Direct Instruction Revisited: A Key Model for Instructional Technology Susan G. Magliaro Barbara B. Lockee John K. Burton Rooted in behavioral theory, particularly the radical or selectivist behaviorism of B.F. Skinner (1953, 1954, 1966, 1968, 1974), the direct instruction (DI) approach to teaching is now well into its third decade of influencing curriculum, instruction, and research. It is also in its third decade of controversy. Our purpose is to present the DI model with the notion that the designer can and should use the model effectively based on appropriate assessment of the learners, content, context, and task at hand. To accomplish our goal, we begin with a general discussion of the basic DI framework, followed by a summary of the major DI models that have been used in live instructional contexts. We then shift to a review of how DI has been used in technology-based learning environments. Finally, we conclude with a look into the future of DI. Rooted in behavioral theory, particularly what Skinner labeled the radical or selectionist behaviorism (see, e.g., Skinner, 1953, 1966), the direct instruction (DI) of Siegfried Engelmann (Bereiter & Engelmann, 1966) is now well into its third decade of influencing curriculum, instruc- tion, and research. It is also in its third decade of controversy (c.f., Gersten, Baker, Pugach, Scanlon, & Chard, 2001). To begin, we offer a definition and our stance related to DI—which has become the whipping post in some pedagogical camps, while the pan- acea in others. For clarity, DI is not a lecture approach (e.g., Freiberg & Driscoll, 2000). It is an instructional model that focuses on the interac- tion between teachers and students. Key compo- nents of DI include “modeling, reinforcement, feedback, and successive approximations” (Joyce, Weil, & Calhoun, 2000, p. 337). Joyce and colleagues specified the instructional design principles, which include the framing of learner performance into goals and tasks, breaking these tasks into smaller component tasks, designing training activities for mastery, and arranging the learning events into sequences that promote transfer and achievement of pre- requisite learning before moving to more advance learning. Essentially, DI is “modeling with reinforced guided performance” (Joyce et al., p. 337). Our intent in this article is to explicate the genesis, components, and permutations of DI as it has evolved in practice, and describe how it is being used in instructional technology. Three purposes undergird this article. (a) First, we believe that DI is a viable, time-tested instruc- tional model that plays an important role in a ETR&D, Vol. 53, No. 4, 2005, pp. 41–55 ISSN 1042–1629 41
Transcript
Direct Instruction Revisited: A Key Model for Instructional Technology
Susan G. MagliaroBarbara B. LockeeJohn K. Burton
Rooted in behavioral theory, particularly theradical or selectivist behaviorism of B.F.Skinner (1953, 1954, 1966, 1968, 1974), thedirect instruction (DI) approach to teaching isnow well into its third decade of influencingcurriculum, instruction, and research. It isalso in its third decade of controversy. Ourpurpose is to present the DI model with thenotion that the designer can and should usethe model effectively based on appropriateassessment of the learners, content, context,and task at hand. To accomplish our goal, webegin with a general discussion of the basic DIframework, followed by a summary of themajor DI models that have been used in liveinstructional contexts. We then shift to areview of how DI has been used intechnology-based learning environments.Finally, we conclude with a look into thefuture of DI.
Rooted in behavioral theory, particularlywhat Skinner labeled the radical or selectionistbehaviorism (see, e.g., Skinner, 1953, 1966), thedirect instruction (DI) of Siegfried Engelmann(Bereiter & Engelmann, 1966) is now well into itsthird decade of influencing curriculum, instruc-tion, and research. It is also in its third decade ofcontroversy (c.f., Gersten, Baker, Pugach,Scanlon, & Chard, 2001).
To begin, we offer a definition and our stancerelated to DI—which has become the whippingpost in some pedagogical camps, while the pan-acea in others. For clarity, DI is not a lectureapproach (e.g., Freiberg & Driscoll, 2000). It is aninstructional model that focuses on the interac-tion between teachers and students. Key compo-nents of DI include “modeling, reinforcement,feedback, and successive approximations”(Joyce, Weil, & Calhoun, 2000, p. 337). Joyce andcolleagues specified the instructional designprinciples, which include the framing of learnerperformance into goals and tasks, breakingthese tasks into smaller component tasks,designing training activities for mastery, andarranging the learning events into sequencesthat promote transfer and achievement of pre-requisite learning before moving to moreadvance learning. Essentially, DI is “modelingwith reinforced guided performance” (Joyce etal., p. 337).
Our intent in this article is to explicate thegenesis, components, and permutations of DI asit has evolved in practice, and describe how it isbeing used in instructional technology. Threepurposes undergird this article. (a) First, webelieve that DI is a viable, time-tested instruc-tional model that plays an important role in a
comprehensive educational program. Theresearch indicates its usefulness in maintainingtime on task, the learning of skilled perfor-mance, and high rates of success when designedcorrectly (e.g., Fisher et al., 1980; Slavin, Mad-den, Dolan, & Wasik, 1996). Therefore, webelieve that instructional designers, softwaredesigners, teachers and the like ought to knowits foundation, essential components, historicaland current uses, and potential for designinginstruction that promotes student success forparticular instructional objectives. (b) Second,and related to the first, our experience with layfaculty (and some instructional technology prac-titioners) who design instruction, especiallyonline education, indicates a dearth of knowl-edge regarding the research and application ofDI. Over the past two decades, DI has been over-used by some, maligned by others, and fre-quently been wrongly equated with a purelecture approach. DI is not for all uses, objec-tives, or learners; no approach is. DI is a usefultool for the appropriate purpose, objectives, andcontext, and the appropriate learners. (c) Finally,while DI has maintained its core principles overtime, it has evolved in response to new under-standings about learners and learning. We willelaborate on these variations (e.g., expositoryteaching) and the research that indicates theirutility.
The DI model was created by Engelmann andhis colleagues in the 1960s at the University ofIllinois at Champagne-Urbana under a ProjectFollow Through grant. The research firstappeared in 1966 (Bereiter & Engelmann). Sci-ence Research Associates published the firstimplementation of the model known as DirectInstruction System for Teaching And Remedia-tion (DISTAR), programs that addressed begin-ning reading, language, and math (Engelmann& Bruner, 1969; Engelmann & Carnine, 1969;Engelmann & Osborn, 1969). Few models havebeen as researched as DI, including the largesteducational evaluation ever conducted compar-ing it with 12 other models, across nearly 30years, and involving nearly 75,000 students at180 sites. In that large evaluation (Bock,Stebbins, & Proper, 1977; Watkins, 1997), as innumerous studies (e.g., Madaus, Airasian, &Kellaghan, 1980; Rosenshine, 1970, 1971, 1985),
DI was found to be effective and superior toother models in everything from learningengagement to achievement to student affect.
