4 Separate Papers. Write a one-page review, single spaced of each of the assigned readings. Article responses should include a description of the article itself and a reaction to the article. I will a

[email protected] 162 stillresults incontroversy, misinformation, and is a source ofgreat confusion forparents attempting to make treatment andeducation decisions for theirchil­ dren. For the purposes of this article, we assume that the reader is familiar with the difficult and complex issues of obtaining anaccurate differential diagnosis for young children, as well as with the critical process of obtaining anassessment of thechild's development (Harris & Handleman, 1994; NewYorkStateDepart­ ment ofHealth, 1999a; Powers & Handleman, 1984; Romanczyk, Lockshin, & Navalta,1994;Schopler & Mesibov, 1988).

A general historical readingin the field of autism quickly results in the impressionthatautism is a severe disability for which little evidenceis found for long­ term positive outcome, thatit is difficult to diagnose, and that incidence andprevalence figures are contro­ versial (California Department ofDevelopmental Ser­ vices, 1999). Autismis also strongly associatedwith a great number of"fads" and"movements" thatover the last several decades havepromised much, but consis­ tently have failed to deliver when the harshlight of objective evaluation isfocused onsupposed break­ through procedures (Delmolino & Romanczyk, 1995; Green, 1996a; Olley & Gutentag, 1999; Smith, 1996).

Often "models" arepromulgated with littleempirical support, but with a wealthof sincerity and enthusiasm.

By studying this history, an appreciationisformed for the complexities ofgenerating aviable model that standsthe test of time and objectiveevaluation.

It is in this contextthatwe first describe anapproach that establishes apriori thattheselection ofassessment and intervention approachesmust bebasedongener­ ally accepted rules of scientific evidencefor efficacy. It is a "bottom up"approach, inthat principles and pro­ cedures withdemonstrated efficacy areassembledinto a coherent modelthatis again subjected toempirical validation. This is quite differentfrom the process of developing aconceptual modelandthen seeking to find confirmatory evidence. at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 163 To date, only one nonmedicalapproachmeets the boundary conditions of this model as applied to autism.

This approach hasconsistently producedoutcomes that are reproducible, describablein precise terms, are tied to a conceptualization thathas strong and extensive experimental support,and uses, as a necessary compo­ nent, continuing objectiveevaluation. Thisapproach, known as applied behavioranalysis(ABA)in the con­ text of autism, but more generically as behaviortherapy or the behavioral approach,was firstappliedin the treatment of autism more than threedecades ago. Its roots are strongly within a research/academicframe­ work, and it has beenapplied with empirically evalu­ ated success to a wide array of humanproblems (Bar­ low, 1988; Barlow, Hayes, & Nelson, 1984; Spiegler & Guevremont, 1993).Interestingly, perhapsbecause of its focus on objective evaluationratherthanon consen­ sus of opinion, it has not becomepopularized and has often beengrossly misunderstood (cf. Maurice, 1993).

Over the past threedecades, systematic researchin­ vestigations have demonstratedthe utility of specific components ofABA. More recently, larger scale out­ come studies again have consistentlydemonstrated that significant impactcan be made for childrenwith autism (New York State Health Department, 1999a). For those who are influencedbyresearch versusanecdotal reports, thereexists a growing and diverse behavioral technology thatcan be applied. One unexpectedout­ come of this extensiveness hasbeen a clustering of be­ havior analysts into several schools with strongly held positions. In ouropinion, theseare divisions based largely on emphasison oneparticular instructional technique oranother, agrouping oftechniques in a certain clustering, or differences in the strategyof ser­ vice delivery. However,all fall within the rubricof ABA. It is our opinionthatno single technique nor collection oftechniques can becorrect (or effective) for every person in every situation. Thus, the databased feedback loop in ABA isinherently aself-correcting mechanism if applied in the contextof clinical decision making. It is our goal in this article to first outlinethe common programmatic andmethodologic elementsof the ABA approach. We discuss some of the features that seemtodefine typical behavioral modelsand within this contextrespond to the many myths that have arisen regardingABA.

