Access to Constructivist and Didactic Teaching: Who Gets It? Where Is It Practiced?
by Becky A. Smerdon, David T. Burkham & Valerie E. Lee - 1999
Calls for the reform of instruction in U.S. classrooms, particularly those in secondary schools, are growing and often strident. Many reformers advocate a move away from traditional, teacher-centered, (didactic) direct instruction, where students are passive receptors of knowledge, toward more student-centered understanding-based (constructivist) teaching that focuses on exploration and experimentation. In this study we investigate the issue of access to these two types of instruction in U.S. high-school science classrooms. We use a nationally representative sample of 3,660 students and their science teachers drawn from the first two waves of the National Educational Longitudinal Study (NELS:88). Although didactic instruction is more common among higher-SES and female students, constructivist instruction is practiced more often among students of lower ability. Constructivist teaching is also more common in both higher-level science courses (i.e., chemistry) and lower-level courses (i.e., basic biology and physical science). The students of average social and academic status appear to be the forgotten majority with respect to constructivist instruction. We offer explanations for the findings and discuss implications for educational policy and social equity in high-school science.
Calls for the reform of instruction in U.S. classrooms, particularly in secondary schools, are growing and often strident. Many reformers advocate a move away from traditional, teacher-centered, (didactic) direct instruction, where students are passive receptors of knowledge, toward more student-centered understanding-based (constructivist) teaching that focuses on exploration and experimentation. In this study we investigate the issue of access to these two types of instruction in U.S. high school science classrooms. We use a nationally representative sample of 3,660 students and their science teachers drawn from the first two waves of the National Educational Longitudinal Study (NELS:88). Although didactic instruction is more common among highersocioeconomic status and female students, constructivist instruction is practiced more often among students of lower ability. Constructivist teaching is also more common in both higher-level science courses (i.e., chemistry) and lower-level courses (i.e., basic biology and physical science). The students of average social and academic status appear to be the forgotten majority with respect to constructivist instruction. We offer explanations for the findings and discuss implications for educational policy and social equity in high school science.
THE CALLS FOR INSTRUCTIONAL REFORM
Recent school reform proposals call for a movement away from teacher-centered, direct instruction toward student-centered, understanding-based teaching. Both the Coalition of Essential Schools (CES) and the National Association of Secondary School Principals (NASSP) in partnership with the Carnegie Foundation for the Advancement of Teaching call for such change. The Coalitions principle number five (CES bases its philosophy on nine pedagogical principles) redefines roles for students and teachers:
The governing practical metaphor of the school should be student-as-worker, rather than the more familiar teacher-as-deliverer-of-instructional-services. Accordingly, a prominent pedagogy will be coaching, to provoke students to learn how to learn, and thus to teach themselves. (Sizer, 1992, p. 226)
According to the NASSP and the Carnegie Foundation:
A problem plaguing high schools has been that too many teachers teach in only one way, that is by the lecture methodthe same way they were taught. They often feel that lecturing is the most expeditious method for covering a large volume of material. Teachers should prepare themselves to offer courses that take advantage of methods that depend less on the teacher as purveyor of all wisdom. . . . Teachers [should] be adept at acting as coaches and as facilitators of learning to promote more active involvement of students in their own learning. (NAASP, 1996, pp. 2223)
This student-centered, student-active instructionoften called constructivistaffords students opportunities to explore ideas and construct knowledge based on their own observations and experiences. In science, this type of instruction may be particularly meaningful for students whose opportunities to experiment with or explore new methods of problem solving outside of the classroom are typically more limitedfemales, economically disadvantaged students, and minority students (Lee & Burkam, 1996; Burkam, Lee, & Smerdon, 1997a; Oakes, 1990).
Politically, these reforms are not always accepted and are sometimes unwelcome. For example, parents, teachers, and students concerns over access to colleges, which is based on high performance on standardized tests of recognized skills and facts, often lead to a dependence on traditional instruction (Talbert & McLaughlin, 1993). Additionally, opposition to reforms may come from parents and teachers who hold traditional views of teaching and learning. This is particularly true of parents who are unfamiliar with the work their children bring home and are unable to help them with it. Several barriers lead to only infrequent instructional reform in the constructivist direction. It is likely that for many teachers and parents this type of instruction remains quite novel (Cohen & Barnes, 1993).
Despite reform rhetoric supporting constructivist instruction, little is known about exactly which types of classrooms and students currently experience this type of teaching. There is little research on this topic, most likely because implementation is recent. The bulk of published writing explores the definition of this innovation (e.g., What is it?) and speculates about where such innovations belong in the larger educational context. Aware of the demonstrated lack of educational opportunity afforded disadvantaged students, writers on this topic often assume that it is more advantaged students who are exposed to innovative teaching methods (Cohen, 1988; Talbert & McLaughlin, 1993). They draw on the evidence of case studies. However, case studies, although interesting and informative, typically provide a series of snapshots of the phenomenon in a limited number of settings (nearly all of which are in elementary schools). Beyond research that focuses on selective samples, there has been little investigation of who is exposed to different types of instruction.
This study pursues the question of access. Specifically, we identify the characteristics of high school science classrooms, and the students in them, where forms of constructivist and didactic teaching occur. We use a nationally representative sample of tenth-grade students from the National Education Longitudinal Study of 1988 (NELS:88) and longitudinal data to investigate this question. We first identify the way we have measured the two forms of science instruction (didactic and constructivist). Using multivariate statistical techniques, we then investigate the link between these methods, the students who experience them, and the classrooms where they occur.
In this study, we draw indicators of constructivist teaching in science from students reports about the frequency of their experience with several science activities: discussing careers in science, making their own choice of science topics to study, designing and conducting their own experiments, and making up methods to solve science problems. Indicators of didactic instruction include students reports of the frequency with which they experience such classroom activities as: listening to the teacher lecture, copying the teachers notes, and watching the teacher demonstrate a science experiment. We identify the social and academic background of students who experience these instructional techniques in their science classes, the characteristics of the teachers who practice them, and the subject areas in which they occur. This study is important in light of the growing concern for learning opportunities (or lack thereof) among economically and socially disadvantaged students. Extant research suggests that both instructional time and the quality of learning opportunities typically favor more advantaged students (Oakes, 1985, 1990).
TRADITIONAL CONCEPTS OF TEACHING
Didactic (or direct) instruction traditionally has been conceptualized as the transmission of facts to students, who are seen as passive receptors. In classrooms where this type of teaching predominates, teachers typically conduct lessons using a lecture format. They often instruct the entire class as a unit, write notes on the chalkboard, and pass out worksheets for students to complete. In traditional classrooms, knowledge is presented as fact, students prior experiences are not seen as important, and students typically are not free to experiment with different methods to solve problems. Moreover, traditional views of teacher-student relationships are characterized as distant, with the teacher as an authority figure (Waller, 1932).
