behavior analytic instruction for learners with autism advances in stimulus control technology-gina...

http://foa.sagepub.comDisabilities

Focus on Autism and Other Developmental

DOI: 10.1177/108835760101600203 2001; 16; 72 Focus Autism Other Dev Disabl

Gina Green Technology

Behavior Analytic Instruction for Learners with Autism: Advances in Stimulus Control

http://foa.sagepub.com/cgi/content/abstract/16/2/72 The online version of this article can be found at:

Published by: Hammill Institute on Disabilities

and

http://www.sagepublications.com

can be found at:Focus on Autism and Other Developmental Disabilities Additional services and information for

http://foa.sagepub.com/cgi/alerts Email Alerts:

http://foa.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

at WESTERN MICHIGAN UNIVERSITY on September 20, 2008 http://foa.sagepub.comDownloaded from

http://www.hammill-institute.org

http://foa.sagepub.com/cgi/alerts

http://foa.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://foa.sagepub.com

Behavior Analytic Instruction forLearners with Autism: Advances in Stimulus Control TechnologyGina Green

Numerous behavior analytic methods developed since the early 1960s have provedeffective for developing a wide range of skills in learners with autism. Recent advancesin stimulus control technology, in particular, offer effective methods for teaching manyimportant skills and for promoting independent, generalized performances. This articlereviews selected stimulus control techniques, including new methods for teaching con-ditional discrimination (matching) skills, stimulus equivalence procedures, prompt andprompt-fading techniques, and incidental teaching procedures.

n 1948 an event occurred that wasto have far-reaching effects: SidneyBijou became director of the Insti-

tute of Child Development at the Uni-versity of Washington. Bijou was a Co-lumbia University-trained psychologistwith a strong interest in child develop-ment, learning theory, and the relativelynew radical behaviorism of B. F. Skinner,with whom he had worked for 2 years atIndiana University. Shortly after Bijouwent to the University of Washington, aresearch program was established withinthe Institute of Child Development tostudy normal and abnormal behavior inyoung children in a laboratory and in anexperimental school. There Bijou and hiscolleagues and students blended Skinner’snatural science approach to behavior-including methods of functionally ana-lyzing individual behavior-with re-

search on child development. Over thenext two decades they conducted land-mark studies of operant behavior in

young children and pioneered numerousmethods for managing problem behav-ior, teaching academic skills to studentswith mental retardation, training parentsto work as therapists with their own chil-dren, and conducting research in naturalsettings. In 1961 Bijou and Donald M.

Baer published their classic text Child De-velopment: A Systematic and EmpiricalTheory) in which they examined devel-opment from a behavioral perspective(Bijou, 1996).How is this history relevant to the

topic at hand? Some researchers might besurprised to learn that much of the sem-inal research on behavior analytic meth-ods for teaching children with autism wasconducted at the University of Wash-ington Institute for Child Developmentunder Bijou’s direction. In the early1960s a local physician asked Bijou ifthe staff at the institute might be able toteach a visually impaired boy with autismto wear glasses. Dicky, then 3V2 years old,exhibited severe tantrums; self-injuriousbehavior; problems with sleeping andeating; and very limited communication,social, and self-care skills. Bijou had seenCharles Ferster and Marion DeMyer useSkinnerian methods to build the reper-toires of children with autism at Indiana

University Hospital (Ferster & DeMyer,1961), so he asked his colleague, behav-ioral psychologist Montrose Wolf, a

graduate student, Todd Risley, and clini-cal psychologist, Hayden Mees to takeon Dicky’s case. Fortunately, they did.After 7 months of intensive behavioral

treatment, Dicky was wearing his glassesmost of the time, and his problem behav-iors had diminished enough that he couldbe returned to his home (Wolf, Risley, &

Mees, 1964). At the age of 5, Dicky wasenrolled in the Institute’s nursery school.He still had few skills; some of the oldproblem behavior had recurred, and ag-gressive behavior had developed. Wolfand Risley trained the nursery schoolstaff to implement behavioral teachingprocedures with Dicky. The goal was todevelop sufficient language and otherskills that Dicky could enroll in publicschool. By the beginning of the nextschool year that goal had been achieved(Wolf, Risley, Johnston, Harris, & Allen,1967). Dicky ultimately graduated fromhigh school. He had been deemed &dquo;un-testable&dquo; as a youngster; at age 26 heachieved a score of 98 on an IQ test.When he was in his 30s, Dicky was livingand working independently and had rela-tively good reading, writing, and socialskills (Bijou, 1996; Wolf, 1999).Wolf went on to start the Journal of

Applied Behavior Analysis and, with Baerand Risley, set out the very first defini-tion of applied behavior analysis, or ABA(Baer, Wolf, & Risley, 1968). They andothers who worked or studied with Bijouat the University of Washington (includ-ing Jay Birnbrauer, Betty Hart, Ivar

Lovaas, Howard Sloane, and Robert

Wahler, among others) made invaluablecontributions in many areas, particularlyapplications of Skinner’s science of be-havior to human development, language,education, mental retardation, autism, andbehavior disorders. Their work estab-



73

lished the foundation for much of thefield of applied behavior analysis, and forthe hundreds of studies of behavior ana-

lytic interventions for autism that havebeen published since the 1960s (see Mat-son, Benavidez, Compton, Paclawskyj, &

Baglio, 1996). That research, in turn, hasmade it possible for people with autismto achieve unprecedented outcomes today.The interventions designed by Bijou,

Wolf, Risley, and colleagues in the 1960shad several key features: (a) integrationof developmental and behavioral ap-proaches ; (b) an emphasis on positive re-inforcement procedures to build usefulrepertoires; (c) functional analysis of in-dividual behavior (i.e., experimentaldemonstrations of functional relationsbetween environmental events and be-

havior) ; (d) use of scientific methods toevaluate the effects of interventions;(e) individualization of goals and in-

structional procedures; (f) gradual, sys-tematic progression from simple to morecomplex skills; (g) training of parents andothers to implement interventions in

multiple environments; and (h) transferof intervention from structured to nat-ural settings. Contrary to assertions thatthese features characterize only certain&dquo;contemporary&dquo; versions of ABA (e.g.,Prizant & Rubin, 1999; Prizant &

Wetherby, 1998; Wetherby, Schuler, &

Prizant, 1997), the facts are that theyhave characterized genuine, comprehen-sive ABA programming for learners withautism since the 1960s, and the work ofBijou and his colleagues has long beenincluded in many graduate training cur-ricula in behavior analysis.

Of course, many additional techniqueshave been developed and integrated intoABA programming for learners withautism since those early days. Indeed, avery large array of behavior analytic pro-cedures have been proved effective fordeveloping a wide range of skills in indi-viduals with autism of all ages. This arti-cle does not provide an exhaustive or de-tailed review of all of them; that wouldtake many pages, and, besides, somegood comprehensive reviews of the rele-vant research have been published re-

cently (e.g., Hall, 1997; Matson et al.,1996). Instead, because all instructional

methods involve manipulations of ante-cedent stimuli, the focus here is on se-

lected recent advances in using stimuluscontrol principles and procedures to buildskills in learners with autism and relateddisorders. Space limitations prohibit ex-haustive coverage of all stimulus control

procedures. Excluded, for example, areinstructional techniques that make use ofestablishing operations, which are ma-nipulations of antecedent stimuli that in-fluence the momentary effectiveness ofreinforcers. (For discussions of establish-ing operations in teaching language tolearners with autism, see Sundberg &

Partington, 1998, 1999.) Because thetopic is skill development, antecedent ma-nipulations designed primarily to man-age or reduce problem behavior are alsoexcluded.

ABA for Autism:An Overview

To provide context for readers, the be-havior analytic framework and basic ABAmethods are first reviewed briefly in thissection. They have been detailed in nu-merous books, chapters, and articles, in-cluding the principal sources used here(Anderson & Romanczyk, 1999; Ander-son, Taras, & Cannon, 1996; Green,1996; Hall, 1997; Keenan, Kerr, & Dil-

lenburger, 2000; Koegel & Koegel,1995; Lovaas et al., 1981; Matson et al.,1996; Romanczyk, 1996; Sundberg &

Partington, 1998).In the behavior analytic view, autism is

a syndrome of behavioral deficits and ex-cesses that have a biological basis but arenonetheless amenable to change throughcarefully orchestrated, constructive inter-actions with the physical and social envi-ronment. Behavior analytic interventionseeks to redress those deficits and ex-cesses by providing multiple planned op-portunities for the learner to develop andpractice skills that are useful in a varietyof situations, and are effective alterna-

tives to less socially acceptable behaviors,such as tantrums, stereotypy, and de-structive behaviors (Green, 1996; Koegel& Koegel, 1995; Lovaas & Smith, 1989;Schreibman, 1988).

Behavior analytic instruction beginswith a comprehensive assessment of eachlearner’s current skills and needs, accom-plished by observing the learner directlyin a variety of situations and recordingwhat she or he does and does not do.

