School choice programs and their scrappy offspring, charter schools, get good press. School choicegives parents and students the option of choosing a public school beyond neighborhood boundaries. Charterschools, formed by parents or teachers who want to create additional options for school choice, are said toproduce high test scores, enthusiastic teachers, satisfied parents, and long waiting lists. Since the early1990s, 37 states have enacted charter school legislation. Nearly 500,000 students attend charter schools.U.S. President George W. Bush and state governors proclaim that charter schools and other school choiceoptions are the most promising approach to public school improvement. In Colorado, the state has approvedlegislation that mandates converting schools with low standardized test scores to charter schools unlessthe schools improve dramatically in two years. Despite this mandate, careful studies of the effects ofschool choice suggest that charters are not necessarily a cure for weak schools.
Proponents of public school choice contend that competition gives parents the power to vote withtheir feet. Schools that perform poorly will lose students and be forced to go out of business or toimprove (Chubb & Moe, 1990). School choice is supposed to promote equity by allowing all students,not just those whose parents have money, to leave failing schools. Critics of school choice argue thatcompetition pits school against school, destroying cooperation, dividing neighborhoods, and promoting inequityby increasing stratification by achievement, income, and ethnicity (Carnoy, 2000; Cobb & Glass, 1999).This controversy provided the general framework for our study of school choice in one district.
The Case of Boulder Valley
The hub of the Boulder Valley School District is Boulder, an affluent community of 96,000 people andhome to the main campus of the University of Colorado and to such high-tech corporations as IBM, SunMicrosystems, Ball Aerospace, and StorageTek. The median household income is $51,000, and the adultresidents are highly educated, with nearly 30 percent holding graduate or professional degrees. Nearly 80percent of students in the school district are white and middle class. The eastern and northern regionsof the district have some concentrations of minority students, mostly Latinos.
Open enrollment—the option for parents to send their children to any district school with spaceavailable after neighborhood children have enrolled—has been available in the district since 1961.Before the 1994–95 school year, however, only five schools provided such choice; they emphasizedbilingual or experiential education.
By 1999–2000, the number of choice schools had grown to 21 of the district's 57 schools.Responding to the demands of a group of parents who believed that the district schools did not paysufficient attention to academics, 8 of the 16 new choice schools emphasized academic rigor and collegepreparation; 5 of these 8 schools adopted the Core Knowledge curriculum. The other eight schools includedtwo charter schools, a Montessori school, a Waldorf school, and four focus schools committed to such themesas science or the arts.
Concerned about the rapid growth of choice schools and the rancor that school choice sometimesprovoked, the school district commissioned us to survey parents' and educators' beliefs and attitudesabout school choice and to analyze records of enrollment patterns, test scores, demographics, funding, andfund raising. Our findings provide more support for the critics of school choice than for its proponents.Choice has increased competition, and schools have struggled to keep enrollments strong by appealing to parents.Although some parents have found better schools for their children, the cost to the district has been high.
Historically, the quality of the district's schools has ranged from very good to excellent. Theexpansion of choice threatens this happy situation. More than 20 percent of students now enroll in schoolsother than their neighborhood schools, a percentage that is unusually high, especially when compared withArizona, the state with the largest charter school movement but with a mere 4 percent of its studentsenrolled in charter schools (Nelson et al., 2000). In many parts of the country, the effects of schoolchoice are diffuse because students move across as well as within districts, but for a variety ofreasons—geography, bonds to the community, and the school system's reputation—most studentsin Boulder Valley's school system stay within the district. As a result, we were able to examine gainsand losses within a district where schools compete for enrollment from the same pool of students.
Data for the Study
The district supplied us with records from the past six school years on open enrollment, test scores,demographics, funding, and fund raising in 55 of the district's 57 schools.
To collect data about attitudes toward school choice, we conducted surveys of principals bytelephone and then held focus group discussions and conducted written surveys with parents and educatorsactive in the district. The principals, parents, and educators totaled 466 individuals representing43 schools.
To ascertain the attitudes of the district's parents who had not participated in open enrollmentand were not active in schools, we called potential respondents selected at random from eight geographicalregions until we obtained 30 completed surveys from each region, for a total of 240 telephone surveys.