In traditional classrooms, it is teachers who are active; they convey facts and inculcate knowledge. Students are passive receptors of this knowledge. These classrooms typically consist of teachers presenting the right way to solve problems (or even the right solution). Knowledge, in this situation, is symbolic and isolated; learning does not typically motivate students or provide them with problem-solving skills they can apply to other situations (Dewey, 1902). High-school classrooms for most of this century have looked this way (Boyer, 1983; Goodlad, 1984; Powell, Farrar, & Cohen, 1985; Sizer, 1984), and most still do.
WHAT IS CONSTRUCTIVISM?
A new theory of teaching.
Underlying recent innovations in teachingalternately called authentic instruction, teaching for understanding, student-centered instruction, or constructivist teachingis the notion of the student as an active learner and the teacher as a guide or coach in the learning process (NAASP, 1996; Cohen, 1988; Conley, 1993; Newmann, Marks, & Gamoran, 1996; Sizer, 1992; Talbert & McLaughlin, 1993). The theory of constructivism is based on the idea that people learn better by actively constructing knowledge and by reconciling new information with previous knowledge.
The theory rests on several assumptions: (1) some of our notion of what constitutes knowledge may be culturally constructed, rather than truth or fact; (2) knowledge is distributed among group members and the knowledge of the group is greater than the sum of the knowledge of individuals; and (3) learning is an active, rather than passive, process of knowledge construction (Conley, 1993). In constructivist classrooms students are encouraged to pose hypotheses and explore ways to test them. They are encouraged to weigh information from these tests with their previous experience or understanding of the topic and then construct an understanding of subject matter. Students develop analytic skills that can be applied to other problems and situations, rather than accept their teachers explanations.
Although a move toward constructivist teaching is now an active reform issue, it has a previous tradition in the philosophy of education. At the beginning of the century John Dewey called for an end to the traditional drill-and-practice method of instruction. Deweys conception suggested that knowledge and instruction should build on students experiences, rather than be viewed as fixed or determined (Dewey, 1902). Even though Dewey wrote nearly a century ago, his notions of learning now appear as radical departures from traditional conceptions of teaching and learning.
Why we dont see more constructivist teaching?
There are numerous barriers to instructional change in American schools. In the first place, teachers generally have considerable autonomy in their classrooms and may easily ignore or dilute educational reforms (Ball, 1990; Cohen, 1988, 1990; Peterson, 1990; Rowan, 1990; Wilson, 1990). In addition, there is little incentive for change among teachers and schools; the survival of schools is not dependent on the adoption of reforms (Cohen, 1988). And teachers success, if it is measured at all, is typically determined by their students standardized test scores. Success on such tests usually requires more knowledge of facts than it does higher-ordered thinking. A growing emphasis on standardized tests often influences teachers practicessometimes they alter subject matter to teach to the test (Rowan, 1990), or use didactic methods in order to get through material quickly. Constructivist teaching, by contrast, is difficult and time-consuming (Talbert & McLaughlin, 1993). Furthermore, teachers who have limited subject-matter knowledge have less flexibility in their instructional choices and employ didactic teaching more often (McLaughlin & Talbert, 1993).
Moreover, because constructivism is a theory of learning, rather than a prescription for teaching, methods of constructivist teaching typically are not spelled out precisely and, moreover, are frequently somewhat ambiguous. The theory does suggest, however, various means to facilitate this kind of learning. For example, teachers should incorporate students prior experiences into the learning process; they should emphasize higher-order thinking and problem solving; students should be allowed numerous opportunities to express themselves, in numerous forms; and classrooms should be characterized as collaborative places where students feel safe to experiment (Newmann et al., 1996). Constructivism is more a philosophical approach to teaching than a given set of particular practices.
WHO GETS WHICH TYPE OF INSTRUCTION?
Instructional choices: The role of student ability.
Despite many possible barriers, empirical evidence suggests that constructivist instruction is practiced in limited settingsbut in which settings? Decisions concerning method and content of instruction are directly influenced by the characteristics of students or groups within the classroom (Barr & Dreeben, 1983; Dreeben & Barr, 1988; Newmann et al., 1996; Herr, 1992; Oakes, 1985, 1990; Raudenbush, Rowan, & Cheong, 1993). Just as traditional views of instruction encourage didactic teaching, traditional views of students intellectual ability may dictate the types of students to whom constructivist teaching is directed.
Many teachers believe that didactic instruction, including drill and practice, may be more effective for students with lower intellectual abilities (Talbert & McLaughlin, 1993). This would suggest that teachers are less likely to use innovative instructional techniques if they believe their students need training in basic skills. There is some evidence that use of traditional instructional techniques is, in turn, more prevalent in lower- as opposed to higher-track classes (Oakes, 1985, 1990; Talbert & McLaughlin, 1993). In lower-level classes, instruction is often characterized by rote memorization, drill, and practice. By contrast, teachers of upper-level courses emphasize higher-order thinking and present more interesting materials (Oakes, 1985, 1990).
Research on which types of students receive which kinds of instruction has yielded inconsistent results. For example, in a study investigating teachers instructional goals for each of their science or mathematics classes, Raudenbush, Rowan, and Cheong (1993) found that teachers differentiated their objectives based on the academic track of their classes; the same teachers who held higher-order objectives for their college-bound students de-emphasized these goals for their nonacademic classes. The authors argued that such differentiation of science and mathematics instruction is highly institutionalized. Furthermore, in a recent study of restructured schools, Fred Newmann and his colleagues (1996) found that exposure to authentic instruction was equitable in the 24 schools they studied; i.e., it was unrelated to race/ethnicity, socioeconomic status (SES), or gender. Students of higher ability were, however, more likely to receive authentic instruction.
In contrast, a study that focused primarily on science courses reported that teachers of higher-level courses (in this case, Advanced Placement, as compared to honors) were more likely to use a strong lecture format with limited laboratory experiments, student projects, or student presentations (Herr, 1992). Teachers used didactic methods in order to cover more material in less time, in order to prepare students to take the Advanced Placement exam. One reason for this discrepancy may be that the studies by Raudenbush, Newmann, and their colleagues investigated teaching goals and emphases, whereas Herr examined teaching methods. Our discussion of instruction so far has combined all of these features.
Instructional choices: The role of school level and subject area.
