Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Climate, Organization, Composition, and Size of Schools Moving from the classroom to the school, in this chapter we summarize the conclusions of various domains of research on how school climate, organization, composition, and size are related to student engagement in learning. The social and economic circumstances of students' families and neighborhoods greatly influence student engagement, and as mentioned earlier, schools cannot "do it all." But the evidence is clear that schools can make a significant difference, and that their effects on students' engage- ment, learning, and future opportunities depend substantially on how they structure the learning environment, and on the values they communicate to students and their families. We chose to examine school climate, organization, composition, and size in part because there was, for each, a fairly substantial research base suggesting effects on student engagement or learning, and because all are amenable to change. As in other chapters, we cast a broad net for evidence and did not limit ourselves to studies that measured student engagement directly. The inclusion of studies that focused on student achievement as the outcome is based on the assumption that gains in achievement imply in- creased engagement. SCHOOL CLIMATE Conceptualizing School Climate School climate refers to the values, norms, beliefs, and sentiments asso- ciated with routine practices and social interaction in schools. Theorists 97
98 ENGAGING SCHOOLS and researchers have used a wide variety of terms to refer to aspects of school climate including atmosphere, culture, environment, morale, school community, and school ethos. Accompanying divergent conceptualizations are different measurement instruments (cf. Gottfredson, HybI, Gottfredson, and Castandeda, 19861. Although the design of the qualitative and correla- tional studies reviewed in this chapter limit the degree to which causal conclusions can be drawn, the evidence is consistent enough to give sub- stantial confidence in a conclusion that qualities and factors related to school climate can affect student engagement and learning. Research-based judgments about desirable practices, however, require a careful examina- tion of the specific definition and measurement of school climate in studies. Investigations of school climate date back as far as Willard Waller's (1932) classic treatise on the sociology of the school, which he conceived as a miniature society as well as a formal organization. Following in Waller's footsteps, Hollingshead's (1949) case study of a small, midwestern commu- nity and its school system, and Gordon's (1957) account of "Wabash High" several years later portrayed the social system and organizational culture of single high schools. Tames Coleman's study of the "value climates" of ten midwestern high schools in the late 1950s, summarized in The Adolescent Society (Coleman, 1961), claimed to document the pervasive influence of the adolescent sub- culture on students' academic values, performance, and social activities. In this provocative book, Coleman concluded that in the typical high school, the adolescent social system channels the energies of students into nonaca- demic directions, and that the system provides a set of social rewards and punishments that supports athletics and other social activities and discour- ages intellectual pursuits. Coleman's study prompted a wealth of commen- taries, subsequent empirical studies, and debate on the antecedents, nature, and consequences of the "adolescent society" in contemporary America (Boocock, 1966, 1972; McDill and Rigsby, 19731. In 1979, Michael Rutter and colleagues published a study of 12 nonse- lective, inner-city, secondary schools in London, which established the no- tion of "ethos" as an important quality of schools (Rutter et al., 1979; see also Grant, 1988, and John Dewey, 1900, who referred to the school as a small society). A school's ethos concerns the coalescence among practices, beliefs. values. and norms. as Driscoll (1995, p. 217) explains: The concept of school ethos . . . is far more than an aggregate collection of individual variables. Rutter et al. recognized that it is the interaction of social processes and not merely their sum that explains the variance in the performance outcomes measured . . . the "ethos" of an effective school is in large measure a reflection of general, schoolwide expectations of con- sistent values and norms that permeate the institution. The ethos of an effective school is characterized by generally shared high expectations of
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 99 teachers and respect for them; positive models of administrators and other teachers for teacher behavior that reflects concern for one another; and some system of feedback through which teachers can evaluate their work. Rutter's stucly was in part a reaction to an earlier highly publicized stucly conclucteci by Coleman and colleagues (Coleman et al., 1966), which conclucleci that nonschoo! variables, especially the racial background and economic circumstances of students, accounted for the vast majority of differences in eclucational outcomes among schools, not differences among schools themselves. Rutter criticized Coleman's survey methods and his focus on tangible resources, arguing that the stucly haci unclerestimateci the magnitude of school effects because it haci not consiclereci the social climate . . anc organization. With the 1980s came the concept of "the school as a community." Implicit in both the ethos and "communitarian" conceptualizations is an assumed synergism in the operation of the school the notion that the whole of a good school is more than the sum of its parts. New concepts like community, democracy, and "an ethic of caring" (Nociclings, 1988) emerged, based on the assumption that students' attachment to school and their academic achievement are contingent on first satisfying teachers' and students' social and personal needs (Phillips, 1997~. Descriptions of communal or communitarian schools are strikingly similar to descriptions of out-of-school environments that appear to sup- port healthy adolescent clevelopment (e.g., McLaughlin, 2000; National Research Council and Institute of Medicine, 2002~. According to a recent review of evidence related to out-of-school environments (National Re- search Council and Institute of Medicine, 2002, pp. 90-91), students are most engaged when the social context promotes physical safety; provides some structure and opportunities for youth to clevelop new skills in the context of warm, supportive relationships; and promotes positive social norms. A relative newcomer to conceptualizations of school climate is the notion of "academic press." In a series of publications, Shouse (1996a, 1996b, 1997) and Phillips (1997) aciciress an ongoing clebate related to what appeared to be competing visions for increasing student engagement and learning a communitarian climate, as already clescribeci, versus a cli- mate characterized by academic press, clefineci succinctly by Shouse (1996a, p. 175) as "the degree of normative emphasis placed on academic excel- lence by members of the organization" (see also McDill et al., 1986~. The notion of academic press is related to what has been referred to in the literature as high expectations for success, and which has been shown to predict relatively high achievement (e.g., Evans, 1997; Hoy and Sabo, 1998; Marks, Secacla, and Doane, 1996; Newmann, 1992~.