As a selectionist model, DI is underpinned bythe basic notion that behavior, like physicalcharacteristics, evolves or is selected by the envi-ronment. Those behaviors that work are selectedby the consequences that follow the behavior.Since there are different consequences for thesame behavior in different environments,behaviors are situated in contexts. (It is impor-tant to note however that the cause of a behavioris not the context but rather the consequence, inthe same sense that high leaves do not cause agiraffe’s neck to grow. Rather the consequenceof longer neck mutations is to be able to eatleaves that few other animals can reach.)
Further, in behavioral-based models such asDI, it is assumed that learners must be active(behaving) to learn. In The Technology of Teaching,Skinner (1968) stated,
It is important to emphasize that a student does notpassively absorb knowledge from the world aroundhim but must play an active role, and also that action isnot simply talking. To know is to act effectively, bothverbally and nonverbally. (p. 5)
Moreover, in such models it is assumed thatlearning is universal, in the sense that the sameselectionist principles are involved in learningfrom planeria to people; from shoe tying to tyingoff the last suture in brain surgery (and every-thing in between). Selectionists might agree thatwhat we call higher level or higher order activi-ties may separate us from the rest of the animalkingdom, but they believe that the way we learnsuch things does not.
Further, behaviorally based models rejectlogical positivism, mentalisms such as mind(although not mental activity), and free will. Asmight be expected, rejection of such conceptscauses passionate reactions even today. Forexample, an editorial in Early Childhood Educa-tion Journal (Jalongo, 1999) reported the author’s(and her classmate’s) first reaction to a movie thatshowed the DI curriculum, DISTAR, as a “harsh,inflexible, and depersonalizing approach” (p.139) that she worried could resurface today. Shesaid that she would “like to see a stake driven inthe heart of DISTAR” (p. 139). Yet, in the same
editorial Jalongo conceded that DI “does have aplace”—that “it is the method of choice for low-level tasks such as learning to cut with scissorsor tying shoes” (p. 139). She also saw it as usefulfor special needs children.
The issue, then, is not whether selectionaccounts for at least some behaviors or whetherbehavioral approaches work with humans, theissue is whether some other type of learningevolved and “kicks in” that is unique to somehigher level behavior in humans. Indeedalthough many critics would argue against DI asa model for higher level learning or perfor-mance, the model does work well in situationswhere motor skills or prerequisite intellectualskills are involved (e.g., Gagné, 1985). Such pre-requisite skills might include learning suchthings as “mathematical procedures, grammati-cal rules, the states of New England, alphabetiz-ing, carburetor overhaul, scientific equations,and the periodic table of elements to name a few”(Gunter, Estes, & Schwab, 1999, p. 79). Moreover,as Gunter et al. put it, “every teacher, in everysubject, at every level of schooling has somelearning objectives related to basic skills thatmust be mastered before the learner can move toother levels of thinking and learning” (p. 79).
In fact, DI has re-emerged in recent years as aviable instructional strategy that can be situatedsuccessfully within a range of tools that promotea range of types of learning within contempo-rary learner-centered pedagogy (e.g., Eggen &Kauchak, 2001; Gersten et al., 2001; Schwartz &Bransford, 1998; Tharp & Gallimore, 1988). Forexample, Schwartz and Bransford reported theirresearch that illustrates that there is a “time fortelling” within a problem-based learningapproach. Gersten and colleagues describedhow recent theoretical frameworks have helpedto elaborate the conceptualization of DI, andoffered examples of contemporary DI researchthat focus on the learning of explicit strategies,concepts, and higher order thinking skills.Tharp and Gallimore situated DI within a rangeof strategies that comprise their “teaching asassisted performance” model. Eggen andKauchak asserted that “the model can also beused to teach other forms of content such asgeneralizations, principles and academicrules” (p. 287).
Nowhere is DI more evident than in com-puter-mediated learning environments. Fromcomputer-aided instruction to distance learningexperiences, the basic tenets of DI are infused—with greater and lesser fidelity. And, althoughDI is no longer the most prominent instructionalframework for the overall design of computer-mediated applications (c.f., Cognition and Tech-nology Group at Vanderbilt, 1996), DI is thestrategy of choice when the learning objectiverequires that the learners have direct practice inwhat must be done, or said, or written (Cazden,1992).
Consequently, our purpose here is not to pro-mote DI as the only instructional framework topromote learning, either in live or computer-mediated learning environments. Our purposeis to present the DI model along with the notionthat the designer can and should use the modeleffectively based on appropriate assessment ofthe learners, content, context, and task at hand(Shambaugh & Magliaro, 1997). To accomplishour goal, we begin with a general discussion ofthe basic DI framework, followed by a historicaltrace of exemplar DI models that have beenstudied and used in live instructional contextsover the past 30 years. We then shift to a reviewof how DI has been used in technology-basedlearning environments. Finally, we concludewith a look into the future of DI.
DI: HISTORY, CONCEPTS, AND MODELS
DI has been used to describe a range of instruc-tional models used in face-to-face learning con-texts—all designed to promote on-task studentbehavior by the teacher’s effort to monitor andcontrol student classroom attention and persis-tence (Corno & Snow, 1986). The various modelshave emerged from primarily behavioral tradi-tions; however, over time the models havereflected the prevailing theoretical orientation toand interpretation of teacher-directed actions ina classroom. Moreover, these models may not beentitled DI per se, but share key components(e.g., Tobias, 1982) that translate very well intodesign features of live, as well as technology-enhanced or technology-driven, instruction.These components are:
1. Materials and curriculum are broken downinto small steps and arrayed in what isassumed to be the prerequisite order.
2. Objectives must be stated clearly and interms of learner outcomes or performance.
3. Learners are provided with opportunities toconnect their new knowledge with what theyalready know.
4. Learners are given practice with each step orcombination of steps.
5. Learners experience additional opportunitiesto practice that promote increasing responsi-bility and independence (guided and/orindependent; in groups and/or alone).
6. Feedback is provided after each practiceopportunity or set of practice opportunities.
The fundamental design principle that con-nects these components is the fact that learnersare actively engaged in the relevant curriculumin order to build knowledge, skills, and disposi-tions related to the goals and objectives of thelesson. This frequent opportunity-to-respondenables ongoing assessment and correctionwhen needed (Delquadri & Greenwood, 1981).The clear goal of this model is that learners willdevelop mastery and automaticity of the targetskills, knowledge, and dispositions.