Studies of instructional practices in classrooms are located in both elementary (e.g., Ball, 1990; Cohen, 1990; Dreeben & Barr, 1988) and secondary schools (e.g., Herr, 1992; Oakes, 1985, 1990). A study that crosses both levels suggests that systematic structural and organizational variations between elementary and secondary schools lead to very different instructional goals and practices between levels (Firestone & Herriott, 1982). Firestone and Herriott (1982) report that elementary schoolteachers share more goals and place greater emphasis on basic skills than do secondary schoolteachers. The authors suggest that departmentalization (and a lack thereof), staff size, and administrative leadership contribute to these differences by level. This may suggest that it is the composition of students that leads to instructional differences between elementary and secondary schools. Teachers, parents, and students educational goals may be very different for young children, compared to adolescents. Adolescents and the adults who care about them may be focusing on opportunities beyond high school.
In addition, level differences in how students are grouped for instruction (i.e., ability grouping within classrooms in elementary schools compared to tracking between classrooms in secondary schools) most likely lead to variation in instructional choices. Unlike elementary school, secondary school students are typically placed in classrooms that vary by ability level. These classrooms are further differentiated by content or subject. For example, tenth graders enrolled in science courses could be studying various levels of biology (e.g., honors, general), chemistry, physical science, or a variety of other science subjects. The choice of instructional practice may vary among classrooms of different subject matter and ability level. These organizational features in turn influence students opportunities to learn in a way that is not found in elementary schools (Stevenson, Schiller, & Schneider, 1994). Also, a student may finish high school never having taken biology or chemistry, and Stevenson and his colleagues (1994) suggest that, for secondary schools, subject matter may be a better indicator of opportunity to learn than curricular track.
Although curriculum is a central feature of secondary schools (e.g., influencing how students and teachers are organized for instruction), little research has systematically explored the role of subject matter in high school teaching. In a comprehensive review of this somewhat fragmented literature, Grossman and Stodolsky (1990) describe subject matter as the context for instructional choices. Such issues as status (students taking, and teachers teaching, particular courses), composition of department (heterogeneous or homogeneousin terms of academic majors and teaching specialties), or nature of the subject (academic or nonacademic) may influence teaching through differences in resources, instructional leadership, or beliefs about student capabilities.
For example, instruction in mathematics may differ from that in social studies because social studies departments are more heterogeneous than math departments in terms of teachers majors and specialties (Grossman & Stodolsky, 1990). Their review suggests that instruction also varies within as well as between subjects. Lower level math courses serve lower-achieving and lower-status students. Students status influences teachers attitudes toward them, as well as the resources they use to teach. These resources and attitudes, in turn, may influence instructional choices that the teachers make. Specific instructional approaches that vary by subject matter were not identified in this review, nor were content areas within each subject area (e.g., biology versus chemistry).
Aside from a limited number of case studies and quantitative studies of selective samples (most of which focus on elementary schools), there is little empirical investigation of instructional techniques in secondary classrooms. A notable exception is a study by Hoffer and his colleagues (1996), who examined instruction in mathematics and science among high school seniors using NELS:88. They divided math and science courses by teacher-reported ability level of the class. Seniors in classes identified as lower ability were less likely to encounter science instruction that emphasized higher-order thinking than their counterparts. Lower-ability classes also had less laboratory time. This sample, however, though representative of 1992 high school seniors, overrepresents certain types of students. Specifically, their sample only included students who enrolled in science in their final year of high school (48% of seniors, most of whom were in higher-ability courses). Students in higher-level courses were of higher SES and more likely to be male and to attend private schools than the U.S. high school population overall. The Hoffer et al. (1996) study did not address the fact that advantaged students are more likely to receive high-quality instruction, because many disadvantaged students are not in this sample (i.e., they did not take science in their last year of high school). Although Hoffer and his colleagues used the same data source as the study described here, those researchers did not conceptualize instruction as didactic and constructivist or examine subject area. They also focused on a more restricted and selective sample.
APPROACH OF THIS STUDY
A focus on science.
Despite the recent interest in moving instruction in U.S. schools toward a more constructivist approach, little is known about who currently receives this type of teaching in the limited number of settings where it is employed. In this study, we investigate instructional approaches (didactic and constructivist) used in tenth grade science classes as a function of the subject matter of the course (e.g., biology or physical science), the characteristics of teachers, the social and academic backgrounds of students, and other classroom practices and characteristics. We use a nationally representative sample of U.S. students who were in tenth grade in 1990.
We center our inquiry on high school, as emphasis on scientific reasoning and experimentation is incorporated in many high school science courses (e.g., chemistry or biology). Because students may choose their science courses in high school, experimentation might not be evenly distributed among high school students. Unlike the study conducted by Hoffer and his colleagues (1996), we chose to focus on tenth graders. This decision allows us to investigate a more diverse sample of high school students, as almost all tenth graders take a science course, most high schools requiring two years of science for graduation.
We focus our investigation on science for two reasons. First, constructivist instruction is already reflected in many science reforms, and these reforms have begun to show up in the instruction employed in some science classrooms. Second, the data we employ, the first two waves of the National Education Longitudinal Study of 1988, include more relevant information about instruction in high school science classrooms (from both students and teachers) than in the other three subjects included in NELS. It is important to note, however, that regardless of the advantages of using NELS data (i.e., a nationally representative sample and a longitudinal design), we recognize that measuring instructional techniques with survey data is difficult.
Our study examines students access to constructivist and didactic instruction. We investigate characteristics of high school students and science classrooms where constructivist and didactic instruction happen. Three research questions drive the study:
Question 1: Subject areas in science.
How are student characteristics, teacher characteristics, and classroom practices linked to different subject areas in science? The large body of empirical work on the subject of tracking and high school course taking leads us to expect that socially and academically disadvantaged students are more likely to be found in lower-level science classes. We also expect that teaching methods differ by course content.
Question 2: Access to different forms of instruction.
Who gets didactic or constructivist instruction? Is access to these practices related to the social and academic characteristics of students or teachers? Based on research that focuses on differentiation of instruction, opportunity to learn, and tracking, we would expect that advantaged students (i.e., higher SES, nonminority, and high achieving) receive constructivist instruction more than their less advantaged counterparts. Furthermore, we expect that teachers who are more knowledgeable and comfortable with their subject matter employ constructivist methods more often.
Question 3: Classroom characteristics related to access.
Where are constructivist and didactic teaching methods practiced? Although the choice of instructional methods is also likely to be related to the subject area of the science course, there is little existing research on this topic. However, we expect constructivist instruction to be related to such classroom characteristics as the availability of laboratory equipment or the number of students enrolled in the science class.
