:. v5 «40%: 330% 0&5 6000238 @033 .6\ 330% 08mm 33.0% 60.503: 30.03.0qu 323% .255 3080K hmxmguflQ 32.0w one» .2 78¢ «a 230 225 0350 00m 0:03:03 000m am 00035 303% 0&5 303% 602.030 0003005 32.0w 008m SWEm 00m 085 N _ _ 5 03 0 m N. o n v m m _ v— v5 nt 220 226 Appendix D Data Analysis Categories for Interviews Interview One INTERVIEWED: CHILDREN: CURRENT GRADE: MARITAL STATUS: INCOME: RACE: MOM’S AGE: MOM’S OCCUPATION: MOM’S EDUCATION: DAD’S AGE: DAD’S OCCUPATION: DAD’S EDUCATION: CURRENT SITUTION’S RELATIONSHIP TO FUTURE SITUTION: SCHOOLS THAT ARE OR WERE CONSIDERED: SCHOOLS THAT WERE NOT CONSIDERED: INFORMATION: TRANSPORTATION: WHERE ARE THEY IN THE DECISION?: IDEA OF A “GOOD” EDUCATION: WHEN THEY STARTEDTHIN KING ABOUT THIS: 227 CHOICE FOR THIS CHILD WITH RESPECT TO THE OTHER CHILDREN: PEOPLE PARENT TALKS TO: FACTORS/ISSUES IN THE CHOICE: VIEWS ABOUT SCHOOL GENERALLY: VIEWS ABOUT CERTAIN KINDS OF SCHOOLS: ARE OR ARE NOT WILLING TO DO: RACE: ROLE OF THE CHILD: ROLE OF THE PARENT: RANDOM FACTS: HUNCHES: QUOTATIONS: OVERLOOKED QUESTIONS: Interview T wo HAVE CHOICE?: NOTES ABOUT RANKINGS: TYPICAL KID?: TYPICAL PARENT?: DESCRIPTION OF CHILD: CHILD IN 10 YEARS: PARENTAL INVOLVEMENT: CHILD OR THE SCHOOL?: 228 NOTES TO MYSELF: REMAINING QUESTIONS: QUOTATIONS: Interview Three LAST YEAR’S ACTIVITIES: THIS YEAR’S ACTIVITIES: SCHOOL SCORES: CHILD SCORES: PARENT’S EDUCATION EXPERIENCE: REPETED FOR CHILD: AVOIDED FOR CHILD: NOTES TO MYSELF: QUOTATIONS: 229 Appendix E Example of First Interview Summary Mrs. Kerry Mrs. Kerry has three children, Tim (11th), Breanna (8th), Barry (5th). Tim is at Protestant High, Breanna is at St. Bede and so is Barry. All the kids are in private school, though early on Tim and Breanna attended the neighborhood public school. Kerry moved all three of her kids to St. Bede at the same time. Mrs. Kerry explained that “Basically, I moved here and I saw the school [St. Bede] and I liked the way it looked. . .everyone around here being that it’s Gaston, mostly had their children in private schools, mostly Catholic. So I didn’t do any investigations, I just knew that that was something I would probably do if I stayed here.” Kerry is the primary person who makes decisions about schoolng for the kids. She views it as her job to work with the teachers and get along. She explained that “teachers know I wanna know what is going on”. She said that they will “call here at 10 o’clock at night” but that she feels very lucky because she really has good kids. She doesn’t feel that she has to deal with a lot of the interactions with schools that some other parents do (because their children are in trouble a lot). The oldest child, Tim, goes to school at Protestant High. Kerry has been extremely happy with Protestant High. Kerry plans for Breanna to attend Protestant High like her brother. Kerry did not consider any other schools, she always knew Breanna would go to Protestant High. Breanna does many activities including, jazz and tap, soccer, cheerleading, volleyball, and track. Mrs. Kerry is unsure how Protestant High will work out for two reasons. First, there is generally declining enrollment and she is 230 wondering what kind of effect that will have on the school. Second, she sees Protestant High as very structured. She is worried that all of the rules will be difficult for Breanna. She explained that Breanna does not have the same king of self-restraint that Tim does. Mrs. Kerry explained that sometimes she wonders about sending them to Protestant High because all of her kids are really sports oriented. She said that Tim is good enough that she wonders if being at Protestant High holds him back. She feels Protestant High is smaller and because it doesn’t have the most winning teams, She wonders if being at a better sports school would expose him to more college scouts. Kerry likes Protestant High for its small class size, community feel, the types of kids that are there, the principal, hard classes, and the way kids treat each other. The issue of being at a school that would give Tim more exposure to scouts is not as important as the community and academic aspects of Protestant High. Kerry explains that she really loves the “kind of community aspect of things” and She loves the classes he is taking and the