A number of empirically supported DI mod-els have appeared in the literature in the lastfour decades. Our historical tour through someof the major models describes the nuances ofeach model and illustrates the richness of DI as auseful instructional strategy across a range oflearning environments. The order of model pre-sentation also reveals a bit of the evolution of DIthat was prompted by the ongoing research onlearning and instruction. We begin with thework of Bereiter and Engelmann (1966). Theseresearchers are often credited with pioneeringthe research on DI, which, at that time, wasbased on principles of behavioral psychologyincluding overt responding, frequent and spe-cific feedback, and contingency management.The next stop in the tour is with exemplar mod-els developed from the effective teachingresearch (e.g., Brophy & Good, 1986), whichbegan to merge the well-established theoreticalconcepts from behavioral psychology with thenewly instantiated principles based on the
research using the information processingmodel of human memory. These models, basedin process-product research, served as the foun-dation of thousands of studies relating teacherbehavior and student achievement in the 1970sand 1980s (e.g., Anderson, Evertson, & Brophy,1979), and continue to current practice withSlavin’s “Success for All” program (Slavin et al.,1996). Next, we address the work of RobertGagné (1977, 1985), whose development ofevents of instruction clearly situated teacher-ledmodels into cognitive psychology and theinstructional design literature. We close this sec-tion with a discussion of variations on the DImodel that includes expository instruction(Jacobsen, Eggen, & Kauchak, 1993) and teach-ing-as-assisted-performance (Tharp &Gallimore, 1988). These variations representimportant examples of the evolution of the DImodel—ones that appropriately advance themodel in light of new understandings of howpeople learn and how to design learning envi-ronments that meet stated learning objectives.
Behaviorally Based Models
Bereiter and Engelmann (1966) and Engelmann(1980) designed their DI approach to be “themost efficient way to teach each skill”(Engelmann, p. xi). The premise was that learn-ers are expected to derive learning that is consis-tent with the presentation offered by the teacher.Learners acquire information through choice-response discriminations, production-responsediscriminations, and sentence-relationship dis-criminations. The key activity is for the teacherto identify the type of discrimination required ina particular task, and design a specific sequenceto teach the discrimination so that only theteacher’s interpretation of the information ispossible. Rapid questioning, frequent testing,continuous interaction, and positive reinforce-ment are all key instructional tools that promotelearning. The perspective of Bereiter andEngelmann was that DI is sufficiently broad ininterpretation to serve as a teaching approachthat has an unlimited number of applications.Over the course of 15-plus years, Engelmann’sDI model framed such successful programs as
DISTAR (e.g., Engelmann & Osborn, 1972), Proj-ect Follow Through (Nero & Associates, 1975;Stallings & Stipek, 1986), and the Tucson earlyeducation model (Rentfrow, 1972).
The initial DI model (Bereiter & Engelmann,1966) established a three-stage, systematic teach-ing design driven by continuous assessment oflearning. The general process included (a) anintroduction to the new content to be learned, (b)the main presentation of the lesson, and (c) prac-tice with immediate feedback. At first, practicewould be teacher directed, with the entire classresponding to quickly paced, strategicallysequenced questions from the instructor. Oncethe teacher was certain that the students wereready to apply the newly learned concepts, thestudents were shifted to independent practice,closely monitored by the teacher to ensure onlycorrect interpretations and applications of thetargeted content. This approach—introductionof new concepts, interactive presentation andapplication of the concepts, and guided prac-tice—would serve as a standard for future varia-tions on the DI model. Table 1 compares themodels, highlighting the consistency of instruc-tional procedure and distinguishing features.
Engelmann’s (1980) DI model “attempts tocontrol every variable in the teaching environ-ment” (p. 80) through scripted tasks and lessons.This releases the teacher to focus on:
• The presentation and communication of theinformation to children.
• Students’ prerequisite skills and capabilitiesto have success with the target task.
• Potential problems identified in the task anal-ysis.
• How children learn, by pinpointing learnersuccesses and strategies for success attain-ment.
• Learning how to construct well-designedtasks.
The students would increase their self-esteemand self-confidence through their academicachievement, providing motivation for subse-quent tasks. Berieter and Engelmann’s (1966)research made important contributions to edu-cational research by illustrating how studentsfrom disadvantaged homes were able toincrease language and school success through
an increased opportunity to respond. In essence,the students succeeded because of the high rateof feedback and subsequent responding. Essen-tially, the idea was that “success breeds success”(Stallings & Stipek, 1986).
The data-driven nature of the DI model, withfrequent opportunities for student response andteacher feedback, reflects the integration of con-tinuous assessment throughout this design.Behavioral assessments of learning focus on thecollection of data related to learning outcomes(Schunk, 2000), that is, how the learner’s behav-ior changes as a result of the instruction. DI les-sons rely on several inherent approaches to datacollection in order for the teacher to monitor stu-dent learning. Oral responses in group-basedinteractions provide the instructor with infor-mation related to how well students are grasp-ing the targeted content, as well as correctingany misconceptions so that only accurate inter-pretations of the new concept are taught. Writ-ten performances and direct observations allowthe teacher to gauge progress and assess thelearner’s ability to apply the newly acquiredconcepts during independent practice. Suchemphasis on continuous assessment suggeststhat DI may have been one of the first teachingmodels to incorporate data-based decision mak-ing, as the teacher based choices related to pre-sentation strategies, timing, examples, andpractice readiness on student response data.
Engelmann (1980) acknowledged two poten-tial problems with this model: external attribu-tions for success, and the need to experience themodel for at least one year in order to accruelong-term benefits.
The “Effective Teaching” DI Models.
During the 1970s and through the mid-1980s, anumber of variations and elaborations ofEngelmann’s (1980) model were designed andtested using a process-product paradigm andcharacterized in the well-known effective teach-ing literature. The general teaching procedureacross these models was to begin with sometype of opening activity, next to enact the mainlesson presentation, and then to give studentsopportunities for practice. Three specific modelsreported high success rates and were widely
Table 1 Comparison of components across three Effective Teaching models.