Every skill that is selected for instruction(often called a &dquo;target&dquo;) is defined in

clear, observable terms and broken downinto its components. Each componentresponse is taught by presenting or ar-ranging one or more specific antecedentsstimuli, such as cues or instructions fromanother person, and/or items of interestto the learner. Often, the kinds of cuesthat are effective with typically develop-ing children (such as spoken instruc-

tions) are not effective when a skill is firstintroduced to a learner with autism; thatis, those cues are not reliably followed by(discriminative for) the desired response.In such cases, another antecedent, calleda prompt, is often added to get the re-

sponse going. Effective prompts are

stimuli that are reliably followed by thedesired responses. For example, giventhe language deficits that are often ob-served in children with autism, spokeninstructions like &dquo;Come here&dquo; or &dquo;Touch

your nose&dquo; may not be followed by theactions they specify initially-that is, theyare not discriminative stimuli for those

responses. But gentle physical guidanceis effective for getting many children tocarry out actions, so it might be added tothe spoken instructions as a prompt dur-ing initial teaching, and gradually re-

duced (faded) over successive learningopportunities.When target responses occur, they are

followed immediately by consequencesthat have been found to function as re-

inforcers ; that is, repeated observationshave verified that when those conse-

quences consistently followed a particu-lar response, the response occurred againand again. Incorrect or interfering re-

sponses are explicitly not reinforced. Eachantecedent-response-consequence cycleconstitutes a learning opportunity, or trial.There are many behavior analytic proce-dures for arranging learning opportuni-ties, some adult-initiated, some learner-initiated, some embedded in typicallyoccurring activities or sequences of re-



Tim Obertein

74

sponses, and some that are hybrids orpermutations of these. Each type of pro-cedure has its uses and advantages. Gen-uine ABA programming uses any and allprocedures required to accomplish thejob of skill development and skill gener-alization with each individual learner; itis by no means just &dquo;discrete trial train-

ing&dquo; (e.g., see Anderson & Romanczyk,1999; Sundberg & Partington, 1999).

Learning opportunities are typicallyrepeated many times until the learner

performs the target response readily andfluently, without prompts from adults.Again, there are many ways to arrangemultiple, repeated learning opportuni-ties ; they need not (and should not) belimited to blocks of massed trials pre-sented as drills. The learner’s responsesare recorded frequently according to

specific, objective criteria. Those data aresummarized and graphed to provide pic-tures of the learner’s progress and to en-able frequent adjustments in instruc-tional procedures when the data showthat the learner is not making the de-sired gains.The timing and pacing of teaching ses-

sions, practice opportunities, and conse-quence delivery are determined preciselyfor each learner and each skill. To maxi-mize the learner’s success, skills are prac-ticed and reinforced in many settings.The long-range goal is to build simple re-sponses systematically into complex andfluid combinations of age-appropriate re-sponses. In quality ABA programming,the overarching goal is to teach learnerswith autism how to learn from typical en-vironments, and how to act in ways thatwill consistently produce positive out-comes for the learner and those aroundhim or her.

Tapping the Power ofStimulus Control

Behavior analysts have long recognizedthat stimuli that accompany or precederesponses that are reinforced can come toinfluence those responses in importantand often complex ways (e.g., Skinner,1938, 1953). A great deal of laboratory

and field research in behavior analysis hasderived from the principle of stimuluscontrol. Because the aim of virtually allinstruction is to get specific responses tooccur reliably under particular antece-dent stimulus conditions and not underother conditions, all instructional tech-

niques involve manipulations of antece-dent stimuli, along with manipulations ofconsequent stimuli. That is certainly trueof ABA techniques for building skills inlearners with autism. Basic proceduresfor establishing stimuli as discriminativefor reinforcement for learners with au-tism are described in some of the sourceslisted at the beginning of the Overviewsection, among others, so they are not re-iterated here. Instead, some specific ap-plications of stimulus control technologywith learners with autism are discussed.

Teaching Discriminations andStimulus Equivalence ClassesFrom the first incarnations in the 1960sto the present versions, many behavioranalytic curricula have included tech-

niques for teaching learners with autismto discriminate among various types of

stimuli, and to match certain stimuli toone another. Discrimination and match-

ing skills are components of many (argu-ably, czll ) cognitive, communication, so-cial, academic, work, and self-care skills,so this emphasis is understandable and ap-propriate. Lovaas, Schreibman, and theircolleagues conducted some seminal re-search on teaching discrimination andmatching skills to learners with autism inthe 1970s. Procedures developed in thoseearly studies seem to have been adoptedby many practitioners, perhaps becausethey have been widely disseminated insuch media as The ME Book (Lovaas et al.,1981). Meanwhile, during the ensuingthree decades, a great deal of stimuluscontrol research has been conducted bybehavior analysts in laboratories andclassrooms affiliated with the Eunice Ken-

nedy Shriver Center, the University ofKansas, and the New England Center forChildren, to name a few. Much of thatresearch has focused on analyzing varioustypes of discriminations and developing

effective and efficient techniques for es-tablishing discriminations and stimulusclasses (sets of mutually substitutable stim-uli) in learners with developmental dis-abilities, including autism. To date, how-ever, those techniques have not enjoyedwidespread adoption, even among ap-plied behavior analysts, perhaps becausethey have not yet been &dquo;packaged&dquo; incomprehensive, user-friendly programsthat are readily available to practitioners(although some steps in that directionhave been taken, e.g., Johnson, White,Green, Langer, & MacDonald, 2000;Serna, Dube, & McIlvane, 1997). Acomplete presentation of the aforemen-tioned stimulus control research is be-

yond the scope of this article, so whatfollows is a summary of principles andmethods drawn from some of that lit-erature.

Types of Discriminations. As notedearlier, virtually all skills involve discrim-.inating among, or responding differen-tially to, environmental events-sounds,colors, shapes, letters, numbers, words,foods, clothing items, people, responses,locations, and the like. At the same time,many functional skills require learners totreat some environmental events as if

they are the same, including some thatbear little or no physical resemblance toone another (such as spoken words andvarious types of visual stimuli). As it turnsout, these latter types of performances alsorequire complex discriminations amongstimuli-discriminations that are made

up of other, simpler discriminations andthat have particular and unique charac-teristics. In short, discriminations vary intheir complexity and in the types of envi-ronmental events they comprise. There-fore, a useful strategy for analyzing andteaching discrimination skills is to breakeach discrimination into its components(R. Saunders & Green, 1999; Sidman,1986).Many important skills require simple

discriminations. A simple discriminationcontingency has three elements: the an-tecedent stimulus (S), the response (R),and the consequence (C). That is, simplediscriminations are established by rein-



Tim Obertein

75

forcing particular responses in the pres-ence of particular antecedents, and not inthe presence of other antecedents. Whenthe defined responses occur reliably inthe presence of the defined stimuli andnot in the presence of other stimuli, thestimuli are said to be discriminative forreinforcement. Some examples relevantto teaching learners with autism include:

~ Oral naming: For example, in thepresence of a spoon (S1), the vocalresponse &dquo;spoon&dquo; (Rl) is reinforcedand &dquo;fork&dquo; ( R2 ) is not; in the pres-ence of a fork (S2), the response&dquo;fork&dquo; (R2) is reinforced and&dquo;spoon&dquo; (Rl) is not.

~ Instruction following: For example, inthe presence of the spoken instruc-tion &dquo;Stand up&dquo; (S1), standing up( Rl ) is reinforced and sitting down(R2) is not; in the presence of thespoken instruction &dquo;Sit down&dquo; (S2),sitting (R2) is reinforced and stand-ing up (Rl) is not.

Many other very important skills re-quire learners to match stimuli to one an-other, that is, to treat them as if they arethe same, in some sense. For instance,language comprehension entails matchingspoken words with corresponding ob-jects, symbols (such as pictures andprinted words), actions, people, and soforth; basic math skills involve matchingcorresponding printed numerals, quanti-ties, and spoken number names; and soon. True matching requires conditionaldiscriminations. Conditional discrimina-tions are established by reinforcing re-sponses to particular antecedent stimuliif and only if they are preceded or ac-companied by particular additional stim-uli. In contrast with simple discrimina-tions, here each antecedent stimulus is

discriminative for reinforcement or not,

depending on (conditional on) the pres-ence of another particular antecedent(Sidman, 1986; Sidman et al., 1982; Sid-man & Tailby, 1982). Conditional dis-crimination contingencies, thus, involvefour (rather than three) elements: con-ditional stimuli, antecedent stimuli, re-sponses, and consequences. Some com-

mon categories and illustrations of con-ditional discriminations are:

~ Conditional identity ~catchin8: Forexample, in the presence of a redpatch (S3), pointing to an identicalred patch (S1) is reinforced, whereaspointing to a blue patch (S2) is not,AND, in the presence of a blue patch( S4 ), pointing to an identical bluepatch (S2) is reinforced, whereaspointing to a red patch (Sl) is not.

~ Object-picture correspondence: For ex-ample, in the presence of a ball (S3),pointing to a picture of a ball (S1) isreinforced, whereas pointing to a pic-ture of a toy car (S2) is not, AND, inthe presence of a toy car (S4), point-ing to a picture of a car (S2) is rein-forced, whereas pointing to a pictureof a ball (S1) is not.

~ Receptive vocctbulary, or &dquo;receptiveidentificstion:&dquo; For example, afterone hears &dquo;nose&dquo; (S3) spoken by anadult, touching one’s own nose (S1)is reinforced, whereas touching one’smouth ( S2 ) is not, AND, after hear-ing &dquo;mouth&dquo; (S4) spoken by theadult, touching one’s mouth (S2) isreinforced, whereas touching one’snose (Sl) is not.

~ Picture-bczsed communication: For

example, when cookies (S3) are avail-able, bringing Mom a picture of acookie (S1 ) is reinforced, whereasbringing Mom a picture of juice (S2)is not reinforced, AND, when juice(S4) is available, bringing Mom apicture of juice (S2) is reinforced,whereas bringing Mom a pictureof a cookie (S1) is not.

Conditional Discrimination Teach-

ing Methods. A handy and reliable setof procedures for potentially teachingconditional discriminations is match-to-

sample (MTS). Typically, each of a seriesof MTS trials begins with the presenta-tion of a designated sample (or condi-tional) stimulus to the learner, who is re-quired to respond to it (e.g., by touchingit if it is visual, or touching a blank cardor a key on a computer screen if the sam-ple is auditory). An array of two or more

other stimuli is then presented, called thecomparison, or choice) stimuli. The learneris required to respond to one compari-son, usually by touching or pointing toit. One comparison is designated correct(discriminative for reinforcement; S+)with each sample; that is, responses toit are followed by reinforcer delivery,whereas responses to the other compar-isons in the presence of that sample arenot reinforced (S-). After a brief inter-trial interval, the next trial is presented insimilar fashion. The sample stimulus typ-ically varies unsystematically from trial totrial, as does the position of the S+ com-parison stimulus on each trial.