SAMPLE AND DATA
We employ data from the first two waves of the National Education Longitudinal Study of 1988 (NELS:88), a broad-based and federally funded study of the educational development of American students available from the National Center for Education Statistics. NELS:88 includes biennial data collected from a large and nationally representative sample of students, their parents, their teachers, and school personnel (teachers and administrators) in the schools they attend. Our study uses 1990 data about tenth grade science classrooms and students, including their high school transcripts. We use data collected at the base year (1988), when these students were in the eighth grade, as statistical controls for their social and academic background as they entered high school.
We employ a subsample of 3,660 NELS students who fit the following data filters: (a) the random subsample of NELS students who have information available from tenth grade science teachers1 (and also were enrolled in science in the tenth grade); (b) the panel sample (the original sample followed from eighth to tenth grades) who completed surveys from both the base year and first follow-up year; (c) students who provided information on instruction in their science classes; and (d) students with transcript data available in the tenth grade (science subject matter is reported only on transcripts). As the original NELS design oversampled certain types of schools (private schools, schools with high proportions of minority students), we use the NELS panel weights for all analyses. This allows us to generalize our results to U.S. tenth graders (enrolled in science courses) in public and private schools in 1990.
The measures we use fall into five conceptual groupings: (1) science instructionconstructivist and didactic (our outcomes); (2) students academic and social characteristics; (3) content area of the science class; (4) science teacher characteristics; and (5) science classroom practices and characteristics. Details of the construction of all measures used in the study, including psychometric characteristics of all composites, are spelled out in the Appendix. The logic guiding our selection of items and construction of composites is laid out below.
MEASURES OF SCIENCE INSTRUCTION
Conceptualizing instruction with survey data.
Our investigation focuses on constructivist and didactic instruction in high school science classes. Current definitions of instruction include two components: (1) method of delivery (i.e., teaching methods), and (2) content (i.e., intellectual quality). We selected a number of survey items that tap students descriptions of the teaching methods employed in their science classes (i.e., how often they experience certain types of instruction). Our aim is to identify instructional techniques that embody the philosophical approaches to teaching we have discussedconstructivist or didactic.
The intellectual quality of particular instructional approaches is more difficult to measure with survey data than are teaching methods. In fact, such measures were not included in NELS surveys of either students or teachers. Interviews and observations over an extended period of time, such as those described in the Newmann et al. (1996) study, are the major sources of data that researchers use to measure instructional quality. Therefore, a limitation of this study is that our measures tap only one component of constructivist teachingteaching methods. In order to strengthen our analyses, however, we investigate teaching methods within the context of contentthe subject matter of the science courses in which students were enrolled.
A factor-analytic approach.
NELS students responded to several items describing the frequency of many activities in their current science classes (responses coded from 1=very rarely, 2=once a month, 3=once a week, 4=almost every day, and 5=every day). Certain instructional practices reflect constructivist principles (e.g., designing and conducting their own experiments, having a choice of the science topic to study), while other practices reflect more traditional and didactic approaches (e.g., listening to the teacher lecture, watching the teacher demonstrate an experiment).
In order to test our hypotheses about the underlying constructs of constructivist and didactic instructional practices, we performed a single principal-components factor analysis of the ten items we used to construct these composites (with VARIMAX orthogonal rotation). As expected, two factors emerged. The principal loadings indicated that one factor reflects the less frequent, constructivist instructional approaches; a second factor reflects the more common, didactic instructional approaches. We present factor loadings and means for each item in Table 1.
Constructivist teaching is conceptualized as an active process for students. In contrast to didactic instruction, students in constructivist classrooms are encouraged to communicate their interests and ideas. Students are also asked to pose and test their own hypotheses about the phenomena under study. This conceptualization is reflected in the first factor. It loads predominantly on five instructional practices involving students who: (1) make up their own problems and work out their own methods to investigate the problems; (2) design and conduct experiments and projects on their own; (3) make their own choice of science topic or problem to study; (4) write up reports of laboratory and practical work; and (5) discuss career opportunities in scientific and technological fields.2 All of these instructional methods are practiced relatively infrequently (means range from 1.31 to 2.24 on a scale of 1 to 5), most less than once a month (coded 2). We created the measure of constructivist instruction as a factor-weighted z-score (mean=0, standard deviation=1); a higher score indicates more constructivist teaching principles.
A second factor reflects more traditional, didactic instruction (i.e., students passive, teachers active). This factor predominantly loads on the five instructional practices where students: (6) listen to the teacher lecture; (7) copy the teachers notes from the blackboard; (8) watch the teacher demonstrate or lead them in an experiment or systematic observation; (9) use a book or other written instructions that show them how to do an experiment; and (10) review work from the previous day. These measures reflect traditional teaching methods, where teachers present material and students are merely expected to absorb it. The means for these items, ranging from 2.48 to 4.00, suggest that they are employed much more often than the practices included in the constructivist composite. Indeed, students report that they listen to the teacher lecture almost every day. Higher scores on the second composite (also a factor-weighted z-score) represent a more didactic approach to teaching.
Are these practices mutually exclusive?
The factor-analytic approach to the construction of our instructional measures and the underlying groupings do not imply that the two sets of practices are mutually exclusive (even though the factors are constrained to be orthogonal with a correlation of 0). As instructional reforms filter down into classrooms, teachers often distort or adapt them to their own teaching styles. It is unreasonable to expect that teachers will adopt constructivist techniques exclusively, particularly since these reforms are relatively new and uncertain. Therefore even where constructivist teaching is employed more often (and the means of these variables suggest that this is seldom), teachers surely use some traditional instruction.
Furthermore, some practices possess both constructivist and didactic elements. For example, the two lab-related activities load on both factors, as do writing up lab reports and use of textbooks. The constructs represented by these factors are perhaps better interpreted as tapping underlying philosophies of instruction, rather than proscribing specific instructional practices as exclusively constructivist or didactic in nature. As such, the two philosophies are reasonably conceptualized as independent motivations to be blended and balanced within each classroom.
MEASURES OF STUDENT CHARACTERISTICS
There is considerable evidence that academically and socially disadvantaged students (i.e., low-achieving, minority, and low-SES students) are often found in lower-level classes. The instruction in such classes is reported to be mostly didactic (Oakes, 1985, 1990). Furthermore, the language status of students may also influence teachers pedagogical choices; teachers may teach more of the basics to students who not fluent in English. We use several measures of students academic and social background (prior achievement, race/ethnicity, gender, socioeconomic status and language minority status) to investigate the kinds of students who receive each type of instruction.
