from: Gardner, R., Sainato, D.M., Cooper, J.O., Heron, T.E., Heward, W.L., Eshleman, J.W., & Grossi, (1994). Behavior analysis in education: Focus on measurably superior instruction. pp. 173-197. The Morningside Model of Generative Instruction A History To develop the model of teaching presented here, Kent Johnson established a private, nonprofit (501 (c) 3) corporation—the Morningside Learning Center—13 years ago in Seattle, Washington. Initially, Mor- ningside offered a broad array of academic and training services, letting the Seattle community determine its focus (Pennypacker, 1992; Skinner, 1971,1981). These services included providing psychoeducational and voca- tional assessment, training human service personnel, teaching time management and study skills to college learners, and improving the academic skills of children and adults. Students, their families, or various social service agencies paid for the services. Within nine months, the parents of tutored children surfaced as the most frequent and stable users of Morningside's ser- vices. Since many of their children needed more intensive academic services than could be offered during weekly tutoring, Morningside Learning Center became Morningside School, a summer program for children with learning and attention problems. That fall, at the request of the parents of the summer-school children, Morningside became a full-fledged, year- round school, accredited by the state of Washington. Morningside Academy, a name its students eventually petitioned the director to use, is a basic skills catchup school; learners participate between one and three years before they are ready to be successful 173
Transcript
from: Gardner, R., Sainato, D.M., Cooper, J.O., Heron, T.E., Heward, W.L., Eshleman, J.W., & Grossi, (1994). Behavior analysis ineducation: Focus on measurably superior instruction. pp. 173-197.
The Morningside Model of Generative Instruction
A History
To develop the model of teaching presented here,Kent Johnson established a private, nonprofit (501 (c) 3)
corporation—the Morningside Learning Center—13years ago in Seattle, Washington. Initially, Mor-ningside offered a broad array of academic and trainingservices, letting the Seattle community determine itsfocus (Pennypacker, 1992; Skinner, 1971,1981). Theseservices included providing psychoeducational and voca-tional assessment, training human service personnel,teaching time management and study skills to collegelearners, and improving the academic skills of childrenand adults. Students, their families, or various socialservice agencies paid for the services. Within ninemonths, the parents of tutored children surfaced as themost frequent and stable users of Morningside's ser-vices. Since many of their children needed moreintensive academic services than could be offeredduring weekly tutoring, Morningside Learning Centerbecame Morningside School, a summer program forchildren with learning and attention problems. Thatfall, at the request of the parents of the summer-schoolchildren, Morningside became a full-fledged, year-round school, accredited by the state of Washington.Morningside Academy, a name its students eventuallypetitioned the director to use, is a basic skills catchupschool; learners participate between one and threeyears before they are ready to be successful
173
students elsewhere. The program teaches them thefundamental component skills necessary to learn suc-cessfully in a content class like history or biology.Learners whocould achieve an "A" in such classes probablywould not need to attend Morningside Academy.
The curriculum is a comprehensive sequence in thebasics: reading, decoding, and comprehension; mathe-matics concepts, calculation and problem solving; lan-guage arts, including handwriting, composing, grammar,spelling, and mechanics; computer basics, including key-boarding, document organization, and word processing;time management, materials organization, and learning tolearn; and critical thinking, reasoning, and argumentation.
The instructional materials represent a combina-tion of Engelmann's Direct Instruction programs inreading (Engelmann & Bruner, 1988a, 1988b; Engel-mann, Carnine, & Johnson, 1978; Engelmann & Han-ner, 1988a; 1988b; Engelmann etal., 1988a; 1988b),writing (Engelmann & Davis, 1991; Engelmann, Ar-bogast, & Davis, 1991; Engelmann & Silbert, 1983,1985,1991,1993; Engelmann & Grossen, 1993), andmathematics (Engelmann & Carnine, 1992a, 1992b,1992c, 1993a, 1993b), Morningside's fluency supple-ments to Engelmann programs (e.g., Johnson & Kevo,1993), Archer and Gleason' s (1989) Direct Instructionprograms in organizational and study skills and our ownprograms in mathematics (Johnson, 1993a, b, c, d, e, f, g,h, i; Johnson & Streck, 1993a, 1993b) and thinkingskills (Layng, Jackson, & Robbins, in preparation). Weplan to formally design and package all of our currentinstructional programs and fluency supplements.
Although many of its school functions, such asorganized physical education during the lunch hour,monthly field trips, and parent potlucks, help create theatmosphere of a typical school, Morningside Academy isessentially a learning laboratory for designing instruc-tional programs and classroom procedures, with alaserlike focus on the essential skills for school successand a 13-year research base (Binder, 1991a).
The Morningside Academy Guarantees AtMorningside Academy, students typically gainbetween two and three grades in each academic skillper year, as measured by national standardized
achievement tests (Johnson & Layng, 1992; Snyder,1992). Table 14.1 presents Morningside students'gains over the past 11 years.
Due to its successes, Morningside Academy nowoffers parents two money-back guarantees. The first isfor children who are two or more years behind inschool. Many children in this group have been offi-cially classified as "learning disabled" by publicschool personnel. These learners, who have rarelygained more than half a year in any one academic year,will gain at least two grade levels per school year ortheir parents will receive a tuition refund in proportion tothe shortfall. The second guarantee is for any otherlearners who are not much behind in grade-levelachievement but who stand apart from their peersbecause they do not coordinate visual and motor skillseffectively, as is most apparent in their handwriting,These students are also highly distractible, hyperac-tive, disorganized, and have poor study and inde-pendent learning skills. Many children in this grouphave been officially classified by their pediatricians orother medical personnel as "attention deficit disor-dered" (ADD). Morningside Academy guaranteesthat these learners will increase their time-on-taskendurance from their typical 1 to 3 minute spans to 20minutes or more—an attention span longer than that
Table 1 Morningside Academy Children's Mean Stand-ardized Achievement Test Grade-Level Gains
of the average college learner (Reese & Johnson,1975). Morningside Academy also guarantees thatstudents with visual or motor coordination deficiencieswill learn word processing skills to communicate inwriting. Furthermore, Morningside guarantees tobring these learners above the 65th percentile of theirpeer group nationally before they transition to anotherschool. In the seven years since the assurances haveformally been in place, Morningside has never had torefund tuition for failure to meet its money-back guarantees(Binder, 1988; Johnson & Layng, 1992).
Morningside Academy began with no particularfocus on learning disabilities or attention deficit dis-orders, populations with which it is now closely as-sociated. Its strengths connected with a sector of thepopulation that needed them in a microcultural evolu-tion. Morningside Academy is now considered amongthe top agencies in the Northwest specializing in "learningand attention disorders," even though none of itsfaculty, including its director, has had any training inneuropsychology, a predominant field in the "treat-ment" of these "disorders." Behavior analysis hasevolved a technology powerful enough to overridecertain hypothesized organic barriers to successfulschool functioning. By so doing, it created the irony ofenvironmental effectiveness in an area most oftenapproached from a physiological perspective.
In 1987 Morningside Academy began its adultliteracy program, providing remedial basic skills educationto minority youth and adults who did not have highschool diplomas and were seriously deficient in reading,mathematics, and writing skills. Some participants tookthe Morningside Academy program concurrently withjob-training skills programs, others as preparation forspecific occupational-skills training courses. The projectwas funded by the Job Training and Partnership Act(1985), a revival of the Comprehensive EmploymentTraining Act (CETA) program sponsored by LyndonJohnson and the Great Society, and proposed by theunlikely duo of Ted Kennedy and Dan Quayle. Fundingfor the project was contingent on Morningside's
proposal that no payment would be requested for anylearner who progressed less than two grade levels.
Morningside's initial literacy project serviced 52African American men and Asian American women.Forty-eight of these learners acquired skills at or abovethe national eighth-grade literacy standard and progressedmore than two grade levels to do so (Johnson & Layng,1992; Snyder, 1992). This was remarkable given theunstable lives of many of the participants, from home-lessness to drug and other criminal activity. Learnersadvanced at an average rate of 1.8 grades per month(20 hours of instruction and practice)—a figure thatcontrasts sharply with the U.S. government standard ofone grade-level progress per 100 hours of instruction(Binder & Watkins, 1989; Johnson, 1990, 1991;Johnson & Layng, 1992; Snyder, 1992). Part of thereason for the speedy progress was perhaps related tothe economic contingencies: the faster learners ad-vanced, the sooner Morningside Academy was paid.
Figure 14.1 illustrates the progress of four repre-sentative learners (KR, WB, DM, and JK) in the adultliteracy project. The dot-dashed lines in each graphrepresent the individual's predicted gains, which werecalculated by dividing the entering grade-level perfor-mances by the number of years spent in school. Thedashed lines drawn on the diagonal of each graphrepresent the standard progress expected of the learner inschool: one year of progress for one year of schooling.The solid lines represent the individual's gains in theMorningside program. In each case the participants' actualprogress far exceeds both the standard and predictedprogress.
Morningside Academy's literacy program attractedmany visitors, including representatives from the U.S.Department of Labor, U.S. Department of Energy, theWashington State Employment Security Department,various Washington county youth employment and trainingprograms, local television and radio personalities, andhundreds of parents, teachers, and professionals in medicaland social service. (Curiously, no one from the U.S.Department of Education has ever visited.)
In January 1991 T. V. Joe Layng became directorof the Academic Support Center at Malcolm X Col-
Figure 14.1. The progress of four adults. SOURCE: From "Breaking the Structuralist Barrier: Literacy and Numeracy withFluency," by K. R. Johnson and T. V. J. Layng, 1992, American Psychologist, 47, 1475-1490. Copyright 1992 by theAmerican Psychological Association. Reprinted by permission.
lege in Chicago. As we had previously collaboratedon other projects (including formalizing the variouselements that define the Morningside Model), andwith the success of the Morningside Academy adulteducation JTPA project, we designed and implemented apilot project.