Programmatic CommonElements It isprobably accuratetostate thatmany models, behavioral analytic and nonbehavioral analytic,share some common programmatic elements.Dawsonand Osterling (1997) reviewed a numberofprograms for children with autism thatmet the boundary condition of having published detaileddescriptions of the pro­ grams and providedintakeandoutcome data(many are reviewed in this special issue of JASH). They pre- sent a series of commonelements thatwere observed that areconsidered triedand true. Moreprecisely, by examining commonelements thatexist across programs that differ significantly in approach,theauthors state that these programmatic commonelements are "un­ likely to reflectanidiosyncratic viewpointor one inves­ tigator's philosophical attitude"(p. 314).Thesepro­ grammatic commonelements are specific curriculum content, highlysupportive teachingenvironments and generalization strategies,predictable routine,func­ tional approach toproblem behaviors, plannedtransi­ tion, and family involvement.

These programmatic commonelements perhapsre­ flect the minimum startingpointforprogram develop­ ment, along with appropriatelytrainedand caring staff, adequate resources, andsupervisory and review mecha­ nisms. Given these as "basics," thenthe task is to utilize a methodology thatallows each of these elementsto be addressed in anindividualized manner,for children, family, and staff. It is at the pointof selecting specific methodology forinstruction that behavioral and non­ behavioral modelssometimes begin to diverge. Methodologic Common Elements There aremethodologic commonelements within the behavioral approach.First, theapproach views be­ havior as functional andpurposeful, even when func­ tion and purposeare notimmediately discernibleby an observer. Behavioris viewed as the result of a complex blend of variables thatinclude theindividual's strengths and limitations, physical status, history, and the current social-en vironmen talcircumstances (Romanczyk & Matthews,1998). As with many complex approaches, there areoften subtle differences betweenspecific methodologies andtheories thatareassociated with the approach. Withrespect to autism, ABA is a specific form of the more generalbehavioral model. AnalysisandMeasurement ABA places stress on understandingthebehavior in question, whetherit is theacquisition of a skillthatis currently absentin aperson's repertoire or the amelio­ ration of aproblem behavior. If emphasis is placed on the analysis level, thenit follows thatthere need to be certain prerequisite steps.

The first of these prerequisitesis the objective mea­ surement ofbehavior. Mostmeasurement systems have technical pros and cons, as well as practicalcost effec­ tiveness parameters. Thereis a large body of literature within the field of science in generaland psychology specifically thatindicates humanobservers areprone to a number of very specific errorsinconducting obser­ vations. We are all subject to influences and biases that limit our objectivity. One credible reasonfor this diffi- at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from 164 AndersonandRomanczyk culty inobjective measurement is that one of our strengths withrespect toinformation processingis the ability todetectpatterns. However, as with all systems, our ability todetectpatterns is not perfect.We are able to perceive specificpatterns where,in fact, none exists.

We can be differentially influencedbyinformation, context, andexperience that have emotional as well as informational content.

Often in clinical and educationalservicedelivery, we violate this basic principleofobjectivity so that the in­ dividuals performing assessment, deliveringservices, and evaluating outcomeare one and the same. It is important to stress that the influences that limit our ability to make objectiveobservations are not primarily based on such factorsassophistication, education,fal­ sification, intentional bias, anddeception. Rather,they are based on aninformation processinglimitation that allhumans share.Byunderstanding theseprocesses, we can guard against potentially inaccurateobservations and conclusions. Thus, the sine qua non of theapplied behavior analyticapproach is that objectivemeasures are taken of the individual's behaviorand that these measures must meet theboundary conditions ofbeing operationally defined,reliable, and valid.

Operational definitionssimplytranslate thenormal colloquial reference that we give to certainbehaviors into more objectively definedobservational terms. A good example would be attempting to assess a child who is "anxious." This is a term that mostindividuals would recognize andbelieve that they have an under­ standing ofwhat itmeans. Difficulty arises with respect to precision and theapplication of suchtermsto spe­ cific individuals. For example, withanxiety, one could view it as a construct,thesummation of a number of different factors that areassumed to becoherent. We can divide the impreciseconstructofanxiety into a number ofcomponents: cognition,self-report, overtbe­ havior, performance, and physiologic.

