Journal Issue: America's High Schools Volume 19 Number 1 Spring 2009
Does it matter how teachers teach? Is there persuasive evidence about "best practices" that can help students learn more and achieve deeper understanding of the curriculum content? Or is good teaching idiosyncratic to the individual teacher, dependent on the educator's philosophy, personality, general intelligence, and subject matter knowledge? Here again we take the broader view of instruction put forth by Cohen and Ball, and use "instructional approaches" as our unit of analysis.38 An instructional approach is characterized by certain regularities in the ways in which teachers and students interact with each other and with instructional materials that can be described, evaluated, and replicated. Among the instructional approaches used by teachers in various disciplines in high schools are interdisciplinary teaching, student teaming or cooperative learning, project-based learning, adaptive instruction, inquiry, and dialogic teaching. We have already discussed interdisciplinary teaching and the evidence on its effectiveness; we next define and discuss the others. While each has ardent advocates, they overlap and can be combined. Project-based learning, for example, typically involves teaming, although the reverse is not necessarily true, and dialogic teaching can be combined with adaptive instruction or inquiry. Nevertheless, each approach represents a distinctive pattern of interaction among teachers, students, and instructional materials, and it is possible to study the effects of each on student learning. We look first at cooperative learning or teaming and project-based learning. For inquiry, dialogic teaching, and adaptive instruction, we examine research on what has been found to work in one or more of three key components of the high school curriculum: language arts and writing, mathematics, and science.
A highly effective instructional approach or "best practice" is one that results in measurable improvements in performance on examinations or standardized tests. In a broader review we would also consider outcomes such as student engagement, effort, persistence, and subsequent success in the subject and in academic work generally. In this short review we cannot consider all the evidence bearing on multiple outcomes, nor can we systematically review the findings from the thousands of studies, largely small in scale, examining the effectiveness of particular pedagogies. Instead we rely heavily on reviews prepared by others to determine whether there is compelling evidence to support the general hypothesis that the specific instructional approaches used by teachers matter and affect student achievement.
Popularly known as cooperative learning after one well-known variant of this approach, student groups or teams are used by many high school teachers. But although this instructional approach is familiar to many teachers and its use is not rare, it remains the exception in high school classrooms. And when it is used, it is often used carelessly, with too little regard to the composition of the groups, the appropriateness of the tasks they are assigned, or the assessment of their work. Effective use of cooperative groups requires attention to these details and training for both the teachers and the students.39
Numerous research reviews have concluded that using various forms of group learning or teaming has improved pupil achievement, social attitudes such as tolerance and acceptance of differences, and classroom climates.40 The studies examined in these reviews typically used experimental designs to look at the effects of structured classroom grouping on student learning, behavior, and attitudes. Researchers carrying out the experiments placed students randomly into grouped and non-grouped classrooms and collected observational and survey data to examine the effects of the grouping strategy on standardized measures of achievement. In one review, Peter Kutnik and several colleagues note some shortcomings of these meta-analyses—namely, that they do not attend sufficiently to differences in curriculum and tasks and to variations in effects across age groups.41 Kutnik and his colleagues also point out that these grouping programs may not always fit well with classroom realities and therefore may be hard for some teachers to implement and sustain. They contend that researchers should give more attention to how both teachers and students are prepared for this kind of instruction and how class size, group composition, seating arrangements, group stability, the number of groups, and other factors influence the effectiveness of this approach. They also present a general theory of grouping that they believe provides teachers with more guidance and more flexibility than many of the current externally designed grouping programs.
Despite these caveats, the evidence shows that using structured student groups is a promising instructional approach. Many independent reviews show that student teams improve student achievement. The effects are so large and so consistent across subjects that group learning would be normative in an evidence-based environment. We shall return to this theme.
Project-based learning (PBL) organizes instruction around student-generated and -managed projects. It emerges from three older traditions of teaching: experiential and problem-based learning, which has been used successfully in higher education for decades; the Outward Bound wilderness expeditions; and the application of research on motivation, expertise, context, and technology to the design of instructional programs. Definitions of project-based learning vary widely, including the degree to which the approach must be student-centered.42 The variations make it somewhat hard to do research on PBL and also hard to summarize research findings, as the latter task requires deciding both what the parameters of PBL should be (do packaged or scripted projects count?) and whether the differences observed in variations of PBL matter and how they matter. There is general agreement, however, that the student projects should be central to the curriculum and focused on questions that direct students to encounter central concepts in a discipline. Most advocates believe the projects should not be teacher-selected or -scripted; rather, the students should have some choice and be expected to design and carry out the project themselves over an extended period of time. Significant portions of the work should be done independently, though the students often work in teams and the teacher may offer advice and guidance and feedback on partially completed or draft products. The projects should be realistic, not academic.43
PBL is often used in technology classes and is often supported by technology when used in other disciplines. The approach is used in many of the small schools funded by the Gates Foundation, such as those developed by Envision Schools and Big Picture Schools, in schools adopting the Expeditionary Learning and Co-nect school designs, and in many schools involved in technology projects. These schools all share a basic instructional approach, with considerable variation in the specifics.