Match-to-sample procedures are com-monly used to teach learners with autismto match identical visual stimuli (e.g., ob-jects, colors, shapes, letters, numerals) andnonidentical visual stimuli (e.g., picturesto objects, pictures to printed words). Itis important to note, however, that vir-tually any kind of stimuli that can beseen, heard, touched, tasted, or smelledcan be used in MTS procedures (with afew commonsense qualifications; e.g., itis not practical to present two or moreauditory stimuli simultaneously as com-parisons). More importantly, as the fore-going examples illustrate, many func-tional skills call for learners to makeconditional discriminations, whether ornot explicit MTS procedures are in-

volved. Skills that are often termed &dquo;re-

ceptive vocabulary,&dquo; &dquo;receptive identifi-cation,&dquo; or &dquo;language comprehension,&dquo;for instance, require learners to respondto each of a number of visual stimuli (e.g.,body parts, objects, pictures, printedwords) if and only if that stimulus waspreceded by a particular spoken word.The spoken words can be thought of assamples, the visual stimuli as compari-sons. When samples and their corre-

sponding correct comparisons are not

physically identical to one another, theprocedure is often referred to as arbi-

trary matching.Analyzing the components of condi-

tional discriminations reveals precisely whatmakes them more complex than simplediscriminations. To fulfill conditional dis-crimination reinforcement contingencies



Tim Obertein

76

consistently (i.e., to respond correctly onmost opportunities), the learner must

(a) discriminate each sample from everyother sample presented across a series oftrials (these are successive simple discrim-inations), (b) discriminate each compari-son from all other comparisons presentedwithin each trial (simultaneous simple dis-criminations), and (c) relate (match) eachcomparison stimulus with one and onlyone sample stimulus. Put another way, inconditional discriminations the samplestimuli exert stimulus control over thefunctions served by the comparison stim-uli (as S+ or S-) from trial to trial. Thiskind of stimulus control develops onlywhen contingencies consistently requirethe learner to observe the sample stimu-lus on each trial, discriminate it from theother samples and from the comparisons,and discriminate each comparison fromevery other comparison and from thesamples (K. Saunders & Spradlin, 1989,1993; R. Saunders & Green, 1999; Sid-man, 1986, 1994). Teaching proceduresthat arrange other kinds of contingen-cies, explicitly or inadvertently, are likelyto establish something other than the de-sired stimulus control.

During the past 20 years or so, manyresearchers have investigated conditionaldiscrimination learning in developmen-tally young learners and those with de-velopmental disabilities. Some analyzedthe kinds of faulty stimulus control thatcan easily arise from certain arrange-ments of trials and trial sequences (e.g.,Harrison & Green, 1990; Johnson &

Sidman, 1993; McIlvane & Stoddard,1985; Stromer & Osborne, 1982). Oth-ers tested procedures for minimizing er-rors and for establishing conditional dis-criminations with learners who did not

readily acquire them via standard trainingprocedures (e.g., Dube & Serna, 1998;McIlvane & Stoddard, 1981; K. Saun-ders & Spradlin, 1989, 1993; Zygmont,Lazar, Dube, & Mcllvane, 1992). Thefollowing recommended general proce-dures for teaching conditional discrimi-nation skills have been extracted fromthat research (cf. Green & Saunders,1998). They are discussed in MTS terms,but it is important to reiterate that thesame basic guidelines can and should

be followed whenever the objective is toestablish conditional discriminations,whether or not the teaching proceduresare explicitly MTS (e.g., for receptive vo-cabulary instruction, picture-based com-munication training, etc.).

1. Within a session or block of MTS

trials, a different sample should be pre-sented on each trial, but the same com-parisons should appear on every trial.

Each comparison should be the S+ withone and only one sample, and should beS- equally often with each of the othersamples.

2. For most purposes, it is preferableto have at least three comparisons onevery trial. Because each comparison isdesignated correct with one sample, thenumber of different sample stimuli pre-sented in a session or block of trialsshould equal the number of comparisonstimuli presented on each trial.

3. Each sample should be presentedequally often within a session or block oftrials, in unsystematic order. A good ruleof thumb is that the same sample shouldnot be presented on more than two con-secutive trials.

4. The position of the S+ comparisonshould vary unsystematically from trial totrial. A good rule of thumb is that the S+

should never appear in the same positionin the comparison array for more thantwo trials consecutively.Some examples of MTS trials that ful-

fill these parameters are illustrated in Ta-ble 1. Each trial has three comparisonspresented in horizontal arrays with posi-tions designated from the learner’s (ratherthan the instructor’s) perspective. Ofcourse, multiple trials (in multiples ofthree, in this case) should be presented ineach teaching session or block of trials,arranged in accordance with the proce-dures described above.

Any deviation from the arrangementsof balanced trial types and teaching se-quences just described creates the poten-tial for one or more types of extraneousstimulus control to interfere with devel-

opment of the desired control by samplestimuli over the selection of comparisonstimuli. For example, if the S+ appearsmore often in one position in the com-parison array than in the others acrosstrials, the learner’s comparison selectionsmay come to be controlled by positionrather than by sample stimuli. Anothertype of extraneous stimulus control canarise from the common practice of pre-senting novel incorrect comparisons (of-ten referred to as &dquo;distractors&dquo;) periodi-

TABLE 1

Examples of Balanced Three-Choice Match-to-Sample Trials

Note. MTS = match-to-sample.



Tim Obertein

Tim Obertein

Tim Obertein

Tim Obertein

77

cally, sometimes on every trial. The riskwith such a procedure is that learners willrespond either to the novel stimuli or awayfrom them simply because they are novelto the context. In either case, learners

may not observe the sample stimuli andtherefore may not learn the intendedconditional discriminations. If the incor-rect (S-) comparisons presented with anygiven sample change from trial to trialwhile the S+ remains constant, learners

may simply learn to respond away fromthe comparisons that are new from trialto trial. If so, performance will seem to behighly accurate-the learner will touchthe S+ consistently-but, again, she or heneed not observe or discriminate the

sample stimuli in order to satisfy the re-inforcement contingencies (Harrison &

Green, 1990; Johnson & Sidman, 1993).Another common practice that may

create unwanted stimulus control is re-

peated presentation of each sample withits designated correct comparison and noother comparisons, often referred to asteaching &dquo;in isolation&dquo; (cf. Lovaas et al.,1981). For example, in an attempt toteach picture-object matching, the in-

structor may start by presenting a pictureof a spoon (the nominal sample) and anactual spoon (the nominal comparison)for many repeated trials, reinforcingtouches to the spoon. Trials to teach&dquo;fork&dquo; are presented in similar fashion,then &dquo;knife.&dquo; In this procedure, the rein-forcement contingencies do not requirethe learner to discriminate among differ-ent sample stimuli, or among differentcomparison stimuli, because all he or shehas to do is touch the one comparisonthat is available on each trial to earn rein-forcers consistently-a simple, not a con-ditional, discrimination. After two or threesamples and their corresponding correctcomparisons are each presented in isola-tion for a number of trials, the stimuli aretypically mixed so that all comparisons(the spoon, fork, and knife) appear onevery trial but the sample (picture ofspoon, fork, or knife) differs from trial totrial, as in the MTS trials described ear-lier. Many learners with autism performat chance levels when mixed trials areintroduced following isolation training,and their performance often does not im-

prove even with repeated presentations.That is probably because the contingen-cies on the mixed trials require condi-tional discriminations rather than simpleones, but the immediate history of re-inforcement for simple discriminations

may predispose learners to simply trackthe comparison stimuli to which re-

sponses were reinforced most recentlyor most often in the training that justpreceded.

Training in isolation may even inad-vertently teach learners not to attend tosample stimuli at all, because attendingto the sample is not required on thosetrials; it is most efficient for the learner tosimply touch the one available compari-son stimulus quickly in order to maxi-mize reinforcement. This same type of

faulty stimulus control may developwhen children with autism are taught touse pictures presented repeatedly in iso-lation to request preferred items, as inthe initial phases of Picture ExchangeCommunication System training (Frost& Bondy, 1994). These are simple dis-crimination contingencies that do not re-quire the learner to observe the item thatcorresponds to each picture, or even toobserve the features of each of the pic-tures, in order to obtain reinforcers con-

sistently and quickly. All he or she has todo is pick up the only picture that is avail-able and hand it to an adult, to immedi-ately receive a preferred item. Followingsuch training, it would not be surprisingto see many learners perform at chancelevels when two or more pictures (com-parison stimuli) are presented simultane-ously, only one of which matches theitem available on any given requestingopportunity.

5. Learners should be required to makean &dquo;observing response&dquo; to the samplestimulus on each trial (e.g., by pointingto the stimulus or to a blank card or

key when the actual sample cannot betouched, such as a spoken word). Ofcourse, requiring learners to touch sam-ple stimuli does not guarantee that theyare actually looking at or listening to thesamples, but it does make those re-

sponses more likely than simply present-ing samples without requiring an observ-ing response.

6. When the sample stimuli are audi-tory (e.g., spoken words, as in receptivelanguage tasks), the sample should bepresented clearly to start the trial, thenrepeated when the comparison stimuliare presented, and again every 2 secondsor so until the learner responds to a com-parison or to some maximum number ofrepetitions. The rationale is as follows: Ifan auditory sample is presented onlyonce to start a trial, learners have just thatone fleeting opportunity to &dquo;observe&dquo; it.There is considerable risk that (a) theywill not hear it, (b) they will hear it butnot discriminate it from samples pre-sented on other trials, or (c) they willhear it but not remember it throughoutthe interval that elapses while they exam-ine the comparison array and respondto one comparison. Repeating auditorysamples reduces those risks.