Thus began the formal dissemination of the Mor-ningside Model. We launched a training program fortutors in the Academic Support Center, who expressedinterest in working in a summer program for peoplewho were behind in mathematics skills. The program was tobe called "Academic Storm: The Mother of All SummerPrograms." Academic Storm was the pilot project andfoundation for the new Precollege Institute, to begin inthe fall of 1991, for learners with high school diplomaswho attempted to register at Malcolm X College buthad reading skills below the sixth-grade level. Themain purposes of the institute would be to acceleratethe building of precollege skills in one to twosemesters, preventing learners from dropping out beforethey achieve college-level status, and to guarantee thesuccess of learners in college courses (Johnson & Layng,1992; Johnson, Layng, & Jackson, 1993; Snyder, 1992).
During five weeks in January, April, May, June, andJuly, Kent Johnson trained the tutors to teach a sequenceof over 250 objectives in mathematics, from simpleaddition through solving ratios and equations with oneunknown, a span of about six grade levels. The tutorsalso learned about the generic components of effectiveinstruction and practice. Between training visits, thetutors practiced their new teaching technologies witheach other at Malcolm X College.
Beginning July 1st, learners and teachers par-ticipated in the Academic Storm Morningside Math-ematics (Johnson & Streck, 1993a) program Mondaysthrough Thursdays, from 9 A.M. to 12 P.M., for sixweeks. Every Friday we gathered the tutors and thefall supervisor-in-training to share charts, discuss suc-cesses, brainstorm problems, and ask the detailedquestions about the model that arise only after directcontact with the system's contingencies. Learnerswith entry math skills at the fifth-grade level whoparticipated in the full Morningside technologygained over six years in math computation and overtwo years in math concepts and problem solving,After only 20 hours in timed reading rate practice,
learners gained an average of 1.1 years in reading(Johnson & Layng, 1992; Snyder, 1992).
Malcolm X College's Precollege Institute (PCI) pro-gram now includes instruction in reading, mathematics,writing, and thinking skills for those who test below thesixth-grade level. Early follow-up data indicate that inaddition to dramatic grade-level gains (Layng &Johnson, in press), following as little as eight weeks inthe program, learners who finish their first semester inthe college credit curriculum are exceeding the statedPCI goal of a 2.5/4.0 GPA and achieving some of thehighest within and semester-to-semester retentionrates in the college. All of these gains with childrenand adults in Seattle and Chicago occurred withoutany homework (to complete or to evaluate)!
The Morningside Model as Applied
The Morningside Model of Generative Instruction isan outgrowth of applied science because it attempts toincorporate built-in, databased self-correcting proce-dures (Bronowski, 1965). It is instructional design,based on scientific findings—as electrical engineering isto physics. Every step a learner takes, every procedure ateacher uses produces data that they use to direct them totheir next activity. The Morningside Model isdesigned so that both the instructional technology andthe students' learning continually improve as more andmore learners and teachers participate in its implementa-tion. Self-correcting mechanisms are crucial to theevolution of a model. Such programs "have a life oftheir own," helping perpetuate their own survival.They contain inherent features that both investigate theprogram's effectiveness and evolve improvements along
As in all design, there is an artistic component—unspecified contingency-shaped designer repertoires—that is not easily subject to obvious self-correction. Thiscomponent is determined by the complex histories ofthe designers. It makes each implementation a uniqueinstance of the model. Each implementation maintainsits stability and integrity, however, by adhering to thebasic instructional design, related scientificfindings, and the use of rate of change in learner
performance to provide continuous self-correctingfeedback.
Experimental psychology had a strong influenceon the measurement features of the MorningsideModel. Some of the learning literature in experimentalpsychology, particularly the experimental analysis ofbehavior, uses frequency measurement, whether ofresponses or time between responses. Much has beenwritten about the sensitivity of frequency measure-ment and the continuous orderly data it produces(Ferster, 1953; Skinner, 1953; Ferster & Culbertson,1982; Ferster & Skinner, 1957). But perhaps the mostimportant feature of frequency is that it can accuratelypredict future action (Ferster, 1953; Ferster & Cul-bertson, 1982). In everyday talk, terms such as habit,disposition, tendency, and personality attempt topredict what someone will do based on the frequencywith which they have behaved that way in the past.For example, a person who "has a habit" of whistlingtunes does so regularly and will probably do so in thefuture. Someone who "has a high disposition" or"tendency" to speak out about politics will do so whenthe occasion arises. Someone who "has an addictivepersonality" engages in behaviors that may bedetrimental at high frequencies, such as drugging orgambling. Indeed, some (see Binder, Chapter 3 in thistext) suggest that frequency may be considered aseparate dimension of behavior along with the com-monly accepted dimensions of duration, intensity, andform (topography).
From the outset, Morningside Model programsmeasured frequencies of accurate performance, andbuilt performance to high frequencies, a technologyCarl Binder called "fluency building" (Binder, 1977-1983, 1988, 1991b). Fluent performance is useful,permanent, and easily applied to new learning, fea-tures to which we return later in this chapter. Perhapsmore important, Morningside Model programs usechanges in frequency—acceleration of accuratelearner performance—to develop permanent and usefullearner success. This focus on acceleration meets theneeds of children who have always progressed"slowly," as well as the concerns of parents who seethe implications of having a child falling further and
further behind (decelerating). A focus on accelerationalso alleviates the frustration of teachers faced withincreasing differences in learner performance levelsand progress within the same classroom, resulting inincreasing teacher work and decreasing reinforcementfor that work (i.e., burnout from ratio strain).
Fortunately, a simple graphical chart — OgdenLindsley's Standard Celeration Chart — is availableto record learning accelerations (and decelerations).This chart is the fundamental teaching tool for aself-identified community of teachers with a distinctset of methods called Precision Teaching (Lindsley,1972, 1990, 1991). The Standard Celeration Chartprovides the foundation for setting fluency aims andfor making students' learning accelerations the basisfor curricular changes — everything from decisionsabout seating arrangements, to effective staffing, theuse of commercially available material, and the design
of materials and teaching methods. The StandardCeleration Chart appears in Figure 14.2.
The animal learning literature in experimentalpsychology influenced not only MorningsideAcademy's measurement practices, but also focusedthe model on acquisition, retention, endurance, andtransfer of learning (Kling & Riggs, 1971). The verballearning and cognitive processes literatureprovided further direction for teaching acquisition,retention, endurance, and transferor application. Cur-rent applications of the model focus on intellectualskills, those concepts, principles, and strategies that gobeyond rote learning and verbal repertoires (Gagne,1970). The distinction between verbal repertoires andintellectual skills is the difference between knowing
Figure 14.2. Daily Standard Celeration Chart. SOURCE: Reprinted by permission of and available from Behavior ResearchCompany, Box 3351, Kansas City, KS 66103. I
about and knowing how (Gagne, 1970; Tiemann &Markle, 1990). We can know about Pythagoras andgeometry, as demonstrated by what we say about him ina historical account (verbal repertoire); and we can alsoknow how to find the length of the hypotenuse of a righttriangle by applying the formula c2 = a2 + b2
(intellectual skill). We can know about reinforcementby saying its definition. We can show reinforcementknow how, by identifying it when we see instances of itin the real world and by applying reinforcementprocedures when we deem them necessary.
Conceptual behavior, identifying instances ofthings and events, is the foundation for intellectualskills. Concept formation research with humans andother animals (Clark, 1971; Gagne & Brown, 1961;Johnson & Stratton, 1966; Tiemann & Markle, 1990)tells about how to teach conceptual behavior with acluster of tasks that includes a range of positive andnegative instances (examples and nonexamples).During discrimination training with a range of ex-amples and nonexamples of a concept, experimentersreinforce learners' identification of examples as ex-amples and nonexamples as nonexamples. Identificationof nonexamples as examples is either ignored (i.e.,extinction) or corrected. Using these differentialreinforcement and correction procedures, learner per-formance comes under the fine-grained control of thestimulus features that embody the concept. Once eachconcept is mastered, the resulting conceptual behaviorcan be related to others in a variety of ways to formmore complex relations or principles (Tiemann &Markle, 1990). In the Morningside Model'sprograms, learners know about their fine-grained in-tellectual skills as well, by learning to answer ques-tions such as "Is this an example? How do you know? Isthis one? Why not?" (Markle, 1991).
An example from Morningside's study skills cur-riculum emphasizes intellectual skill building, begin-ning with conceptual behavior (Layng, Jackson, &Robbins, in preparation). Learners are taught togenerate a special type of question as they study textsand notes. The questions they generate promote thecontinual hypothesis-testing behavior shown by ex-perts within a given discipline. These "focusing"questions are special because they focus on materiallearners have not read but are about to read. Specifi-
cally, a focusing question must contain the followingcharacteristics or attributes: (1) it must make aresponse request, (2) it must be about what one ispreparing to read (not what one has already read), and(3) it must be a complete sentence.
If any one of these features is missing, the questionor statement would not qualify as a focusing question.For example, a section heading or topic sentence of aparagraph might begin: "Almost everyone wantseconomic growth." Examples of a good focusingquestion might be "Why does almost everyone wanteconomic growth?" or "Distinguish between thosewho want economic growth and those who do not."Notice that although only one example is stated as aquestion, both make a response request, both askabout the text about to be read, and both are completesentences. Nonexamples include, "People typicallywant more of everything" (no response request);"What do most people want?" (asks about materialalready read); and "Why economic growth" (incom-plete sentence).
The teaching sequence begins by providing stu-dents with scripted instruction featuring the rules foridentifying focusing questions, and exercises usingexamples and nonexamples of the concept. So-calledgeneralization is taught, not left to chance. Examplesare drawn from a wide variety of subject matters, types ofthings questioned (e.g., topic sentences, headings,captions, and charts), and complexity of text untilstudents can identify new examples not previouslyencountered in instruction. Nonexample work con-cludes when students reject all new statements orquestions that do not include all three of the criticalfeatures of the concept. As a result, the students learn todiscriminate focusing questions from nonfocusingquestions. This sequence provides them with automaticreinforcement (Skinner, 1957) from a discriminativerepertoire with which they are able to judge whetherthe questions they later produce are focusing ques-tions. That is, the learners are able to correct themselvesor at least know when they need assistance.