MEASURES OF SCIENCE SUBJECT MATTER
Certain courses in high school science (e.g., physics, chemistry, honors biology) are intended for college-bound students (and these courses also have prerequisites), whereas others (e.g., basic biology, physical science) are courses intended for students of average or low ability. In high school, students are sometimes separated into classrooms and courses by ability level and in science, such classrooms typically differ by subject matter, as well as instructional rigor. We argue that it is important to investigate teaching methods within the context of course content. Here we examine the subject matter of science courses tenth graders take. To investigate the potential relationship between subject matter and instruction, we include measures indicating students tenth grade science course: basic biology, general biology, honors biology, chemistry, physical science, or other science courses (e.g., earth science, environmental science, unified science, zoology).
MEASURES OF TEACHER CHARACTERISTICS
Research indicates that teachers personal and professional characteristics are related to how they teach. For example, teachers who have limited subject matter knowledge are reported to be less flexible in the type of instruction they use, and thus are more likely to employ didactic teaching (McLaughlin & Talbert, 1993). To explore this we include several items in which teachers describe themselves and their preparation for teaching: whether they feel unprepared to teach, their gender, teaching experience, and whether their college degree(s) were in science.
MEASURES OF SCIENCE CLASSROOM CHARACTERISTICS AND PRACTICES
Many classroom features influence the type of instruction teachers employ. For example, science classrooms that are characterized by more laboratory time and better lab equipment are likely to be places where constructivist teaching is more common. We include several classroom practices that may influence teaching style in our investigation: time spent on whole-class instruction, emphasis on science skills and topics, noninstructional time, laboratory time and quality of equipment, grading criteria (i.e., relative vs. absolute performance), and teaching goals (affective vs. academic). To explore evidence that structural and social characteristics of classrooms also influence the type of instruction teachers use we included measures of class size and the percentage of racial, ethnic, and language minority students in the class. We converted many of the composites we investigated here into z-scores: emphasis on science topics, noninstructional time, goals, grading, and amount/condition of lab equipment.
Before we take up our major analytic task, linking constructivist and didactic instruction with student, teacher, and classroom characteristics, we categorize high school science classrooms by their subject matter (i.e., basic biology, general biology, honors biology, chemistry, physical science, and other) and link them to student, teacher, and classroom characteristics [Question 1]. We employ one-way analysis of variance to test these relationships.
Our major analyses link instructionas a dependent variableto the characteristics of students, teachers, and classrooms. To carry out our analysis we employ ordinary least squares (OLS) regression methods to explore research questions 2 and 3 (Who gets didactic and constructivist instruction? and Where are constructivist and didactic teaching methods practiced?) and to estimate the relationship between student, teacher, and classroom characteristics and didactic and constructivist instruction. We present a separate model of constructivist and didactic instruction for each outcome, each one having two steps. The first step investigates: (a) who has access to either of these instructional practices; and (b) the characteristics of teachers who practice each instructional type. The second step examines where these practices are employed (i.e., the subject matter of science classes and the characteristics of particular classrooms). This model structure allows us to investigate whether differences in subject matter and classroom characteristics explain initial demographic differences (Cohen & Cohen, 1983).
Tenth grade science courses.
In this study, we examine constructivist and didactic teaching methods within the context of course content. Our measure of content includes the major science courses that enroll sampled tenth graders, which we divide into six categories: (1) basic biology; (2) general biology; (3) honors biology; (4) chemistry; (5) physical science; and (6) other science courses (a variety of life and physical science courses, each enrolling fewer than 3% of the sample). Enrollment in these courses varies widely (see Table 2). General biology, the most common tenth grade science course for students of broad abilities, enrolls a majority of tenth graders (57%). By contrast, physical science enrolls very few (4%).
Prior research indicates that students typically enroll in biology courses (i.e., basic, general, and honors) when they are tenth graders (Burkam et al., 1997b). Chemistry typically enrolls eleventh graders, whereas physical science typically enrolls ninth graders. When students take a course is often indicative of their ability level and future learning opportunities (Stevenson et al., 1996). Consequently, tenth graders who enroll in chemistry are somewhat advanced in science compared to their peers, and tenth graders who enroll in physical science are somewhat behind in terms of their exposure to science material. This assumption is supported in our study by the prior achievement of students in these courses.
Who takes which science courses?
Table 2 displays means on all variables in this study for the 3,660 students in our sample, who are accounted for separately according to the subject of their tenth grade science class. Group means (for continuous variables) and percentages (for categorical variables) are presented for each of six content areas: basic biology, general biology, honors biology, chemistry, physical science, and other science.3 Variables are separated into logical groups: instructional practices, student and teacher characteristics, and classrooms practices and characteristics. We organize our discussion around these groupings.
Both constructivist and didactic teaching methods in science are related to science subject matter. Because our measures of instructional methodsor instructional philosophiesare conceptualized as independent, students can potentially experience high (or low) levels of one or both approaches in the same subject. Somewhat higher levels of constructivist instruction characterize classes in basic biology, chemistry, and physical science. Two of these courses, basic biology and physical science, enroll the fewest tenth graders, as well as the least able and most socially disadvantaged students. Higher levels of didactic instruction, on the other hand, characterize classes in chemistry and physical science. Note that higher levels of both didactic and constructivist instruction are employed in chemistry and physical science classes. Higher levels of constructivist and lower levels of didactic instruction are employed in basic biology (a typical tenth grade course for lower-ability students).
Sophomores in basic biology, physical science, and other science courses display substantially lower prior and current science achievement, whereas students in honors biology and chemistry display substantially higher prior and current science achievement. Language minority and African American students are represented equally among the subject areas. However, tenth graders enrolled in basic biology, physical science, and other science courses are more likely to be Hispanic and come from families of lower social class than their peers enrolled in the remaining science classes. It may be surprising that there is no relationship between science subject and student gender. There is evidence that female students complete more life science coursework over the four years of high school than males, as well as enroll more often in honors biology (Burkam, Lee, & Smerdon, 1997b).
Teachers are not randomly assigned to science subjects. Female teachers are less likely than males to instruct honors biology and chemistry (i.e., courses with the highest-achieving tenth graders). Physical science and basic biology (i.e., courses with the lowest-achieving tenth graders) are least likely to be taught by individuals with degrees in science. Unsurprisingly, the most experienced instructors teach honors biology (17.2 years), and the least experienced teach physical science (13.8 years). It is notable how much experience all of these teachers have. Honors biology instructors report having the most control over their teaching, and physical science teachers report the least control. Teachers of other science feel least prepared to teach their subjects.
Science classroom practices and characteristics.
In addition to constructivist and didactic instructional techniques, other features of science classrooms are also related to content. Here we consider the emphasis teachers place on science topics and skills and time on noninstructional activities. The latter is an indicator of the amount of time content is not taught. Not surprisingly, teachers emphasize science topics and skills more in advanced science classes (i.e., honors biology and chemistry) and less in lower-achieving classes (i.e., basic biology, general biology, and other science). An exception is with physical science, a course enrolling lower-achieving tenth graders, where emphasis on science topics is high. Teachers of basic biology and physical science, however, spend more time on noninstructional issues than teachers in science subjects taken by higher-achieving sophomores (e.g., honors biology and chemistry).