Another desirable practice is to limitthe auditory stimulus presented to starteach trial to the word to which one ofthe comparisons is to be matched (e.g.,&dquo;spoon,&dquo; &dquo;fork,&dquo; or &dquo;knife&dquo;), rather thanstarting each trial with a nominal in-

struction like &dquo;Touch ,&dquo; or&dquo;Point to &dquo; If every sample con- -sists of a series of sounds, the first ofwhich are exactly the same from trial totrial while only the last differs, it is likelyto be very difficult for many learners todiscriminate among those sounds and re-

spond to the critical one on each trial. In-stead of using spoken instructions like&dquo;touch&dquo; or &dquo;point to,&dquo; it is preferable toteach learners to point to comparisonsthrough nonverbal methods, such as

physical guidance or modeling, and tohelp them discriminate among auditorysamples by presenting each one distinctlyand repeatedly, unaccompanied by re-

dundant and extraneous words.7. Instead of rearranging the compar-

ison stimuli in front of the learner be-tween trials, prepare the comparison ar-ray for each trial on a mat or board outof the learner’s sight (e.g., behind a screen,on a chair next to you) during the in-tertrial interval. Then present the whole

comparison array after the sample hasbeen presented and responded to. Thishelps reduce the likelihood that thelearner will respond to extraneous cues,



Tim Obertein

Tim Obertein

Tim Obertein

Tim Obertein

Tim Obertein

Tim Obertein

78

such as the comparison stimulus the in-structor touched first or last when she or

he rearranged the comparisons.8. Teach learners how to perform MTS

tasks before trying to teach them newconditional discrimination using those

procedures. Several skills are required toperform these tasks: sitting quietly be-tween trials, orienting to the stimuli,making observing responses to samplestimuli, scanning arrays of multiple com-parisons, and responding to just onecomparison on each trial. If the learnerlacks one or more of these skills, she orhe is likely to make errors on MTS tasksthat might be erroneously interpreted toreflect deficient conditional discriminationskills (Johnson et al., 2000). Similarly,auditory-visual matching and visual-visual matching procedures differ some-what. If the learner does not know howto do tasks that employ both types ofprocedures, his or her performance onone or the other may lead to erroneousconclusions about discrimination skilldeficiencies (Kelly, Green, & Sidman,1998).One very important skill involved in

performing MTS is pointing to the sam-ple stimulus (or to a designated sample&dquo;key&dquo; or card) that starts each trial, andpointing to one comparison stimulus oneach trial. It is strongly recommendedthat learners be taught to point- to sam-ples and comparisons rather than to placesamples on top of or alongside compar-isons (&dquo;put with same;&dquo; Lovaas et al.,1981 ), for a couple of reasons: (a) Point-ing is a much more generally useful skillthan &dquo;putting with.&dquo; Many stimuli thatlearners are required to match in every-day situations (such as spoken words andcorresponding objects, people, or actions)cannot be physically &dquo;put with&dquo; one an-

other, but learners can readily indicatethat such stimuli go together by point-ing to them; (b) teaching matching skillsby requiring learners to put identical orphysically similar stimuli on top of or

alongside one another may establishstimulus control by identical outlines orother identical features of the stimuli, be-cause that is all the learner needs to at-tend to in order to obtain reinforcement

consistently. That could block the devel-

opment of stimulus control by the rele-vant features of stimuli and put thelearner at a distinct disadvantage when heor she is confronted with stimuli thatcannot be readily matched by aligningidentical features.

9. Use errorless teaching methodsrather than trial-and-error procedures toteach conditional discriminations. Thereis a large body of research on errorlessdiscrimination learning that goes backmany years (e.g., Sidman & Stoddard,1967; Terrace, 1963) and provides richinformation about the benefits of, andmethods for, minimizing errors early intraining. (Few, if any, truly &dquo;errorless&dquo;

training procedures exist, but there aremany that can maximize the probabilitythat correct responses will occur.) Someof that research has shown that errorslead to further errors and emotional re-

sponses that can interfere with acquisi-tion, can be difficult to correct, and caninhibit skill generalization (see Hecka-mon, Alber, Hooper, & Heward, 1998;MacDuff, Krantz, & McClannahan, inpress ) .

Errorless teaching methods generallyentail the addition of stimuli that reli-

ably control the target response-that is,prompts-to the target antecedents at

the beginning of instruction. Promptsare reduced (faded) systematically acrosssuccessive trials in an effort to transferstimulus control to the target antece-

dents, resulting in final performancesthat are unprompted. Put another way,errorless teaching methods are most-to-least prompt- and prompt-fading meth-ods : The learner is given the most assis-stance necessary on initial trials to ensurethat the target response occurs so thatit can be reinforced frequently, and tominimize errors; the amount of assistance isthen systematically decreased as long asthe learner continues to respond cor-rectly. If an error occurs, there is typicallya provision for &dquo;backing up&dquo; to the pre-ceding prompt level on the very next trialto reduce the likelihood that anothererror will occur. Following a correct

response, the systematic fading processresumes.

Errorless discrimination training meth-ods can and should be used to teach all

kinds of new skills to learners with au-

tism, because virtually all skills entail dis-criminations. Strategies for minimizingerrors and prompt dependency whileteaching such things as communicationand play skills are discussed further in thesection on prompting and prompt fad-ing. Here, some specific methods for teach-ing discriminations with few errors arehighlighted briefly. One category of er-rorless discrimination training methods iswithin-stimulus prompts. These methodsinvolve altering the physical characteris-tics of the stimuli to be discriminated to in-crease the likelihood that correct responseswill occur early in training. For example,simple or conditional discriminations canbe taught by making the stimuli differ inintensity on initial teaching trials. Withvisual stimuli, such as printed letters orline drawings, the S- might be made toappear very faint and the S+ dark. If the

intensity difference is an effective prompt,the learner will respond to the more in-tense stimulus. Over successive trials, theintensity difference is gradually decreasedin a series of graded steps-that is, the S-gradually becomes darker (or louder)-until both the S+ and the S- are pre-sented at the same intensity. If the learnercontinues to respond reliably to the S+and not to the S-, it is evidence that stim-ulus control has transferred from the

prompt to the target stimulus. This pro-cedure is usually termed intensity fading.Size difference can also serve as a within-stimulus prompt. For example, the S-might be made very small and the S+large on initial teaching trials. Over suc-cessive trials the size difference is gradu-ally decreased, analogous to intensity fad-ing, until the S+ and S- are the same size(sometimes called size fadin8).Another within-stimulus prompting

procedure involves altering the physicalappearance of only those features of thestimuli that differentiate them from oneanother. An example is exaggerating thevertical parts of the lowercase letters b

and p on early training trials and gradu-ally making them appear more and morealike over succeeding trials. Such proce-dures are often referred to as criterion-related prompting. In general, criterion-related prompting has been found to be



Tim Obertein

Tim Obertein

79

more effective for teaching discrimina-tions to learners with disabilities thannoncriterion-related prompting, but theefficacy of the various within-stimulusprompting methods can vary with thecharacteristics of the stimuli involved, thelearner’s skills, and the complexity of thediscriminations being trained (for reviewssee Demchak, 1990; Etzel & LeBlanc,1979; MacDuff et al., in press).A unique within-stimulus prompting

procedure for teaching arbitrary MTShas been termed sample stivnulus controlshaping. In this procedure, the initial

teaching trials are identity MTS trials pre-senting samples that are physically iden-tical to their corresponding S+ compar-isons (e.g., identical line drawings). Oversuccessive trials, the sample stimuli aregradually transformed into entirely dif-ferent stimuli (e.g., line drawings changeinto printed letters). Sample stimuluscontrol shaping has proven efficaciousfor teaching arbitrary conditional discrim-inations to learners who can do condi-tional identity MTS but have failed to ac-quire arbitrary conditional discriminationsthrough other training methods (Carr,Wilkinson, Blackman, & McIlvane, 2000;Zygmont et al., 1992).One of the major drawbacks of within-

stimulus prompting procedures is that agreat deal of advance stimulus prepara-tion is required, which can make themdifficult to use in typical instructional set-tings. (Many recent studies on this topictook advantage of computer technologyand rather sophisticated software to ac-complish very precise stimulus transfor-mations, timing, backing up and advanc-ing through prompt hierarchies, etc.). Analternative errorless teaching method ise.~^a-stimulus prov~cptin8, in which promptsare separate from and external to the stim-uli to be discriminated (Etzel & LeBlanc,1979; MacDuff et al., in press). One ex-ample of an extra-stimulus prompt in thecontext of teaching conditional discrimi-nations is an instructor’s gesture (such asa point) toward the correct comparisonstimulus on initial teaching trials. In

other contexts, extra-stimulus promptscould be models provided by the in-

structor ; for example, to teach picturenaming to a learner who reliably imitates

spoken words, the instructor might holdup pictures one at a time and simultane-ously model the name of each picture.

Extra-stimulus prompts can be faded

along any of several dimensions, such asintensity, distance relative to the targetstimulus, or time relative to presentationof the target stimulus. When prompts arefaded along the time dimension, the pro-cedure is often referred to as time delay(although delayed cue or delczyed promptis more accurate, as it is the prompt ratherthan time that is delayed). For instance,when an instructor’s point to the S+comparison is used as an extra-stimulusprompt within MTS procedures, oninitial trials the prompt is provided im-mediately upon presentation of the com-parison array. If the learner follows theprompt and responds correctly for a spe-cified number of trials, on subsequenttrials the instructor begins to delay theprompt for a short period of time (say,1 sec) after presentation of the compari-son array. After a specified number ofcorrect responses at that delay interval,the delay is gradually increased (e.g., in1-sec increments) over successive trials.Such a procedure is referred to as pro-gressive delczy. Alternatively, after a speci-fied number of trials on which the promptis provided immediately ( 0-delay) or

after a brief delay (e.g., 1 sec), the delayinterval might be increased to a certainvalue (say, 5 sec) that is maintained for allsubsequent trials. This is termed a con-stant or fixed delczy procedure. If delayedprompting is effective, the learner beginsat some point to respond correctly beforethe prompt is provided (sometimes re-ferred to as &dquo;anticipations&dquo;), probably be-cause preempting the prompt reduces thedelay to reinforcement. Delayed prompt-ing has been shown to be an effectivemethod of transferring stimulus controlfrom extra-stimulus prompts to targetstimuli in some studies, but it is not uni-versally effective. The main risk in usingthis method is that learners will not an-

ticipate the prompt but will simply waitfor it on trial after trial, probably becausewaiting for the prompt maximizes theprobability of reinforcement. Progressivedelay procedures in particular can shapewaiting behavior. Therefore, like all prompt

and prompt-fading procedures, theseshould be used judiciously (Etzel & Le-

Blanc, 1979; MacDuff et al., in press;Oppenheimer, Saunders, & Spradlin,1993).