Respondents' Views of School Choice
Most of the more than 700 people whom we surveyed agreed that school choice was an effective means ofresponding to students' diverse interests and needs but that various inequities existed in the currentsystem. For example, almost all agreed that lack of transportation and information effectively reduced theopportunities to participate in school choice.
Respondents were divided on the scope, seriousness, and cause of the perceived inequities. Peoplein neighborhood and bilingual choice schools saw serious and direct outcomes of expanded school choice:increased inequities in available resources, the tendency of academically strong students to move tohigh-achieving schools (called skimming), and stratification by race and income as whitestudents chose schools outside of predominantly minority areas. Some individuals also complained about unfaircompetition between neighborhood and choice schools and about the extent to which the schools' need tomarket themselves detracted from their education mission.
By contrast, parents of students who attended the new choice schools that emphasized academicrigor and college preparation had few concerns about unfair competition or the market imperative.They pointed to the schools' high parent satisfaction ratings and believed that the new schoolswere spurring increased academic achievement. They also thought concerns about skimming and whiteflight were overblown or attributable to other causes, such as shifts in where students live.
The weight of the evidence from our study, however, is not on the side of these school choiceproponents. Although market competition appears to be working, creating the greatest demand for theschools with the highest test scores and parent satisfaction ratings, concerns about inequitiesassociated with skimming, stratification, unfair competition, and unequal resources are well founded.
Open Enrollment Patterns
Demand for schools. Parent satisfaction ratings and test scores—the Colorado SchoolAchievement Program tests for elementary and middle school students and the Terra Nova Comprehensive Test ofBasic Skills for high school students—were the two strongest reasons for parents demanding a particularschool. At the elementary level, the greater demand was for schools with the highest test scores; for middleschools, test scores and parent satisfaction ratings were equal; for high schools, parent satisfaction ratingshad the strongest association with demand.
The only notable exception to these preferences was the attitude of Latinos, who were apparentlyless motivated by test scores and satisfaction ratings than whites were or were willing to trade thesemeasures for bilingual programs (Howe & Eisenhart, 2000).
Giving parents the power to choose the curriculum and the methods of instruction that they deembest for their children seems to explain the pattern of demand for open enrollment, but, in fact, parentsatisfaction had a closer association with test scores than with curriculum.
Skimming. In general, students requesting open enrollment in 6th and 9th grade hadhigher test scores than their district cohorts and applied disproportionately to schools with higher testscores. Some schools drew a disproportionate number of students from the high-scoring pool—some schoolsdrew all their students from this pool—whereas other schools lost a disproportionate number of high-scoringstudents.
Stratification by race and income. Race and ethnicity were prominent features of openenrollment patterns. Students left regions with higher percentages of minorities, in the eastern and northernportions of the district, and went to regions with lower percentages of minorities, in the southern and southeasternportions of the district. They also left regions with lower enrollments relative to school capacity and wentto regions with higher enrollments relative to capacity. Further, whites left high-minority schools throughopen enrollment at a disproportionate rate—in one case, at a rate nearly double their proportion of theschool's population.
The repetition of this pattern since the mid-1990s has led the district's schools to becomesignificantly more stratified with respect to race and ethnicity. Three percent of elementary schools hada minority enrollment of more than 50 percent in 1994–1995; by 1999–2000, 15 percent of the schoolshad a minority enrollment of more than 50 percent. And it was the new choice schools, not bilingual choiceschools, that accounted for the change.
Because more whites than minorities chose to move out of certain schools, schools with thesharpest drops in white enrollment also had sharp drops in enrollment overall. This pattern closelyresembles what happened in New Zealand's school choice system, where schools that were relatively highin minority enrollment at the outset came to have higher minority enrollments as a consequence of choice(Fiske & Ladd, 2000).
Stratification of the district's schools has also increased with respect to income, as measured byfree and reduced-price lunch percentages, in a pattern that is remarkably similar to that associated with raceand ethnicity.
Open enrollment procedures and practices. School district procedures and practices wereimportant factors in creating these stratification patterns. The practice of prominently displaying test scoresin the local newspaper, as well as on district and school Web sites, contributed to the prominence of testscores in the demand for district schools. Some of the district's other practices—expecting parentsto obtain their own information on open enrollment; providing most information in English only; requiringparents to visit schools in which they to enroll their students; and insisting that parents provide their owntransportation—also help explain why open enrollment has had a stratifying effect. These practices favorparents with savvy, time, and resources.