things he is learning (“The stuff that he is learning just blows me away.”). She also believes that the church, the school, and the families all sort of “hang together”. She explained that even if she and her husband were to die tomorrow, people in their church and school community would take care of her kids for her. She loves this. She explained, “It’s like I might live in the city, in Gaston but where my children go to school and where they go to church are in their own little community. And that’s the way it makes me feel. SO it gives me that sense that they’re completely taken care of. I don’t think I could give them any more than what they’ve got if I had more money. I’m very happy with the whole thing.” 231 Mrs. Kerry values education and expects all of her children to get an education. She explained, “I know how hard it is out there. Me and his father, we didn’t even finish high school and I know how hard it is out here. ...You’re gonna go to college cause you know, it is very difficult. . ..I believe if you give yourself every opportunity, it will pay off for you. I guess if we made it rich overnight I might feel different but I really feel that no matter what you do, you have to get an education. That is just my number one obsessive goal. And I constantly stay on them about that.” 232 Appendix F The Excluded Variables There are four variables the reader might be left contemplating. SO that my thinking may be made visible, I will now share the reasoning for the exclusion of these variables. The variables sibling ’s school status and distance to previous school have substantively important, statistically significant relationships with final school status. The variables knowledge of child 's homework and parental volunteerism do not have substantively important or statistically significant relationships with final school status however, they are an important part of the stated preference literature. We turn now to the reasoning behind the omission of these four variables. Sibling ’s School Status In order to take account of the role of siblings there were two variables which could have been included in the model, sibling’s school status and sibling’s school selectivity. Sibling’s school status is the status (failing or non-failing) of the Older sibling’s school.“ Sibling’s school status has a strong relationship with final school status (x2 = 6.90, p < .01, (p = .48). But it also was strongly correlated with previous school, suggesting the two are measuring similar constructs. Sibling’s school status is, afler all, a variable Similar to previous school status. Both reflect parents’ prior school choices. Thus to produce a parsimonious model, previous school status was included because it captured parents’ preferences for the target child. ‘4 For sibling’s school status and sibling’s school selectivity, if there was not an older sibling, the closest younger srbling’s school was used. If there was not a school age (K-12) sibling in the family, the family was dropped from the analysis. 233 Distance to Previous School There has been a great deal of discussion in the literature about the role of convenience in parental choice. Scholars have reported stated preference data which suggests parents prefer schools that are close to their homes (e.g., Wells, 1996). The Gaston parents were no different. In interviews parents talked about wanting to find schools that were not too far away or “on the other side of the city.” There was however, only a modest correlation between distance to school and final school status (er = 0.29, p = .057) that did not achieve statistical significance. In the city group of parents, distance had a slightly stronger relationship to final school status which was statistically significant (er = 0.33, p < .05). The range of the distance variable was checked to determine if outliers were underestimating the strength of the relationship. There was one outlier, a parent who drove almost 27 miles to take her daughter to a high school in the suburbs. But that school was non-failing so if anything, that outlier made the relationship appear stronger than it actually is. Though convenience matters to parents it is possible that it has little bearing on ending up in a failing school. With the higher density of schools in urban areas, perhaps parents are equally likely to have both failing and non-failing school close to them. It is also possible that distance seems relatively unrelated to final school status because of the group of parents we are dealing with. For parents with fewer income and transportation resources who are less engaged in their child’s education, distance may matter a great deal."3 We do not however, see that with the parents in this study. ‘5 Distance likely matters in the decision between two particular schools. This is not however, what is being predicted in this model. 