100 ENGAGING SCHOOLS The evidence we will review suggests that these two conceptualizations of school climate are not, in fact, incompatible. Effect of School Climate on Engagement and Learning Building on previous research on private and public high schools (Coleman, Hoffer, and Kilgore, 1982), which had been highly criticized for not addressing problems of selection bias, Coleman and Hoffer (1987) studied school communities in private and public schools using the second wave of data from the High School and Beyond Longitudinal Study (National Center for Education Statistics, 19821. They reported Catholic high schools as having higher average achievement than public schools, lower dropout rates, and greater success in placing their graduates in some type of post-secondary institution. Although selection bias could not be ruled out altogether, they controlled for the variables most likely associated with selection bias, and attributed the better outcomes in Catholic schools in part to a strong, positive, disciplinary climate and conforming student behavior. Also using data from the High School and Beyond Longitudinal Study (National Center for Education Statistics, 1982), Bryk and Driscoll's (1988) subsequent study examined in greater detail the elements of a communally organized school, and demonstrated the importance of the school commu- nity to the quality of both public schools and Catholic schools. In contrast to much of the earlier literature, it provided quantitative evidence on the effects of school community and has served as the prototype for many of the more recent efforts to demonstrate the importance of school climate for adolescents' school attachment, engagement, and achievement. The authors posited three crucial components of communally organized schools: (1) a shared value system that pervades the school and derives from a shared history; (2) a common agenda for school members involving coursework, activities, rituals, and traditions that function as a unifying factor; and (3) an ethic of caring that permeates relationships among students and staff and between staff and students. Bryk and Driscoll's most important finding was that communally organized schools (schools with these three sets of qualities), whether public or private, had better attendance, higher morale, and better mathematics achievement. Using data from the National Education Longitudinal Study of 1988 (NELS 88), Lee with Smith (2001) examined whether reforms promoting communally organized schools were associated with indicators of student engagement and achievement. The NELS 88 database includes information on 800 high schools and a nationally representative sample of 25,000 students. These students were first studied when they were in eighth grade in 1988. Follow-up data were collected for the cohort in 10th grade and
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 101 12th grade.1 Additional data from the students' teachers and on the charac- teristics of their high schools also were collected and used in this study. Using hierarchical linear modeling in part to control for possible con- founds, the research design compared traditional "bureaucratically" orga- nized schools with communally organized schools. According to Lee with Smith (2001, p. 1031: The bureaucratic structure of most high schools relies on affectively neutral social relationships to facilitate the administration of standardized rules and procedures. Strong personal ties among adults, or between adults and students, make it more difficult for staff to comply with standard practices and procedures. Communal schools, in contrast, were structured to facilitate the cre- ation of emotional bonds between teachers and students and also among teachers. Lee with Smith (2001, p. 104) describes the characteristics of communally organized schools as follows: Rather than formal and affectively neutral relationships, members of com- munally organized schools share a common mission. Staff and students interact outside the classroom; adults see themselves as responsible for students' total development, not just for the transmission of lessons. Teachers share responsibility for students' academic success, often ex- changing information and coordinating efforts between classrooms and across grades. Analyses revealed that communal schools had better outcomes than bureaucratic schools for both teachers (satisfaction, morale, absenteeism) and students (less class cutting, less absenteeism, lower dropout rates; see also Marks, 2000~. Lee also found that disparities in these outcomes associ- ated with ethnicity and class were smaller in communally structured schools than in others. Additional studies suggest that supportive and caring schools may be especially beneficial for disadvantaged students (Battistich et al., 1995; Battistich et al., 1997; Bryk and Driscoll 1988; Bryk et al., 1993~. Lee with Smith (2001) examined alternative explanations of the advan- tages she observed of communally structured schools, such as differences in the curriculum and instruction. The most potent predictor of student out- come differences was teachers' collective responsibility for learning (see also Lee, Dedrick, and Smith, 1991~. This factor reflects teachers' views on their students' abilities and willingness to learn, their sense of responsibility for the learning of their students, whether teachers believe they can "get 125 percent of the 8th grade student participants in the study were not surveyed 4 years later when their cohort reached 12th grade because they either had been retained in grade or had dropped out of school.