Some choice schools have contributed to stratification by instituting enrollment preferences andrequirements, including preferences given to siblings of graduates, children of teachers and staff members,or students previously enrolled in a tuition-based preschool program; additional application requirements,such as interviews and supplementary forms; and expectations for parent participation, formalized inwritten agreements.
Funding and fund raising. Choice schools generally have high per-pupil costs, primarilybecause of their relatively small size. Determining whether small schools are inherently inequitable becauseof higher per-pupil costs is difficult, but issues of equity certainly arise with the advent and protectionof small, relatively expensive choice schools that serve predominantly white, middle-class students. Thehigher costs associated with these schools make less money available for other, needier schools.
In addition to the district's contribution, all district schools generate money through variousfund-raisers. Schools use these additional dollars to pay for such crucial items as library and classroombooks, curriculum materials, computers, art supplies, physical education equipment, adjunct faculty, guestspeakers, field trips, building improvements, staff development for teachers, and stipends for teachers toattend out-of-state professional meetings.
As a school's percentage of low-income students increases, its ability to raise funds decreases, andvice versa. For example, the most successful of the elementary schools with a high free and reduced-pricelunch percentage raised $75 per pupil, whereas the most successful of the elementary schools with alow free and reduced-price lunch percentage raised $278 per pupil.
Student achievement. The data show that school choice has not resulted in improvedachievement in the district. For example, when we examined three years of data for 3rd grade reading and four yearsfor 4th grade reading and writing scores on the Colorado Student Assessment tests, we found that the district'sscores showed steady improvement over the period of school choice expansion in a pattern typical with theintroduction of any new test (Linn, 2000) and that the district's improvement was comparable toimprovement in Colorado generally, where school choice is considerably more limited. We also examinedthree years of SAT data from six district high schools. Overall, SAT scores were flat or went down.
Contrary to the achievement improvement hypothesis, our evidence indicates that open enrollmentis a zero-sum game with respect to achievement—some schools do better at the expense of othersdoing worse.
Rather than boosting achievement overall, open enrollment in this district is merely redistributingit. The schools that gain high-scoring students thrive, whereas others languish or decline. Schools in thefirst group win awards for excellence, receive coverage in the press for their exceptional curriculum andteaching, and are the darlings of school choice advocates. Schools in the second group find themselves withmore students who are harder to teach and more parents who do not have access to extra resources; theseschools then suffer lower test scores, negative press coverage, and declining morale. Thus, in the shortspan of seven years, school choice has produced—or at least contributed mightily to—a two-tieredsystem of advantaged and disadvantaged schools that did not previously exist in this district.
What Can Be Done?
Despite these results and similar ones from other parts of the country and abroad, the political appealof school choice makes it unlikely that the movement will abate in the near future. Taking lessons fromBoulder's situation, however, we believe that limiting the more pernicious effects of school choiceis possible. We recommend, for example, that a school district centralize school choice so that proceduresand requirements are consistent and fair across all public schools, holding to a minimum the specialpreferences, opportunities, and arrangements for choice schools. The district should make its services, suchas bus transportation and informational materials, accessible to all students and their parents. The district'sfunding formula should monitor and address differences in the characteristics of school populations andschools' abilities to raise funds. Finally, and perhaps most important, the district should take measuresto ensure that school choice does not result in stratification by race, ethnicity, and income.
Provisions that address these recommendations may enable school districts to eliminate some ofthe problems with school choice and help equalize the school resource differences that school choiceexaggerates. But taking these steps does not diminish the fact that school choice and charter schools, forall their popularity, have widened differences in achievement, social characteristics, and resources acrossthe schools that we studied. In a district blessed by relatively high incomes, access to resources, and,until recently, a reputation for uniformly good schools, school choice has enabled some schools to achieveexceptionally high test scores and parent satisfaction, but only at the expense of other schools. The selectfew have garnered favorable press, but the others have lost many of their best students, their most involvedparents, and their good reputations. Yet for all the increased differentiation, school achievement overall hasnot improved. Worse, when the costs to equity are figured in, school choice in this district must be judgeda loss.