234 Knowledge of Child ’3 Homework and Parental Volunteerism Other scholars have documented differential levels of parental involvement across parents (Clark, 1983; Epstein, 2001; Godwin et al., 1998). It makes sense that parents who are more involved in their child’s education will have more knowledge of their child’s needs and be more capable of choosing a school in which the child can be successful. It was surprising that neither the knowledge of a child’s homework or the amount of time a parent Spent volunteering in her child’s school had a substantively important relationship with the Odds of winding up in a non-failing school, (x‘ = 1.22, p = .27, (p = .17, OR = 1.98 and x‘= 1.22, p = .32, (p = .09, OR = 1.41, respectively). It is possible that the high levels of parental involvement among Gaston parents resulted in so little variance, a pattern is undetectable in a sample of 48 people. It is also possible, and perhaps more likely, that knowledge of homework and parental volunteering are not particularly good proxies for the kind of preferences and skills that might be required for choosing a non-failing school. It is quite possible that other model variables (such as school visits and sibling’s school selectivity) better capture the type of parental interest that is necessary to choosing a non-failing school. The lack of a substantive or statistically significant relationship together with questions about the ability of these variables to capture the characteristics of interest led me to exclude these two variables from the model, despite their prevalence in the state preference literature. There are other variables that the reader might be interested in but for the sake of brevity, I will conclude my explanation of the omission of variables. A correlation matrix for the use of the curious reader is in Appendix A. 235 Appendix G Description of Model Variables Variable Description Scale Income Education Potential set size Cohort consensus School visits Parents’ years gross income. Self-reported. Low < $59,399. High > $59,400. Primary parent’s highest level of educational attainment. If there was no “primary” parent, the highest level of education between the two parents was used. Self reported. Less = any amount less than a college or technical degree. More = any amount including a college degree and more. The number of grade appropriate schools within two miles of primary parent’s home. Using Arcview and the state’s list of private, religious, charter, magnet, and neighborhood schools. Researcher calculated. Range = 1-24. In a given feeder school, the percentage of non-failing schools in the set of schools that were in more than half of the parents’ choice sets. Researcher calculated. Range = 0-1. The percentage of schools in the choice set the parent visited. Visiting included a range of behaviors—attending an open house, going on a pre-ananged tour of the school, and “looking around” during a visit in which the parent also dropped off the application. Researcher calculated. Range = 0-1. 236 Binary Binary Continuous Continuous Continuous Previous school status Choice set status Choice set size Sibling’s school’s selectivity Status of the target child’s previous school. The school was Binary non-failing if it was not on the State’s failing list, had > 45% of the students at grade level on the state exam (reading and math combined), or was accredited (self-reported). The school was failing if it was on the list, had < 45% of students at grade level on the state test or it was unaccredited.‘6 0 = Failing. 