Basic Direct Instruction Engelmann’s Direct Instruction model Rosenshine’s Explicit Teaching model
Introduction 1. Introduction of new concept based 1. Review: Review homeworkon previously mastered skills and Review relevant previous learningknowledge Review prerequisite skills and
knowledge for the lesson
Main Presentation 2. Presentation: Fast-paced, scripted 2. Presentation:of the Lesson explanation or demonstration State lesson goals and/or provide
designed to elicit only one interpre- outlinetation of concept. The target concept Teach in small stepsmust be reinforced with Model proceduresappropriate examples and Provide concrete positive and negative nonexamples. examples
Use clear languageCheck for student understandingAvoid digressions
Practice 3. Students are provided with 3. Guided practice: More time High opportunities to verbally respond, frequency of questions or guided either through a set of questions or practicetasks, in order to indicate their All students respond and receive learning of the concept and their feedbackability to connect it to further High success rateexamples. Continue practice until students are
4. Feedback: Teacher either confirms fluidcorrect student response or provides 4. Corrections and feedback:corrections and repetition of the Give process feedback when answers missed items. are correct but hesitant
5. Independent practice: After group Give sustaining feedback, clues, or work, students engage in self- reteaching when answers are incorrectdirected practice in workbooks. Reteach when necessaryTeacher monitors progress and 5. Independent practice Students receive provides guidance when needed. help during initial steps or overview
Practice continues until students are automatic (where relevant)Teacher provides active supervision (where possible)Routines are used to give help to slower students
Basic Good & Grouw’s Direct Instruction Strategies for Effective Teaching model Hunter’s Design of Effective Lessons model
Introduction 1. Daily review (first 8 minutes except 1. Anticipatory set:Mondays): Provide a mental set that causes Review concepts and skills associated students to focus on what will be with the homework learned Use to glean diagnostic infor- Collect and deal with homework mation about students’ ability to assignments connect with topicAsk several mental computation 2. Objective and purpose:exercises Present objective to students to clearly
communicate what they are supposed to learn from the lessonPresent purpose to students so they know why the information is relevant to them
Main Presentation 2. Development (about 20 minutes): 3. Input:of the Lesson Briefly focus on prerequisite skills and Conduct a task analysis on final
concepts objective to determine the knowledge Focus on meaning and promoting and skills that need to be acquiredstudent understanding using lively Use pedagogies that will facilitate the explanations, demonstrations, process kinds of learning intended (e.g., explanations, illustrations, etc. discovery, discussion, reading, Assess student comprehension using listening, lecture, observation)process/product questions (active 4. Modeling:interaction); using controlled practice Demonstrate the processes and products Repeat and elaborate on the meaning that facilitate learning—these can be portion as necessary live or filmed, but must enable
students to perceive directly what is to be learned
Practice 5. Seatwork (about 15 minutes): 5. Checking for understanding: Provide uninterrupted successful Determine if the students understand practice what they are supposed to do in the Momentum—keep the ball rolling— lesson’s task through questioningget everyone involved, then sustain 6. Guided practice: Practice the new involvement knowledge or skill under direct teacher Alerting—let students know their work supervisionwill be checked at end of period 7. Independent practice: Assigned only Accountability—check the students’ after teacher is reasonably sure that work students will not make serious errors
6. Homework assignment:Assign on a regular basis at the end of each math class except FridaysShould involve about 15 minutes of work to be done at homeShould include one or two review problems
7. Special reviewsWeekly review & maintenance: conduct during the first 20 minute each Monday, focus on skills and concepts covered during the previous weekMonthly review & maintenance: conduct every fourth Monday, focus on skills and concepts covered since the last monthly review
integrated into practice in K–12 settings: (a)Rosenshine’s (1979) explicit teaching model,Good and Grouws’s (1979) strategies for effec-tive teaching model, and Hunter’s (1982) designof effective lessons model. In addition to repre-senting DI models that are supported by soundresearch and practice, these models wereselected for this chapter because they representDI variations that were sensitive to contextualneeds, discipline, and the changing landscape ofeducational theory and practice. Rosenshine’smodel was designed to be sensitive to differ-ences in student ability and complexity of sub-ject matter. Good and Grouws’s model wasdesigned for the teaching of mathematics. Thismodel is included to illustrate how DI was mod-ified for a particular subject matter context.Hunter’s model was designed to incorporate thenew cognitive principles, as well as to becomemore user-friendly for K–12 teachers with acloser alignment with well-established educa-tional practices.
Rosenshine’s explicit teaching model. The centraltheme in Rosenshine’s (1979) and Rosenshineand Stevens’s (1986) model is that teachers needto enact intentionally clear and well-defined les-sons. The six functions of each lesson include (a)review, (b) presentation, (c) guided practice, (d)corrections and feedback, (e) independent prac-tice, and (f) weekly and monthly reviews. Themajor instructional strategies include teachingin small steps with student practice after eachstep, guiding students during initial practice,and providing all students with a high level ofsuccessful practice. The specific steps of thisteaching model are listed in Table 1.
Note the strong parallels between Engel-mann’s and Rosenshine’s models. Both have aclear emphasis on frequent teacher-studentinteraction to present information, ask ques-tions, guide practice, and provide feedback andreinforcement. Rosenshine, in extending themodel, added guidelines to suit different stu-dents and difficult material. To meet studentneeds, Rosenshine suggested the teacher pro-vide slower students with more review, less pre-sentation, more guided practice, and moreindependent practice. For faster students, hesuggested less review, more presentation, lessguided practice, and less independent practice.
Rosenshine also brought attention to theneed to modify lessons based on the material orcontent to be taught. His modification for diffi-cult content emphasized additional monitoring,with the lesson cycle focused on presentation,guided practice, and supervised independentpractice.
Good and Grouws’s strategies for effective teachingmodel. Focusing on the teaching and learning ofmathematics, Good and Grouws’s (1979)research resulted in a scripted procedure thatincluded both instructional and managementstrategies. While following the basic DI proce-dure, this model offered suggested lesson man-agement strategies and time allotments for eachphase of the lesson, including weekly andmonthly practice intervals (see Table 1).
Although there are clear alignments with theaforementioned DI procedures, Good andGrouws’s (1979) model began to emphasize thecognitive dimension of learning. In Table 1, notethe focus on meaning and conceptual under-standing of mathematics, along with the devel-opment of automaticity with computation andprocedures. Good and Grouws’s (1981) researchindicated that teachers who used this modelused more problem-solving procedures. More-over, there were significant differences in favorof the students who were taught using thismodel in terms of problem-solving scores andgains in achievement.
Hunter’s design of effective lessons model. Mad -eline Hunter became a “household name” inteacher professional development in the 1980s.She developed and disseminated a widelyknown and used teaching model that mergedthe well-engrained and more highly regardedfeatures of DI with more current ideas and ver-biage from cognitive psychology. Her lessoncycle model (Hunter, 1976; 1982) became thecenterpiece of K–12 professional developmentand teacher evaluation programs because of itsresonance with practitioners and ease of use foradministrators who were conducting observa-tional evaluations of teachers.