Although anxietyis a useful term for thepurpose of communication concerningaproblem theindividual is experiencing, from abehavioral perspective it would be further defined within the abovecategories. This allows highly individualized assessmentfor a given personas to how specifically anxietyismanifested for the indi­ vidual. ABA emphasizes addressing the specific, unique expression ofbehavior by theindividual.

Reliable observations referto the degree to which the various specific behavioral observations conducted by different observers are inagreement. This is typi­ cally done by having two independentobserversper­ form observations and then compareveryprecisely the degree to which they agreeanddisagree on the specific temporal pattern of the behaviors observed.To be in­ dependent, individualsshould not be given specific ex­ pectations such as"medication is beingconsidered," or "we are seeingproblems with rising anxiety,"or"it's clear he's anxious and we need todocument that."Op- erationalized, unbiased,andreliable observation serves as the basis for hypothesistestingas to factors that may be of importance andinfluence theindividual, and thus leads to the processofconducting afunctional analysis (a point to bediscussed next).

Assessing the Child Assessment is a crucialcomponentof any clinical/ educational model.Because thereare as many differ­ ences between youngchildren withautism as similari­ ties among them,assessment must be aconstantfocus point when developing andimplementing acompre­ hensive intervention program.Traditional assessment methods such as the administration ofstandardized psychological, speech,andachievement tests,"survey" assessments such asratingscales, and behavioralassess­ ment all haverelativestrengths andweaknesses.

Within thepractice of ABA, therearevarious sub­ components ofassessment. First,assessment of an in­ dividual with autism, particularlya young child, can be a very difficult task. While assessmentisoften some­ what arbitrarily dividedintostandardized psychometric evaluation, socialhistory/family statusassessment, in­ formal observation, and much more rarely,functional assessment, ABA focuses strongly onfunctional assess­ ment (functional analysis). It is not andshould not be seen asincompatible with the assessment methods mentionedabove. For example,standardized assess­ ment, iffeasible andproperly conducted, providesim­ portant information. Suchassessment allows theestab­ lishment of a "marker" as to the currentrepertoire of the child with respecttovarious developmental do­ mains and allows a comparisontoother individuals, as well as relative strengths andweaknesses within the individual. It also serves as a standardizedformatto assess the ability to interactin a social manner with respect to thevarious directions anddemands and in­ terchanges that occur during standardized assessment.

Standardized assessment,however,is not necessarily directly useful in the selectionof specific, immediate, short-term goals. Nor is it typically useful in determin­ ing the specific intervention methodology that will be utilized. It is beyondthe scope of this articletoexamine in detail thevarious aspects of the assessmentprocess.

We focus on that aspect most specific to the behavioral model, that of functional analysis.

Functional analysis is an oftenmisunderstood term, partly because different disciplines have varying defini­ tions. Within ABA, functional analysis is the processof ascertaining empiricallythecontrolling variables that enhanceorinhibit theexpression of abehavior. It is not done byobservation, filling out abehaviorchecklistor scale, nor by consensus amonginvolved parties.Rather, these sources ofinformation are used to form hypoth­ eses as to what factorsmay be involved, and then to test these hypotheses (Iwata,Vollmer, Zarcone, & Rod­ gers, 1993; Miltenberger, 1998). It is aprocessof ob- at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 165 serving, hypothesizing, testing,evaluating, refininghy­ potheses, andrepeating.

The essential aspect of functional analysis is the ex­ plicit testing of factors presumedto beimportant, and to conduct the testing in a mannerthathas the potential to clearly disprovethehypothesis. There are many uses for this very powerfulmethodology beyondunder­ standing problem behaviors, and it isparticularly useful for assessment ofchildren withoutverballanguage.

While it is oftentechnically difficult and time consum­ ing to perform, the accuracy of the informationob­ tained through functionalanalysis and its directappli­ cability to intervention makeit a most important com­ ponent of ABA. Developing anIndividualized Curriculum The word curriculum has various connotations and meanings for various professionals. In thecontext of this article, curriculummeansasequence of goals: (1) organized fromboththelong-term andshort-term per­ spective' (2)resulting from focused assessment,(3) re­ flecting the collective prioritiesof involved adults, and (4) tempered by thecurrent developmental level of child.