Other measures of classroom practices and characteristics are also related to content. In honors biology and chemistry, a greater percentage of time is spent on whole class instruction than in the other science courses. Students of chemistry and physical science spend more time conducting laboratory work and have more (and better) lab equipment at their disposal. Teachers grading criteria (affective or absolute) and their instructional goals also differ based on content. Teachers of basic biology, chemistry, physical science, and other science emphasize affective and absolute goals less than general and honors biology teachers do. Honors biology and chemistry teachers use more absolute criteria for grading, and teachers of lower-achieving science classes (i.e., basic biology, general biology, and other science) use more affective grading criteria.
The demographic and structural characteristics of classrooms we investigated are also related to subject. Basic biology is taught in smallest classes (20.7 students); honors biology is taught in the largest (24.7 students). Basic biology enrolls more language-minority students than the other subjects (3%); whereas honors biology is taken by almost no language-minority students. Physical science is chosen by substantial proportions of minority tenth graders (33%), as are basic biology (26%) and other science (24%).
Tables 3 and 4 present results of the regression models describing characteristics of (a) students who receive and (b) teachers who employ constructivist and didactic instruction [Question 2], as well as (c) characteristics of classrooms where these methods are practiced [Question 3]. We present the results for constructivist instruction in Table 3, and the results for didactic instruction in Table 4. We dropped some of the teacher and classroom characteristics displayed in Table 2 from the models presented in Table 3 and 4, as they were not significantly related to either outcome.
The multivariate regression results are shown in two steps. In step 1, we display student and teacher characteristics that are associated with the outcomes; in step 2 we add classroom practices and characteristics. All continuous measures (outcomes and predictors) are z-scores. Results are presented in effect size units for the purpose of comparison.4 It is important to remember that science classes are not exclusively taught with only one of these methods or the other. The outcome variables tap tenth graders assessments of their science teachers philosophical position more than the frequency of a simple list of techniques.
Several student and teacher characteristics are related to constructivist instruction in these multivariate models (see Table 3). Female students have less access to constructivist teaching than males (ES = -.23). After controlling for other characteristics of students and teachers, language-minority, lower-achieving, and slightly higher-SES students have more access to constructivist instruction than their peers. However, SES and language-minority status differences disappear once science subject matter and classroom characteristics are entered into the model. This suggests that socially and academically differentiated access to instruction is mediated, to some extent, by the fact that language-minority and higher-SES students are more likely to be in particular classrooms for tenth grade science.
Students who experience more constructivist instruction are taught by teachers with less teaching experience. After controlling for classroom characteristics, teacher gender becomes statistically significant (ES=.08); students taught by female teachers are exposed to constructivist methods slightly more often than their peers. In the final model, female students still have less access to constructivist teaching (ES=.25), and less able students have more (ES=-.13). Student and teacher characteristics alone explain only 3% of the variance in constructivist instruction.
Constructivist instruction is related to classroom characteristics. Reflecting the bivariate results from Table 2, students who experience constructivist instruction are more likely to be enrolled in basic biology, chemistry, and physical science courses than in general biology (the reference group). The size of these effects, ranging from .17 to .21, is small to moderate, but among the larger effects in the model. Students who are exposed to more constructivist teaching are enrolled in classrooms characterized by more lab time (ES=.08) and less whole class instruction (ES=-.10). Furthermore, emphasis on science topics is positively associated with constructivist teaching. Classroom practices and characteristics explain nearly twice as much of the variance in constructivist instruction as student and teacher characteristics. However, the total explained variance in the outcome is modest (10%).
Similar to constructivist teaching, student and teacher characteristics are related to didactic teaching methods (see Table 4). In contrast to constructivist teaching, females and higher-achieving students receive more didactic instruction than their counterparts. However, prior achievement is not associated with this outcome after controlling for the types of classrooms and courses in which students are enrolled. Higher-SES students experience more didactic instruction than their counterparts (ES=.08), even when courses and classroom characteristics are taken into account. Hispanic and African American students experience more didactic instruction than their white counterparts, although the difference between white and African American students disappears once classroom characteristics are taken into account. Furthermore, students with male teachers and teachers with degrees in science receive didactic teaching methods more than their peers. The teacher gender effect is explained once subject matter is taken into account. Student and teacher characteristics account for only 4% of the variance in didactic instruction.
The frequency of didactic teaching methods is influenced by course subjects and classroom characteristics. Students enrolled in chemistry and physical science, compared to general biology, receive more didactic instruction (ES=.20 and .24, respectively). Sophomores enrolled in general biology receive more didactic instruction than their peers enrolled in other science courses (ES=-.11). Not surprisingly, students who experience more didactic instruction are more likely to attend classes characterized by more whole class instruction. Like constructivist teaching, students who experience more didactic instruction are enrolled in classrooms with more laboratory time (ES=.09) and more emphasis on science topics (ES=.36). In addition, students enrolled in classrooms with fewer minority students receive more didactic instruction. The full model explains 19% of the variance in didactic instruction.
SUMMARY OF MAJOR FINDINGS
Instruction in high school science classes.
Although we recognize that instructional strategies brought together under the umbrella of constructivism actually reflect a theory of learning rather than a set of teaching methods per se, we have investigated instruction in a more prosaic manner. Specifically, in this paper we have used reports from a large and nationally representative sample of high school sophomores about the frequency with which their teachers use several instructional strategies in their science classrooms. We used contemporary theoretical writings about teaching to guide the construction of our measures of these two types of instruction. Furthermore, we argue that these instructional measures tap (student-perceived) teachers relative allegiances to the philosophical principles involved: namely, student-centered and teacher-centered instruction. Our analyses have investigated the characteristics of students, teachers, and classrooms that are related to each type of instruction. We explored identical analytic models for the two types of instruction in order to make comparisons between them. In this summary we highlight the most important findings related to our research questions.
Which types of students receive which type of instruction?
Although not much of the variance in the two forms of instruction is explained solely by student or teacher background, we did identify important relationships between background and instruction. Our most surprising finding here is that access to constructivist teaching is related to students ability in scienceless able students get more of it. One demographic characteristicstudent genderis strongly related to both forms of instruction: girls receive more didactic instruction and boys receive more constructivist teaching, even after controlling for student background and course content. Access to didactic instruction is influenced by social class, with more affluent students receiving this type of teaching more often. In general, our findings concerning social access to constructivist teaching are contrary to much of the writing on this topic. In high school science, our findings suggest that teachers are using constructivist methods among less able, rather than more advantaged and more able, students.