The Hunter model follows the basic DI proce-dure (see Table 1), however, she connectsobservable behavior with internal processinginferences. For example, her initial phase is
called an “anticipatory set” and provides theintroduction to the lesson by trying to connectwith a “mental set” that the children alreadyhold. When she explains the nuances of model-ing, Hunter alerts teachers to pay attention tostudent perception and knowledge acquisition.Hunter’s model represents a direct instructionalmodel that was designed for practitioners whowere trying to infuse the concepts and terminol-ogy used by cognitive psychologists into theexisting strategies of earlier DI models.
Gagné’s Events of Instruction Model
Robert Gagné (1977, 1985) made enormous con-tributions to the instructional theory literature.His work consistently merged current learningtheory and instructional practice. His events ofinstruction model has provided instructional
designers with a framework for creating instruc-tional lessons in which every component speaksdirectly to empirically based principles of learn-ing. Although the major components of Gagné’smodel fit the basic DI procedure (i.e., introduc-tion, presentation, practice), the full model pro-vides clear direction for lesson design and, morerelevant for this discussion, the design of tech-nology-enhanced or technology-driven instruc-tion. However, whereas Gagné recommendedfollowing the sequence of events as published,Frieberg and Driscoll (2000) purported that thesequence could be modified based on the needsof the learner, context, and content.
Table 2 outlines Gagné’s (1977, 1985) eventsorganized according to the basic DI procedure,and elaborated with the learning processes andprinciples that his events support. It is for thisreason that we separate Gagné’s model from the
Table 2 Gagné’s Events of Instruction model.
Basic Connections with Possible Connections with Direct Instruction Events of Learners and Design Features of Procedure Instruction Instruction Instructional Technology
Introduction 1. Gaining attention Motivation phase: Attention gained through use 2. Informing the learner expectancy of auditory and/or visual
of the objective Apprehending phase: stimuliattention and Presentation of information selective perception through appropriate media
types
Main Presentation 3. Stimulating recall of Acquisition phase: Prerequisite knowledge can of the Lesson prerequisite learning coding and storage be assessed through quizzing
4. Presenting the stimulus entry or assessment toolsmaterials Retention phase: Information presented through
5. Providing learner memory storage appropriate media typesguidance
Practice 6. Eliciting the performance Recall phase: retrieval Guided practice can be 7. Providing feedback with Generation phase: teacher-led through tele-
performance correctness transfer communications or self-8. Assessing the perform- Performance phase: paced through computer-
ance responding based interactions.9. Enhancing retention and Feedback phase: Corrective feedback can be
transfer reinforcement provided, based on learners’ responsesAutomated remediation can be designed into programRetention and reinforcement through application of new knowledge in scenarios through various media types
effective teaching models and highlight hiswork as foundational to the design of instruc-tional technologies. His work reflects a blendingof the behavioral and cognitive frameworks, andfor these authors, serves as a bridge across per-spectives on learning. Also in Table 2 are specificexamples from instructional software or dis-tance education course activities that supporteach event.
Expository Teaching
Expository teaching is a teacher-centeredapproach to learning content that parallels thegoals and features of its predecessor, DI(Jacobsen et al., 1993), yet clearly is orientedtoward cognitive- or information-processing–based learning. Rather than strengthening stu-dent behaviors, this model is designed tostrengthen students’ cognitive structures (Joyceet al., 2000). Ausubel (1968) is often cited as theoriginator of this model (e.g., Freiberg &Driscoll, 2000).
Expository teaching is used in order to helpstudents learn concepts, principles, generaliza-tions, and rules. Aligned with the DI tradition,the two major advantages of the model are time
and control. Lesson objectives are clearly deline-ated, and questioning is convergent to ensurethat the objectives are met. Examples areplanned to ensure that students are gradually“scaffolded” toward the target concept, abstractrelationship, or generalization that was articu-lated in the objective.
The teacher follows a sequence of steps thatare designed to bring students closer and closerto defining the concept in terms that make senseto them (Eggen & Kauchak, 1993). Concepts areclarified through the development of definitionsand connections with students’ prior knowl-edge. Active participation is encouraged toensure that the teacher can assess student prog-ress. Students must provide their own examplesto promote practice. Feedback is renderedimmediately to ensure that misconceptions arenot developed. Table 3 illustrates the compo-nents of expository teaching as they relate to DI.
Teaching as Assisted Performance
Research that emphasizes the social constructionof knowledge has continued to include and elab-orate on the contribution that DI makes to thecreation of successful learning environments
Table 3 Current direct-instruction–related models.
Basic Direct InstructionProcedure Expository Teaching Teaching as Assisted Performance
Introduction 1. Visual presentation of targeted Instructionconcept, abstraction, or generalization
2. Inform learner of intended learning outcome
Main Presentation of 3. Define concepts, abstractions, Modelingthe Lesson or generalizations Questioning
4. Link to prior knowledge Feeding-back5. Provide positive and negative Contingency management
examples
Practice 6. Classify or explain teacher examples Cognitive structuring Types 1 & 27. Provide additional examples Feeding-back
(e.g., Tharp & Gallimore, 1988). The theoreticaland empirical work of Vygotsky (1978) has iden-tified two critical elements of DI as essential tolearning from a social perspective (Eggen &Kauchak, 2001). Vygotsky’s use of the notion ofscaffolding and his construct of the zone of prox-imal development (ZPD) are instantiated in theteaching as assisted performance model (Tharp& Gallimore). Although the model expandsbeyond scaffolding and the ZPD1, these con-cepts are addressed here to illustrate the impor-tant elaborations of DI into a new perspective onlearning and teaching.
For Vygotsky (1978), scaffolding refers to theinstructional support provided to students asthey learn new skills, content, and dispositions.Information is broken down into manageable,smaller chunks of recognizable knowledge;skills are broken down to subskills to ensure asequential, step-by-step acquisition of the targetobjectives aided by teacher guidance, question-ing, hints, and so forth. Essential in this processis a task analysis that thoroughly examines whatis to be learned, and the trajectory of the devel-opment of knowledge to meet that objective.
The ZPD is, according to Vygotsky (1978), the“distance between the actual developmentallevel as determined by individual problem solv-ing and the level of potential development asdetermined through problem solving underadult guidance or in collaboration with morecapable peers” (p. 86). Eggen and Kauchak(2001) asserted that the ZPD is “instructionalpaydirt” (p. 278) in that it is within this time,place, and space that teachers are most effectivein helping students learn. From a DI perspective,teachers are striving to meet each student withinthe zone by a clear analysis of the task, constantassessment of understanding and provision ofsupport when and as needed, and practice firstwith the teacher, then with peers, then indepen-dently. Tharp and Gallimore’s (1988) notion of
teaching as assisted performance makes explicitthe need for teachers to directly plan and inter-vene with teacher-directed instruction based onstudent needs as evidenced in their practice. SeeTable 3 for the explicit connections betweenteaching as assisted performance and DI.