1 = Non- failing. The percentage of schools in the choice set that are “non- Continuous failing.” Researcher calculated. Range = 0-1. The number of schools the parent mentioned to the researcher. Continuous This includes schools the parent eventually did not choose. Researcher calculated. Range = 1-12. The next oldest sibling’s school’s selectivity. If the sibling’s Binary school was a magnet, private, religious, or homeschool, it was considered selective. If there was not an older school age (K- 12) sibling, the next youngest sibling was used. If there were no siblings, the family was not included in these analyses. Self-reported. 0 = Failing. l = Non-failing. ’6 The state’s list of failing schools changed three times in as many years. The list that was used was the one available to parents during the fall and winter of data collection (2002—2003). Very few of the choice set schools were caught in the controversy. Across the three different lists, all but one school stayed in the category to which they were originally assigned. 237 Race Primary parent’s reported race divided into “people of color” Binary (African American, Latino, “mixed” and “other”) and “white” (European American). 0 = People of color. 1 = White. 238 Appendix H Study Participants Parent’s Name Children’s Name School Parent’s Primary choosing Race Parent Natalie Bear Nema Middle African- Mother American Nikita & Michael Bond James High White Mother Emily & Derrick Borden Clyde & Johnny Middle White Mother Katerina & John Bounder Ella & George High White Mother Connie and Barry Brown Sebastin Johnson Middle Afi'ican- Mother American Diane Carol Denzel & Jennifer High African- Mother American Susan Cartwright Carla, James, Jessica, Donald, Middle Mixed Mother/ Sheryl, & Sarah Aunt Lisandra & Larry Cats Latoya, Alexis, Lydia, Justin, High African- Mother Kobe, Tiffany, & Netta American Ernestine Conwell Jimmy & Monique High African- Mother American Rochelle Davis Antoine, Sam, & Dave High Afiican- Mother American Dana and Dan Denver David & Darlene High White Mother Roger & Tara Dish Tasha, Janet, & Terrell Middle White Father Sasha Dorsey Spirit & Austin Middle African- Mother American Amelia & Lance Erhardt Faye, John, Joe, Rosie, Anise High White Both Lottie & Mark Feagin Michaelangelo & Ben Middle African- Mother 239 *_ _._._' ads-— Shelly Fields Megan Foster Diamond & Richard Gooden Leola Gunnison Liyah & Jacob Hawill Slick & Catherine Holley Charles and Louise Jones Robert Jones Barbara & Charles Jordan Diamontina Martinez Ann & Frank Moore Dalia Morales Sandra Morgan Sara Ronald & Alex Shalanda Jewel, Doug, & Randy Alecia & Brandon Andrea, Soujomer, & J .Lee Martin Orlando & Kelly Beverly Ashley Rose, Clifton, Cassandera, & Dushawn Erika Mike, Julie, & Scott Maria Grisson & Eugene Norman Stacey Norman, Shawn Susan and Sam Parker Grisson, Tracy Norman, Charles Norman, & Melissa Norman Sam Jr., Seth, Sara, Serena, 240 Middle Middle Middle Middle High Middle Middle High Middle Middle High Middle High High Elem/ American African- American African- American African- American African- American African- American African- American White African- American African- American Hispanic Afi'ican- American Hispanic White African- American White Mother Mother Mother Mother Mother Grandfather Mother Father Mother Mother Mother Mother Mother Mother Mother Renee & David Robinson Kaye Shirley Lynn & Maurice Smith Joyce & Jerry Smith Bree & Jack Smothers Emma Springfield Carolyn & George Steward Mary and James Thomas Dexter and Shelese Walker Lillian Watson Susan Webb Teresa & Tom Wembez Chiffon West Lynn & John Whitaker Chester & Clara Williams Sally Raven & Robert Breara, Breanna, Brent & Brandon Jason & Amber Dylan & Anna Barbie, Bruce Aleah, India, & Brianna Frank Washington & Chaquira Wilson Mark & Steven Brandon & Natalie Shaniqua Alexandra Sylvia, Isabelle, & Samantha Jerry & Suzanne Elizabeth, Becky, Christina, & Peter Josephine, Jonathan, & 241 Middle Middle Middle High Middle Middle Middle High High Middle Middle High Middle High Middle Middle African- American African- American African- American White African- American African- American African- American African- American African- American African- American African- American White African- American White African- Mother Mother Mother Mother Mother Mother/ Aunt Mother Mother Both Mother Mother