Building on this discriminative repertoire, studentsare introduced to the more complex skill of writing afocusing question by simply showing them how to add aquestion word and an action word (verb) to thematerial being questioned and then check to see that
it is a grammatically correct complete sentence. Theythen practice writing focusing questions for a widerange of written material drawn from a variety of topics.Next, students apply their ability to write focusingquestions to entire passages of text with and withoutheadings. Once question writing is established,similar procedures are used to teach the students togenerate brief answers to their questions, to read toanswer their questions, and finally to compare theiranswers to the answers provided by the text—hypothesis testing.
Defining Academic Behavior II: CumulativeProgramming of Concurrent Intellectual Skills
Like student like teacher. The skills that we teachteachers to implement with their learners require thesame intellectual skill-building technologies thatwe use with the learners themselves. The faculty atMorningside learn concepts embodying positive ap-proaches to decelerating behavior, such as Goldiamond' s(1974) constructional approach, differential rein-forcement of other behavior (DRO), differentialreinforcement of alternative behavior (Alt-R), differentialreinforcement of low rates of behavior (DRL), anddifferential reinforcement of diminishing rates ofbehavior (DRD) (Sulzer-Azaroff & Mayer, 1991).Each procedure is taught as conceptual behavior throughdiscrimination training with examples and nonexamples.As each successive concept is mastered and added tothe cumulative mix of scenarios, the faculty are re-quired to discriminate each one from all of the others.Less formal applications of cumulative programmingof concurrent intellectual skills occur whenteachers require learners to integrate (i.e., dis-criminate among) the skills they are learning fromtypical textbooks. For example, most mathematicstexts treat topics in isolation. Learners may solveaddition story problems in one section and subtractionproblems in the next section. However, learners rarelypractice solving both types together. Teachers mayapply cumulative programming principles by writingpractice worksheets that successively and continuouslyadd each new math topic to the preceding topics.Teachers have applied cumulative programming toother subject matters, such as grammar, social science,and natural science.
At about the same time that Kent Johnson foundedMorningside Academy, T. V. Joe Layng was col-laborating independently at the Behavior AnalysisResearch Laboratory of the University of Chicago,with Paul Andronis and Israel Goldiamond, its director,on basic animal research on the conditions responsiblefor combining previously trained simple nonsocialbehaviors, such as pecking a disk and walking back-and-forth between other disks, into complex untrainedsocial patterns (aggression in this case), such asincreasing the number of times another bird had topeck its disk for food (Andronis, Goldiamond, &Layng, 1983). This and other basic research both atChicago and elsewhere (Epstein, 1985, 1991) led toformalizing some of the key practices at work in themodel from the outset and later provided researchprotocols we are using to formally investigate thehuman intellectual skills that lie at the heart of themodel.
Both Morningside Academy's methods and theChicago laboratory investigations were preceded byyet other contributing observations by both Johnsonand Layng and their colleagues in the mid-1970s,These included Johnson's work with Carl Binder andBeatrice Barren teaching developmentally disabledadults the activities of daily living skills, such asshowering, brushing teeth, and washing hands. Theirwork used Eric and Elizabeth Haughton's observation ofthe importance of establishing very basic com-ponent motor skills before training more complexcomposite skills. The Haughtons called their basicmotor skills list the big 6 + 6, which included reach,point, touch, grasp, place, release, push, pull, twist,squeeze, tap, and shake. Layng's work in inpatientpsychiatry with Paul Andronis led to the observationthat so-called pathological composite behavior couldoften be traced to everyday component behaviors thathad been established earlier in other contexts (seeLayng & Andronis, 1984). A patient's "symptomchoice," such as pulling up clotheslines while screamingthat they were blasphemous representations of the cross,could often be understood as a reinforced composite ofpreviously learned religious and other componentsocial patterns that resulted in loved ones rallyingto the patient's assistance in the past.
This relation between basic behavioral com-ponents and more complex composites of these com-ponents may be the single most-important reason theMorningside Model has had the successes reported todate. Component/composite analysis is the foundationof the Morningside Model.
By combining (1) component/composite analysisof behavior with (2) cumulative instruction that in-cludes a range of positive and negative instances ofthe component skills and (3) procedures that increasethe frequency of the component behaviors, a powerfultechnology of instruction emerges—one that addressesperformance acquisition, retention, and application.When practitioners apply this technology, new andcomplex repertoires emerge as a function of simplypresenting a context for their combination, such as anactivity, game, or simulation; component behaviorstaught in the instructional sequence combine into new,untaught complex behavior as a result of the require-ments imposed by the game, activity, or simulation.The application of fluent skills to new contexts andcombinations without the need for instruction is theresult, and therefore the meaning, of generative in-struction (Alessi, 1987; Epstein, 1991).
Let us take a brief look at the teaching of threecomplex academic skills—factoring equations in al-gebra, sentence writing, and debating—to illustratecomponent/composite analysis, cumulative teachingwith positive and negative instances, and building thecomponent skills thus acquired to higher frequencies.The composite behavior, factoring an equation like4x2 - 10xy + 4y2, stumps many beginning algebralearners. Much has been written about teaching suchmathematical problem solving, including some direct,programmed instruction produced by people of a be-havioral bent (Eraut, 1970). Nearly all of this literaturesupposes that teaching the complex skill requires acomplex algorithm and fairly elaborate methods.Cognitive science research in education, the whole lan-guage movement, and even the elaborate programmedinstruction procedures that incorporate fading and chainingprocedures based on task analysis, attempt to teach skillsat the level of their observed complexity.
In contrast, the Morningside Model builds the com-ponent skills involved in factoring equations from thebottom up. Number writing, addition facts (i.e., num-bers with sums to 18; Johnson, 1993a), subtractionfacts (differences between numbers through 18 minus9; Johnson, 1993a), multiplication facts (through 9times 9; Johnson, 1993b), isolating and solving for Xin a simple linear equation (Johnson, 1993e), squaringand factoring squared numbers (Johnson & Streck,1993), and certain organization and sequencing skillsare first firmly established and then practiced untilthey occur at high frequencies. With high-frequency(fluent) component skills, learners learn how to factorwithin minutes by simply learning which operationsgo in which position within which set of parentheses toproduce the answer: (4x - 2y)(x - 2y). This complex skillis not learned by complex teaching procedures, but bya quick, three-step method: (1) show and tell theoperation, (2) provide mnemonics to remember thesequence (i.e., FOIL—first numbers, outside numbersplus inside numbers, last numbers give the equation),and (3) discrimination training/testing withexamples and nonexamples (e.g., "Is this equationfactored correctly? Why not? Is it factored correctlynow? How do you know?" and so on; Johnson,1993b).
English textbook authors usually spend severalpages teaching learners how to write and correctlypunctuate appositives like that illustrated in the middleof the following sentence:
John F. Kennedy, president of the United Statesin the early 1960s, was assassinated in 1963.
However, at Morningside Academy, when webrought the component skills—relative pronouns(e.g., who, which), nonrestrictive relative clauses(e.g., who was president of the United States in theearly 1960s), and the linking "to be" verb (e.g., am,was)—to fluency first, instruction in appositives forour middle school students was errorless after two-rule sentences:
Appositives are like nonrestrictive relativeclauses using which or who and a linking verb.However, in an appositive, the relative pronounand linking verb are left out.
With only three examples, all five learners so in-structed immediately began fluency building in ap-positives. No learner ever made an error.
How about debating? Morningside begins withargumentation rules from one of Engelmann's DirectInstruction programs, like "Just because two thingshappen at the same time doesn't mean one causes theother" (Engelmann, Manner, & Haddox, 1980). Onerule is taught at a time. Learners apply the rules toshort passages that follow or violate the rules:
The rooster crows on my farm every morningbefore the sun rises. I think I better bring himwith me on my trip to New York City. Theydon't have roosters there, so the sun may riselater than usual, if at all.
Argumentation rules are cumulatively programmed.Learners eventually catch the violations of many rulessimultaneously in lengthy prose passages through dis-crimination training procedures (e.g., "What argu-ment rule does the passage violate? How do youknow? Is a rule broken in this passage? How about thisone? Which one?" and so on).
Learners also build these skills almost to theirreading rates. For example, a learner who can read100 words per minute should be able to detect faultylogic in passages at his reading rate divided by 1.2, toaccount for covert reasoning skills (a little "thinkingtime"). Thus, 100 divided by 1.2 = 80 words of faultylogic detection per minute. Such automaticity guaranteesretention of the skill and thereby its immediateavailability for combination with related debating skills toform a new and complex composite skill. When skills arenot at our fingertips they may be temporarily forgotten, ormay occur only after a long delay, as when we get intobed and remember how we could have countered acolleague's argument earlier that morning.
The debating instructional program graduallyshifts from written passages to vocal statements re-quiring listening skills. Finally, the rules of debatingprotocol are learned by discrimination training withexamples and nonexamples and cumulative program-ming. Learners serve as debate referees who catchviolations of debating rules at higher and higher fre-quencies in live simulations, before they practice themon their own.
As in the algebra factoring problem, the complexrepertoire, seen in the act of actual debating, is afunction of simpler component skills. In our ap-proach, as skills get more complex they actually geteasier to learn, because they are combinations ofsimpler high-frequency repertoires. Because themore complex repertoires are easier to learn, they arealso easier to teach. The tendency is to add moreinterventions, more directions, more teaching to com-plex situations. But our experience has taught us thatthis is not necessary. Morningside teachers do notfollow elaborate algorithms or programmed sequencesto teach debating. In its later stages, teachingdebating is about guiding or coaching—like insertingbrief tips and quips to steer an ongoing, complex,generative repertoire when it goes off-course. Inten-sive, teacher-directed instruction occurs in the earlystages, when component skills are learned.