Developing Stimulus EquivalenceClasses. One of the most exciting areasof research in contemporary behavior

analysis is stimulus equivalence. Most ofthe work on this topic has been con-ducted within a conceptual and analyticalframework developed by Murray Sidmanand his colleagues (Sidman, 1971, 1994;Sidman et al., 1982; Sidman & Tailby,1982). In a nutshell, the Sidman modelprovides methods for analyzing howphysically dissimilar stimuli come to betreated as equivalent to, or substitut-able for, one another in certain contexts.Scores of experiments using these meth-ods have shown that after certain arrange-ments of conditional discriminations aretrained (usually, though not always, withMTS procedures), other conditional dis-criminations emerge without any further

training or direct reinforcement. In hisfirst experiment on stimulus equivalence,Sidman (1971) set out to investigate sight-word reading in a young man with severemental retardation. When he entered the

experiment, the young man could match20 common pictures (e.g., a bed, a hat)to their corresponding dictated Englishnames, but he did not match corre-

sponding printed words to those samepictures, or to the dictated words; that is,he did not have a sight-word readingrepertoire with those words. After he wastrained with standard MTS proceduresto match the printed words to the dic-tated words, the young man provedimmediately capable of matching the

pictures to their corresponding printedwords and vice versa, without further in-struction or reinforcement. Those out-comes documented the development of20 classes of equivalent stimuli, each con-sisting of a corresponding picture, dic-tated word, and printed word. This basicfinding has been replicated hundreds oftimes by many investigators using a

wide array of visual, auditory, tactile, andolfactory stimuli with a wide array of

learners, including nonhumans, typically



80

developing children and adults, and in-dividuals with various disabilities (e.g.,see Green & Saunders, 1998; Sidman,1994).Much of the interest in the stimulus

equivalence model stems from the factthat it enables experimental analyses ofphenomena that were long consideredoutside the realm of behavior analysis,such as concept formation, transitive in-ferences, symbolic learning, and genera-tive behavior (Mcllvane, Dube, Green,& Serna, 1993). For practitioners, stim-ulus equivalence methods can provide alot of &dquo;bang for the buck,&dquo; because afterjust a few conditional discriminations areestablished through direct training, manyothers typically emerge &dquo;for free,&dquo; with-out any additional instruction whatso-ever. Additionally, generative (untrained)performances, which have been notori-ously difficult to produce in learners withautism and related disorders, occur quitereliably when stimulus equivalence pro-cedures are used.

For example, in one experiment an ado-lescent with severe autism who could matchthe printed numerals 1 through 6 to cor-responding dictated number names wastrained with MTS procedures to matchquantitites of one to six dots (each ar-ranged in three different configurationson cards) to those same dictated numbernames. He then matched the quantitiesto corresponding printed numerals andvice versa on unreinforced probe trials,making very few errors. Similarly accu-rate performances were demonstrated onunreinforced probes with quantities ofpennies and of pictures of apples andhouses-stimuli that had never been pre-sented in training-appearing as compar-ison stimuli with dictated number namesas samples, as comparisons with printednumerals as samples, and as samples withthe numerals as comparisons. Collectively,these performances documented the de-velopment of six equivalence classes, eachcontaining a dictated number name, thecorresponding printed numeral, and cor-responding quantities of dots, pennies,pictures of apples, and pictures of houses(Green, 1992). Equivalence classes ofGreek letters and their dictated nameswere demonstrated by high-functioning

preschoolers with autism in another ex-periment (Eikeseth & Smith, 1992).Numerous other experiments with par-

ticipants with learning difficulties havedemonstrated the efficacy and efficiencyof stimulus equivalence procedures fordeveloping classes of stimuli relevant tosuch skills as reading, spelling, math, andaugmentative or alternative communi-cation (for overviews see Remington,1994; Sidman, 1994; Stromer, Mackay,& Stoddard, 1992). Like the aforemen-tioned discrimination training methods,the rich technology for developing stim-ulus equivalence classes does not seem tohave made its way from laboratory re-search into widespread application withlearners with autism, so this is an areathat is ripe with possibilities for field re-search and practice.

Promoting Independenceand Initiations

Prompting and Prompt Fading.Prompting procedures were discussedearlier in the context of teaching dis-

criminations, but they have broad utilityfor teaching all sorts of new skills to

learners with autism. Prompts, or auxil-iary antecedent stimuli, can take manyforms: physical guidance, gestures, mod-els, verbal cues, other types of auditorystimuli, colors, pictures of various kinds,written text, and tactile stimuli. Promptslike manual guidance are often necessaryto get new responses to occur in learn-ers with autism who are just entering in-struction, and it is often necessary to useprompting procedures to make the mostefficient use of instructional time. Butwhenever prompts are used, there is therisk that learners will become dependenton them so that their responding does notcome under the control of the relevantenvironmental stimuli (Demchak, 1990;MacDuff et al., in press). Therefore, con-siderable research in applied behavior

analysis has focused on instructional

techniques that preclude or minimize

prompt dependence and promote spon-taneous, independent performances in

learners with autism. Because adult-delivered prompts (especially verbal cues)can be particularly difficult to fade, some

stimulus control research has focused on

techniques that either fade adult-deliveredprompts very rapidly after they are intro-duced, or never introduce them in thefirst place. Again, an exhaustive review isbeyond the scope of this article, so just afew recent advances and novel applica-tions are presented here as examples.

Activity schedules. A very powerfuland flexible technology for promotingindependent performances in learners withautism has been developed primarily byscientist-practitioners at the Princeton

Child Development Institute. Learnersare taught to use activity schedules con-sisting of photos or printed wordsarranged in notebooks to guide extendedsequences of behavior in the absence ofadult instruction or supervision. Initially,adults deliver nonverbal prompts (usuallymanual guidance) and reinforcers frombehind the learner. In this fashion, thelearner is taught to touch a photo or textthat represents a particular task or action(such as assembling a puzzle), obtain therelevant materials, complete the task oraction, turn to the next page in the note-book, and repeat the process with thephoto or text shown there. The manualprompts are faded rapidly as long as cor-rect responding occurs; when they havebeen completely removed, adult proxim-ity is also faded. Adult-delivered rein-forcers are also thinned, and shifted fromprompted to unprompted responses.This training is designed to establish eachof a series of photos or textual stimuli(rather than adult-delivered instructions)as discriminative for lengthy chains of be-havior. Once a basic schedule-followingrepertoire has been established, the orderof stimuli in the activity schedule is var-ied. Activity schedules can incorporatestimuli that are discriminative for inde-

pendent performance of a host of self-care, domestic, vocational, leisure, aca-demic, and social activities in a varietyof settings. Research and practice haveshown that individuals with autism can

readily learn to choose and sequence theirown activities, and that activity schedulescan be transformed into the types ofschedules that many of us use routinely,such as calendars, appointment books,class schedules, and &dquo;to do&dquo; lists (Krantz,



81

MacDuff, & McClannahan, 1993; Mac-Duff, Krantz, & McClannahan, 1993;McClannahan & Krantz, 1999). It is im-portant to note that this technology dif-fers considerably from other uses of vis-ual schedules with learners with autism,such as depictions of daily activities for anentire class of students, and pictures orprinted words on cards that the learnermoves from a centrally located envelopeto the physical locations in which each ofa series of activities takes place (as in theTEACCH model; e.g., Schopler, Mesi-bov, & Hearsey, 1995).

Other nonverbal prompts. As notedpreviously, just about any type of stimu-lus that reliably sets the occasion for tar-get behaviors can be used in prompt and

prompt-fading procedures. For instance,once learners with autism have reliable

reading repertoires, prompts in the formof written or typed words can be usedeffectively to get them to engage in de-sirable behaviors without verbal or otherinstructions from adults. Such textual

prompts can be as simple as single wordsor short phrases embedded in an activityschedule (such as &dquo;Take a shower&dquo; or &dquo;TellDad a joke&dquo;), or as complex as lengthyscripts that specify a series of interactionsamong two or more people. In one ex-ample of the latter, adolescents with au-tism were taught to initiate and respondto verbal statements (i.e., to engage inconversational exchanges) in the contextof everyday activities, such as completingclassroom art projects (Krantz & Mc-

Clannahan, 1993). In another, youngchildren with autism and minimal read-

ing skills learned to initiate comments toothers by responding to such textual

prompts as &dquo;Look&dquo; and &dquo;Watch me&dquo; in

their activity schedules (Krantz & Mc-

Clannahan, 1998). Textual prompts areoften faded by gradually removing com-ponents (words or letters) from the lastcomponent to the first.One innovative procedure makes use

of a tactile prompt in the form of a vi-

brating beeper. Learners with autism havebeen taught to initiate verbal play bidsand conversations with adults, then withpeers, in response to activation of the

beeper. Initially, the learner’s hand is

placed on the device, and when the de-

vice is activated, an adult models a verbalstatement for the learner to imitate as aninitiation to another person. The verbalmodel is faded until the learner reliablyinitiates responses when he or she feelsthe device vibrate, without an adult-

delivered prompt. The device is then

placed in the learner’s pocket and used toprompt initiations to peers. This type of

prompt has the multiple advantages ofbeing portable and unobtrusive so that itcan be used in various settings, such asclassrooms, and functional in the absenceof adult cues or reinforcer delivery (Hol-berton, Taylor, & Levin, 1998; Taylor &