Science subject matter, classroom characteristics, and instruction.
In high school the frequency of these instructional methods is related to science subject matter, as are student and teacher characteristics and classroom practices. Several findings are noteworthy. Results regarding the practice of constructivist and didactic instruction follow somewhat unexpected patterns. Both teaching methods are more prevalent among students enrolled in chemistry (an advanced course for tenth graders) and physical science (a remedial or repeated course for tenth graders). In many ways, these two courses represent opposite ends of the science course taking spectrum. One enrolls very high achieving and high-SES tenth graders (chemistry), and the other enrolls very low achieving and low-SES sophomores (physical science).
Physical science and chemistry also differ markedly on some important teacher and classroom characteristics. Students in chemistry and physical science classes had teachers who either reported having much control over their teaching (chemistry) or little control (physical science). This may suggest that chemistry teachers elect to employ constructivist methods, but physical science teachers may be pressured to do so. In addition, compared to physical science teachers, more chemistry teachers have science degrees (80% versus 66%). Therefore we can assume that for chemistry teachers, constructivist methods are often coupled with a solid foundation of scientific knowledge, though the same may not be true of many physical science teachers. Yet despite their differences, these two courses also share one important featurestudents in these courses spend the most time on laboratory work and have access to the best (and most) lab equipment.
In terms of typical sophomore-level science courses, students enrolled in basic biology (which enrolls the lowest achieving sophomores) experience more constructivist teaching and less didactic instruction than tenth graders enrolled in other biology courses. Of all the courses, students in honors biology receive the most didactic instruction. Overall, the basic biology and physical science courses are similar in terms of student and teacher characteristics, both courses enrolling low-achieving and low-SES students. Basic biology and physical science teachers are less likely to have science degrees and have less teaching experience.
WHAT DO THESE FINDINGS MEAN?
The focus of some recent reforms targeted at high schools is on instruction. Much of the writing about school reform highlights the need to move teaching in the direction of allowing students more exploration, of encouraging them to use their own observations and experiences to construct rather than reconstruct knowledge. In other words, reformers are calling for schools and teachers to move instruction away from didactic and toward more constructivist teaching. Despite these calls for reform, however, many who have observed U.S. high school classrooms do not see much of this type of teaching.
Why has the reform of instruction taken hold so sparingly, and only in low-or high-status settings, within U.S. secondary schools? Prior research suggests that political support for the new forms of instruction is probably very weak, especially among parents (Talbert & McLaughlin, 1993). Such important outcomes as access to college and success on standardized testsgateway issueslead parents to resist the use of any instructional methods that might impede their childrens success in vaulting these well understood hurdles. Research suggests that coverage of content may be more thorough in didactic than constructivist classrooms (Herr, 1992). Moreover, following your nose, expressing yourself in writing, pursuing collaborative projectsall these activities take a lot of time if students are to learn from them. Students can be led down many blind alleys before they discover a good solution to a difficult problem. Feeling free to experiment means being allowed to make lots of mistakes without being chastised for being wrong.
Because of the inclination toward constructivist teaching among scholars, it might seem logical to interpret our findings about the types of students who have access to constructivist teaching and the sorts of classrooms where it is most likely to occur as positive evidence of a move toward more social equity in schools. After all, constructivist teaching is more common for less able students and in lower-level science courses. However, constructivist teaching is also more common for advanced sophomoresthose enrolled in chemistry. This indicates that the average sophomores, those enrolled in the typical tenth-grade science courses, were least likely to receive constructivist instruction.
We suspect that constructivist teaching is employed to meet different educational goals for chemistry students than for their counterparts in lower-level science courses. Constructivist teaching methods in chemistrya course taught by teachers with solid scientific backgrounds and more control over their teachingmay be coupled with instruction of higher intellectual quality and content. We would expect that chemistry in the tenth grade, a course typically enrolling college-bound eleventh graders, is more cognitively and academically demanding than physical science or basic biology, which enroll students of lower ability in science. Therefore, we are not encouraged by the fact that constructivist teaching in science is common among more academically disadvantaged studentsin courses where half to a third of teachers do not have science degrees. Rather, we suspect that teachers who use this type of instruction in lower-level courses are not employing them effectively (i.e., together with strong scientific content). This supports Newmann et al.s (1996) arguments that teaching methods alone do not guarantee high quality educational experiences.
A major American philosopher of education, John Dewey, was passionate, eloquent, and dedicated in his advocacy of experience-based learning, especially when in place of drill and practice. It is ironic that almost a century after Dewey (1902) wrote about these practices they are still so rare today. It is equally ironic that constructivist instruction is more prevalent at either end of the science spectrum, leading to questions about the intellectual quality of some of these instructional endeavors. We argue that teaching methods should be examined within the context of contenteither in terms of subject matter, as we have done, or in terms of more direct measures of intellectual quality. We also posit that as long as educational reforms focus on students and families that demand the most attention (those at either end of the spectrum), the unspecial (those in the middle of the spectrum) are likely to remain untouched in terms of instructional innovation.
PROS AND CONS OF STUDYING INSTRUCTION WITH SURVEY DATA
Although we recognize that studying instruction is best accomplished by trained observers visiting the same classroom several times over an extended period in a modest number of schools (the method followed by Fred Newmann and his colleagues ), we suggest that there is also considerable value in data drawn from NELS. The fact is that empirical studies that link instruction to anything are few. Rather than waiting for perfect conditions to study this phenomenon, we suggest that using current and nationally representative data to provide important information about where these types of instruction are happening in U.S. high schools, and which students have access to them, is useful and timely. We do, however, discuss the major limitations and advantages of this approach below.
Learning about classroom instruction through survey data is far from ideal. The set of NELS items describing instruction is limited, where students are asked to make global statements about their science classes. Their judgments about instruction in their science classes are probably influenced by their liking of the subject, their performance in the class, their relationships with the teacher, and their classroom peers.
Another limitation of information about instruction drawn from NELS is that very few students were sampled in each school, so the number of students from a particular classroom is very limited. It would be preferable to characterize the instruction in a particular science class through the reports of several students in the same class and evaluate their agreement about this. Were NELS students sampled by classroom, rather than by school, we would analyze these data hierarchically (i.e., using Hierarchical Linear Modeling), with students nested in classrooms. However, in this study we did not explore instructional effects on students, but rather investigated instruction as an outcome, as a function of characteristics of students, teachers, and classrooms. Thus, the absence of hierarchical methods did not seem problematic in this case.