Summary
The DI model has enjoyed a more than 30-yearhistory of framing successful learning experi-ences. The model has evolved to address currentunderstandings about learners and learning, butmaintains the central purpose of promoting stu-dent on-task behavior through explicit instruc-tion, ongoing support, and student engagementin successful practice. The DI model is wellsuited to the design of technology-enhanced andtechnology-based instruction because of its clearstructure and potential for providing learnerswith opportunities for practice and immediatefeedback, especially in asynchronous learningenvironments.
TECHNOLOGICAL APPLICATIONS OF DI
DI continues to hold potential as an effectiveteaching method, particularly in technology-mediated learning environments. Computer-based programs have been designed to modelinstructor-led DI approaches while leveragingthe technological ability to provide feedback,remediation, and guided practice, all essentialcomponents of the DI process and all of whichcontribute to its effectiveness. The following sec-tion provides examples of computer-basedimplementations of DI that demonstrate the par-ticular advantages of technology to instantiatethis model.
Successful Applications ofComputer-Mediated DI
One of the first technology-based programs toimplement the DI approach was developed bythe originators of the DI method. Core Concepts,a reading, math, and language videodisc pro-
1 Tharp and Gallimore (1988) identified six components totheir teaching model: (a) modeling, (b) contingencymanaging, (c) feeding back, (d) instructing, (e) questioning,and (f) cognitive structuring. Although all these teachingstrategies have clear connections to DI, the two selected forthis article are highlighted to illustrate how DI hastranscended current theory on learning and emerged as a keymodel for the design of successful learning environments.
gram, was developed by the originators of theDI approach (Hofmeister, Engelmann, & Carn-ine, 1986, 1988). In this instructional program,brief segments with narration and animationwere used to break down complex skills intosmall steps, model problem-solving strategies,present a wide range of examples, review rele-vant preskills, provide discrimination practiceand cumulative review, and frequently assessstudent learning through weekly progresschecks. Five experimental studies over sevenyears have demonstrated the program’s effec-tiveness for low-achieving students. A naturalis-tic study found that the constant review wasessential for low-achieving students, and thevariety of activities within each lesson helpedkeep students interested and motivated (Adams& Engelmann, 1996).
Another basis for the creation of such pro-grams is the teaching of complex skills or sub-jects, an example of which is a programdesigned to teach the solution of mathematicalword problems (Steele & Steele, 1999). ProjectDiscover is an intelligent tutoring system com-prising 11 independent programs, based on theDI instructional approach. As students workthrough each program, the system collects infor-mation about their performance and makes rec-ommendations regarding sequencing andpractice options. In accordance with the DImodel, the first program provides a pretest toassess their existing knowledge and an introduc-tion to the process of solving word problems.The next three programs teach the eight stepsinvolved in the solution of word problems, eachaspect incorporating practice and correctivefeedback. The next five programs provide prac-tice opportunities particularly related to theeight steps, with problems automatically basedon the steps that the learner finds problematic.The next program gives students in-depth prac-tice, with incorrect responses being met withhints or coaching. Successful performance onthis program (more than 90% correct) directs thelearner to the posttest, less than 90% correctleading students to more practice based on theirindividual needs.
Merging Pedagogies ThroughTechnology
Although viewed as an “instructive” designstrategy (Rieber, 1992), DI can be combined withmore open-ended strategies to provide dynamicand meaningful learning experiences (Fitzger-ald & Semrau, 1998; Rieber; Sfondilias & Siegal,1990). Fitzgerald and Semrau described TheClassroom Behavior Record, a hypermedia pro-gram that implements DI methods for trainingeducators and health care professionals in obser-vational skills. While the use of hypermedia toengage students in DI may sound paradoxical,the program is based on the stages of learningmodel (Gagné, 1977) and facilitates learnerprogression through the hierarchical phases(acquisition, fluency, generalization, and main-tenance) with the inherent flexibility of a nonlin-ear system. Rieber contended that constructivistand instructivist strategies are not mutuallyexclusive, and described how the two are inte-grated within a microworld program calledSpace Shuttle Commander, a computer-basedlearning environment modeled after the LOGOsystem (Papert, 1980).
Sfondilias and Siegal (1990) utilized a uniquecombination of DI and discovery methods in acomputer-based program to teach learners theprocess to determine equations for parabolicgraphs. Learners are guided through the cogni-tive routine for creating the correct equation, andthen presented with an exploration situation inwhich they attempt to make their graph resemblea target graph, based on their inputs into the cor-relating equation. Feedback allows them to learnfrom their mistakes, and errors in the cognitiveroutine will prompt the system to repeat the rou-tine until the learner has mastered it.
Recent Iterations of DI
The current emphasis on accountability andhigh-stakes testing in education has opened thedoor for commercial software products based onthe DI model. Courseware packages such asSuccessMaker Enterprise by Pearson DigitalLearning (http://nclb.pearsonedtech.com) andPLATO Learning (http://www.plato.com)claim to use DI methods to address student per-
formance mandates of the No Child Left Behind(NCLB) Act. Both packages provide pre- andpostassessment of learners with integrated prac-tice and additional assessments throughout theinstructional programs, designed to closelymonitor student progress and customize thelearning experience based on individual needs.While PLATO has evolved over the past 40 yearsfrom a large-scale mainframe program to a com-prehensive courseware system for K–12 andadult education, it has maintained the aforemen-tioned features that signify the DI approach toinstruction.
The Future of DI
Recently, DI seems to have fallen out of favor interms of philosophical trends of learning andinstruction (Duffrin, 1996; Edmondson & Shan-non, 2002). However, this model still serves as aviable and effective teaching approach in manyclassroom settings, and has been shown toincrease students’ problem-solving skills (Good& Grouws, 1981). Maintaining the tradition ofbeing one of the most empirically tested forms ofinstruction, research on the DI model continues(i.e., Cashwell, Skinner, & Smith, 2001; Swanson,2001; Viadero, 2002). In an effort to support thedissemination of such studies, the Journal ofDirect Instruction was established in 2001 as apeer-reviewed forum to disseminate contempo-rary research regarding DI (Slocum & Marc-hand-Martella, 2001).