Component skills, initially shaped under separatecircumstances, may be recruited in a substantiallydifferent context into a new composite skill. Andronisand Layng call this sudden combination of componentelements contingency adduction (Andronis, 1983;Andronis, Goldiamond, & Layng, 1983; Layng &Andronis, 1984; Johnson & Layng, 1992). The ex-amples of factoring equations, sentence writing, anddebating illustrate contingency adduction. The Mor-ningside Model of Generative Instruction is itself aninstance of adduction from teaching repertoires thatembodied applied experimental science, frequencymeasurement, a focus on multiple learning processes,and cumulative programming of sequences of com-ponent skills, derived from component/compositeanalyses, and established through discrimination training.All of these examples of contingency adductionillustrate the rearranging of existing repertoires (afterFerster, 1965) through rules, prompts, and activities.
When teachers use the Morningside Model, theyneed to be on the lookout for an increase in unintendedcontingency adduction. Learner repertoires may com-bine in ways other than those planned. The schoolenvironment may intersect with and adduce repertoireswithout any rules or instruction! An example of unin-tended contingency adduction occurred in our adult
literacy program at Malcolm X College (Johnson &Layng, 1992; Snyder, 1992).
At one point in Morningside Mathematics(Johnson & Streck, 1993), learners are introduced toword problems involving fractions. This is a terriblydifficult juncture for most learners, as many of us whonever really mastered the fractions word problem skillcan attest. Universal, defective conventional instruc-tional practices prevailed in our thinking that surelyno adduction would occur at this juncture. Indeed, aswe reported earlier (Johnson & Layng, 1992), on acourse pretest of problem solving with fractions, fourlearners' performance ranged from 3 to 7 problemscorrect out of 14. Significant component skills thatwere weak or nonexistent at the time of the pretest,however, were now occurring at high frequencies.These repertoires included the elements of problemsolving with whole numbers and fractional computa-tion. Now, with no instruction or even a mention ofproblem solving with fractions, learner performanceranged from 13 to 14 correct out of 14. Assessing forcontingency adduction saved teachers and learnersmany hours of instruction! The learners simply com-pleted fluency-building exercises to assure the retentionof this adduced repertoire.
This is why every lesson in the Morningside Math-ematics program (Johnson & Streck, 1993) contains ashort adduction exercise, consisting of each of the com-ponent and composite skills to be taught in the lesson, tobe sure that the learner really needs the instruction before itoccurs. Since different learners bring different historiesto the classroom, different numbers of them skip theinstructional portion of each lesson, and eitherpractice the skills taught in the lesson to high frequen-cies, or move to the next lesson if already automatic.As the program proceeds, more and more skippingoccurs as an ongoing, complex, generative repertoiresnowballs. We call these contingency-adduced reper-toires curriculum leaps (Johnson & Layng, 1992).
The Morningside Model of Generative Instructionis an adduction of many findings in experimentalpsychology over the last five decades. Increasedlearner contingency adduction was one unintended
(Haughton, 1980, 1981) led Precision Teachers—mostnotably, Eric and Elizabeth Haughton, Bea Barrett, andCarl Binder—to coin and use the term fluency todescribe behavior that meets REAPS. Thus began adistinction between accurate and fluent performance.Accuracy, unlike fluency, rarely predicts whether per-formance will be retained, endure, transfer to morecomplex situations, combine with other repertoiresunder the same contingencies, and remain stableduring distracting conditions. For example, learnersmay be taught to spell words that follow the rule:double the final consonant before adding an endingthat begins with a vowel. The typical teacher com-plaint is that many learners, even those who spell thewords with 100% accuracy on the Friday spelling test,misspell them when writing their compositions. Thisoccurs because the new skill is accurate but not fluentand can therefore be disrupted when it must occuralong with other skills or in new situations.
The discovery of REAPS led many Precision Teachers toabandon goal setting and competency defining by norm-based frequency criteria. Three norm-based criteria havebeen used most often. One derives the average perfor-mance in a school or work setting, such as the rate ofmath facts per minute of all fifth-graders at a certainschool or district. A second method derives the averagefrom people judged to be "truly competent," such as allkids scoring above the 90th percentile on the CaliforniaAchievement Test in Math Computation at a certainschool or district. A third method derives the averagerate of math facts per minute of people who choose acertain career, such as all tellers at all branches of U.S.Bank. REAPS significantly altered the selection ofcriterion rates by focusing on rates of math facts thatensure retention, endurance, transfer, and stability.
Seven Tenets of the Morningside Model These twodiscoveries—contingency adduction andREAPS—were unintended adductions that crys-
Figure 14.3. The Morningside Model of Generative Instruction. SOURCE: From "Breaking the Structuralist Barrier: Literacy andNumeracy with Fluency," by K. R. Johnson and T. V. J. Layng, 1992, American Psychologist, 47, 1475-1490. Copyright 1992 bythe American Psychological Association. Reprinted by permission.
tallized the seven tenets of the Morningside Model ofGenerative Instruction, illustrated in Figure 14.3.
1. Identify the component elements of instructionalobjectives.
2. Measure their frequency until true mastery, definedby REAPS, reached (Binder, 1988).
3. Establish a component behavior through highlyinteractive, contingent exchanges between learnerand teacher, until behavior stays accurate at graduallyincreasing frequencies.
4. Build the component skills to fluency aims to en-sure remembering.
5. Build the endurance of component skills that arerepeated in succession en masse in the real world.
6. Include application activities that allow multiplecomponent skills to combine in ways that define thehigher-level complex activities of an expert in afield.
7. Alter the procedures for implementing the Mor-ningside Model according to the data collected.
These are seven tenets, not procedures. A range ofpossible procedures can be used to achieve each step.The arrows next to the boxes in Figure 14.3 alsoindicate that procedures for meeting the goal of eachstep could be implemented in overlapping fashion.For example, some learners could begin buildingfluency while still establishing their skills; others need towait to build fluency until establishment is complete.Individual-learner performance data should dictatewhich procedures to use to meet the goals of a step,when work on a new step should begin, and when returnto a previous step for more work is warranted.
The simple-to-complex sequence of tasks learnersencounter in the Morningside Model contrasts sharplywith currently popular cognitive science and wholelanguage approaches. These methods attempt to usecomplex activities to teach component elements. Forexample, debating is used to teach argumentation, andconducting a scientific experiment is used to teachobservation skills. Indeed, cognitive science ap-proaches may not bother with the component elements
at all (e.g., Engelmann, 1992; Resnick, 1988)! Learningby activities, as illustrated by whole language andother cognitive science approaches to teaching, standthe Morningside Model on its head!
Self-Correction and the Three-Term ContingencyTenet 7, the dynamic mechanism, is what makes theMorningside Model applied science. The self-correcting mechanism operates at the level of thethree-term contingency: the presentation of a moment ofinstruction, the learner's performance at that moment ofinstruction, and the immediate consequences of thelearner's performance: confirmation and praise, orcorrections, depending on the adequacy of thelearner's performance (Sherman & Ruskin, 1978).The self-correcting mechanism is defined by reverberationsthrough the three-term contingency. Any givenpresentation of information, materials, tasks, or problems,including course procedures, may or may not be ap-propriate to the learner's current knowledge and skills—learner performance helps teachers adjust the presentationuntil the learner is successful. Likewise, learner perfor-mance determines the nature of consequences: praise orcorrections. The consequences also help adjust the nextmoment of instruction, and so on, through cycles ofpresentation-performance-consequence reverbera-tions that make the system viable for each learner.
More frequent opportunities for learner perfor-mance increase the interactivity of the teacher and thelearner, reverberations through the three-term contin-gency, and thus self-correction (see Heward, Chapter21 in this text). Optimal interactivity, defined byreverberation frequency, depends on the complexityof the task presentation, learner performance, and per-formance consequence. Optimal interactivity alsodepends on the efficiency of learning—reverberationsproduced by unnecessary tasks and learner responses thatoccur too frequently don't alter the presentation-perfor-mance-consequence cycle and thus delay mastery ofskills. The most efficient use of the model occurswhen teachers and learners take the largest instruc-tional step that produces successful performance.
By operating under the assumption that complexperformance can be understood as a product of
simpler elements, the Morningside Model is an ex-ample of a selection science, like evolutionary biologyand paleontology. Selection sciences seek to find themore molecular processes that are sufficient to producemolar order (Donahoe, 1986, 1991; Palmer & Donahoe,1992). Such sciences are in stark contrast to thosePalmer and Donahoe (1992) called essentialist in character.Essentialist sciences, often typified by cognitivescience, view complexity as an expression of inherentprocesses operating at the level of observed events.Every observation has its proposed mechanism toexplain it. For example, some cognitive theories explainthe activity of remembering by ascribing to it a structure(memory) that is composed of highly organized semanticnetworks (Palmer & Donahoe, 1992). A selectionistscience, often typified by behavior analysis, explainsremembering as just that—an activity that describesparticular occasion-behavior relations that were initiallyreinforced and now occur at a later time (Palmer, 1991).
Essentialist-designed instruction attempts to teachcomplex skills at the same level of complexity that isobserved in an expert or highly skilled performer.Indeed, instruction that is essentialist in nature ofteneschews any attention paid to simple components ortheir sequence (Resnick, 1988). Cognitive scienceapproaches to instructional design (e.g., Stahl &Miller, 1989; Stepich, 1991), and the educationalmovements on which they are loosely based (e.g.,whole language), exemplify this approach.
By contrast, instruction designed according to prin-ciples of selection science focuses on the simple ele-ments that comprise complex behavior (Layng, 1988,1989, 1991). When the elements occur at high fre-quencies, activities, games, and simulations can bearranged to promote the adduction of complex reper-toires. The Morningside Model exemplifies this ap-proach. It assumes that learning and teaching geteasier, not harder. It is only the accumulation of weakcomponent skills that makes learning harder and har-der. As we noted in an earlier paper (Johnson &Layng, 1992), of the nearly 10 million learners whomake it to high school mathematics courses each year,only 800 go on to earn doctorates in math (Mullis etal., 1991). Cumulative ignorance may be a primarycause of the fallout. For most of us, studying calculus islike climbing a mountain with a bag of bricks on our
back, each brick representing a weak component skill.When instruction is designed from the bottom up,people may be able to learn calculus at a cocktail partyon a napkin!