Because thisshould be aninteractive process, and typically involves individualsat various levels of exper­ tise, it is useful to have an outlineordocument that serves as a map (Romanczyk,1996).However, caution must be expressed inthat acurriculum shouldnot be seen as a specific sequenceoflearning and skill activi­ ties that allchildren willprogress throughin asequen­ tial manner.

One of the important characteristics ofchildren with autism is unevenlearning ability and skill levels. Thus, individualization ofintervention cannotbeoverstated.

While generally anexcellent startingpoint,it is not necessarily most effective to teachall skills in a "typi­ cally developing" sequence.Use of acurriculum must occur within a very tight feedbackloop that assesses not only the logic and priorityof a goal and its subcompo­ nents, but also its interactionwithassessment informa­ tion which includes a child's currentrepertoire, moti­ vation, andpreferences. A goodcurriculum should have a conceptual structure(we suggest a developmen­ tal sequence), offergreat detail (operalization), and be used in a child specific manner(nonlinear branching).

Selecting andSystematically UsingReinforcers It is a truism concerning humanbehavior thatmoti­ vation is animportant component oflearning and main­ taining skills.Motivation can come from a numberof sources. For most individuals,this diversity providesa rich context foracquiring andmaintaining skills. It is also the case thatsome individuals, such aschildren with autism, have impairmentinmotivation.

At times, motivation may be quite idiosyncraticandlimited in its extensiveness. Anexample would be childrenwho are not motivated by socialattention andpraise, physical contact, and the sense of accomplishmentforcomplet­ ing a task or solving a problem.Rather,theseindividu­ als might find theirownrepetitive andstereotyped be­ havior moreinteresting andenjoyable, and thusengage in it disproportionately comparedtoprosocial behav­ ior. The termreinforcer describesafunctional relation­ ship thatisempirical innature, not speculative. This is a critical aspectof thebehavioral model:procedural or technique components are not to be used in isolation, detached from the critical processofongoing assess­ ment. The stereotyped andincorrect reinforcement procedure "forchildren withautism who fail to make eye contact, each time they look at you, reinforcethem with a fruit loop,"iscompletely erroneousand misses the point entirely. Alsoerroneous would be the con­ clusion that"eye contact is not getting better even though wekeep reinforcing them with fruit loops." Even at its most basic level, ABA isintimately tied to continuous assessmentof theindividual and not simply the application ofmisperceived standardtechniques.

Promoting Generalization Generalization is a keyconcept. It isoften viewed as the degree to which a behavioror skilllearned under particular conditions andsettings will beexpressed in other conditions and settings. An examplemight be taking pianolessons and being able to performa par­ ticular musical piece quiteadequately athome with the piano teacher, andthen being asked topresent that same piece duringarecital whereperformance may be observed to be significantly impaired. From theinception of ABA, generalization hasbeen a focal concept thatisbound directly to goal selection, teaching, andevaluation ofbehavior. Intheir classic article thatappeared in the first issue of the Journal of Applied BehaviorAnalysis, Baer,Wolf, &Risley (1968) stated thatgeneralization is acentral component of ABA and that "... generalization should be pro­ grammed,ratherthanexpected orlamented" (p. 97).

That is, animportant characteristic of ABA isthat the intervention process must explicitly addressstrategies and procedures toteach andpromote generalization across time, setting,people, and tasks. There is anextensive generalization methodologyto be found in thepublished literature. It involves knowl­ edge of stimulus control,stimulus generalization, rein­ forcement schedules,prompthierarchies and fading, setting events, antecedent conditions,responsevariabil­ ity, contingency criteria,use ofmultiple exemplars, set­ tings, people, andcontexts, as well as task analysis and response repertoire assembly.Becausegeneralization can be problematic forcertain individuals, and given the complex factorsthatinfluence generalization, a spe­ cific and detailedplan for generalization shouldbe a part of allintervention programs. at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from 166 AndersonandRomanczyk Selecting Intervention Techniques With Documented Effectiveness Froma clinical perspective, selectionofintervention techniques has twocomponents. The first is to ascertain controlled researchevidence concerning specific skills, behaviors, orconditions forindividuals with a similar diagnosis/characteristics that appear inpeer reviewed journals andthat meet generally acceptedcriteriafor well controlled clinical studies. Anecdotes,casestudies, "expert opinions," theoretical arguments, andappeals to "clinical experience" are not acceptablesubstitutes.