With the exponential growth of distance edu-cation, DI holds potential as a teaching methodthat can be effectively implemented on a widescale in distributed learning environments, par-ticularly through Web-based instruction. Aspreviously described, computer-based instruc-tion can efficiently execute all phases of the DIapproach in an individualized and self-pacedmanner. However, networked systems can addincreased flexibility in that instruction can becomputer based, instructor led, or a combinationof both. Asynchronous course template systemssuch as Blackboard or WebCT possess the abilityto conduct preassessments of knowledge, pres-ent content information in a variety of formats,and provide varied levels and types of practiceand postassessment with corrective feedback to
customize the experience to the needs of theindividual learner. Synchronous systems suchas CentraOne allow for assessment of contentknowledge prior to engagement in a live, audio-conference with a teacher. In such a session, theinstructor has the ability to present relevantinformation verbally, supported by a variety ofvisual tools, such as a shared whiteboard, text-based chat spaces, and software applications.Conferencing systems such as these also allowthe teacher to conduct real-time questioning andprovide appropriate feedback, addressing thehuman component often missing from asyn-chronous instruction. The tool’s capacity to sup-port asynchronous practice is based on quizzingand testing instruments designed by the instruc-tor, which can also provide automated, correc-tive feedback. Features of the aforementionedtypes of Web-based systems can be blended tooffer a hybrid approach to DI, in which someaspects are live and some asynchronous,depending on the needs and constraints of theparticipants.
DI will likely see the pendulum swing back toits favor in the near future, especially given fed-eral and state mandates related to standards-based performance in schools. Advances inlearning technologies are ready to support andimplement this long-standing teaching methodin more efficient and personalized ways. Stand-alone computer-based systems offer the flexibil-ity of either supplementing DI in the classroomor providing entire self-contained units ofinstruction. Networked distance delivery sys-tems supply the same possibilities to geographi-cally dispersed learners, with the added abilityof interacting with an instructor either synchro-nously or asynchronously. When the instruc-tional task calls for the teaching of discrete skillsand knowledge in an interactive and guided for-mat, DI remains a proven approach. With theability to exemplify the DI method throughinnovative, mediated experiences, instructionaltechnology may hold the key to the continuingevolution of the DI method.
Susan G. Magliaro [[email protected]] and Barbara B.Lockee are Associate Professors and John K. Burtonis Professor, all in the Department of Teaching andLearning at Virginia Tech.
Adams, G. L., & Engelmann, S. (1996). Research onDirect Instruction: 25 Years beyond DISTAR. Seattle:Education Achievement Systems.
Anderson, L. M., Evertson, C. M., & Brophy, J. E.(1979). An experimental study of effective teachingin first grade reading groups. The Elementary SchoolJournal, 79, 527–32.
Ausubel, D. P. (1968). Educational psychology: A cogni-tive view. New York: Holt, Rinehart, and Winston.
Bereiter, C., & Engelmann, S. (1966). Teaching disadvan-taged children in the preschool. Upper Saddle River,NJ: Prentice Hall.
Bock, G., Stebbins, L., & Proper, E. (1977). Education asexperimentation: A planned variation model (Volume IV-A & B). Effects of Follow Through models. Washington,D.C.: ABT Associates.
Brophy, J. E., & Good, T. L. (1986). Teacher behaviorand student achievement. In M. C. Wittrock (Ed.),Handbook of research on teaching (3rd ed.) (pp. 328–375). New York: Macmillan.
Cashwell, T. H., Skinner, C. H., & Smith, E. S. (2001).Increasing second-grade students’ reports of peers’prosocial behaviors via direct instruction, groupreinforcement, and progress feedback: A replicationand extension. Education and Treatment of Children 24(2), 161–175.
Cazden, C. B. (1992). Whole language plus: Essays on lit-eracy in the United States & New Zealand. New York:Teachers College Press.
Cognition and Technology Group at Vanderbilt.(1996). Looking at technology in context: A frame-work for understanding technology and educationresearch. In D. C. Berliner & R. C. Calfee (Eds.),Handbook of educational psychology (pp. 807–840).New York: Macmillan.
Corno, L., & Snow, R. E. (1986). Adapting teaching toindividual differences among learners. In M. C.Wittrock (Ed.), Handbook of research on teaching (3rded.) (pp. 605–629). New York: Macmillan.
Delquadri, J., & Greenwood, C. (1981). The importanceof “opportunity to respond” in the education of theblack minority students. Paper presented at TheCouncil for Exceptional Children Conference on TheExceptional Black Child, New Orleans, LA. (ERICDocumentation Reproduction Service No.ED204893)
Duffrin, E. (1996). Direct instruction making waves.Catalyst 8 (1), 1–11. Available: http://www.cata-lyst-chicago.org/09-96/096main.htm. Accessed onJuly 3, 2002.
Edmondson, J., & Shannon, P. (2002). The will of thepeople. The Reading Teacher 55, (5), 452–454.
Eggen, P. D., & Kauchak, D. P. (2001). Strategies forteachers: Teaching content and thinking skills. Boston:Allyn & Bacon.
Engelmann, S. (1980). Direct instruction. EnglewoodCliffs, NJ: Educational Technology.
Engelmann, S., & Bruner, E. C. (1969). DISTAR reading
Engelmann, S., & Carnine, D. (1969). DISTAR arithme-tic I (Teacher’s presentation book, student material,and teacher’s guide). Chicago: Science ResearchAssociates.
Engelmann, S., & Osborn, J. (1969). DISTAR language I(Teacher’s presentation book, student material, andteacher’s guide). Chicago: Science Research Associ-ates.
Engelmann, S., & Osborn, J. (1972). DISTAR languageprogram. Chicago: Science Research Associates.
Fisher, C. W., Berliner, D. C., Filby, N. N., Marliave, R.,Ghen, L. S., & Dishaw, M. M. (1980). Teachingbehaviors, academic learning time, and studentachievement. In C. Denham & A. Lieberman (Eds.),Time to learn. Washington, DC: National Institute ofEducation.
Fitzgerald, G. E., & Semrau, L. P. (1998). Hypermediaand direct instruction: Do the paradigms fit? A demon-stration of a learner-centered hypermedia program inclassroom observation skills. Paper presented at theannual meeting of the American EducationalResearch Association (San Diego, CA, April 13–17,1998). (ERIC Document Reproduction Service, ED419527)
Frieberg, H. J., & Driscoll, A. (2000). Universal teachingstrategies (3rd ed.). Boston: Allyn & Bacon.
Gagné, R. M. (1977). The conditions of learning (3rd ed.).New York: Holt, Rinehart & Winston.