Despite the limitations of using survey data with this structure to study instruction, there are several countervailing strengths in using NELS to investigate this topic. Most important, students are linked to their teachers, so that information about classes drawn from teachers (e.g., track level, laboratory equipment, teacher preparation) can be linked to the students in those teachers classes. Even more fortunate is the fact that students are linked with teachers in particular subjects. Although each NELS student has data available from only two of his or her teachers, the subsample of students with full data from science teachers is substantial and random. Previous federally sponsored longitudinal education data (i.e., High School and Beyond, NLS-72) did not include information from teachers that could be linked to students.
There are other obvious advantages as well. NELS is a longitudinal study, so that data about students academic status at the beginning of high school are available. Of particular value to this study is the science achievement test score. The survey items about science instruction are specific in describing instruction in this subject, and these items lent themselves to the constructivist/didactic constructs most readily. Moreover, it is also helpful that reliable data are available from students transcripts. We used the transcript data to identify the science courses in which tenth graders were enrolled.
Ball, D. L. (1990). Reflections and deflections of policy: The case of Carol Turner. Educational Evaluation and Policy Analysis, 12(3), 247259.
Barr, R., & Dreeban, R. (1983). How schools work. Chicago: University of Chicago Press.
Boyer, E. L. (1983). High school: A report on secondary education in America. New York: Harper & Row.
Burkam, D. T., Lee, V. L., & Smerdon, B. A. (1997a). Gender and science learning early in high school: Subject matter and laboratory experiences. American Educational Research Journal, 34(2), 297331.
Burkam, D. T., Lee, V. L., & Smerdon, B. A. (1997b). Debunking the myths: Exploring common explanations for gender differences in high-school science achievement. Unpublished paper, University of Michigan.
Cohen, D. K. (1988). Teaching practice: Plus que Ha change . . . In P. W. Jackson (Ed.), Contributing to educational change: Perspectives on research and practice (pp. 2784). Berkeley, CA: McCutchan Publishing Corporation.
Cohen, D. K. (1990). A revolution in one classroom: The case of Mrs. Oublier. Educational Evaluation and Policy Analysis, 12(3), 311329.
Cohen, D. K., & Barnes, C. A. (1993). Conclusion: A new pedagogy for policy? In D. K. Cohen, M. W. McLaughlin, & J. E. Talbert (Eds.), Teaching for understanding: Challenges for policy and practice. San Francisco: Jossey-Bass Publishers.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlational analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Conley, D. T. (1993). Road map to restructuring: Policies, practices and the emerging vision of schooling. University of Oregon. (ERIC Document Reproduction Service number ED409603)
Dewey, J. (1902). The child and the curriculum. Chicago: The University of Chicago Press.
Dreeban, R., & Barr, R. (1988). Classroom composition and the design of instruction. Sociology of Education, 61, 129142.
Firestone, W. A., & Herriott, R. E. (1982). Prescriptions for effective elementary schools dont fit secondary schools. Educational Leadership, 4(2), 5153.
Goodlad, J. (1984). A place called school: Prospects for the future. New York: McGraw-Hill.
Grossman, P. L., & Stodolsky, S. S. (1990). In L. Darling-Hammond (Ed.), Review of Research in Education, 20, 179221. Washington, DC: American Educational Research Association.
Herr, N. E. (1992). A comparative analysis of the perceived influence of advanced placement and honors programs upon science instruction. Journal of Research in Science Teaching, 29(5), 521532.
Hoffer, T. B., Moore, W., Quinn, P., & Suter, L. E. (1996). High school seniors instructional experiences in science and mathematics. National Center for Education Statistics: Statistical analysis report #95278.
Lee, V. E., & Burkam, D. T. (1996). Gender differences in middle grade science achievement: Subject domain, ability level, and course emphasis. Science Education, 80(6), 613650.
Lee, V. E., Smith, J. B., & Croninger, R. J. (in press). Understanding how high school organization influences the equitable distribution of learning in mathematics and science. Sociology of Education.
McLaughlin, M. W., & Talbert, J. W. (1993). Introduction: New visions of teaching. In D. K. Cohen, M. W. McLaughlin, & J. E. Talbert (Eds.), Teaching for understanding: Challenges for policy and practice. San Francisco: Jossey-Bass Publishers.
National Association of Secondary School Principals. (1996). Breaking ranks: Changing an American institution. Reston, VA: NASSP.
Newmann, F. M. (1996). Authentic achievement: Restructuring schools for intellectual quality. San Francisco: Jossey-Bass Publishers.
Newmann, F. M., Marks, H. M., & Gamoran, A. (1996). Authentic pedagogy and student performance. American Journal of Education, 104(4), 280312.
Oakes, J. (1985). Keeping track: How schools structure inequality. New Haven, CT: Yale University Press.
Oakes, J. (1990). Opportunities, achievement, and choice: Women and minority students in science and mathematics. In C. B. Cazden (Ed.), Review of Research in Education, 16, 153222. Washington, DC: American Educational Research Association.
Peterson, P. L. (1990). Doing more in the same amount of time: Cathy Swift. Educational Evaluation and Policy Analysis, 12(3), 261280.
Powell, A. G., Farrar, E., & Cohen, D. K. (1985). The shopping mall high school: Winners and losers in the educational marketplace. Boston: Houghton Mifflin.
Raudenbush, S. W., Rowan, B., & Cheong, Y. F. (1993). Higher order instructional goals in secondary schools: Class, teacher, and school influences. American Educational Research Journal, 30(3), 523553.
Rosenthal, R., & Rosnow, R. L. (1984). Essentials of behavioral research: Methods and data analysis. New York: McGraw-Hill.
Rowan, B. (1990). Commitment and control: Alternative strategies for the organizational design of schools. In C. B. Cazden (Ed.), Review of Research in Education, 353389.
Sizer, T. R. (1984). Horaces compromise: The dilemma of the American high school. Boston, MA: Houghton Mifflin.
Sizer, T. R. (1992). Horaces compromise: The dilemma of the American high school. Boston: Houghton Mifflin.
Stevenson, D. L., Schiller, K. S., & Schneider, B. (1994). Sequences of opportunities for learning. Sociology of Education, 67 (July), 184198.
Talbert, J. E., & McLaughlin, M. W. (1993). Understanding teaching in context. In D. K. Cohen, M. W. McLaughlin, & J. E. Talbert (Eds.), Teaching for understanding: Challenges for policy and practice. San Francisco: Jossey-Bass Publishers.
Waller, W. (1932). The sociology of teaching. New York: J. Wiley & Sons.
Wilson, S. M. (1990). A conflict of interests: The case of Mark Black. Educational Evaluation and Policy Analysis, 12(3), 293310.