Such sources canprovide potentially usefulinformation in order to test hypotheses abouteffectiveness compari­ son of different approaches andprocedures in a re­ searchcontext, butshould not be used as a substitute for controlled researchevidence for clinical practice.

Second, the selected intervention must meet the boundaryconditions of theoriginal research param­ eters. Sadly, interventions areoften implemented in name only,thatis,terms are used to label what is being offered, but the specifics of the interventionas actually applied are not consistent with the specifics of the origi­ nal intervention research.Procedural integrityis mea­ sured andevaluated as is theobjective evaluation of the child's progress. Specifically, anevaluative process known as single subjectmethodology isemployed (Bar­ low et aI., 1984; Hersen & Barlow1981;Sidman, 1988). There are many very powerfultoolscurrently avail­ able to parents, educators, and clinicians who wish to avail themselves of theempirical literature. One cau­ tion that should beraised isthat it isessential in this process toread and review researchreportsintheir original form, rather thanassummaries. Inparticular, the explosion ofinformation on theinternet has suf­ fered greatly bymisrepresentation andinaccuracy. In­ formation isoften "packaged" toprovide noncritical support for aparticular pointof view. In readingthe original research report,one is able to ascertainthe specific characteristics of thechildren whoparticipated, the specifics of the proceduresutilized, the adequacyof the research design, and the degreeandmagnitude of the outcomes. Certainly,it ispossible to have a research study thatdemonstrates a significantstatisticaldiffer­ ence between intervention procedures, but thatdoes not necessarily mean that thissignificant difference rises to the level of clinical significance. We require both statistical significance and substantialchangein the child's cognitive, social, and family and community life. The task of reviewing suchresearch can seem daunt­ ing. Because ABA isbased on basic principles of hu­ manbehavior, thereis awealth ofresearch available.

The published literatureofprofessional journalswas searched forresearch studiesconcerning applied be­ havior analysis and autism(Palmieri, Valluripalli, Arn­ stein, & Romanczyk, 1998). Given the varying termi­ nology, thereareabout 19,000 published articles if one uses applied behavior analysis and its synonyms. While not all this literatureis directly relevantto ABA as an intervention for autism, it underscoresthe vast base of research thatserves as the foundationfor the ABA approachand itsbroadapplicability to a wide range of populations, skills, andbehaviors.

Five hundred articles specific to both ABA and au­ tism were found. If we narrowthe focus to research with young childrenwith autism, conductedafter1980, and employ a single subjectresearch methodology, ap­ proximately 90published researchstudies were identi­ fied. These provide supportfor abroad continuum of behavioral techniques thatfocus on the development of skills in social, cognitive, self-help, independence,emo­ tion, language, self-control, attachment,recreation,and academic areas.

In short, thereis a large base of researchliterature that addresses specificpopulations, ages,characteris­ tics, and specific educational,clinical, social, and physi­ cal emphases, as well as a substantialbase ofresearch specific to ABA and young childrenwith autism. A full review of thesearticles is not possible here(for exten­ sive reviews, see Matson,Benavidez, Compton,Paclaw­ skyj, & Baglio, 1996; New YorkState Department of Health, 1999b).However, thereareseveral large-scale studies thatbase their conceptualization andproce­ dures on the large researchbase, which we will briefly review. Theyrepresent theimportant endeavorof con­ ducting controlled clinical trials.

Six studies havebeen published thatevaluated the benefits ofintensive homebased intervention for chil­ dren with autism. Each of these studies involved at least 1 year of intervention, includedabroad rangeof be­ havioral techniques, andevaluated its effects of a vari­ ety of developmental outcomes(intellectual function­ ing, language, socialinteractions, adaptivefunctioning).