102 ENGAGING SCHOOLS through" to even the most difficult students, whether teachers assess their own effectiveness, and whether they change how they teach depending on whether their students are learning. In schools where teachers assumed collective responsibility for their students' learning, students learned more, and differences associated with race and class were less prominent. Schools that had developed practices associated with the communal mode! were relatively high on the measure of teachers' collective responsibility for learn- ing, but the causal order between school restructuring and collective re- sponsibility is unclear. The authors point out that schools may be able to implement particular reform practices effectively because their teachers have these attitudes. Whether teacher collective responsibility (or self-efficacy) is a consequence or a cause of communal practices, evidence from many studies suggests its value in promoting student engagement and learning (e.g., Bandura, 1993, 1997; Lee, Dedrick, and Smith, 1991; Tschannen- Moran et al., 1998~. Findings from Bryk and Schneider (2002) support Lee's findings re- garding the importance of teacher attitudes for student outcomes, and fo- cused attention on another aspect of a communal social climate (see also Marks, 20001. In a longitudinal study documenting the effects of school reform in Chicago between 1991 and 1997, Bryk and Schneider found that "relational trust" was essential to school improvement. The extent to which relational trust was present in hundreds of Chicago schools was determined by teacher responses to survey questions about their interactions with the principal, fellow teachers, and parents. The questions addressed issues such as respectfulness, belief in each others' competence and the willingness to fulfill obligations, caring about each other professionally and personally, and putting the interests of kids first. Bryk and Schneider found that in schools with the highest achievement, nearly all teachers reported strong relational trust characterizing their inter- actions with the principal; three-fourths reported strong relational trust with fellow teachers, and 57 percent reported trusting relationships with parents. In bottom-quartile schools, fewer than half had trusting relation- ships with the principal, only a third with fellow teachers, and 40 percent reported trusting relationships with parents. Schools with strong relational trust had a 50-50 chance of making significant advances in mathematics and reading achievement; only one in seven schools with weak relational trust made similar advances. Relational trust was higher in small schools (fewer than 350 students) and in schools in which administrators, teachers, and students chose to be there instead of having been assigned. A few findings suggest strategies for promoting relational trust. In the study by Lee with Smith (2001), teachers in schools high on communality were likely to have a common planning time and to work in interdiscipli- nary teams. Opportunities to work collaboratively may have contributed to
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 103 respectful, trusting relationships among teachers and between teachers and administrators. The high levels of parent involvement in communally orga- nized schools may have promoted more trusting relationships with teachers and administrators, although parent involvement is also likely to be a result of feelings of trust. Studies of alternative high schools created for students at risk of drop- ping out provide additional support for the value of a communal social climate. These programs operate either within regular schools or as sepa- rate, alternative schools. They generally provide a complete, but alterna- tive, educational program separate from the one found in regular, compre- hensive schools. There have been several evaluations of alternative programs for at-risk students at the secondary level that bear on the importance of the educa- tional climate. Stern, Dayton, Paik, Weisberg, and Evans (1988) evaluated 10 within-school academy programs in California high schools; Wehiage et al. (1989) evaluated 12 alternative and 2 comprehensive schools; and Dynarski and Gleason (1998) evaluated 3 within-school and 6 alternative schools. Although the programs differed in the types of students they en- rolled, the curricula and services they provided, and the way they were structured, there appear to be several common features among programs with relatively low dropout rates:2 · a caring and committed staff who accepted personal responsibility for student success; · a school culture that encouraged staff risk taking, self-governance, and professional collegiality; and · a school structure that provided for a low student-teacher ratio and a small class size to promote student engagement. Taken together, the evidence suggests that student engagement and learning are fostered by a school climate characterized by an ethic of caring and supportive relationships; respect, fairness, and trust; and teachers' sense of shared responsibility and efficacy related to student learning. The evi- dence is clear, however, that a communitarian climate is not sufficient to increase academic engagement and learning. Equally important is a focus on learning and high expectations for student achievement, or "academic press. " 2Effective programs were programs in which there was a statistically significant difference in dropout rates between program participants and those in the control group computed at the 10 percent level. For example, dropout rates for participants in the Seattle Middle College program were 27 percent compared to 42 percent for students in the control group (regular high school students).
104 ENGAGING SCHOOLS The term "academic press" does not mean "pressure." Press means focusing students' attention on genuine learning (rather than simply going through the motions). Teachers "press" students to learn by encouraging them, by paying attention to their work and giving constructive feedback, by not accepting low or half-hearted effort, by holding them accountable, by providing assistance when they need it, and by not giving up on them. Simply having informal conversations with students about college and ca- reer goals is a form of academic press.3 A strong press for achievement might seem to contradict the notion that engagement is promoted by a supportive, communal school climate. But the evidence suggests that the combination is both achievable and desirable. Shouse (1996a, p. 184) explains: Any achievement effects associated with a sense of community stem pri- marily not from consensus and cohesion, but instead from the strength and transmission of particular organizational values related to the impor- tance of academic endeavor. Without such commitment, commonality of beliefs, activities, and traditions and care for students as individuals are unlikely to affect achievement levels in positive ways, and may even work to impede them. An important corollary follows from this, however: the most powerful impact on student achievement should occur when a school's sense of community is built around a solid structure of academic press. Consistent with this hypothesis, Shouse, using NELS:88 data on high schools serving economically disadvantaged children, found that achieve- ment effects were particularly strong when high communality was accom- panied by high academic press (Shouse, 1996a, 1996b, 1997; see also Marks, 2000; Phillips, 19971. Boyd and Shouse (1997) conclude, "average achievement in low-SES Esocioeconomic status] schools having high levels of both academic press and communality, in fact, rivaled that of schools serving more affluent students; the least academically effective low-SES schools were those that combined strong communality and weak academic press. " Lee and Smith (1999) also assessed the relative effects of the communal organization mode! (social support for learning) and academic press (focus- ing on content reform directed at increasing academic standards) on stu- dent engagement in school and achievement in math and reading. They 3Students in a national study who claimed that this occurred frequently in their school had, on average, higher educational expectations for themselves and higher postsecondary educa- tion participation (Wimberly, 2002). Compared to white students, African-American students claimed that such conversations occurred less frequently in the schools they attended.
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 105 analyzed 1994 data from the Annenberg Institute for School Reform, which included 30,000 sixth- and eighth-grade students in 304 public elementary schools (K-8) in Chicago. The findings indicated that the relationships between social support and performance on standardized achievement tests in mathematics and reading were stronger in schools with greater academic press (Lee and Smith, 1999, pp. 934-9351. The authors conclude: Our point is that reforms that focus on both the academic and the social domains are important.... To succeed in schools that press them to learn, students need support from the people with whom they interact (Lee and Smith, 1999, p. 935~. The findings of the complementarily of academic press and a com- munitarian school context provide important guidance for efforts to make urban high schools more engaging and effective as learning environments. These findings on school climate, moreover, are consistent with research on achievement motivation, discussed in Chapter 2, which reveals that stu- dents are more engaged in contexts in which they fee! socially connected and in which they are held to high standards. Taken together, the evidence suggests that high schools need to convey a clearly articulated and coherent set of values that focus on learning and achievement in the context of close and caring relationships with adults and peers. Policies for "Trouble-Makers" Policies for how to deal with behavior problems have direct implica- tions for the school climate as well as for the trouble-making students' own engagement especially their persistence in school. An understandable re- sponse to adolescents who violate the rules, undermine a climate focused on learning, and threaten students' and teachers' feeling of safety is to remove them to suspend them for some number of days or to expel them alto- gether. This form of disengagement has been described as "discharge"; "students drop out of school, schools discharge students" (RichI, 1999, p.2311. The students who are asked to leave are typically low performers with poor attendance (Lee and Burkam, 1992; Rumberger and Larson, 1998). Being prohibited from attending school puts them further behind, and makes it more likely that they will drop out (Lee and Burkam, 1992; Rumberger, 1995; Rumberger and Larson, 1998; Swanson and Schneider, 19991. More- over, being suspended or expelled may not seem like much of a punishment for students who are already disengaged from school. Thus, although sus- pension and expulsion may create a more hospitable environment for the teachers and students remaining in school, they are more likely to exacer- bate than to help the disengagement of the students who are misbehaving.