A selectionist approach also calls into serious questionmost tutoring practices and remedial approaches tolearning. Learners who are tutored almostinvariably have deficient component repertoires, yetare kept afloat in their studies by tutors who focus onthe complex repertoires needed for next week'schemistry or history test and neglect the verycomponent skills that would make adduction of thecomplex skills likely. The problem is that learnersmove through the educational system based on theirage and on the credits they have accumulated, not ontheir competencies. The time available to study fornext week's chemistry test makes it unlikely thateither the tutor or the learner will focus on learningany deficient component skills needed for the test.
An insidious problem results from the educationalestablishment's developmentalism (see Stone, Chapter6 in this text). Learners react emotionally to thesuggestion that they are not appropriately placed in aparticular academic sequence based on their age or the
time they have spent in school. Eric and ElizabethHaughton met much resistance from McGill Universitystudents to their approach to helping learners who weredoing poorly in calculus. Students wanted help indifferential equations, not number-writing skills,math facts, adding, and the like. However, with notutoring in calculus and instead focusing on simplecomponent elements that were deficient, learners sub-stantially improved their calculus grades. The problempresented is not always the problem to solve!
Figure 14.4 shows a typical arrangement of a Mor-ningside classroom. At station 1, learners work with ateacher in small groups to establish new componentskills. There is a desk or at least bookshelves at station 1for the teacher to keep materials for immediate use. Atstation 5, learners build fluency individually, in pairs,or in triads, and may work with a fluency coach, whomay be the teacher, a teacher's aide, or an ad-
Figure 14.4. Recommended floor plan.
vanced learner. Somewhere in the classroom, eachlearner has a personal shelf or cubby space to storematerials. In larger settings, learners may have per-sonal desks at which to study or an area for self-study. Inapplications with children and youth, there is anarea, 6-7, for leisure activities. The area containsactivities for reinforcement, such as construction toys,computer games, cards, books, puzzles, and boardgames. Children earn points throughout the class orschool day that can be exchanged throughout the classor day for minutes in the leisure-time activities area.
In addition to designing instruction and teachingnew skills, instructors usually serve as supervisors ofthe system, making sure materials are readily avail-able, recordkeeping is organized and up-to-date, anddaily operation is smooth, especially in makingdatabased adjustments for learners. Supervisors alsomonitor their staff in vivo and provide written and oralfeedback on their performance.
In addition to the instructor, learners are assistedby coaches before working by themselves and witheach other to build their skills to fluency. In programsfor children and youth, coaches can be the teacher,teacher aides, parents, or other paraprofessionalsskilled in the subject matter. Coaches can also belearners in higher grades or otherwise advanced in thesubject matter, in a kind of cross-age tutoring (seeMiller, Barbetta, & Heron, Chapter 20 in this book).Ultimately, coaches need to come from the ranks ofthe learners in the program.
In programs for college and adult learning, coachescan be the teachers themselves or tutors hired by auniversity, college, or learning center. As withchildren and youth, coaches can be advanced studentsworking for credits; eventually learners in the coursecan dominate coaching.
Much has been written about peer coaching, par-ticularly at the college level (Johnson & Ruskin, 1977;Sherman, Ruskin, & Semb, 1982). Strain (1981)edited a comprehensive book on the use of childrenand youth peers as classroom behavior-change agents.Unfortunately, almost nothing has been written aboutthe minutiae of coaching to increase skill frequency tofluency aims. The procedures involved in training
peer coaching are beyond the scope of this paper. Thejob description is clear, however. Coaches increasethe opportunity for practice, measurement, and feed-back and contribute significantly to the self-correctingreverberations of the system. They help maximizelearner and teacher productivity by increasing thenumber of performances that can be monitored,praised, and corrected. Coaches can also be the in-cidental locus for learning organizational, observa-tional, and learning-to-learn skills. Coaches are alsothe source of interpersonal social reinforcers in thesystem, and are often the best source for adjusting thesystem for the learner's benefit. Coaches can be for-mally trained to provide all these services.
The best peer coaches have skills that overlap withthe learners' current skills to a far greater degree thanan expert such as the instructor (Skinner, 1957;Johnson, 1977; Johnson & Ruskin, 1977). Overlap-ping repertoires produce more fine-grained, frequentchanges, and learning proceeds at a more rapid pace.Athletes show such judicious selection of coacheswhen they select their practice partners from thosewho play just a little better than they do. Many ofthe staff in the adult learning program at Malcolm XCollege built their component teaching skills inshifts to positions of more responsibility in thesystem. One person recently moved in steps fromcoach to supervisor, another from adult literacy studentto coach.
The following is a thumbnail sketch of 14 dailyoperational procedures employed in a typical Mor-ningside Model classroom. The suggested timeframes assume that 60 minutes are available each day.
1. Lesson plan. Divide up the daily course or skill areaclass period into three parts. All three parts can occursimultaneously, but for different learners at differenttimes. Each of these three daily segments can bedivided into several specific tracks, each focusing on adifferent skill. Every day the teacher must map thecourse the learner is to follow the next day, based oncareful analysis of the learner's performance that day.We strongly recommend daily classes over classesthat meet only one or two days per week. The more
infrequently a course or skill is studied, the less effi-cient the learner's daily map may be, due to variations indysfluent skills from one class period to the next, andthe slower the learner's progress will be.
2. Precision placement. Pretesting learners frequentlykeeps the instructor current on what skills they need tolearn. Learners bring entering repertoires to acourse or class, and new repertoires may be adducedduring a course of study. A beginning course orprogram pretest will give information about enteringrepertoires. Individual lesson pretests will tell moredetails about entering behavior, plus any new reper-toires that have been adduced since the course began.No learners should participate in an instructionalepisode if they already have the skills to be taught.The composition of instructional groups should changefrom day to day, according to pretesting results. Carefulpretesting can greatly improve the efficiency of learningand teaching, and may play a large role in thegrade-level gains that we have reported thus far.
3. Public progress records. Entering and adducedbehavior can be marked on a public wall chart that liststhe learners on the vertical axis and the lessons to bemastered on the horizontal axis. Teachers can refer tothe chart to call only those learners to the instructionaltable who need to learn the skills taught in the lesson.As learners establish and build fluency on the skills inthe lessons, additional marks can be made on the chart.At Malcolm X College, learners, with their instructorsor coaches, make a small circle in the cell of a lessonwhen they have participated in instruction. Theymake the circle larger after they have successfullybuilt some frequency with coach support and thencompletely color in the cell when they reach thefluency aim. The progress chart is also used as theoccasion for daily meetings between the instructor andcoaches. Public progress records also help maintainthe pace (Johnson & Ruskin, 1977) and quality (VanHouten, 1980) of performance.
4. Tool skill fluency. Each course or skill has a numberof fundamental component elements or tool skills. Forexample, an analysis of addition and subtraction ofwhole numbers reveals nine tool skills that are
fundamental to the majority of their concepts andoperations, including number writing and reading,math facts, and math language skills with symbols andterms (Johnson & Streck, 1993). Reading has prereq-uisite pointing, decoding, and scanning skills. Writinghas critical transcriptive and dictation tool skills, withboth pen and paper and computer keyboards.Studying has critical organizational and cooperativelearning tool skills.
Learners should be given ample time to practicethree to four tool skills daily. Typically, five to tenminutes is sufficient practice to make a noticeableimprovement in one tool skill compared to the pre-vious day.
5. Establishing new skills. In the Morningside Model,instructors provide highly interactive lessons to establishnew skills and knowledge. Lessons are taught to smallgroups of no more than 12, and only to those whoneed them. Instructors examine the public wall chartfor a particular lesson and call only the learners whohave blank cells. The other learners practice theobjectives until they are fluent and work on otherfluency assignments.
At least three instructional procedures contain theingredients necessary for establishing new skills effi-ciently. The first is Engelmann's Direct Instruction (DI)(e.g., Engelmann & Carnine, 1982; Kinder & Carnine,1991). The second is Programmed Instruction (PI)(e.g., Markle, 1991; Gilbert, 1962; Evans, Homme, &Glaser, 1962). The third is Keller and Sherman'sPersonalized System of Instruction (PSI) (Keller,1968; Sherman & Ruskin, 1978; Sherman, Ruskin, &Semb, 1982). We will describe our applications ofeach of these instructional strategies in turn.
In Direct Instruction at Morningside, a small group oflearners gather with a teacher at a horseshoe-shapedtable for 20 to 30 minutes. During that time, theteacher stands in front of the group, either in front of ablackboard or next to an overhead projector. Next tothe teacher on a music stand is a scripted instructionalpresentation, which he or she refers to whennecessary. The script provides the exact lines theteacher will use for instruction. The lines arerelatively brief, allowing learners to chorally respondto the teacher on signal approximately 10 times a
bite-sized as the tasks presented in programmed in-struction sequences—there may be only 2 or 3 taskspresented per textbook page. The study guide alsocontains clarifications, corrections, and additions totext, and is most successfully presented in a personal-ized framework, as if a tutor or study partner werepresent when the learner is studying the text. In mostPSI courses learners do not receive confirmation oftheir responses to the teacher's tasks; they must takethe initiative to clarify any problems they have withcourse personnel before testing. This relativelysophisticated format for learning puts PSI proceduresat the other end of the continuum of establishingprocedures: Direct Instruction-Programmed Instruc-tion-Personalized System of Instruction.
6. Sprinting. Immediately following Direct, Programmed,and/or Personalized Instruction, learners put the finishingtouches on establishing new skills by engaging insprinting exercises for five to ten minutes with a coach(Johnson & Layng, 1992). The coach may be theinstructor who continues this work with the learners,or another person who steps in. A number of learnersmay test out of Direct, Programmed, or PersonalizedInstruction, and go directly to sprinting.