The most comprehensive study of homebasedinter­ vention forchildren withautism waspublished by Lovaas (1987). Lovaas assigned preschoolaged chil­ dren to one of two groups: an intensivetreatment group that received anaverage of 40hours ofone-on-one treatment per week or a minimaltreatment control group thatreceived 10 hours or less per week.Each child in the experimental groupwas assigned several well trained therapists whoworked with the child and the parents in thehomefor 2 or moreyears. Pretreat­ ment measures revealedno significant differencesbe­ tween the treatment and control groups. However, posttreatment dataindicated that9 of 19 (47%) chil­ dren in the experimental grouprecovered. Thesechil­ dren were reported to haveachieved normalintellec­ tual and educational functioningin the first grade. In contrast, only 20/0 of thechildren in thecontrol group met this criterion. A follow-up study was conducted when thesechildren reachedamean of age 13 years (McEachin, Smith, & Lovaas, 1993). Evaluationwas done by clinicians blind to the children'spriorhistory at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 167 and intermixed withchildren who had no history of developmental orpsychological disturbance. At this point, 8 of the 9 recovered childrenwere still indistin­ guishable from thecomparison group.

Several other investigators (Birnbrauer & Leach, 1993; Sheinkopf & Siegel, 1998;Smith, Eikeseth, Klevstrand, & Lovaas, 1997) have partiallyreplicated the intensive behavioral intervention modeldescribed by Lovaas. In each case, therewereimportant devia­ tions from the modelincluding the fact thatall of these studies provided fewerhours per week (18 to 30 hours rather thanthe 40 hours providedbyLovaas). In each of these studies, children whoreceived thebehavioral interventions showedgreaterimprovement thanchil­ dren in the control groupswhoreceived eitherless in­ tervention oranother type ofintervention. A sixth study byAnderson, Avery,DiPietro, Edwards, and Christian (1987)appears tosupport thesefindings as well, but did not includeacontrol group of comparable children receiving an alternativeapproach.

The results of these group studies when mergedwith the results of single subjectresearch methods offer con­ siderable supportfor the positive effects of intensive behavioral interventions for youngchildrenwith au­ tism. Nevertheless, manyquestions remain.AsGreen (1996b) pointed out, it is still unclearwhatvariables are critical tointervention intensity (number ofhours, length of theintervention, proportionofone-to-one to group instruction) and what are the expectedoutcomes when intervention intensityvaries. It is also unclear what particular behavioral techniques (discretetrials, incidental teaching,pivotalresponse training)are most likely to be successful for a given child and in what proportions particulartechniques shouldbe used. The current research islimited inthat it does not allow us to draw comparisons across studies. At thispoint, we can conclude thatthe best outcomes haveoccurred when the children received at least 30 hours of behavioral intervention. Continuum-Based BehavioralModels Since thepublication of Let Me HearYourVoice by Catherine Maurice(1993),requests for ABA services has grown rapidly. Parents,armedwithempirical stud­ ies, often approach localgatekeepers of special educa­ tion services by strongly advocatingfor ABA. The re­ quests have challengedlocal capacity to providequali­ fied personnel and tomeettheextraordinary demand for the number of hours recommended. Theseissues are exacerbated byresistance from some key officials, sometimes basedoninaccurate, butstrongly held be­ liefs about behavioral interventions.

Contrary to the beliefs of some, ABA isnot a stag­ nate, single continuum ofprescribed methods.It em­ phasizes the use of methodsthatchange behavior in systematic andmeasurable ways. Theuniquecontribu- tion of this approachis itsinsistence on analysis, repli­ cation, socialimportance, andaccountability (Baeret al., 1968; Sulzar-Azaroff &Mayer, 1991). Arguably,any intervention strategycould be studiedandembraced by behavior analysts,if it can be describedinprecise terms, reproduced, anddemonstrated to be effective.