106 ENGAGING SCHOOLS Students in urban schools are more likely than others to be suspended. Nationwide, more than 3 million students were suspended from school during the 1997-1998 school year, nearly 7 percent of all students (Office for Civil Rights, 20001. But suspension rates vary widely among students and schools, even schools within the same district (The Advancement Project and The Civil Rights Project, 20001. Black students are more than twice as likely to be suspended as white students nationwide (13.2 percent versus 5.5 percent). As a result, urban schools, with high concentrations of stu- dents of color, generally have higher suspension rates than other schools. The recently growing policy of zero tolerance for violations of school safety rules is likely to increase the number of students who are forced to disen- gage from school (Skiba and Peterson, 1999), and these policies are more often applied in urban schools serving low-income students than in schools serving more affluent students (Kaufman et al., 2001, Table All. One strategy that has been used to deal with seriously misbehaving students is to send them to alternative schools. But this too can aggravate any inclination towards disengagement. In their longitudinal study of 100 Latino high school students in Austin, Romo and Falbo describe the "spe- cial" schools the school district created for difficult students. Youth who had broken school rules were mixed in these schools with those who had committed serious, even violent crimes. In the district's own evaluation, staff reported that students who attended these special schools were more likely to drop out of school than were comparable students who stayed in the regular high school (Romo and Falbo, 1996,p.87~. Clearly prohibiting students from attending school, either for some fixed amount of time or permanently, will not increase their engagement in school. From this perspective, the policy makes little sense. When alterna- tive arrangements are made for students who have difficulty functioning in a regular school or are making it impossible for other students to be en- gaged, care must be given to the composition of the alternative arrange- ments, and attention needs to be paid to the social climate of these schools in which the students are the most in need of a caring, communally oriented context focused on achievement. "Zero-tolerance" policies and alternative school programs for disrup- tive students do not address the day-to-day challenges related to discipline that classroom teachers in many urban schools encounter. To a substantial degree, implementing the kind of engaging instruction described in Chapter 3 in a caring, respectful social context should reduce discipline problems. Consistent with the notion of creating a respectful social context, re- cent efforts to develop effective discipline strategies have emphasized col- laboration instead of authoritarian, punishment-oriented approaches with students having an active voice in deciding on matters of classroom man- agement (Charles, 2000, 20021. An emphasis on democratic classroom
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 107 principles and student responsibility moving students from being "tour- ists" in the classroom to being "citizens" (Freiberg, 1996, p. 32) is also being touted. Creating a climate of trust also has been proposed as a pro- active strategy for maintaining order and cooperation (Charles, 20001. Charles (2002, p. 9) suggests that to be effective in today's classroom, teachers need to make "all students fee! welcome, replacing reward and punishment with strong values, asking students to contribute ideas for resolving problems . . . moving beyond the usual conception of discipline in favor of developing a sense of community in the classroom." There is currently no evidence on the effectiveness of these more egali- tarian approaches to student discipline. They are, however, compatible with findings suggesting that students are more engaged in academic tasks when they are in a trusting, caring and respectful social context. SCHOOL ORGANIZATION By organization we mean how teachers and students are sorted and how instruction is delivered. The way high schools are organized can affect engagement in learning by the messages the organization conveys and by the opportunities it creates for students to experience a climate that pro- motes engagement. High schools are currently experimenting with many aspects of school organization to facilitate closer relationships and more personalized teach- ing, in part by enabling teachers to see fewer students and students to see fewer teachers each day. One organizational feature that is touted by many experts is scheduling some courses in longer blocks (at least 90 minutes) to allow deeper and more sustained engagement and more individualized pac- ing. These courses are often multidisciplinary and team taught. A second feature is referred to as "looping" teams of teachers working with cohorts of students for at least two continuous years. Looping allows more sus- tained relationships to develop among teachers and students, as well as more individualized instruction. These organizational changes are usually embedded in efforts to con- vert large high schools into smaller units or to create new small high schools. Consequently, it is impossible to untangle the effects of any particular organizational feature. As a package, however, they seem to have consider- able value. The Coalition Campus Schools Project (CCSP), for example, redesigned two large and very troubled New York City high schools, Julia Richman High School in Manhattan and James Monroe High School in the Bronx (see Darling-Hammond et al., 20021. The design was derived from success- ful schools launched earlier Central Park East Secondary School, Interna- tional High School, and the Urban Academy. Teachers work in interdisci-
108 ENGAGING SCHOOLS plinary teams with a group of 40 to 80 students, and block scheduling and looping are implemented. The schools were able to create smaller teacher- student ratios by putting a relatively high proportion of their resources into hiring teachers, and by having students take fewer and longer courses each day. Analyses of 7 years of student data conducted by Darling-Hammond et al. (2002) suggest that this package of organizational changes, along with a small size and curricular reforms that included individualized programs and portfolio-based, authentic assessment, substantially improved students' at- tendance rates, lowered the incidence of disciplinary problems, increased performance on reading and writing assessments, and increased graduation and college-going rates, despite serving a more educationally disadvantaged population of students than the high schools they replaced. Variations of these design features are currently being encouraged by the Gates Founda- tion, and implemented in many small high schools throughout the coun- try.4 Although they are relatively new to the world of secondary education, they show considerable promise of achieving the kind of school climate and learning environments that the evidence suggests should engage students in academic work. Another organizational feature of high schools, tracking, has been stud- ied extensively. Tracking is particularly relevant to motivation because it is likely to convey the message that some students can achieve at high levels and some cannot. If tracking is rigid, in the sense that each year in a lower track makes it more difficult to move to a higher track, students may learn that they are not really expected to improve their academic status. Short of a clear-cut tracking system, course requirements can also limit students' opportunities. For example, some courses, such as 9th-grade algebra, can serve as gate-keeper courses; without them, students are prevented from ever getting into college preparatory courses. Taking low-track or lower level courses also disadvantages students because they are likely to get the least experienced, least expert teachers. McLaughlin and Talbert (2001) report from their case studies of high schools that teachers' subject expertise was often matched to students' competence in the subject. The best-prepared teachers were assigned to the high-achieving students and the least prepared teachers taught the classes with students having the most academic difficulty (see also Oakes, 19901. Using High School and Beyond, Talbert and Ennis (1990) found that teacher 4For further information on these design features, see http://schoolredesign.net; http:// www.lab.brown.edu/public/pubs/ic/block/block.shtml (for information on block scheduling); and http://www.lab.brown.edu/public/pubs/ic/looping/looping.shtml (for information on loop- ing).