During sprinting, learners gradually increase thefrequency of the new component skills that they havelearned, and the coach ensures that these skills don'tbreak down as frequency is increased from 10 to 20or more per minute. Coaches introduce activities suchas, "How many can you complete in twenty seconds?Let's see. Then we'll see if you can get more done inthe next round," or "Let's see how long it takes you tocomplete three of these. Say 'done' when you'redone. I'll tell you the time it took you. You write thattime down. We'll see if you can beat that time in thenext round." As learners practice, they record theirfrequencies and/or durations. Coaches rotate amonglearners as they put some rate behind their skills,noting and changing elements of their practice thatslow them down, and further firming their accuracy iferrors begin to occur as their rates increase. Coacheslet the learners who are doubling their frequenciesleave the group to fluency-build with one another;they continue to coach only those learners who needfurther guided practice.
Establishing involves not only making componentskills firm (accurate at 10 or so per minute), but alsosetting them in celeration motion, to new frequenciesthat predict retention, endurance, and application.Thus, establishing is a combination of Direct Instructionand Precision Teaching, with their foci on accuracyand charted daily frequencies of performance (Kinder& Carnine, 1991; Binder & Watkins, 1990; Lindsley,1972,1990). Establishing charts the course of initiallower frequencies of performance and helps guaranteethat steep celerations will occur with further practice.
7. Fluency building with peer coaches. Once learnershave doubled or tripled their rates in short spurtswithout making errors, they engage in timed practicewith each other to build their frequencies until theymeet fluency aims. Fluency aims are those frequencyranges that predict retention, endurance, and applica-tion. For example, the fluency aim for math facts is80 to 100 per minute; for oral reading 200 to 250words per minute; for solving math word problems 12to 15 per minute, for writing paragraphs 20 to 25words per minute. Miller and Reward (1992)provided a detailed description of timed practice pro-cedures (see also Miller, Barbetta, & Heron, Chapter20 in this text). Visitors to a Morningside classroomliken sprinting and fluency building to an academicgymnasium, complete with warm-ups, sprints, longerendurance training, monitoring, coaching, stretching,and resting.
For 15 to 20 minutes each day, each pair of learnersfluency-builds the skills learned in the day's instruc-tion, along with other recently learned skills that needmore practice to be fluent. They also keep track oftheir progress on recording sheets. Peer-coachingprocedures allow learners to become independent ofteachers by the end of the course. Under these proce-dures, star learners emerge as fluency coaches (andinstructors!). Peer coaching requires training, adescription of which is beyond the scope of this paper.
Fluency building is a combination of PrecisionTeaching (PT), with its focus on daily timings andcharting, and the Personalized System of Instruction,with its reliance on the peer as a proctor to guideindividual learning. As such, it is about instruction
and practice intertwined, but differently than the waythey are intertwined during establishment. There, theaccuracy of accelerating performance is at issue; here,the issue is celeration itself and how to increase it.
Instructional materials. Sprinting and fluency buildingrequire multiple timers with features that allow thetiming of duration and the length of a specific period.The timers must electronically signal the latter, with abeep or ring, to prevent the learner from having tocontinually check the timer while completing the ex-ercise. (So stop watches and sandglasses are out.)
Teachers must also make a great many copies ofpractice exercises, because each learner will completemultiple timings, over several days, to reach a fluencyaim. An alternative is to cover a practice sheet withMylar. By using a hard-tipped marking pen, eachtimed practice can be easily erased with a moistsponge.
Design. As in establishing skills, many instructionaldesign considerations should be followed to constructpractice sheets and set fluency aims for them. Furtherdescription of sound instructional materials design isbeyond the scope of this paper. To keep the focus onREAPS and to avoid using norm-referenced fluencystandards, we offer the following beginning points.
To design a practice sheet and set a fluency aim for it,an instructor sums the basic reading and writing toolskills that the learners will be required to use whilecompleting the tasks to be practiced and divides thisnumber by 1.2 to allow for a little thinking time.Consider our earlier example of building fluency withargumentation rules, in which learners identify faultyreasoning in text passages. One design option is towrite 50-word passages that contain 0-2 broken rules.Learners can identify the broken rules by writing thenumber of the rule that the argument breaks next tothe sentence that violates it. The tool-skill aim forwriting numbers is 160-180 per minute. The tool skillaim for reading text is 200-250 words per minute. In a2-minute timing, learners could read 4-5 passages andwrite their numbers. To allow a little think time, theteacher could set the aim at 3-4 passages perminute.
tivities specified by the Morningside Model ofGenerative Instruction. Fluency building is by far thepredominant activity, involving 70% or more of thelearner's time, depending on what the data show.Conventional classroom time is proportioned inreverse, with about 70% spent establishing skills, 20%practicing, and 10% testing. When the authors pro-vide training in the Morningside Model to schools,districts, and colleges, many teachers' first reaction to
the "happy learner" is disbelief: "How can you spendso much time practicing? I barely cover the year'scurriculum (or the course's objectives) with muchmore time devoted to establishing. How can youproduce two to three year gains this way?" Theanswer lies in the inefficiency of instruction that is top-heavy in establishment. Unless learners becomefluent in the skills established, they lose them. Thereis also far less likelihood of curriculum leaps.Learners in classrooms that are top-heavy inestablishing recycle through learning and forgettingand relearning during "review" sessions. The learn-forget-review-forget cycle is far more time-consumingthan Morningside's learn-it-once-and-for-allapproach.
8. Standard celeration charting. In Morningsideprograms, learners use standard celeration charts tomonitor fluency building. Learners plot the dailyfrequency of each skill. They also draw miminumceleration lines on their charts to show daily progressthat doubles in frequency each week. They first ex-amine their charts for yesterday's performance fre-quency and record it on a practice record form. Next,they set the day's progress aim by determining whatfrequency would keep them on or above their mini-
mum celeration line, and record this information.Next, learners practice until they meet the day'sprogress aim. Once this occurs, the instructor or fluencycoach is called to give a final timing, which verifies theachievement. The learner then plots the final timingon the chart.
Learning how to use the standard celeration chart iseasy. The interested reader may consult several sourcesfor good training (e.g., Precision Teaching Project,1984; Pennypacker, Koenig, & Lindsley, 1972). In onearticle, 5-year-old Stephanie Bates tells her friendshow to chart while her father takes notes (Bates &Bates, 1971)!
9. Decision-making and fluency-building interven-tions. In most cases, timed practice itself is theintervention needed to build fluency. Sometimes,however, learners will experience performance locks(Elizabeth Haughton, personal communication,August 1978). Try as they may, they don't achieve x2 (read "times 2") weekly progress through repeatedtimings. This is more often the experience of learnerswith learning and attention problems. There are amultitude of procedures to help learners buildfluency when their performance is locked, but theseare beyond the scope of this chapter. The interestedreader should consult a variety of resources togain these competencies; most notably, Haughton(1980) and Binder's (1977-1983) Data-SharingNewsletter.
At Morningside, fluency-building interventionsoccur whenever learners fail to double their accelerationeach week. Some of these interventions require theinstructor's skills, others a coach's skills, and stillothers can be implemented by the learnersthemselves. An important goal of any Morningsideprogram is to teach all learners to make their owninterventions.
10. Clearing a backlog. It may take several days oreven a week or more to reach the fluency aim of agiven skill. Periodically, learners will accumulatemore skills established in instructional lessons thatneed fluency building than there is time in the day orclass period to practice, leaving no time to attend newestablishing lessons. This is okay. At Morningside
Academy and Malcolm X College, learners may skip aday or two of new instruction until they reachfluency on the skills established in previous days. Thenumber of skills that constitutes a backlog and hencean instruction-skip day will vary according to theamount of time devoted to daily practice. As ageneral rule of thumb, each skill requires about 10 to15 minutes of daily practice. If 45 minutes of dailypractice is allotted, learners would skip instructionif they needed to practice more than four skills thatday.
11. Endurance building. Some skills need to be repeated athigh frequencies for longer periods than others. Forexample, column addition and subtraction lasts 15minutes or more when balancing a monthly check-book statement. Reading a chapter in a textbook maylast 45 minutes or more. Editing a paper may take twohours or more. Writing or typing may occur for awhole workday! We set endurance aims by multiplyingthe fluency aim by the number of additional minutesrequired for endurance. For example, our silent-readingfluency aim is 500 words per minute; our 15-minute silent-reading endurance aim is 500 x 15 =7,500 words in 15 minutes.
Although reaching a fluency aim often predictsthe skill will endure for longer periods, this is notalways the case. Sometimes reaching a higher fluencyaim produces endurance; other times learners need topractice the skill over the anticipated time period.(See Johnson & Layng, 1992, for a more in-depthdiscussion.) Judicious examination of the curriculum,particularly at composite skill junctures, is necessaryto identify skills slated for endurance building.These skills should be written in their appropriateplaces in the sequence on the public progresscharts so that the work won't be overlooked.
Sometimes learners will show endurance problemseven while building fluency. For example, somechildren and adults diagnosed with so-called attentiondeficit disorder slack off in performance frequencyduring a one-minute timing. Teachers can ascertainsuspected endurance problems during fluency buildingby keeping track of frequencies during successive ten-second intervals of the minute. If frequencydiminishes, learners need to gradually build their rates
over successively longer intervals up to one minute.Binder, Haughton, and Van Eyk (1990) have describedexcellent procedures for accomplishing this.
12. Applying. Once several component skills arefluent, instructors can schedule planned activities topromote the complex and creative behaviors charac-teristic of experts in a field. There are at least threekinds of activities to promote such contingency ad-duction. The first kind promotes a relatively simpletransfer of training from one context to another, aswhen learners apply cumulative decoding skills learned incontrolled basal readers to reading newspapers andmagazines. A second kind of activity promotes chainingof separately taught repertoires, as when learnersconsult a dictionary or thesaurus for an adjective todescribe a feeling as they write a composition aboutthe winter holiday season.