At this point, ABA is anemerging technology thathas consistently producedthe bestoutcomes forchildren with autism. It will continuetoexpand itsbenefit to children withautism if itmaintains itscommitment to studying changes in behaviorinsystematic andmeasur­ able ways. At the same time, it is importanttoremain abreast ofdevelopments outsidethebehavioral com­ munity in biology, medicine,andneuroscience thatmay provide a clue to moreeffective interventions. It is pos­ sible thatacombination ofbehavioral andother inter­ vention strategies mighteventually maximizeoutcomes for some childrenwith autism. For example, it has been empirically demonstrated thatbehavioral interventions and medications aremore effective whencombined than either is alone for many childrenwithattention deficit hyperactivity disorder(Pelham, 1989;Pelham & Murphy, 1986). In short, ABA includesa largenumber of conceptually consistenttechniques thatcan be used in various combinations across manydifferentcontexts.

This diversity of proceduraltechniqueshascreated some confusion regardingtheprecise characteristics of quality ABA programs and the limits of its applicabil­ ity. It has sometimes polarizedpractitioners who be­ lieve one particular combination ofbehavioral tech­ niques is moreeffective thananother for allchildren concerned. Compounding theproblem areinexperi­ enced professionals whooften apply a very narrow, rigid range ofbehavioral techniques to verycomplex situations. Parents,administrators, andother profes­ sionals have beenleft to wade throughamyriad of options, oftenwithout sufficient information and tech­ nical ability to evaluatequality.Theseissuescombined with historical misrepresentations havecreated many myths abouttheapproach. There are program modelsthatshare thecommon elements of ABA described earlierandappear to uti­ lize a broad continuum oftraditional behavioral tech­ niques (e.g., Anderson,Campbell, & O'MallyCannon, 1994; Handleman & Harris, 1994;McClannahan & Krantz, 1994;Romanczyk et al., 1994). It is our sense that these types of behavioral programsdraw from the large base of researchstudies specifically addressingthe young child with autism(approximately 90 single sub­ ject designs and six groupstudies) and use a rangeof behavioral techniques (e.g.,incidental anddiscrete tri­ als teaching) andcontexts forlearning (e.g., home basedandintegrated preschoolsettings). At the same time, each programisunique inthat itcombines tech­ niques in differentways and emphasizes, moreor less, one technique oranother. It isimpossible todescribe exactly a "typical"behavioral program.Nevertheless, at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from 168 AndersonandRomanczyk we will attempt toidentify somefeatures thatseem to define (distinguish) typicalmodels (hereinafter re­ ferred to ascontinuum basedmodels). We have chosen to do this while simultaneouslyaddressingsome of the myths thathave developed about ABA. Common Features of Continuum BasedModels Multiple Contexts forLearning Myth: ABA is exclusively a homebased intervention model. The most promising outcomesforchildren with autism havebeen reported by Ivar Lovaas in his semi­ nal 1987 paper (described earlier). That study when taken in combination with the partial replications (Anderson et al., 1987; Birnbrauer & Leach,1993; Sheinkopf & Siegel, 1998; Smith et aI., 1997) provides the most detaileddescription ofmultiyear behavioral interventions forchildren with autism. One common element of all ofthesestudies is thatmost interventions occurred in the child's homewith a latertransition to school. Particularly whenaddressing theneeds of very young children, compelling argumentscan bemadefor providing services in the home: (1) it is a familiar and comfortable settingfor the child, (2) it encouragesac­ tive parent involvement, and (3) it is a naturalcontext for learning many skills (e.g., self-help, independent play skills). However,untilthere aredirect compari­ sons, the strongest statement we canmakeisthat the best outcomes forchildren with autism have been re­ ported byprograms employing a significant periodof home based intervention. On theotherhand, many of the elements commonly found inhomebased programs also can be found in school programsthatuse similar methods (i.e., highly structuredenvironments, careful analysis and measurement ofprogress, functional as­ sessment, individualized curricula,planneduse of rein­ forcers, active programmingofgeneralization, and the use of intervention techniqueswithdocumented effec­ tiveness).