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 109 tracking (some teachers teaching almost exclusively low-track courses and some teachers teaching almost exclusively classes with relatively higher achieving students) was more extensive in schools with relatively high pro- portions of low-income and minority students (see also Lucas, 19991. More- over, low-track teachers reported significantly lower levels of administra- tive support, control over instructional choices, and influence over school decisions. Finley (1984), similarly, reports also that the low-track teachers in her study were accorded less respect by their peers and were less involved in professional development. A few studies have examined the effect of tracking on student engage- ment. Ethnographers in both Great Britain and the United States have concluded that tracking polarizes students into positive and negative aca- demic attitudes and behavior (see Berends, 19951. It is possible, however, that the differences in attitudes existed before students entered a tracking system. Wiatrowski, Hansell, Massey, and Wilson (1982) analyzed attitudes toward school in a longitudinal data set and found no differences, based on students' track, in 12th graders' attachment to and misbehavior in school, after controlling for social background variables and initial attitudes. Using High School and Beyond longitudinal data from 1980 to 1982, however, Berends (1995) found modest but significant effects of track placement on students' college expectations, absenteeism, disciplinary problems, and en- gagement to school (reported interest in school and time spent on home- work), with students in the academic track showing more favorable out- comes. Coleman and Hoffer (1987) attributed the lower dropout rates and higher average achievement that they found in Catholic schools to less curricular tracking. Similarly, in their study of 4,450 sophomores in 160 public and Catholic high schools from the High School and Beyond data set, Bryk and Thum (1989) found that students were more likely to persist in schools that had less curricular differentiation and more students in academic programs, even after controlling for the academic and social class background of students. In their study of 11,794 sophomores in 830 high schools from the NELS:88 data set, Lee and Smith (1995) found that schools where more students took advanced coursework had higher levels of stu- dent engagement, defined in terms of attitudes and self-reported effort, after controlling for student academic and social class background. A few other studies suggest that being in a low track reduces the likelihood of students taking challenging, college-preparatory classes, even after control- ling for socioeconomic status and previous achievement (Braddock and Dawkins, 1993). Low-track courses may undermine students' motivation unnecessarily. In mathematics, for example, most teachers believe that there is a clear
110 ENGAGING SCHOOLS sequence for learning, and that students cannot process high-level concepts, such as algebra, until they have mastered basic computational skills (Grossman and Stodolsky, 1994, 1995; Stodolsky and Grossman, 20001. As a consequence, students who enter high school with poor basic skills are likely to receive drill in basic concepts rather than more motivating instruc- tion that engages them in challenging, open-ended problem solving. Re- search discussed in Chapter 4 shows clearly that even students who have relatively poor skills can engage in instruction that allows them to engage in deep analysis, and there is considerable evidence now that algebraic and other mathematical concepts can be introduced at many levels and at al- most any age (RAND Mathematics Study Panel, 20021. Formal and comprehensive tracking is less common now than in the past. High schools more often offer courses that are differentiated by level of achievement or perceived ability of students. High-achieving students take honors and Advanced Placement courses, but they are not distin- guished formally as being in a college-bound track. A common strategy to make advanced courses accessible to low-in- come students and students of color who have in the past been dispropor- tionately placed in remedial or vocational tracks is to give students the choice of taking more advanced courses. Although choice theoretically pro- vides opportunities for all students to be held to high expectations and to be engaged in challenging learning experiences, in reality, students end up being sorted, or they sort themselves along the same class and ethnic lines that applied to more formal tracking structures. For 3 years Yonezawa, Wells, and Serna (2002) studied 10 racially and socioeconomically mixed secondary schools that had voluntarily implemented detracking, giving stu- dents the choice of taking more advanced courses. Few of the Black and Latino students took advanced classes for a variety of reasons, including lack of information, failure to meet prerequisites, and a lack of encourage- ment from counselors or teachers. One student in their study described a counselor's weak encouragement (p. 521: " . . . she said . . . 'Who knows, if you work real hard, you can probably . . . keep up.' " Some of the most significant barriers were psychological. The students themselves often did not take advantage of opportunities because they had experienced too many years of being perceived to have low skills, and they lacked the confidence to take advanced courses, or they were anxious about being uncomfortable in classes with predominantly white and Asian and more affluent students. They feared being uncomfortable or disrespected in classrooms where they were different from the other students. Some also expressed concerns about the burden of demonstrating their own capability as well as countering racial stereotypes. As one student in an honors math class explains (p. 561:
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 111 I was like, "Oh man, I don't even belong in here," because it was like 30 Caucasian kids and one African student. I felt like I had to prove myself and prove that Blacks aren't stupid. fI felt like] if I were to get a problem wrong and raise my hand, they would look at me and say, "Ah, that R1~ ink ~~ T born ~ ~ learn via 1 1 n r1 tar Arts 1 ret ~m ~~ r--~--~ ~~ · · · I transferred to just The] advanced level. Although eliminating a rigid tracking system is a step in the right direction, allowing choice is not enough to eliminate the status quo with course-taking patterns differentiated along social class and race lines. To be effective, such a policy decision would need to be accompanied by con- certed efforts to create a social climate that informed and encouraged low- income students and students of color to participate in a rigorous and challenging curriculum. Special efforts also would have to be made at the classroom level to ensure that all students felt valued, respected, and in- cluded. Most important, strategies would need to be implemented to ensure that all students in a class are able to handle the material covered and the academic demands of the course a difficult task in a class of students who have dramatically different skill levels. One organizational strategy that has been used successfully at the college level for students who are not profi- cient in English is to link tutoring and small-group learning, often with a specialist in reading or English as a Second Language (ESL), to a regular class (Benesch, 1988a). Thus poor readers or English-language learners have the advantage of being integrated in regular classes, but with the extra support they need to master the material. In one program at University of California, Los Angeles (UCLA), English-language classes were linked to content courses (e.g., in political science or psychology; Snow and Brinton, 19881. The language class assisted students in developing reading, writing, and study skills that were directly related to the content course. City Uni- versity of New York created a block of three linked courses freshman social science, ESL reading, and ESL writing (Benesch, 1988b). Like the UCLA program, the links were designed to help students develop basic skills in the context of developing content knowledge. Hostos Community College in New York connected small tutor-led groups to content courses, such as biology, business, and early childhood education (Hirsch, 19881. Although we did not find examples of this strategy being applied at the high school level, the approach addresses some of the problems of the typical strategy of creating remedial or ESL courses that isolate low-skilled students and English-language learners from high-skilled and native En- glish-speaking peers. The strategy gives students access to regular courses, and special instruction designed to help them master the material of the course. Although not proven, the approach merits some experimentation in urban high schools.
2 ENGAGING SCHOOLS A key concern in classrooms with cultural, linguistic, and academic diversity is that students create social rankings within the classroom based on "status" differences, including academic ability (Cohen, Lotan, and Leechor, 19891. Differences in literacy resources at home, the use of stan- dard English, and knowledge of academic discourse are often perceived by both students and teachers to reflect intellectual ability. These perceptions can be translated into a status hierarchy, which can influence students' individual behavior and contributions to the group. For example, those students with high academic status typically talk more and express their opinions more freely, while "low-status" students defer to their leadership (Cohen, 19971. Cohen and Lotan's mode! of "complex instruction" (a mode! of col- laborative learning) was designed to address these status differences in the classroom (Cohen and Lotan, 19971. Their studies suggest that teachers can reduce the influence of academic status by publicly recognizing the contri- butions and value of many different kinds of intellectual competencies in the classroom and by emphasizing that every student brings valuable and different strengths which are required for the group's success. Although most of the research on complex instruction has been done at the elemen- tary level (Cohen and Lotan, 1997), Bower (1997) found that students in llth-grade social studies classes in which the approach was used gained more in achievement than the control group. In summary, although the evidence suggests that tracking and other policies that limit students' access to a rigorous curriculum have negative effects on some students' engage- ment, simply eliminating tracking or formal prerequisites is not sufficient. These organizational changes need to be supplemented with others that provide students with the encouragement, support, and structures they need to be able to achieve success. SCHOOL COMPOSITION Policy decisions at the district level can affect the distribution of stu- dents among schools, and the evidence suggests that those policies affect student engagement and learning. Although we know little about the mecha- nisms involved, correlational evidence suggests that the social composition of students in a school influences student achievement above and beyond the effects of student characteristics at an individual level . Student compo- sition may also influence achievement indirectly, through its relationship with other school characteristics, such as resources and practices. Kahienberg (2001) proposes that school composition affects student learning in part by influencing three different peer mechanisms the influ- ence of peers on learning through in-class and out-of-class interactions (e.g., cooperative work-groups, study groups), the influence of peers on the
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 113 motivation and aspirations of fellow students, and the influence of peers on the social behavior of other students (Kahienberg, 2001; see Chapter 3 for a more detailed discussion of peer effects). Student composition might also affect some of the climate variables discussed earlier, such as teachers' expectations for student learning and the degree to which they press for excellence or develop supportive, personalized relationships with students. Research has shown that several aspects of student composition are associated with student performance: the average socioeconomic status of the students in the school, their average academic skills, and the schools' racial and ethnic composition. Schools serving relatively more students from high socioeconomic backgrounds and with high academic skills have lower school dropout rates, lower absentee rates (Bryk and Thum, 1989), and lower student mobility rates, after controlling for the individual effects of student background characteristics. An analysis of 26,425 sophomores attending 968 schools from the High School and Beyond data indicated that the effect of school socioeconomic composition on dropout rates was about one-third to one-half as strong as the effect of individual socioeco- nomic status (Mayer, 1991). Because the extent and quality of resources, such as the proportion of qualified and experienced teachers, are confounded with the proportion of low-income students, it is not possible to untangle these confounds from the effect of the student body composition itself on student engagement. In future research it will be important to identify the reasons for the effects of the student body composition on student engagement and learning. We suspect that in addition to differential resources, differences in standards, expectations, and in the curriculum mediate school composition effects. SCHOOL SIZE The research evidence on small schools suggests that reducing substan- tially the size of schools is a promising strategy for achieving the kind of personalized education that engages youth. Definitions of "small" for the purpose of research varies, with some studies considering a high school enrolling 400 or fewer students as "small" (Howley, 2002) and others considering "small" to be any number up to 1,000 students (e.g., Lee, 20011. Howley (2002) suggests that the terms "smaller" and "larger" are more useful, allowing for relative comparisons to be made (see also McLaughlin and Drori, 20001. Researchers studying school size also use many different definitions of academic engagement including attendance, persistence, graduation rates, and sense of belonging and some use stan- dardized test scores. Despite some inconsistencies in definitions and find- ings, a careful review of the evidence supports three tentative conclusions: (1) smaller school size is associated with higher achievement under some