The research on PBL consists largely of small-sample, non-experimental studies.44 Most rely on observations and interviews of students and teachers. Some use surveys. Although the findings suggest that participating in PBL increases student motivation and engagement, reduces absenteeism, strengthens cooperative behavior and improves higher-order thinking skills, the methodologies employed typically do not support such causal inferences. A series of studies of the Expeditionary Learning/Outward Bound (ELOB) and Co-nect school designs reported modest but significant gains in academic outcomes and changes in school climate.45 A more recent analysis of the effectiveness of comprehensive school reform designs found the research evidence for ELOB to be promising while indicating the need for more research on Co-nect.46 A review conducted in 2006 found only limited evidence that ELOB was effective and did not rate Co-nect.47
Although teaming, project-based learning, and interdisciplinary teaching are used in many subjects, the specifics of instruction usually are closely connected to curriculum content. That is, the pedagogies used in mathematics differ somewhat from those used in science or in language arts. Therefore, much of the research on instruction has been domain-specific, and evidence about instruction is typically examined domain by domain. We follow the pattern here and examine the evidence about instructional effectiveness in mathematics, science, and language arts below.
The effectiveness of various instructional approaches in mathematics has been heavily debated in recent years without much regard to empirical evidence about what works. Simply put, traditionalists, led by some respected mathematicians, and progressives have disagreed, among other things, over whether school mathematics should place more emphasis on algorithms (procedures for solving problems) or concepts, and whether discovery (constructivist methods) or direct instruction is more appropriate and effective and for whom. In 2005, representatives of the two groups issued a manifesto called "Reaching for Common Ground" to try to resolve some of their apparent differences over content and pedagogy.48 In the manifesto, leaders from the two groups agreed on three fundamental premises: students need proficiency with computational procedures, students must develop the ability to reason using mathematical language with precision, and students must be able to formulate and solve problems. They also agreed that automatic recall of certain basic procedures and algorithms was desirable, that calculators could be useful but should be used carefully in order not to impede fluency with basic procedures, that students should understand and be able to use basic whole-number algorithms and fractions fluently, that teachers should use methods appropriate to the goals, and that teachers should understand the mathematics they teach and how to make mathematics accessible to students.49
However, the debates continue, and evidence from a rich body of research on mathematics education does not resolve them. Most studies of mathematics teaching are small in scale, and many are observational, although some are small experimental and quasi-experimental studies. And most of the research has focused on the elementary and middle grades. Reviewing all this research would be beyond the scope of this article, but, fortunately, excellent reviews have been conducted in recent years.50 The reviews tend to agree that the practice of American mathematics teachers is not in line with the vision of reformers who want to see more emphasis placed on conceptual understanding and more student-centered and hands-on pedagogy in mathematics classrooms. In spite of decades of professional development and introduction of more constructivist curricula, the IRE (initiation-response-evaluation) pattern of teacher-student interaction prevails. The reviewers also note the lack of well-developed pedagogical theory to guide research and methodological difficulties associated with linking specific practices to student learning. Nevertheless, the reviewers see patterns in the research evidence suggesting the importance of teaching mathematical concepts explicitly, regular opportunities for student discussion, and collaborative work by students. Megan Franke, Elham Kazemi, and Daniel Battey point out that "simply using manipulatives, putting students in cooperative groups, or asking higher order questions does not lead to classrooms that support the development of mathematical understanding. How teachers and students engage with higher order questions, engage students in groups, or use manipulatives matters."51
These recent research reviews also emphasize the importance of discourse in mathematics classrooms. They cite many small studies that report that open discourse helps teachers understand their students' mathematical thinking and that when students have opportunities to express their ideas, they develop greater understanding. Franke and her colleagues describe four key strategies for effective discourse: revoicing, assigning worthwhile tasks, having students participate, and interrogating meaning. These strategies, of course, have been found to be components of most effective instructional approaches.52 Again, most of this research has been conducted in elementary or middle schools, and its implications for high school teaching are not clear.