Mother Mother Mother Father Sabrina & Jonathan Winslow Delores Wise Kathy and Scott Wolf Leonard Richard, Ronald, & Ross Seymour Jackson Scott & Alexander 242 American Middle Afi’ican- Mother American Middle African- Mother American High White Both REFERENCES Allison, G., & Zelikow, P. (1999). Essence of decision: Explaining the Cuban missile crisis. New York: Addison-Wesley. Anderson, C. (2003). Teaching science for motivation and understanding. Unpublished manuscript, East Lansing. Armor, D., & Peiser, B. (1997). Competition in education: A case study of interdistrict choice. Boston: Pioneer Institute. Arsen, D., Plank, D., & Sykes, G. (1999). School choice policies in Michigan: The rules matter. East Lansing, MI: The Education Policy Center at Michigan State University. Asimov, N. (2003, October 9). Few parents seize chance to transfer schools 'No child left behind' made offer mandatory. San Francisco Chronicle. Bast, J. L., & Walberg, HI. (2004). Can parents choose the best schools for their children? Economics of Education Review, 23, 431-440. Beales, J. R., & Wahl, M. (1995). Private vouchers in Milwaukee: The PAVE program. In T. Moe (Ed.), Private Vouchers. Stanford, CA: Hoover Institution Press. Blank, R. K., Levine, R.E., & Steel, L. (1996). After 15 years: Magnet schools in urban education. In B. Fuller, Elmore, R.F. & Orfield, G. (Eds), Who chooses? who loses? .' Culture, institutions, and the unequal eflects of school choice. New York: Teachers College Record. Bogdan, R. C., & Bilden, SK. (1992). Qualitative research for education: An introduction to theory and methods (2nd ed.). Needham Heights, MA: Allyn and Bacon. Boudon, R. (1973). Education, opportunity and social inequality. New York: I. Wiley & Sons Bourdieu, P. (1984). Distinction: A social critique of the judgment of taste (R. Nice, Trans). Cambridge, MA: Harvard University Press. Bourdieu, P. (2000). The aesthetic sense as the sense of distinction. In J. B. Schor, & Holt, D.B. (Eds), The consumer society reader (pp. 205-211). New York: The New Press. Bowles, S. (1998). Endogenous preferences: The cultural consequences of markets and other economic institutions. Journal of Economic Literature, 36(1), 75-111. 243 Bowles, S., & Gintis, H. (1976). Schooling in capitalist America. New York, NY: Basic Books, Inc. Bridge, G. (1995). Gentrification, class and community: A social network approach. In A. Rogers, & Vertovec, S. (Eds), The urban context: Ethnicity, social networks and situational analysis. Washington, DC: Berg Publishers. Brown, J. S., & Durgid, P. (2000). The social lifi: of information. Cambridge, MA: Harvard Business School Press. Bruner, J. S. (1990). Acts of meaning. Boston, MA: Harvard University Press. Buckley, J ., & Schneider, M. (2003). Shopping for schools: How do marginal consumers gather information about schools? The Policy Studies Journal, 31(2), 121-145. Cameron, T. (1992). Combining contingent valuation and travel cost data for the valuation of nonmarket goods. Land Economics, 68, 69-75. Campbell, D. T., & Stanley, J.C. (1963). Experimental and quasi-experimental designs for research. Boston, MA: Houghton Mifflin Company. Charles, C. Z. (2001). Processes of racial residential segregation. In A. O'Connor, Tilly, C., & Bobo, L. (Eds), Urban inequality: Evidence fiom four cities. New York: Russell Sage Foundation. Chubb, J. E., & Moe, TM. (1990). Politics, markets, and America 's schools. Washington, DC: Brookings Institution. Clark, R. M. (1983). Family life and school achievement: Why poor black children succeed or fail. Chicago: The University Of Chicago Press. Cohen, D. K. (1988). Teaching practice: Plus ca change. In P. W. Jackson (Ed.), Contributing to educational change: Perspectives on research and practice (pp. 27-84). Berkeley, CA: McCutchan. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbatun. Cohn, E. (Ed.). (1997). Market approaches to education: Vouchers and school choice. Tarrytown, NY: Elsevier Science Inc. Coleman, J. S. (1990). Foundations of social theory. Boston: Harvard University Press. Coons, J. E., & Sugarman, SD. (1978). Education by choice: The case for family control. London: University of California Press. 