4 ENGAGING SCHOOLS conditions; (2) smaller schools improve achievement equity; and (3) smaller schools may be especially important for disadvantaged students (Howley, 2002~. One of the first challenges to the effectiveness of large, comprehensive high schools in the United States came from Roger Barker and Paul Gump's 1964 book, Big School, Small School: High School Size and Student Behav- ior. Their study of more than 200 schools revealed that a much higher proportion of students participated in school activities in smaller schools than in large ones. In the nearly 40 years since Barker and Gump's book was published, many researchers have sought to examine the effects of school size on student engagement and achievement. Barker and Gump's original challenge to the assumption that bigger is better has been sup- ported and expanded, and although questions remain about school size, we can conclude that bigger usually is not better. Whereas some studies find no difference between the achievement lev- els of students in large and small schools (e.g., Caldas, 1987 Louisiana schools]; Haller, Monk, and Tien, 1993 flOth graders in math and science]; Meinick, Shibles, Gable, and Grzymkowski, 1986 Enonurban Connecticut schools]), many studies find students in small schools doing better than those in large schools (e.g., Eberts, Kehoe, and Stone, 1982 Enational data sample]; Eichenstein, 1994 t10 New York City public high schools]~. In a relatively large study of the approximately 300 high schools in New Jersey, Fowler and Walberg (1991) found that students in smaller high schools performed better on achievement tests than students in larger schools (see also Walberg and Fowler, 1987; Walberg, 1989, 1992~. Data tracking students from large schools in Philadelphia that converted into small schools suggest that small schools can also increase attendance, course passage, persistence, and graduation rates (McMullen, 1994~. Wasley and colleagues (2000) likewise found more persistence in small schools in Chicago, in restructured existing large schools as well as autonomous, free-standing small schools. Fetler's (1989) study in California similarly reported lower dropout rates as well as higher achievement in smaller schools. In contrast to these studies, which found a linear relationship between school size and achievement, Lee's analyses of the National Education Longitudinal Study of 1988 (NELS 88) showed a curvilinear relationship (Lee, 2001; see also Lee, V. E., 20001. Schools of about 300 to 900 students performed better than both larger and smaller schools. The advantages of being small appeared to diminish substantially below a threshold of about 300 students. We suspect that the national data set, which included a substantial number of very small (<300 students) schools, explains why Lee found a curvilinear relationship when other researchers found a linear rela- tionship. The data sets used by researchers who studied relatively populated states such as California and New Jersey may not have had enough very
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 115 small high schools to affect the statistical findings. Taken together, the findings suggest that smaller might be better up to a point, but below that point small schools may have disadvantages, such as in course offerings and other resources, which affect student achievement. The evidence suggests that relatively small schools may also promote achievement equity. A few studies have shown dramatically higher levels of academic engagement among poor and working-class youth of color in small schools relative to their peers in large schools (e.g., Ancess, 2000; Fine, 1994; Lee and Smith, 1995, 1997; Meier, 1998; Wasley et al., 2000; for reviews see Fowler, 1995; Gladden, 1998; end Raywid, 19981. Lee (2001) likewise found that disparities in achievement gains based on SES were smallest in the three smallest categories of school size (up to 1,200 students). With regard to promoting equity, being very small did not appear to create any disadvantages. The evidence on greater achievement equity in small schools suggests that school size may be especially important for economically disadvan- taged students. Taking a somewhat different approach to examining pos- sible differential effects of school size on students from different economic backgrounds, Howley and Bicke! (2000) found that small school size re- duced the impact of poverty on student achievement. In their analysis of data from three states (Texas, Georgia, and Ohio), they found that "the lower the income in the community, the more student achievement is benefited by smaller schools" (p. 41. They describe a possible "differential excellence effect," an interaction effect of socioeconomic status and school size, in which larger sizes are particularly harmful in low-income commu- . . nltles. Schools-Within-A-Schoo! Findings suggesting the benefits of small schools have stimulated the creation of schools-within-a-school (SWAS) dividing schools into small learning communities, sometimes associated with academic or vocational themes (see Chapter 6, this volume). The research supporting the advan- tages of this structure is not as strong as research comparing whole schools that vary in size (e.g., Cotton, 1996; Howley, 2002; Meier, 19951. The evidence does, however, show some benefits for students in small learning communities with regard to both academic achievement and attitudes toward school. In an evaluation of the Kansas City SWAS program, Robinson- Lewis (1991) found that achievement test scores for SWAS students increased during their years in the SWAS program. Students also showed improved attendance and grades, and believes the program had helped them improve their basic skills. In a similar evaluation, Levine and Sherk (1990) found evidence of benefits for the SWAS programs in Kansas City,
116 ENGAGING SCHOOLS New York, and Orlando urban middle and high schools. Students in all three metropolitan areas showed improved reading comprehension during participation in the SWAS programs. Finally, Eichenstein (1994) evaluated the Project Achieve program for at-risk students in New York City public high schools. She found that restructuring existing schools into "houses" (averaging 250 students) was associated with improved attendance, student responsiveness, and student satisfaction. Inside Small Schools What Really Matters? Smallness by itself is not likely to promote greater engagement (Dar- ling-Hammond et al., 20021. More likely a smaller number of students makes it easier to implement policies and create the kind of climate that studies suggest are conducive to high levels of engagement. Achievement motivation theory and findings discussed in Chapter 2 provide some clues as to why small schools might typically engage youth better than large schools. For example, small schools should facilitate mean- ingful faculty-student relations, a sense of belonging and attachment, more individualized instruction that can create optimal levels of challenge for all students, and opportunities for both students and teachers to exercise au- tonomy. In the absence of these effects, we suspect smallness in itself has little value. Very few studies have documented directly these kinds of possible me- diators of the effect of size on student engagement. One exception is a study of Chicago elementary schools, in which students in small schools were more likely to report "academic press" (the feeling that teachers challenged them to reach high levels of academic performance), "peer academic sup- port" (their friends tried hard to get good grades, attended classes, believed paying attention in class was important, treated homework assignments seriously, and followed school rules), positive "classroom behavior" (class- mates treated each other with respect and care), and "safety" (perception of personal safety in and around school; Sebring et al., 1996~.5 Lee with Smith (2001) also found that a communal climate is more likely to be seen in small schools, which, according to the research discussed earlier, supports student engagement. Small schools that do not also adjust their organization and instruction may have few advantages. Note that in studies finding positive effects of small schools on student engagement and learning, the schools also imple- 5Two federally funded studies are currently under way, looking for the "active ingredients" in effective small learning communities (Gambone, Klem, Moore, and Summers, 2002).