As James Hiebert and Douglas Grouws point out, the empirical links between particular patterns of discourse and student learning have not been established.53 They note the primacy of two learning goals in mathematics— teaching skill efficiency or fluency and teaching conceptual understanding—and note there are no empirical studies that set out to examine which instructional approaches are associated with which of these outcomes. They argue that some features of instruction emphasize one and some the other, but that they overlap. Reviewing the process-product research, they conclude that "teaching that facilitates skill efficiency is rapid-paced, includes teacher modeling with many teacher-directed product-type questions, and displays a smooth transition from demonstration to substantial amounts of error-free practice. Noteworthy in this set of features is the central role played by the teacher in organizing, pacing, and presenting information to meet well-defined learning goals."54
They then examine the research findings about conceptual development and conclude that the keys are: teachers and students attending explicitly to the concepts, and students struggling with important mathematics ideas. They conclude that features of teaching that are often associated with conceptual development—use of concrete materials or higher-order questioning—are too closely tied to particular classroom conditions to make general claims about their efficacy. They also note that in many of the studies showing conceptual development, students also gained greater skill efficiency.55
Perhaps the most compelling evidence regarding the link between specific instructional approaches in high school mathematics and student learning is found in the Best Evidence Encyclopedia (BEE). A meta- analysis of research on middle and high school math programs examined studies with randomized or matched control groups, a study duration of at least twelve weeks, and pretest data that were roughly equal for non-randomized studies. The programs evaluated fell into three main categories: mathematics curricula, which mainly consisted of standard and alternative textbooks; computer-assisted instruction, which included programs that used technology, such as instruction or practice on computers; and instructional process programs that focused on the use of specific instructional approaches, including cooperative learning, individualized instruction, mastery learning, and comprehensive school reform. The latter does not seem to meet the usual definition of an instructional approach as comprehensive school reform models typically include structural, programmatic, and curricular changes as well as changes in instruction.
The programs associated with the most gains in student achievement, as measured by standardized tests, were those that focused on instructional processes, particularly cooperative learning, which had a median effect size of 0.3. Those linked with the smallest gains were the mathematics curricula programs, with a median effect size of 0.07.56
The mantra of reformers in science education is inquiry, and the past two decades have witnessed significant efforts to introduce the inquiry approach into high school science classrooms. Inquiry is built into the national science standards and used to describe good practice in the state standards for science. Inquiry is often used in other subject areas, but, like project-based learning, which might be viewed as a special form of inquiry, definitions and practices vary widely across and within subjects.57 There is, however, a common understanding of inquiry in science because it is central to the discipline. Although pure constructivists define inquiry as an activity in which students pursue answers to questions that they generate, more typically inquiry is viewed as the conduct of investigations selected by the teacher to help students understand key concepts in the discipline. Such "guided" inquiry is featured in many of the instructional materials used in science and social studies classes and is the focus of much of the professional development provided for teachers. There is general agreement that inquiry involves active learning and should reflect what scientists actually do.
Researchers do not, however, agree about how effective inquiry instruction is, or which forms are most effective, or how much of it is needed. Should teachers be using pure inquiry or guided inquiry? Should they be using inquiry all of the time or only occasionally? Does inquiry work better for certain students or for certain content? And even more fundamentally, is inquiry more effective at helping students master scientific concepts and processes than more traditional forms of instruction are? With the support of the National Science Foundation, the Education Development Center undertook a rigorous review of research on the effectiveness of the inquiry approach.58 The results of their analysis of more than 400 studies will be released in 2009, and the reader is advised to look for that report. In the interim, we must rely on other, less rigorous reviews conducted in the 1980s and 1990s that reported modest, positive effects of inquiry on achievement, process skills, and attitudes toward science. These reviews are often cited in support of constructivist arguments that students need "hands-on" experiences in classrooms or that they need to "do" science rather than simply read textbooks, listen to lectures, or watch demonstrations. However, the development of virtual laboratories offers another, less expensive option, and raises questions about the conventional wisdom about inquiry and good science teaching. A study by David Klahr and several colleagues has found that virtual labs are a viable alternative for elementary and middle school students, although particular domains of science such as life science might require direct experience with physical objects.59
Reformers often connect inquiry to the use of student teams in the classroom, noting that scientists work in communities of inquiry. A review of research found that using cooperative learning in science classrooms was linked with improved student learning, as well as more positive attitudes, more engagement in tasks, and higher motivation.60 These findings are consistent with the larger body of research on grouping or teaming discussed earlier. However, most of the science studies were small in scale; few had comparison groups, and most were in biology, so the evidence simply suggests that this approach is promising.