244 Cusick, P. A. (1992). The educational system: Its nature and logic. New York: McGraw- Hill, Inc. David, M., West, A., & Ribbens, J. (1994). Mother's intuition? Choosing secondary schools. London: The F almer Press. Dewey, J. (1902/1990). The child and the curriculum. Chicago: The University of Chicago Press. Dillman, D. A. (2000). Mail and internet surveys: The tailored design method (2nd ed.). New York: J. Wiley. Edin, K., & Lein, L. (1997). Making ends meet: How single mothers survive welfare and low-wage work. New York, NY: Russell Sage Foundation. The Education Trust (2004). Afiican American achievement in America. Retrieved June 1, 2004, from The Education Trust's website: http://www2.edtrust.org Epstein, J. L. (2001). School, family, and community partnerships: Preparing educators and improving schools. Boulder, CO: Westview Press. Erickson, F. (1986). Qualitative methods in research on teaching. In M. Wittrock (Ed.), Handbook of research on teaching (3rd ed., pp. 119-161). Washington, DC: American Educational Research Association. Figlio, D. N., & Lucas, ME. (2000). What's in a grade? School report cards and house prices. Unpublished manuscript, Cambridge, MA. Fiske, E. B., & Ladd, H. F. (2000). When schools compete: A cautionary tale. Washington, DC: Brookings Institute Press. Fowler, F. J. (1995). Improving survey questions: Design and evaluation. Thousand Oaks, CA: Sage. Friedman, M. (1962). Capitalism and fieedom. Chicago: The University Of Chicago Press. Friedman, M., & Friedman, R. (1980). Free to choose: A personal statement. New York: Avon Books. Fuller, B., Elmore, R.F. & Orfield, G. (1996). Who chooses? Who looses?: Culture, institutions, and the unequal efibcts of school choice. New York: Teachers College Press. 245 Fuller, B., Holloway, S.D., Rambaud, M., Eggers-Pierola, C. (1996). How do mothers choose child care? Alternative cultural models in poor neighborhoods. Sociology of Education, 69(2), 83-104. Gauri, V. (1998). School choice in Chile: Two decades of educational reform. Pittsburgh, PA: The University of Pittsburgh Press. Geertz, C. (1983). Local knowledge: Further essays in interpretive anthropology. New York: Basic Books, Inc. Gelpi, C. (2004). Coping with Endogeneity. Retrieved July 18, 2004, from the course website: http://www.dee.edu/~gelpi/p8233 .lecture21 .DDt Gewirtz, 8., Ball, S.J., & Bowe, R. (1995). Markets, choice, and equity in education. Buckingham, England: Open University Press. Gill, B. P., Timpane, P.M., Ross, K.E., & Brewer, DJ. (2001). Rhetoric versus reality: What we know and what we need to know about vouchers and charter schools. Santa Monica, CA: Rand. Gilligan, C. (1982). In a di/fkrent voice: Psychological theory and women 's development. Cambridge, MA: Harvard University Press. Glaser, B., & Strauss, A. (1979). The discovery of grounded theory: Strategies for qualitative research. Hawthorne, NY: Aldine. Glatter, R., Woods, P., & Bagley, C. (1997). Diversity, differentiation and hierarchy: School choice and parental preferences. In R. Glatter, Woods, P., & Bagley, C. (Eds), Choice and diversity in schooling: Perspectives and prospects (pp. 7-28). London: Routledge. Godwin, R. K., Kemerer, F.R., & Martinez, V.J. (1998). Comparing public choice and private voucher programs in San Antonio. In P. E. Peterson, & Hassel, B.C. (Eds), Learning from school choice. Washington, DC: Brookings Institution Press. Grench, D. (2002, November 3). Many reject vouchers, return to public schools The Miami Herald. Gupta, R. (2004, July 20). Few take advantage of options under US. school law. Palm Beach Post. Henig, J. R. (1994). Rethinking school choice: Limits of the market metaphor. Princeton, NJ: Princeton University Press. 246 Henig, J. R. (1996). The local dynamics of choice: Ethnic preferences and institutional responses. In B. Fuller, Elmore, R.F., & Orfield, G. (Eds), Who chooses? who loses?: Culture, institutions, and the unequal efibcts of school choice. New York: Teachers College Press. Hensher, D., Louviere, J., & Swait, J. (1998). Combining sources of preference data. Journal of Econometrics, 89, 197-221. Herriges, J. A., Kling, C.L., & Azevedo, C. (1999). Linking revealed and stated preferences to test external validity. Unpublished manuscript, Ames, IA. Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms organizations and states. Cambridge, MA: Harvard University Press. Hohne, J. J. (2002). Buying homes, buying schools: School choice and the social construction of school quality. Harvard Educational Review, 72(2), 177-207. Hossler, D., & Gallagher, KS. (1987). Studyng student college choice: A three-phase model and the implications for policymakers. College & University, 62(4), 207- 221. Howell, W. G., & Peterson, PE. (2000). School choice in Dayton, Ohio: An evaluation after one year. Cambridge, MA: Program on Education Policy and Governance, Harvard University. Jackson, G. (1982). Public efficiency and private choice in higher education. Educational Evaluation and Policy Analysis, 4(2), 237-247. Kozol, J. (1991). Savage inequalities: Children in America 's schools. New York: Crown. Lareau, A. (1989). Home advantage: Social class and parental intervention in elementary education. Philadelphia: F almer Press. Lee, V. E., Croninger, R.G., & Smith, J .B. (1996). Equity and choice in Detroit. In B. Fuller, Elmore, R.F., & Orfield, G. (Eds), Who chooses? who loses?: Culture, institutions, and the unequal eflects of school choice. New York: Teachers College Press. Leman, K. (1998). The new birth order book: Why you are the way you are. New York: Revell Publishing. Litten, L. (1982). Different strokes in the applicant pool: Some refinements in a model of student choice. Journal of Higher Education, 4, 383-402. Loureiro, M. S., McCluskey, J .J ., & Mittelhammer, RC. (2003). Are stated preferences good predictors of market behavior? Land Economics, 79(1), 44-55. 247 Loury, G. C. (2002). The anatomy of racial inequality. Cambridge, MA: Harvard University Press. MacLeod, J. (1995). Ain't no makin' it: Aspirations and attainment in a low-income neighborhood. Boulder, CO: Westview Press, Inc. Massey, D. S., & Denton, NA. (1993). American apartheid: Segregation and the making of the underclass. Boston, MA: Harvard University Press. Maxwell, J. A. (2004). Causal explanation, qualitative research, and scientific inquiry in education. Educational Researcher, 33(2), 3-11. McDonough, P. M. (1997). Choosing colleges: How social class and schools structure opportunity. Albany, NY: State University of New York Press. Metcalf, K. K. (1999). Evaluation of the Cleveland scholarship and tutoring grant program: 1 996-1 999. Bloomington, IN: The Indiana Center for Evaluation. Meyers, D., Peterson, P., Mayer, D., Chou, J ., & Howell, W.G. (2000). School choice in New York City afier two years: An evaluation of the School Choice Scholarship Program. Washington, DC: Mathematica Policy Research, Inc. Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook (2nd ed.). Thousand Oaks, CA: Sage. Moe, T. M. (Ed.). (1995). Private Vouchers. Stanford, CA: Hoover Institution Press. Moses, A. R. (2004, February 4). Few use school transfer rule. The Detroit News. NCES. (2003). Trends in the use of school choice 1993 to 1999. Washington, DC: US. Department of Education. Newman, S. (2001). Parenting an only child: The joys and challenges of raising your one and only. New York: Broadway Books. Ni, Y., & Donahue, T. (2004). Social stratification through inter-district choice: The Michigan perspective. Paper presented at the annual meeting of the American Educational Research Association, San Diego, CA. NO Child Left Behind Act Of 2001, Pub. L. No. 107-110, 115 Stat. 1425 (2002). O'Connor, A., Tilly, C., & Bobo, L.D. (Eds). (2001). Urban inequality: Evidence fiom four cities. New York: Russell Sage Foundation. 248 Okey, T. N. (1990). The family perspective on the individual ’s decision to drop out of high school. Unpublished Dissertation, Michigan State University. Orfield, G., Losen, D., Wald, J ., & Swanson, C. (2004). Losing our future: How minority youth are being lefi behind by the graduation rate crisis. Cambridge, MA: The Civil Rights Project at Harvard University. Orfield, G. A., & Yunn, J .T. (1999). Resegregation in American schools. Cambridge, MA: The Civil Rights Project, Harvard University. Paine, T. (1915). The Rights of Man. New York: E.P.Dutton. Peterson, P. E. (1998). An evaluation of the New York City School Choice Scholarships Program: The first year. Cambridge, MA: Program on Education Policy and Governance, Harvard University. Peterson, P. E., Myers, D., & Howell, W.G. (1999). An