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 117 mented a variety of other innovations, such as block scheduling, looping, multidisciplinary courses, and teachers seeing fewer students and students seeing fewer teachers (Darling-Hammond et al., 20021.6 A noteworthy and probably essential component of small schools that produces learning gains is a great deal of attention to instruction including developing a curricu- lum that challenges students to understand concepts deeply, adjusting modes of teaching to individual students' skills and learning styles, and providing extra supports for students who need them to succeed. The small schools that are most effective with respect to academic engagement, persistence, and graduation rates have also developed and implemented thoughtful and rigorous alignment of curriculum, pedagogy, and assessment systems in which youth and faculty are held publicly ac- countable to high standards (Fine, 1994; Wasley et al., 20001. The small schools that are least effective are those that are simply small in size, but have developed neither curricula nor assessment systems that demand rig- orous engagement and performance by all (Wesley et al., 20001. Some caution in interpreting the research on school size is in order because selection bias most likely contributes to the positive findings dis- cussed. Many small schools are open to all students within a district or geographical area larger than the existing neighborhood high school, and require students and families to apply or at least put themselves into a lottery. Some make particular requirements of families and students, and admit students to some degree on a judgment about whether the "fit" is good. Teachers and administrators also often choose to teach in small schools, and some small schools are able to select from among a large pool of teachers and administrators. Selection bias also can be relevant in studies of schools within schools, because these are sometimes theme based and attract students who have interests that match the school theme. Tracking also occurs, although not necessarily by design, with low- or high-achieving students systematically self-selecting into different schools or learning com- . . munltles. Although selection bias may play a role in some studies of the effect of school size on students' engagement and learning, we suspect that it does not explain all of the differences that have been found. Studies do show that small schools have advantages for students who typically perform very poorly in school, and if selection played a role, it may not be entirely because of "creaming," but because choice and matching have their own benefits. We recommend, however, that future researchers pay careful at- 6Intensive experimentation with small high schools sponsored by the Gates Foundation has resulted in 10 design principles for effective small schools; see http://www.stanford.edu/dept/ SUSE/csrn.
118 ENGAGING SCHOOLS tension to the possible role of selection bias of students, teachers, and administrators. CONCLUSIONS A review of the evidence presented in this chapter provides convincing, if not conclusive, evidence that school climate, organization, composition, and size can have important effects on student engagement and learning. Effective schools communicate high expectations for their teachers and students in the form of academic press and an atmosphere of trust and caring relationships among teachers, administrators, parents, and students. To achieve this kind of climate, teachers need opportunities to develop relationships with each other and with their students. Promising strategies for promoting caring relationships include decreasing the size of schools and the number of students seen by each teacher and the number of teach- ers seen by each student, block scheduling' and looping. Research on school organization suggests that tracking undermines engagement for students in the lower tracks, but that merely offering more choices to those students does not eliminate differential course taking. Stu- dents who have a history of poor achievement or who were previously excluded from challenging courses often do not have the confidence to take the more challenging courses. Teachers and administrators will need to overcome these psychological barriers by providing strong encouragement and eliminating the option of watered-down courses. In addition, they need to make sure that supports are provided for students to be able to succeed in challenging courses. Student composition within a school is associated with student engage- ment; the more high-achieving and high-SES students in a school, the more individual students are engaged and learning. Special efforts need to be made to increase the resources of schools serving a high concentration of low-income students and to maintain a climate that is conducive to high engagement. To the degree that there is diversity in students' economic backgrounds' efforts should be made to mix students from different back- grounds. There is also convincing evidence that small schools can confer an advantage for those students most at risk, and may help achieve greater equity in achievement outcomes. But the committee does not recommend creating small schools without consideration of the qualities of schools that have been shown to promote student engagement and learning challeng- ing and clear standards, personalization, meaningful and rigorous peda- gogy and curriculum, and professional learning communities. Moreover, it is not clear from the evidence that these qualities can be achieved only in small schools, although it appears more difficult to achieve them in large
CLIMATE, ORGANIZATION, COMPOSITION, AND SIZE OF SCHOOLS 119 schools. We do not yet know whether the school-within-a-school strategy will show the same benefits that are seen in some studies of autonomous small schools. But the evidence suggests that this is a strategy worth inves- tigating, and studies are underway that will provide useful information on the potential of this school reform strategy.