A third kind of activity promotes a recombinationof components, an intertwining of behavior, as whenlearners debate, solve a math problem, write a re-search paper, create a stock exchange simulation intheir classroom, or publish a classroom newspaper.Planning these latter adduction activities involves as-sembly of skills from multiple domains, as componentskills from seemingly disparate subject areas can bebrought into play. For example, creating a stock ex-change involves component skills learned from math-ematics, language arts (e.g., persuasive writing),organizational skills, and reference skills. With suchextended sequences of instruction, learners may takesix months of instruction before they are ready todesign a classroom newspaper or a personal budget;they may take two or more years of a course of studybefore it will be fruitful to arrange a debate, or thedesign of a school yearbook or scientific experiment.
Games and simulations can also be scheduled tobring together component repertoires. Thiagarajanand Stolovitch (Stolovitch & Thiagarajan, 1980;Thiagarajan, 1990; Thiagarajan & Stolovitch, 1978)have written extensively on this motivating approach toapplying component skills.
13. Unintended contingency adduction: Monitoring forcurriculum leaps. As mentioned earlier, curriculumleaps will occur at junctures other than during planned
activities, games, and simulations. Component skillswill combine at odd times, in complex and unusualways, additively or multiplicatively, to the highercomplex skills exhibited by experts. Curriculum leapingwill accelerate as learners progress through acourse of study. Careful pretesting before each lessonwill ensure that a learner doesn't miss a chance to skipinstruction and accelerate progress.
14. Integrity between criteria and progress. Mor-ningside instructors make the contingencies betweenreaching criterion and progress through the programexplicit at each step, every minute of the day. Eachreverberation through a three-term contingency, ateach learning step—establishing, fluency building,endurance building, and applying—tells instructors toeither go forward or to provide the extra instruction orpractice necessary for learners to achieve success.Skinner (1968) estimated that 42,000 contingenciesprobably occurred during a 50-minute observation ofhis daughter's elementary school math lesson! Adjustingthe learner's contact with the Morningside system onthe basis of even a small fraction of the three-termcontingency reverberations will make a big difference.
A Summary and Conclusion
Table 14.2 summarizes 14 daily operational steps to asuccessful implementation of the Morningside Modelof Generative Instruction. Several places in the over-view contained in this paper need to be magnifiedwith greater attention to detail. Direct Instructiontraining, Standard Celeration Chart training, fluency-building procedures training, peer-coaching training,and instructional materials design training may all benecessary.
Besides the usual classroom materials, teacherswill need timers for practice; one-third the number oflearners in the program will be sufficient. Standardceleration charts are required, as well as a wall chartfor marking progress through the program. If copying isat a premium, the teacher will also need sponges,Mylar, and marking pens. Since learners will perhapsbe more self-directed than teachers have been used to,
Table 14.2 Summary of daily procedures in the MorningsideModel of Generative Instruction
1. Arrange each class period to allow for new instruction, tool skillfluency building, and sprinting/fluency building of skills recentlyestablished.
2. Pretest learners at the beginning of the course and before eachlesson. Place students in instruction precisely matched to theircurrent performance levels.
3. Keep public progress records to direct daily learner activities andreinforce progress.
4. Build prerequisite tool skills to high frequencies, to facilitate themastery of curriculum objectives.
5. Use Direct Instruction to establish new skills with learners whodon't have fluent reading, writing, and studying skills; useprogrammed instruction, or texts with study guides in a PSIformat with experienced learners.
6. Coach the gradual beginning of skill acceleration with sprintingexercises.
7. Use peer coaching and testing to build skills to fluency aims, thoselevels of performance that guarantee retention, endurance, andapplication following significant periods of no practice.
8. Monitor fluency building on standard celeration charts.9. Use acceleration criteria: change fluency-building procedures
whenever skills are not doubling their rates per week.10. Include catchup days of all practice and no instruction whenever a
learner gets backlogged.11. Build skill endurance on certain skills and sets of component
skills.12. Arrange application activities for contingency adduction.13. Skip instruction when students make curriculum leaps.14. Guarantee that the contingencies for reaching criterion are met
before learners move forward in the program—every minute ofthe day.
the teachers should make sure that the learners don'thave to ask for materials.
Instructors have a threefold job during class ses-sions: alternating between teaching from scripts,troubleshooting with learners who need special fluency-building procedures, during sprinting or peerfluency building, and reinforcing learner-coach andlearner-learner interactions. For a year the teacherwill be busier teaching than ever before. Eventually,however, the system will perpetuate itself and theteacher will find that very little daily preparation isneeded. If the teacher has never taught with others,there will be a need to adjust to the cooperative natureof the Morningside Model. Instructors have threelevels of meetings with their coaches: daily meetings,using the wall chart as a point of departure; weekly
work sessions, with the wall chart and standard celerationcharts governing their interactions; and monthlytraining sessions to teach new skills.
Anything can be taught with the MorningsideModel. Its use, however, may be circumscribed bythree factors: (1) local policies and procedures thatplace constraints on curriculum, teaching methods, orthe availability and use of paraprofessionals and ad-vanced learners; (2) courses in which goals cannot bespecified; and (3) courses designed to select, rank, orsort learners (Sherman & Ruskin, 1978).
We have argued that certain generative effects-occurring as they do without direct intervention andoften described as intuitive leaps, insight, problemsolving, sudden realization, and expert knowledge—are a product of the contingency adduction of interactingalternative contingency sets that have in the pastoccurred independently at high frequencies. Al-though we have accumulating evidence to support ourargument, the research effort and its application is in itsinfancy. Many questions still need to be answered, andmore still have to be asked. The model we present hereis designed to promote this research effort, increaseits efficiency, and make it available to everyone.
References
ALESSI, G. (1987). Generative strategies and teaching forgeneralization. Analysis of Verbal Behavior, 5, 15-27.
ANDRONIS, P. T. (1983). Symbolic aggression by pigeons: Con-tingency coadduction. Unpublished doctoral dissertation. TheUniversity of Chicago, Department of Psychiatry and BehaviorAnalysis, Chicago.
ANDRONIS, P. T., Goldiamond, I., & Layng, T. V. J. (1983)."Symbolic aggression" by pigeons: Contingency coadduction.Paper presented at the meeting of the Association for BehaviorAnalysis, May, Milwaukee.
ARCHER, A., & Gleason, M. (1989). Skills for school success,Books 3-7, North Billerica, MA: Curriculum Associates.
BATES, S.,& Bates, D.F. (1971)."... and a child shall lead them":Stephanie's chart story. Teaching Exceptional Children, 3,\l\-
crisis. Future Choices, 1, 33-39.BINDER, C. V., & Watkins, C. L. (1990). Precision teaching and
direct instruction: Measurably superior instructional technologyin schools. Performance Improvement Quarterly, 3, 2-14.
BINDER, C. V., Haughton, E., & Van Eyk, D. (1990). Increasingendurance by building fluency: Precision teaching attentionspan. Teaching Exceptional Children, 22, 24-27.
BRONOWSKI,J.(1965). Science and human values. New York:Harper & Row.
CALIFORNIA Achievement Tests. (1985). Monterey, CA:CTB/McGraw-Hill.
CARNINE, D. (1992). The missing link in improving schools-reforming educational leaders. Association for Direct InstructionNews, 11,25-35.
CLARK, D. C. (1971). Teaching concepts in the classroom: A set ofteaching prescriptions derived from experimental research.Journal of Educational Psychology, 62, 253-278.
DENAPOLI, J., & McKinny, P. (1992). Becoming an engineer ofinstruction: Advanced correction and firming procedures forReading Mastery and Corrective Reading. A videotaped inser-vice program. Baldwinsville, NY: J/P Associates.
DONAHOE, J. W. (1986). How shall we understand complexity? InP. N. Chase & L. J. Parrott (Eds.), Psychological aspects oflanguage: The West Virginia lectures (pp. 36-43). Springfield,IL: Charles C. Thomas.
DONAHOE, J. W. (1991). Selectionist approach to verbal behavior:Potential contributions of neuropsychology and computersimulation. In L. J. Hayes & P. N. Chase (Eds.), Dialogues onverbal behavior (pp. 119-145). Reno, NV: Context Press.
ENGELMANN, S. (1991). Why I sued California. Direct InstructionNews, 10(2), 4-8.
ENGELMANN, S. (1992). War against the schools' academicchild abuse. Portland, OR: Halcyon House.
ENGELMANN, S., & Bruner, E. (1988a). Reading mastery LChicago: Science Research Associates.
ENGELMANN, S., & Bruner, E. (1988b). Reading mastery II.Chicago: Science Research Associates.
ENGELMANN, S., & Carnine, D. (1982). Theory of instruction:Principles and applications. New York: Irvington Publishers.
ENGELMANN, S, & Carnine, D. (1992a). Connecting math con-cepts A. Chicago: Science Research Associates.
ENGELMANN, S., & Carnine, D. (1992b). Connecting math con-cepts B. Chicago: Science Research Associates.
ENGELMANN, S., & Carnine, D. (1992c). Connecting math con-cepts C. Chicago: Science Research Associates.
ENGELMANN, S., & Carnine, D. (1993a). Connecting math con-cepts D. Chicago: Science Research Associates.
ENGELMANN, S., & Carnine, D. (1993b). Connecting math con-cepts E. Chicago: Science Research Associates.
ENGELMANN, S., & Davis, K. L. S. (1991). Reasoning and•writing A. Chicago: Science Research Associates.
ENGELMANN, S., & Grossen, B. (1993). Reasoning and writingE. Chicago: Science Research Associates.
ENGELMANN, S., & Hanner, S. (1988a). Reading mastery III.Chicago: Science Research Associates.
ENGELMANN, S., & Hanner, S. (1988b). Reading mastery IV.Chicago: Science Research Associates.
ENGELMANN, S., & Silbert, J. (1983). Expressive writing I.Chicago: Science Research Associates.