Levin, 1998).Automated auditory prompts. Ver-

bal prompts from adults or other chil-dren can impede spontaneity on the partof the learner with autism and can be dif-ficult to fade, but verbal prompts pre-sented via automated devices can be veryeffective because they do not require theconstant presence of another person; learn-ers with autism can be taught to activatethe devices themselves. Such promptscan also be faded very systematically andprecisely and (unlike textual prompts)can be used with nonreaders.One useful automated system for es-

tablishing and fading verbal prompts isthe Language Master (Stevenson, Krantz,& McClannahan, 2000; Taylor, in press).Spoken statements or questions that areto serve as models for the learner are re-corded on cards with electromagneticstrips. The cards are run through theLanguage Master machine, which playsthe recorded message. These auditoryprompts can be faded gradually by re-cording several versions of the message,each of which eliminates one word fromthe end of the previous version, thusfading the prompt across successive ver-sions. Alternatively, the prompt can beremoved all at once by activating a block-ing option on the Language Master ma-chine that renders the recorded messageinaudible when the card is run throughit. Studies have shown that after learn-ers with autism learn to imitate verbalmodels provided via Language Masters,prompt and prompt-fading techniquesusing those devices are effective for

teaching the learners to initiate requests

or comments to peers (e.g., &dquo;Susie, let’steeter-totter&dquo; ) and to engage in con-

versations. Another automated auditoryprompting technique involves recordingspoken instructions on audiotape andteaching the learner with autism to de-liver them to himself via a self-operatedpersonal tape player and headphones(e.g., Taber, Seltzer, Heflin, & Alberto,1999). Like the vibrating beeper, this

type of prompting system is portable andcan be used in various settings withoutdisturbing others or drawing undue at-tention to the learner with autism. The

utility and generality of automated audi-tory prompting systems remain to be in-vestigated fully.

Video Modeling. A few studies havedemonstrated that learners with autismcan imitate various skills after viewingvideotaped segments of themselves orothers performing those skills. For ex-

ample, three young adults with autismwere taught purchasing skills in class-rooms and in either their school cafeteriaor a nearby convenience store (Haring,Kennedy, Adams, & Pitts-Conway, 1987).Probes conducted in three communitysettings indicated that very little skill

generalization occurred. After the youthsviewed videotapes of same-age nonhan-dicapped peers making purchases in threecommunity settings and responded to aninstructor’s questions about what theywere viewing, generalization increasedsubstantially. Another set of investiga-tors reported acquisition and generaliza-tion of conversational speech (three briefquestion-and-answer exchanges with

therapists) in three young children withautism after the children observed video-

tapes of familiar adults engaging in suchconversations, and then practiced theconversations with a therapist (Charlop& Milstein, 1989). More recently, the ef-ficacy of video self-modeling was evalu-ated with three children with autism. Thechildren viewed specially prepared video-tapes of themselves answering a series ofquestions presented by adults. The in-vestigators reported that after viewingthe videotapes several times, all threechildren showed increased accuracy of

responding to the same questions in vivo



82

(Buggey, Toombs, Gardener, & Cervetti,1999).Video modeling is another potentially

promising technique that is ripe for fieldand laboratory research. Some questionsthat need to be addressed in careful ex-

perimental analyses include (a) whetherthe initial studies on this topic can bereplicated by independent investigators,(b) the entry skills required for learnerswith autism to benefit from video in-

struction, (c) the efficacy of video mod-eling for establishing an initial imitativerepertoire (including generalized imita-tion), (d) factors that may account forthe rapid skill acquisition and broad gen-eralization reported in the few studiespublished to date (if indeed those find-ings can be replicated), (e) whether

simply viewing video models (withoutpracticing the target skills) suffices to es-tablish new skills, and (f) the compara-tive efficacy of in vivo and video model-ing for producing skill acquisition andgeneralization.

Priming. Another antecedent manip-ulation that bears further careful investi-

gation with learners with autism has beentermed &dquo;priming.&dquo; This term generallyrefers to providing learners with a pre-view of upcoming events by presentingthe stimuli and/or activities involved inthose events under low-demand, high-reinforcement conditions (Wilde, Koe-gel, & Koegel, 1992). In one study, forexample (Zanolli, Daggett, & Adams,1996), each of two preschool boys withautism was prompted by a teacher to ini-tiate verbal and nonverbal interactionswith a typically developing peer. The peerprovided reinforcers to the child withautism contingent on correct responses.These priming sessions were followed im-mediately by sessions in which the childwith autism was given opportunities toengage in preferred activities with a

trained peer, without adult-delivered

prompts or reinforcers for the child withautism. Unprompted initiations by bothboys with autism, which were at near-

zero rates during a baseline phase, werereported to increase after the priming in-tervention. Recently, priming in the formof viewing a commercial toilet-trainingvideo was reported to increase the fre-

quency of initiations to use the toilet andof dry diapers with a preschooler withautism (Bainbridge & Myles, 1999). Inanother study, previewing specially pre-pared videos of settings in which threeyoung children with autism had exhib-ited disruptive behavior was reported tosubstantially decrease the problem be-havior exhibited by all three children inthe actual settings (Schreibman, Whalen,& Stahmer, 2000).Many important questions about prim-

ing remain to be addressed in well-controlled studies, some of them parallelto the questions about video modelingraised earlier. First, the handful of pub-lished studies on this topic need to bereplicated. Replications and extensionsshould specify the entry skills of thelearners and address questions about the&dquo;active ingredients&dquo; in this intervention.For example, is it necessary for learnerswith autism to practice target skills orotherwise respond actively during prim-ing sessions (as in Zanolli et al., 1996),or is mere exposure to the priming stim-uli (as in Schreibman et al., 2000) suffi-cient to change behavior? What is the re-lation between the rate or probability ofreinforcer delivery in priming sessionsand behavior change in the subsequentactivity? It has been hypothesized thatpriming works by making future eventspredictable for learners with autism

(Schreibman et al., 2000). That raisesquestions as to the comparative efficacyof priming and other, perhaps briefer andsimpler methods of previewing what isabout to happen, such as providing a

simple spoken description or showingthe learner a picture. The predictabilityhypothesis also suggests that the degreeof correspondence between the primingactivity and the actual activity might berelated to the occurrence of the targetbehavior in the actual situation: Do prim-ing activities that are very similar to

the actual situations produce behaviorchange more reliably than priming activ-ities that bear less resemblance to the ac-tual situations?

Incidental Teaching and Other&dquo;Naturalistic&dquo; Techniques. No articleon behavior analytic methods for pro-moting independent, unprompted be-

havior in learners with autism would be

complete without a discussion of inci-dental teaching and related techniquesfor promoting spontaneous communica-tion. Some popular misconceptions arethat traditional applied behavior analysiseschews everything but highly structured,adult-directed instructional procedures,and that only the more &dquo;contemporary&dquo;behavioral approach has incorporatednaturalistic methods drawn from the de-

velopmental social-pragmatic approachto communication enhancement (e.g.,Prizant & Wetherby, 1998; Wetherby,Schuler, & Prizant, 1997). An even moreegregious misrepresentation is that &dquo;ap-proaches to enhance spontaneous lan-guage use, including incidental teachingand pivotal response training, were de-veloped primarily because of concernsabout children’s inability to generalize com-municative use of language ‘learned’ indiscrete trial training&dquo; (Prizant & Rubin,1999, p. 204).The fact is that incidental teaching

methods were developed in the 1960s byBetty Hart and Todd Risley, who begantheir long collaboration at the Instituteof Child Development at the Universityof Washington under the leadership ofSidney Bijou, as recounted in the intro-duction to this article. Working withdisadvantaged preschoolers, Hart and

Risley found that the children’s sponta-neous use of color-noun combinationsincreased when access to snacks and playmaterials (naturalistic consequences) weremade contingent on language produc-tion. They went on to specify incidentalteaching procedures for elaborating lan-guage, as follows: (a) Make available sev-eral items of interest to the child; (b) waitfor the child to show interest in an

item and initiate an interaction about it;(c) ask the child for approximations tospeech, or for more elaborate language ifhe or she offers some speech (providinga model for the child to imitate, if neces-sary) ; (d) when the child makes the re-quested response, provide her or himwith the item for which she or he initi-ated (Hart & Risley, 1968, 1982, 1995).As mentioned at the beginning of this ar-ticle, Hart and Risley were among the pi-oneers who integrated the principles andmethods of the experimental analysis of



Tim Obertein

Tim Obertein

83

behavior with work on typical and atypi-cal child development in the early 1960s,establishing the foundations of appliedbehavior analysis. Their incidental teach-ing methods have been part and parcel ofapplied behavior analysis since that time.The efficacy of those methods for teach-ing learners with autism was first docu-mented in 1983 (McGee, Krantz, Ma-son, & McClannahan, 1983); several otherstudies followed (Fenske, Krantz, & Mc-

Clannahan, in press; McGee, Morrier, &

Daly, 1999).An approach to promoting sponta-

neous speech in learners with autism thatbears a strong resemblance to incidental

teaching is the Natural Language Para-digm, or NLP (e.g., Koegel, O’Dell, &

Koegel, 1987). In this and similar mod-els, various age-appropriate items foundin the child’s natural environment aremade available, and the interventionistwaits for the child to choose one. The in-terventionist may label the item or ex-

plicitly prompt speech production by thechild. All attempts at speech, includingapproximations, are reinforced immedi-ately by providing the selected item tothe child. These teaching trials are em-bedded in a series of reciprocal interac-tions between interventionist and child

(see Koegel, 1995).It should be noted that although inci-

dental teaching and NLP procedures areoften characterized as child-initiated be-cause learning opportunities begin whenthe child shows interest in a preferreditem, it is often necessary (and desirable)for interventionists to prompt a response.At that point the instruction becomesadult-directed. Adults also arrange ante-cedents and control access to reinforcersin typical incidental teaching and NLPinteractions. Therefore, the difference be-tween these naturalistic procedures anddiscrete-trial procedures for teachingcommunication skills to learners with au-tism is not as substantial as some have

suggested. Both incidental-teaching anddiscrete-trial procedures have been shownto be effective for teaching communica-tion skills to children with autism, andboth are important because they teachdifferent types of verbal behavior (Fenskeet al., in press; Sundberg & Partington,1999).