evaluation of the Horizon Scholarship Program in the Edgewood Independent School District, San Antonio, Texas: The first year. Cambridge, MA: Mathematic Policy Research and Program on Education Policy and Governance, Harvard University. Plank, S., Schiller, K.S., Schneider, B., & Coleman, J .S. (1993). Effects of choice in education. In E. Rasell, & Rothstein, R. (Eds), School choice: Examining the evidence (pp. 111-134). Washington, DC: Economic Policy Institute. Pleasants, H. M. (2000). Defining and seeking a "good education ": A qualitative stuay of black parents who select charter schools for their children. Unpublished Dissertation, Michigan State University, East Lansing. Quinn, N., & Holland, D. (1987). Culture and cognition. In D. Holland, & Quinn, N. (Eds), Cultural models in language and thought (pp. 3-40). Cambridge, England: Cambridge University Press. Rasell, E., & Rothstein, R. (Eds). (1993). School choice: Examining the evidence. Washington, DC: Economic Policy Institute. Robelen, E. (2002, August 7). Few choosing public school choice for this fall. Education Week. Rose, L. C., & Gallup, A.M. (2003). The 35th annual Phi Delta Kappa/Gallup poll of the public's attitudes toward the public schools. Phi Delta Kappa. Rubinowitz, L. S., & Rosenbaum, J .E. (2000). Crossing the class and color lines: From public housing to white suburbia. Chicago: The University of Chicago Press. 249 Sanders, W. L., & Horn, S. (1994). The Tennessee Value-Added Assessment System (TVAAS): Mixed-model methodology in educational assessment. Journal of Personnel Evaluation in Education, 8, 299-311. Schemo, K. (2002, August 28). Few exercise new right to leave failing schools. New York Times. Schneider, M., Teske, P., & Marschall, M. (2000). Choosing schools: Consumer choice and the quality of American schools. Princeton, NJ: Princeton University Press. Schneider, M., Teske, P., Roch, C., & Marschall, M. (1997). Networks to nowhere: Segregation and stratification in networks of information about schools. American Journal of Political Science, 41, 1201-1223. Simpson, J. A., & Weiner, E.S. (Eds). (1991). The Compact Oxford English Dictionary (2nd ed.). Oxford: Clarendon Press. Stier, H., & Tienda, M. (2001). The color of opportunity: Pathways to family, welfare. and work. Chicago, II: The University of Chicago Press. Tufie, E. R. (1983). The visual display of quantitative information. Cheshire, CT: Graphics Press. Tufte, E. R. (1990). Envisioning information. Cheshire, CT: Graphics Press. Tufte, E. R. (1997). Visual explanations: Images and quantities, evidence and narrative. Cheshire, CT: Graphics Press. Tyack, D. B. (1976). Ways of seeing: An essay on the history of compulsory schooling. Harvard Educational Review, 46(3), 355-389. Walker, V. S. (1999). Culture and commitment: Challenges for the future training of education researchers. In E.C. Lagemann, & Shulman, L.S. (Eds), Issues in education research: Problems and possibilities. San Francisco: Jossey-Bass. Wells, A. S. (1993). Time to choose: America at the crossroads of school choice policy. New York, NY: Hill and Wang. Wells, A. S. (1996). African-American students' view of school choice. In B. Fuller, Elmore, R.F. & Orfield, G. (Eds), Who chooses? who looses?: C ulture, institutions, and the unequal eflects of school choice. New York: Teachers College Press. Williams, E. R. (2004). Lifting every voice: African American parental perceptions of school efficacy. Paper presented at the annual meeting of the American Educational Research Association, San Diego, CA. 250 Willis, P. (1977). Learning to labor: How working class kids get working class jobs. New York: Columbia University Press. Wilhns, J. D., & Echols, EH. (1993). The Scottish experience of parental school choice. In E. Rasell, & Rothstein, R. (Eds), School choice: Examining the evidence (pp. 49-68). Washington, DC: Economic Policy Institute. Witte, J. F. (2000). The market approach to education: An analysis of A merica 's first voucher program. Princeton, NJ: Princeton University Press. Woods, P. A., Bagley, C., & Glatter, R. (1998). School choice and competition: Markets in the public interest? London: Routledge. 251 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII llullllllllllllllllll