ENGELMANN, S., & Silbert, J. (1985). Expressive writing II.Chicago: Science Research Associates.
ENGELMANN, S., & Silbert, J. (1991). Reasoning and writing C.Chicago: Science Research Associates.
ENGELMANN, S., & Silbert, J. (1993). Reasoning and writing D.Chicago: Science Research Associates.
ENGELMANN, S., Arbogast, A. B., & Davis, K. L. S. (1991).Reasoning and writing B. Chicago: Science Research Associates.
ENGELMANN, S., Carnine, D., & Johnson, G. (1978). Wordattackbasics. Chicago: Science Research Associates.
ENGELMANN, S., Hanner, S., & Haddox, P. (1980). Conceptapplications. Chicago: Science Research Associates.
ENGELMANN, S., Osborn, J., Osborn, S., & Zoref, L. (1988a).Reading mastery V. Chicago: Science Research Associates.
EPSTEIN, R. (1985). The spontaneous interconnection of threerepertoires. Psychological Record, 35, 131-143.EPSTEIN, R. (1991). Skinner, creativity, and the problem ofspontaneous behavior. Psychological Science, 2, 362-370.ERAUT, M. (1970). Fundamentals of elementary algebra. New
York: McGraw-Hill.EVANS, J. L., Homme, L. E., & Glaser, R. (1962). The RULEGsystem for the construction of programmed verbal learningsequences. Journal of Educational Research, 55, 513-518.FERSTER, C. B. (1953). The use of the free operant in the analysis
of behavior. Psychological Bulletin, 50, 263-274.FERSTER, C. B. (1965). Verbal behavior as magic. Paper
presented at the 50th Anniversary Conference of the GraduateSchool of Education, the University of Pennsylvania, May. InC. B. Ferster, S. A. Culbertson, & M. C. P. Boren. (1975).Behavior principles. 2nd ed. (pp. 563-568). EnglewoodCliffs, NJ: Prentice-Hall.
FERSTER, C. B., & Culbertson, S. A. (1982). Behavior principles.3rd ed. Englewood Cliffs, NJ: Prentice-Hall.FERSTER, C. B., & Skinner, B. F. (1957). Schedules of reinforcement.Englewood Cliffs, NJ: Prentice-Hall.GAGNE, R. M. (1970). The conditions of learning. 2nd ed. New
York: Holt, Rinehart & Winston.GAGNE, R. M., & Brown, L. T. (1961). Some factors in theprogramming of conceptual learning. Journal of Experimental
Psychology, 62, 313-321.GAGNE, R. M., & Foster, H. (1949). Transfer to a motor skill frompractice on a pictured representation. Journal of Experimental
Psychology, 39, 342-355.GILBERT, T. F. (1962). Mathetics: The technology of education.Journal of Mathetics, 1, 7-73.
GOLDIAMOND, I. (1974). Toward a constructional approach tosocial problems: Ethical and constitutional issues raised byapplied behavior analysis. Behaviorism, 2, 1-84.
HAUGHTON, E. C. (1972). Aims: Growing and sharing. In J. B.Jordan & L. S. Robbins (Eds.), Let's try doing something elsekind of thing (pp. 20-39). Arlington, VA: Council on Excep-
tional Children. HAUGHTON, E. C. (1980). Practicing practices: Learning by
of teaching concepts. Journal of Educational Psychology, 57,48-53.JOHNSON, K. R. (1977). Proctor training for natural control.
Journal of Personalized Instruction, 2, 230-237.JOHNSON, K. R. (1990). Literacy training: JTPA adults and
youth at risk. Presented at the Ninth International PrecisionTeaching Conference, November, Boston.
JOHNSON, K. R. (1991). About Morningside Academy. FutureChoices, 3,64-66.
JOHNSON, K. R. (1993a). Morningside mathematics fluency. Vols1-3, Addition and subtraction math facts. 2nd ed., Seattle:Morningside Academy.
JOHNSON, K. R. (1993b). Morningside mathematics fluency. Vols4-6, Multiplication and division math facts. 2nd ed. Seattle:Morningside Academy.
JOHNSON, K. R. (1993c). Morningside mathematics fluency. Vols7 & 8, Reading and writing whole numbers and decimals. 2nded. Seattle: Morningside Academy.
JOHNSON, K. R. (1993d). Morningside mathematics fluency. Vol.9, Place value, rounding, greater than less than. 2nd ed. Seattle:Morningside Academy.
JOHNSON, K. R. (1993e). Morningside mathematics fluency. Vol. 10,Addition and subtraction equations. Seattle: Morningside Academy.
JOHNSON, K. R. (1993f). Morningside mathematics fluency. Vol.11. Addition and subtraction problem-solving. Seattle: MorningsideAcademy.
JOHNSON, K. R. (1993g). Morningside mathematics fluency. Vol.12. Addition and subtraction computation of whole numbers 2nded. Seattle: Morningside Academy.
JOHNSON, K. R. (1993h). Morningside mathematics fluency. Vol.13. Multiplication and division computation of whole numbers. 2nded. Seattle: Morningside Academy.
JOHNSON, K. R. (1993i). Morningside mathematics fluency. Vol.14. Multiplication and division tool skills. Seattle: MorningsideAcademy.
JOHNSON, K. R., & Kevo, H. (1993). Morningside phonics fluency.Seattle: Morningside Academy.
JOHNSON, K. R., & Layng, T. V. J. (1992). Breaking the struc-turalist barrier: Literacy and numeracy with fluency. AmericanPsychologist, 47, 1475-1490.
JOHNSON, K. R., & Ruskin, R. S. (1977). Behavioral instruction.Washington, DC: American Psychological Association.
JOHNSON, K. R., & Streck, J. (1993a). Morningside mathematics.Seattle: Morningside Academy.
JOHNSON, K. R., & Streck, J. (1993b). Addition and subtractionequations: Equations and problem solving. Seattle: Morningside
what it is becoming. Journal of Behavioral Education, 1, 193-214.KLING, J. W., & Riggs, L. A. (1971). Woodworth & Schlosberg's
experimental psychology. 3rd ed. New York: Holt, Rinehart &Winston.
LAYNG, T. V. J. (1988). Problem solving: The role of metaphoricalextension. Paper presented at symposium, Problem Solving:The State of the Science. Association for Behavior Analysisconvention, May, Philadelphia.
LAYNG, T. V. J. (1989). Problem solving: Elements made absent.Paper presented at symposium, Problem Solving: The State ofthe Science II. Association for Behavior Analysis, convention,May, Milwaukee.
LAYNG, T. V. J. (1991). A selectionist's approach to verbalbehavior: Sources of variation. In L. J. Hayes & P. N. Chase(Eds.), Dialogues on verbal behavior (pp. 146-150). Reno, NV:Context Press.
LAYNG, T. V. J., & Andronis, P. T. (1984). Toward a functionalanalysis of delusional speech and hallucinatory behavior. Be-havior Analyst, 7, 139-156.
(Eds.),Let's try doing something else kind of thing (pp. 1-11). Reston,VA: Council for Exceptional Teaching.
LINDSLEY, O. R. (1990). Precision teaching: By teachers forchildren. Teaching Exceptional Children, 22, 10-15.
LINDSLEY, O. R. (1991). Precision teaching's unique legacy from B.F. Skinner. Journal of Behavioral Education, 1, 253-266.
MARKLE, S. M. (1969). Good frames and bad: A grammar offrame writing. 2nd ed. New York: Wiley & Sons.
MARKLE, S. M. (1991). Designs for instructional designers.Champaign, IL.: Stipes Publishing Company.
MILLER, A. D., & Heward, W. L. (1992). Do your students reallyknow their math facts? Using daily time trials to build fluency.Interventions in School and Clinic, November, 98-104.
MULLIS, I., Dorsey, J., Owen, E., & Phillips, G. (1991). The state ofmathematics achievement: Executive summary. Washington, DC:National Center for Education Statistics, U.S. Department ofEducation.
PALMER, D. C. (1991). A behavioral interpretation of memory. In L.J. Hayes & P. N. Chase (Eds.), Dialogues on verbal behavior (pp.261-279). Reno, NV: Context Press.
PALMER, D. C., & Donahoe, J. W. (1992). Essentialism andselectionism in cognitive science and behavior analysis.American Psychologist, 47, 1344-1358.
Handbook of the standard behavior chart. Lawrence, KS: The
izedsystem of instruction: 48 seminal papers. Lawrence, KS: TRIPublications.
SKINNER, B. F. (1953). Some contributions of the experimentalanalysis of behavior to psychology as a whole. AmericanPsychologist, 8, 69-78.
SKINNER, B. F. (1954). The science of learning and the art ofteaching. Harvard Educational Review, 24, 86-97.
SKINNER, B. F. (\951).Verbal behavior. Englewood Cliffs, NJ:Prentice-Hall.
SKINNER, B. F. (1968). The technology of teaching. EnglewoodCliffs, NJ: Prentice-Hall.
SKINNER, B. F. (1971). Beyond freedom and dignity. New York:Knopf.
SKINNER, B. F. (1981). Selection by consequences. Science, 213,501-504.
SNYDER, G. (1992). Morningside Academy: A learning guarantee.Performance Management Magazine, 10, 29-35.
STAHL, S., & Miller, P. (1989). Whole language and languageexperience approach for beginning reading: A quantitativeresearch synthesis. Review of Educational Research, 59, 87-
TIEMANN, P. W., & Markle, S. M. (1990). Analyzing instructionalcontent: A guide to instruction and evaluation. Champaign, IL:Stipes Publishing Company.
VAN HOUTEN, R. (1980). Learning through feedback: A sys-tematic approach for improving academic performance. NewYork: Human Sciences Press.
WATKINS, C. L. (1988). Project Follow-Through: A story of theidentification and neglect of effective instruction. Youth Policy,70,7-11.
WHITE, O. R., & Haring, N. G. (1980). Exceptional teaching. 2nded. Columbus, OH: Charles E. Merrill, Inc.