Conclusions

The historical antecedents to current ap-plications of behavior analysis with learn-ers with autism included research and

theory in child development as well asthe principles and methods of the exper-imental analysis of behavior. The ensuingfour decades of field research have pro-duced a very wide array of behavior ana-

lytic techniques for building a very widearray of useful repertoires in learners withautism. Other techniques have been de-veloped in laboratory research, particu-larly in the area of stimulus control, buttheir potential for application in autismremains largely unrealized. Transfer ofbehavioral technology for teaching sim-ple and complex discriminations fromthe laboratory to the field, together withfurther field research on some stimuluscontrol techniques that have been de-veloped recently in applied settings, areamong the most richly promising ave-nues for the development of increasinglyeffective instructional methods for learn-ers with autism.

ABOUT THE AUTHOR

Gina Green, PhD, is director of research at theNew England Center for Children and re-search associate professor, E. K. Shriver Center,University of Massachusetts Medical School.Her current interests include stimulus control,autism and related disorders, and disseminat-ing behavioral science to the public. Address:Gina Green, New England Center for Chil-dren, 33 Turnpike Road, Southborough, MA01772-2108.

AUTHOR’S NOTES

Preparation of this manuscript was sup-ported by USPHS Research Grants No.

POIDD03610 from the National Institute onDeafness and Other Communication Disordersand No. P01HD25995 from the National In-stitute of Child Heczlth cznd Human Develop-ment to the E. K. Shriver Center, University ofMassachusetts Medical School, and by the NewEngland Center for Children.

REFERENCES

Anderson, S. R., & Romanczyk, R. G.

(1999). Early intervention for young chil-

dren with autism: Continuum-based be-havioral models. Journal of the Associationfor Persons with Severe Handicaps, 24, 162-173.

Anderson, S. R, Taras, M., & Cannon, B.

(1996). Teaching new skills to young chil-dren with autism. In C. Maurice, G. Green,& S. C. Luce (Eds.), Behavioral interven-tion for young children with autism: A man-ual for parents and professionals (pp. 181-194). Austin, TX: PRO-ED.

Baer, D. M., Wolf, M. M., & Risley, T. R(1968). Some current dimensions of ap-plied behavior analysis. Journal of AppliedBehavior Analysis, 1, 91-97.

Bainbridge, N., & Myles, B. S. (1999). Theuse of priming to introduce toilet trainingto a child with autism. Focus on Autism andOther Developmental Disabilities, 14, 106-109.

Bijou, S. W. (1996). Reflections on some earlyevents related to behavior analysis of childdevelopment. The Behavior Analyst, 19, 49-60.

Buggey, T., Toombs, K., Gardener, P., &

Cervetti, M. (1999). Training respondingbehaviors in students with autism: Usingvideotaped self modeling. Journal of Posi-tive Behavior Interventions, 4, 205-214.

Carr, D., Wilkinson, K. M., Blackman, D., &

McIlvane, W. J. (2000). Equivalence classesin individuals with minimal verbal reper-toires. Journal of the Experimental Analysisof Behavior, 74, 101-114.

Charlop, M. H., & Milstein, J. P. (1989).Teaching autistic children conversational

speech using video modeling. Journal ofApplied Behavior Analysis, 22, 275-285.

Demchak, M. (1990). Response promptingand fading methods: A review. Americanjournal on Mental Retardation, 94, 603-615.

Dube, W. V., & Serna, R. W. (1998). Re-evaluation of a programmed method toteach generalized identity matching to sam-ple. Research in Developmental Disabilities,19, 347-379.

Eikeseth, S., & Smith, T. (1992). The de-velopment of functional and equivalenceclasses in high-functioning autistic children:The role of naming. Journal of the Experi-mental Analysis of Behavior, 58, 123-133.

Etzel, B. C., & LeBlanc, J. M. ( 1979). Thesimplest treatment alternative: The law ofparsimony applied to choosing appropriateinstructional control and errorless-learningprocedures for the difficult-to-teach child.Journal of Autism and Developmental Dis-orders, 9, 361-382.

Fenske, E. C., Krantz, P. J., & McClannahan,L. E. (in press). Incidental teaching: A



84

not-discrete-trial teaching procedure. InC. Maurice, G. Green, & R M. Foxx (Eds.),Making a difference: Behavioral interven-tion for autism. Austin, TX: PRO-ED.

Ferster, C. B., & DeMyer, M. K. (1961). Thedevelopment of performances in autistic

children in an automatically controlled en-vironment. Journal of Chronic Diseases, 13,312-345.

Frost, L. A., & Bondy, A. S. (1994). The Pic-ture Exchange Communication Systemtraining manual. Cherry Hill, NJ: PyramidEducational Consultants.

Green, G. ( 1992 ). Stimulus control technol-ogy for teaching number/quantity equiva-lences. Proceedings of the Conference ofthe National Association for Autism (Aus-tralia), pp. 51-63. Melbourne, Australia:Victorian Autistic Children’s & Adults’ As-

sociation.

Green, G. (1996). Early behavioral interven-tion for autism: What does research tell us?In C. Maurice (Ed.), G. Green, & S. Luce

(Co-Eds.), Behavioral intervention foryoung children with autism: A manual forparents and professionals (pp. 29-44).Austin, TX: PRO-ED.

Green, G., & Saunders, R R. (1998). Stimu-lus equivalence. In K. A. Lattal & M. Per-

one (Eds.), Handbook of research methodsin human operant behavior (pp. 229-262).New York: Plenum.

Hall, L. J. (1997). Effective behavioural

strategies for the defining characteristics ofautism. Behavior Change, 14, 139-154.

Haring, T. G., Kennedy, C. H., Adams, M. J.,& Pitts-Conway, V. (1987). Teaching gen-eralization of purchasing skills across com-munity settings to autistic youth usingvideotape modeling. Journal of Applied Be-havior Analysis, 20, 89-96.

Harrison, R. J., & Green, G. (1990). Devel-opment of conditional and equivalence re-lations without differential consequences.Journal of the Experimental Analysis of Be-havior, 54, 225-237.

Hart, B., & Risley, T. R, (1968). Establishinguse of descriptive adjectives in the sponta-neous speech of disadvantaged preschoolchildren. Journal of Applied Behavior Anal-ysis, 1, 109-120.

Hart, B., & Risley, T. R. (1982). How to useincidental teaching for elaborating lan-

guage. Austin, TX: PRO-ED.Hart, B., & Risley, T. R. (1995). Meaningful

differences in the everyday experience of youngAmerican children. Baltimore: Brookes.

Heckamon, K. A. Alber, S., Hooper, S., &

Heward, W. L. (1998). A comparison ofleast-to-most prompts and progressive time

delay on the disruptive behavior of studentswith autism. Journal of Behavioral Educa-tion, 8, 171-201.

Holberton, L., Taylor, B. A., & Levin, L.(1998, October). The use of a tactile promptwith a child with autism to increase play-related comments to peers. Paper presented atthe annual meeting of the New York Asso-ciation for Behavior Analysis, Saratoga, NY.

Johnson, C., & Sidman, M. (1993). Condi-tional discrimination and equivalence rela-tions: Control by negative stimuli. journalof the Experimental Analysis of Behavior, 59,333-347.

Johnson, C., White, D., Green, G., Langer, S.,& MacDonald, R. (2000, May). A dis-crimination and stimulus equivalence cur-riculum for individuals with severe learningdifficulties. Paper presented at the annualmeeting of the Association for BehaviorAnalysis, Washington, DC.

Keenan, M., Kerr, K. P., & Dillenburger, K.(2000). Parents’ education as autism thera-pists. London: Jessica Kingsley.

Kelly, S., Green, G., & Sidman, M. (1998).Visual identity matching and auditory-visual matching: A procedural note. Jour-nal of Applied Behavior Analysis, 31, 237-243.

Koegel, L. K. (1995). Communication andlanguage intervention. In R. Koegel &

L. Koegel (Eds.), Teaching children withautism (pp. 17-32). Baltimore: Brookes.

Koegel, R. L., & Koegel, L. K. (1995). Teach-ing children with autism. Baltimore: Brookes.

Koegel, R. L., O’Dell, M. C., & Koegel,L. K. (1987). A natural language paradigm for teaching non-verbal autistic children.

Journal of Autism and Developmental Dis-orders. 17, 187-199.

Krantz, P. J., Mac Duff, M. T., McClannahan,L. E. (1993). Programming participation in family activities for children with autism:Parents’ use of photograhic activity sched-ules. Journal of Applied Bahavior Analysis,26, 137-138.

Krantz, P. J., & McClannahan, L. E. (1993).Teaching children with autism to initiate topeers: Effects of a script-fading procedure.Journal of Applied Behavior Analysis, 26, 1121-132.

Krantz, P. J., & McClannahan, L. E. (1998).Social interaction skills for children withautism: A script-fading procedure for be-ginning readers. Journal of Applied Behav- 1

ior Analysis, 31, 191-202.Lovaas, O. L, Ackerman, A., Alexander, D.,

Firestone, P., Perkins, M., Young, D. B.,Carr, E. G., & Newsom, C. (1981). Teach-

ing developmentally disabled children: TheME book. Austin, TX: PRO-ED.

Lovaas, O. 1., & Smith, T. (1989). A com-prehensive behavioral theory of autistic

children: Paradigm for research and treat-ment. Journal of Behavior Therapy and Ex-perimental Psychiatry, 20, 17-29.