A study committee appointed by the National Research Council looked at the traditional laboratory activities found in high schools and concluded that labs are usually disconnected from the content of lectures. They argued for a more integrated curriculum that allows students to engage in the practices of science (for example, ask questions, make observations, analyze data, and construct explanations) and to support and deepen their understanding of science principles and concepts.61
The bottom line is that the evidence in support of using the inquiry approach in science is modest at best and that researchers must do more rigorous work to answer the questions raised above. Furthermore, new applications of technology are altering the meaning of inquiry and changing the debate about the reform of science instruction.
Reading and Writing
Because so many students enter high school lacking basic skills in reading and writing, these two areas have received considerable attention from researchers. Writing, in particular, has been the subject of hundreds of studies, perhaps because poor writing skills among high school graduates have been a major complaint of college faculty for decades.
Arthur Applebee and Martin Nystrand developed conceptual frameworks defining high-quality instruction in reading and writing that have guided subsequent research in this area.62 The frameworks define quality in terms of quantity, content, coherence, and student voice. Quantity denotes the time devoted to written and oral analysis of text, the content of which must be rich enough to support sustained discussion. Coherence denotes how well lessons relate to various parts of the curriculum. Student voice refers to the use of "dialogic" instruction, with students engaging in free-flowing discussions and expressing their own ideas and questions rather than merely responding to teacher monologues or questions.63 Nystrand reports that an observational study of twenty-five high schools found that students receiving such dialogic instruction outperformed peers receiving monologic instruction on assessments in which they were asked to critique literary passages.64 Nystrand and Adam Gamoran report similar findings from an analysis of hundreds of language arts lessons.65 In a subsequent study of forty-four classrooms in twenty-five schools, Judith Langer found that both high-performing and low-performing students who regularly engaged in dialogic discussions outperformed peers who did not.66
Researchers have also carried out some big-picture studies of instruction in the language arts. Gamoran and William Carbo-naro, examining data from the 1990 National Education Longitudinal Survey, found that both students and teachers reported that most students were not receiving instruction in the language arts that met the expectations of reformers in terms of the amount of time allocated, the coherence or content of the curriculum, or the opportunities for students to express themselves.67 They also found that students in honors classes were more likely than others to receive high-quality instruction. In a related analysis of national data from the National Adult Literacy Survey, Carbonaro and Gamoran found that student voice and the content of the curriculum were related to reading achievement but that quantity of assignments and coherence of instruction were not.68
Researchers at Johns Hopkins University recently conducted a systematic review of evidence on the effectiveness of various approaches to teaching reading to adolescents, as well as the effectiveness of instructional materials in reading and of computerized reading programs. The biggest gains in achievement were associated with instructional process programs involving cooperative learning (a median effect size of 0.28) and for mixed-method programs, such as Read 180 and Voyager Passport, that combined large-group and small-group instruction with computer activities. No studies of reading curricula or textbooks met the criteria for the analysis.69
Steve Graham and Dolores Perin conducted a rigorous meta-analysis of 123 studies of instruction in writing, all of which used experimental or quasi-experimental designs.70 The authors categorized the instructional approaches into four groups: process writing, such as writers' workshops; explicit teaching of skills, processes, or knowledge; strategies for "scaffolding" students' writing, such as pre-writing, peer assistance, and feedback; and alternative modes of composing, such as using word processing. The largest effect size (0.82) was linked with instructional approaches that explicitly taught strategies for planning, revising, and editing writing.71 Teaching students how to summarize reading material had a similarly large effect on writing quality. Using grouping arrangements that allowed students to work together to plan, draft, revise, and edit had an effect size of 0.75.72
It seems clear even from this unsystematic review of the evidence that the instructional approach teachers choose matters for student learning. And interesting commonalities in the evidence across disciplines suggest the power of well-designed grouping strategies, of classroom discourse that allows students to express their ideas and questions, and of offering students challenging tasks. Some evidence also suggests that inquiry approaches may add value. But although researchers look for routines in classroom practice that are linked to achievement, teachers, who have great discretion in their choice of instructional strategies, appear to pay little heed to the evidence that researchers amass. As a consequence, less than half of American high school students report working in groups. An even smaller share reports being engaged in any inquiry.73