Gagné, R. M. (1985). The conditions of learning (4th ed.).New York: Holt, Rinehart & Winston.
Gersten, R., Baker, S., Pugach, M., Scanlon, D., &Chard, D. (2001). Contemporary research on specialeducation teaching. In V. Richardson (Ed.), Hand-book of research on teaching (4th ed.) (pp. 695–722).Washington, DC: American Educational ResearchAssociation.
Good, T. L., & Grouws, D. A. (1979). The Missourimathematics effectiveness project. Journal of Educa-tional Psychology, 71, 355–362.
Good, T. L., & Grouws, D. A. (1981). Experimentalresearch in secondary mathematics classrooms:Working with teachers. Final Report. Washington,DC: National Institute of Education. (ERIC Docu-ment Reproduction Service No. ED219261)
Gunter, M. A., Estes, T. H., & Schwab, J. (1999). Instruc-tion: A models approach (3rd ed.). Boston, MA: Allyn& Bacon.
Hunter, M. (1976). Improved instruction. El Segundo,CA: Theory Into Practice.
Hunter, M. (1982). Mastery teaching. El Segundo, CA:Theory Into Practice.
Hofmeister, A. M., Engelmann, S. & Carnine, D. (1986).Mastering fractions, Core concepts in science and mathe-matics videodisc program. Washington, DC: SystemsImpact Inc.
Hofmeister, A. M., Engelmann, S., & Carnine, D.(1988). Developing and validating science education
Hunter, M. (1982). Mastery teaching. El Segundo, CA:Instructional Dynamics.
Jacobsen, D., Eggen, P., & Kauchak, D. (1993). Methodsfor teaching; A skills approach (4th ed.). Columbus,OH: Merrill.
Jalongo, M.R. (1999). Editorial: On behalf of children.Early Childhood Education Journal 26,(3), 139–141.
Joyce, B., Weil, M., & Calhoun, E. (2000). Models ofteaching (6th ed.). Boston: Allyn & Bacon.
Madaus, G. F., Airasian, P. W., & Kellaghan, T. (1980).School effectiveness: A review of the evidence. NewYork: McGraw-Hill.
Nero and Associates. (1975). A description of followthrough sponsor implementation processes. Portland,OR: Nero & Associate, Inc. (ERIC Document Repro-duction Service No. ED128926)
Papert, S. (1980). Mindstorms: Children, computers, andpowerful ideas. New York: Basic Books.
Rentfrow, R. K. (1972). Intensive evaluation of Head Startimplementation in the Tucson early education model.Arizona Center for Educational Research and Devel-opment, University of Arizona. (ERIC DocumentReproduction Service No. ED071778).
Rieber, L.P. (1992). Computer-based microworlds: Abridge between constructivism and direct instruc-tion. Educational Technology Research and Develop-ment, 40(1), 93–106.
Rosenshine, B. (1970). The stability of teacher effectsupon student achievement. Review of EducationalResearch, 40, 647–662.
Rosenshine, B. (1971). Teaching behaviours and studentachievement. London: National Foundation for Edu-cational Research.
Rosenshine, B. (1979). Content, time, and directinstruction. In P. Peterson & H. Walberg (Eds.),Research on teaching: Concepts, findings, and implica-tions. Berkeley, CA: McCutchan.
Rosenshine, B. (1985). Direct instruction. In T. Husen &T. N. Postlethwaite (Eds.), International encyclopediaof education, Vol. 3 (pp. 1395–1400). Oxford: Per-gamon.
Rosenshine, B., & Stevens, R. (1986). Teaching func-tions. In M. C. Wittrock (Ed.), Handbook of research onteaching (3rd ed.) (pp. 376–391). New York: Macmil-lan.
Schunk, D. H. (2000). Learning theories: An educationalperspective. Upper Saddle River, NJ: Merrill.
Schwartz, D. L., & Bransford, J. D. (1998). A time fortelling. Cognition and Instruction, 16(4), 475–522.
Sfondilias, J.S., & Siegel, M.A. (1990). Combining dis-
covery and direct instruction strategies in computer-based teaching of mathematical problem-solving.Journal of Computer-Based Instruction, 17(4), 130–134.
Shambaugh, R. N., & Magliaro, S. G. (1997). Masteringthe possibilities: A process approach to instructionaldesign. Boston: Allyn & Bacon.
Skinner, B. F. (1953). Science and human behavior. NewYork: Macmillan.
Skinner, B. F. (1954). The science of learning and the artof teaching. Harvard Educational Review, 24 (86–97),99–113.
Skinner, B. F. (1966). The phylogeny and ontogeny ofbehavior. Science, 153, 1205–1213.
Skinner, B. F. (1968). The technology of teaching. Engle-wood Cliffs, NJ: Prentice Hall.
Skinner, B. F. (1974). About behaviorism. New York:Vintage Books.
Slavin, R. E., Madden, N. A., Dolan, L. J., & Wasik, B.A. (1996). Every child, every school: Success for all.Thousand Oaks, CA: Corwin.
Slocum, T. A., & Marchand-Martella, N. (2001). Intro-ducing the Journal of Direct Instruction. Journal ofDirect Instruction, 1 (1) 1–2.
Stallings, J. A., & Stipek, D. (1986). Research on earlychildhood and elementary school teaching pro-grams. In M. C. Wittrock (Ed.), Handbook of researchon teaching (3rd ed.) (pp. 727–753). New York: Mac-millan.
Steele, M. M., & Steele, J. W. (1999). DISCOVER: Anintelligent tutoring system for teaching studentswith learning difficulties to solve word problems.The Journal of Computers in Mathematics and ScienceTeaching, 18 (4), 351–359.
Swanson, H. L. (2001). Searching for the best model forinstructing students with disabilities. Focus on Excep-tional Childre,n 34 (2), 1–15.
Tharp, R. G., & Gallimore, R. (1988). Rousing minds tolife: Teaching, learning, and schooling in social context.Cambridge, UK: Cambridge University Press.
Tobias, S. (1982). When do instructional methods makea difference? Educational Researcher, 11, 4–10.
Viadero, D. (2002). Studies cite learning gains in directinstruction schools. Education Week 21 (31), p. 15.
Vygotsky, L. S. (1978). Mind and society: The develop-ment of higher psychological processes. Cambridge,MA: Harvard University Press.
Watkins, C. (1997). Project Follow Through: A casestudy of contingencies influencing instructionalpractices of the educational establishment. Cam-bridge, MA: Cambridge Center for Behavioral Stud-ies.