MacDuff, G. S., Krantz, P. J., & McClanna-

han, L. E. (1993). Teaching children withautism to use photographic activity sched-ules : Maintenance and generalization ofcomplex response chains. Journal of Ap-plied Behavior Analysis, 26, 89-95.

MacDuff, G. S., Krantz, P. J., & McClanna-

han, L. E. (in press). Prompts and prompt-fading strategies for people with autism. InC. Maurice, G. Green, & R M. Foxx

(Eds.), Making a difference: Behavioral in-tervention for autism. Austin, TX: PRO-ED.

Matson, J. L., Benavidez, D. A., Compton,L. S., Paclawskyj, T., & Baglio, C. (1996).Behavioral treatment of autistic persons: A

review of research from 1980 to the pres-ent. Research in Developmental Disabilities,17, 433-465.

McClannahan, L. E., & Krantz, P. J. (1999).Activity schedules for children with autism:Teaching independent behavior. Bethesda,MD: Woodbine House.

McGee, G. G., Kranz, P. J., Mason, D., &

McClannahan, L. E. (1983). A modifiedincidental-teaching procedure for autistic

youth: Acquisition and generalization of re-ceptive object labels. Journal of Applied Be-havior Analysis, 16, 329-338.

McGee, G. G., Morrier, M. J., & Daly, T.(1999). An incidental teaching approach toearly intervention for toddlers with autism.Journal of the Association for Persons withSevere Handicaps, 24, 133-146.

Mcllvane,W. J., Dube, W. V., Green, G., &

Serna, R. W. (1993). Programming con-ceptual and communication skill devel-

opment: A methodological stimulus classanalysis. In A. P. Kaiser & D. B. Gray(Eds.), Enhancing children’s communica-tion (Vol. 2, pp. 242-285). Baltimore:

Brookes.

Mcllvane, W. J., & Stoddard, L. T. ( 1981 ).Acquisition of matching-to-sample perfor-mances in severe retardation: Learning byexclusion. Journal of Mental Deficiency Re-search, 25, 33-48.

MvIlvane, W. J., & Stoddard, L. T. (1985).Complex stimulus relations and exclusionin mental retardation. Analysis and In-tervention in Developmental Disabilities, 5,307-321.



85

Oppenheimer, M., Saunders, R R., & Sprad-lin, J. E. (1993). Investigating the general-ity of the delayed-prompt effect. Researchin Developmental Disabilities, 14, 425-444.

Prizant, B. M., & Rubin, E. (1999). Con-temporary issues in interventions for autism

spectrum disorders: A commentary. jour-nal of the Association for Persons with SevereHandicaps, 24, 199-208.

Prizant, B. M., & Wetherby, A. (1998). Un-derstanding the continuum of discrete-trialtraditional behavioral to social-pragmaticdevelopmental approaches in communica-tion enhancement for young children with

autism/PDD. Seminars in Speech and Lan-guage, 19, 329-353.

Remington, B. (1994). Augmentative andalternative communication and behavior

analysis: A productive partnership? Aug-mentative and Alternative Communica-

tion, 10, 3-13.Romanczyk, R G. (1996). Behavioral analy-

sis and assessment: The cornerstone to ef

fectiveness. In C. Maurice, G. Green, &

S. C. Luce (Eds.), Behavioral interventionfor young children with autism: A manualfor parents and professionals (pp. 195-217).Austin, TX: PRO-ED.

Saunders, K. J., & Spradlin, J. E. (1989).Conditional discrimination in mentally re-tarded adults: The effects of training thecomponent simple discriminations. journalof the Experimental Analysis of Behavior,52, 1-12.

Saunders, K. J., & Spradlin, J. E. (1993).Con-ditional discrimination in mentally retardedsubjects: Programming acquisition and

learning set. Journal of the ExperimentalAnalysis of Behavior, 60, 571-585.

Saunders, R R., & Green, G. (1999). A dis-crimination analysis of training structure

effects on stimulus equivalence outcomes.Journal of the Experimental Analysis of Be-havior, 72, 117-137.

Schopler, E., Mesibov, G. B., & Hearsey, K.(1995). Structured teaching in the

TEACCH system. In E. Schopler &

G. Mesibov (Eds.), Learning and cognitionin autism (pp. 243-268). New York:Plenum Press.

Schreibman, L. (1988). Autism. NewburyPark, CA: Sage.

Schreibman, L., Whalen, C., & Stahmer,A. C. (2000). The use of video priming toreduce disruptive transition behavior in

children with autism. Journal of Positive Be-havior Interventions, 2, 3-11.

Serna, R. W., Dube, W. V., & Mcllvane, W. J.(1997). Assessing same/different judg-

ments in individuals with severe intellectualdisabilities: A status report. Research in De-

velopmental Disabilities, 18, 343-368.Sidman, M. ( 1971 ). Reading and auditory-

visual equivalences. Journal of Speech andHearing Research, 14, 5-13.

Sidman, M. (1986). Functional analysis ofemergent verbal classes. In T. Thompson& M. D. Zeiler (Eds.), Analysis and inte-gration of behavioral units (pp. 213-245).Hillsdale, NJ: Erlbaum.

Sidman, M. (1994). Equivalence relations

and behavior: A research story. Boston: Au-thors Cooperative.

Sidman, M., Rauzin, R, Lazar, R, Cun-ningham, S., Tailby, W., & Carrigan, P.

(1982). A search for symmetry in the con-ditional discriminations of rhesus monkeys,baboons, and children. Journal of the Ex-perimental Analysis of Behavior, 37, 23-44.

Sidman, M., & Stoddard, L.T. (1967). The effectiveness of fading in programming a

simultaneous form discrimination for re-

tarded children. Journal of the Experimen-tal Analysis of Behavior, 10, 3-15.

Sidman, M., & Tailby, W (1982). Condi-tional discrimination vs. matching to sam-ple : An expansion of the testing paradigm.Journal of the Experimental Analysis of Be-havior, 37, 5-22.

Skinner, B. F. (1938). The behavior of orga-nisms: An experimental analysis. New York:Appleton-Century-Crofts.

Skinner, B. F. (1953). Science and human be-havior. New York: Macmillan.

Stevenson, C. L., Krantz, P. J., & McClanna-

han, L. E. (2000). Social interaction skillsfor children with autism: A script-fadingprocedure for nonreaders. Behavioral In-terventions, 15, 1-20.

Stromer, R, Mackay, H. A., & Stoddard,L. T. (1992). Classroom applications of stim-ulus equivalence technology. Journal of Be-havioral Education, 2, 225-256.

Stromer, R, & Osborne, J. G. (1982). Con-trol of adolescents’ arbitrary matching-to-sample by positive and negative stimulus re-lations. Journal of the Experimental Analysisof Behavior, 37, 329-348.

Sundberg, M. L., Partington, J. W. (1998).Teaching language to children with autismor other developmental disabilities. Danville,CA: Behavior Analysts.

Sundberg, M. L., & Partington, J. W. (1999).The need for both discrete trial and natural

environment language training for childrenwith autism. In P. M. Ghezzi, W. L. Wil-liams, & J. Carr (Eds.), Autism: Behavior

analytic perspectives (pp. 139-156). Reno,NV: Context Press.

Taber, T.A., Seltzer, A., Heflin, L. J., & Al-

berto, P. A. (1999). Use of self-operatedauditory prompts to decrease off-task be-havior for a student with autism and mod-erate mental retardation. Focus on Autism

and Other DeveGopmental Disabilities, 14,159-166.

Taylor, B. (in press). Teaching peer socialskills to children with autism. In C. Mau-

rice, G. Green, & R. M. Foxx (Eds.), Mak-ing a difference: Behavioral intervention forautism. Austin, TX: PRO-ED.

Taylor, B. A., & Levin, L. (1998). Teaching astudent with autism to make verbal initia-tions: Effects of a "tactile prompt." journalof Applied Behavior Analysis, 31, 651-654.

Terrace, H. S. (1963). Discrimination learn-ing with and without errors. Journal of theExperimental Analysis of Behavior, 6, 1-27.

Wetherby, A. M., Schuler, A. L., & Prizant,B. M. (1997). Enhancing language andcommunication development: Theoreticalfoundations. In D. J. Cohen & F. R Volk-mar (Eds.), Handbook of autism and per-vasive developmental disorders (2nd ed.,pp. 513-538). New York: Wiley.

Wilde L., Koegel, L., & Koegel, R (1992).Increasing success in the school through prim-ing: A training manual. Santa Barbara:University of California.

Wolf, M. M. (1999, August). Division 25 Ap-plied Research Award presentation. Paperpresented at the annual meeting of theAmerican Psychological Association, Boston.

Wolf, M. M., Risley, T. R, Johnston, M. K.,Harris, F. R, & Allen, K. E. (1967). Appli-cation of operant conditioning proceduresto the behavior problems of an autistic

child: A follow up and extension. BehaviourResearch and 7-herapy, 5, 103-111.

Wolf, M. M., Risley, T. R, & Mees, H.(1964). Application of operant condition-ing procedures to the behavior problems ofan autistic child. Behavior Research and

Therapy, 1, 305-312.Zanolli, K., Daggett, J., & Adams, T. (1996).Teaching preschool age autistic children tomake spontaneous initiations to peers usingpriming. Journal of Autism and Develop-mental Disorders, 26, 407-421.

Zygmont, D. M., Lazar, R. M., Dube, W V.,& Mcllvane, W. J. (1992). Teaching arbi-trary matching via sample stimulus controlshaping to young children and mentally re-tarded individuals: A methodological note.Journal of the Experimental Analysis of Be-havior, 57, 109-117.



behavior analytic instruction for learners with autism advances in stimulus control technology-gina...

Health & Medicine

behavior behavior problems

autism skills bijou

abnormal behavior

oldproblem behavior

behavior analytic instruction

behavioral teaching

abalyzing individual

aggressive behavior