Several studies have documentedoutcomesfor chil­ dren served in comprehensive centerbasedprograms using ABA methods. In one study, Fenske,Zalenski, Krantz, andMcClannahan (1985)reported treatment outcomes for agroup ofchildren whoreceived inten­ sive behavioral programming in a school setting. Al­ though the study had some design limitations,the re­ sults indicated thatchildren weremore likely to achieve positive outcomes (regularschoolplacement) if they enrolled in theprogram beforeage 60months and par­ ticipated at least 24 months. Another study byHarris, Handleman, Gordon,Kristoff, and Fuentes(1991) pro­ vided intensive behavioral intervention in a special or inclusive classroom. Changesinintellectual and lan­ guage functioning wereevaluated afterabout 1 year of participation. Onposttesting, thechildren with autism achieved significant improvement intheir ability as measured byformal testing. Since children spendmost of theirwaking hours at home withtheir families, it makes sense to provide most initial trainingintheir home with active parent involvement. Although not all families are willing and able to participate in a homebasedprogram, thereis no reason tosuspect thatquality behavioral interventions applied at school would have significantly differentout­ comes, particularly ifattention was given to the issues of generalization andparent involvement.

In summary, homeand school approaches are all found in continuum basedmodels for childrenwith au­ tism. It is also commonto findcombinations ofhome and school approaches in which the child progresses gradually from ahomeprogram to a school program (Anderson, et aI., 1994) over severalmonths. Other models are even moreflexible and the child and staff move continually and do not occupyself-contained classrooms (McClannahan & Krantz,1994).

Progression FromIndividual toGroup Instruction Myth: ABA always is characterized byone-to-one intervention. Althoughmanybehavioral models have emphasized theneed forintensive one-to-one to build initial skills (Andersonet al., 1987; Lovaas, 1987; Ro­ manczyk, Matey, & Lockshin, 1994), most programs actively build the requisiteskills for small and large group participation. Individualinstruction makessense when the studenthaslimited attention, respondsidio­ syncratically, requiresphysicalguidance most of the time, lacks basic group readinessskills, and when the child is first introducedtoteaching situations. Although the best outcomes havebeen reported byprograms us­ ing intensive one-to-one trainingfor at least 2 years (Anderson et al., 1987; Lovaas, 1987), therehavebeen no direct comparisons of thebenefits of group versus individual instruction.

The potential benefitsof group instruction are fun­ damentally clear and are unlikely to producemuch dis­ agreement amongprofessionals: (1)preparing the child for kindergarten andelementary classrooms, (2) pro­ viding potential forincidental orobservational learn­ ing, and (3) providinganopportunity for social and language interactions withother children. Whatoften seems to be at issue are when to introducegroup in­ struction andwhether there areprerequisite skills.

Again, we recommend thatthisbecome an issue of individualization andreasonable balanceratherthan rigid practice. It is also possibletoblend methods by providing one-to-one within agroupcontext (Kamps et al., 1991). In this format, one studentreceives instruc­ tion while the same teachersupervises otherstudents who are working independently(typically on a "back­ ground task").Thismethod allows the teacherto move back and forth betweenindividual and group instruc­ tion and gives the child an opportunitytoobserve the behavior of other children.In our opinion,flexible at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 169 models that allow fluid changefrom one context to another and movement between individual and group instruction are mostlikely to be successful. Another factor that shouldbeconsidered inchoosing a context forlearning concerns the number ofinstruc­ tional opportunities. Childrenwithautism appear to learn most readily through repetition orpractice on a small set of targeted skills for learning. The number of instructional opportunities areinfluenced by many fac­ tors including the total number ofhours ofintervention per day and the length of theintervention period in months oryears. Inaddition, the size of the instruc­ tional group may exert significant influenceon the number ofopportunities. Typically,as a child moves from individual to smallgrouptolarge group, the num­ ber ofopportunities decreases.This issometimes rea­ sonable andappropriate, particularlywhen the goals are concernedwith the acquisition ofmany kindergar­ ten readiness skills (e.g., followinginstructions givento the group). Although eachsituation must beevaluated individually, it isnecessary tostrike abalance between the level of individualandgroup instruction that each child receives, continually evaluating and prioritizing the goals to be achieved.

In summary, the continuum basedmodels oftenem­ phasize one-to-one instructiontobuild aninitial reper­ toire of skills. Every effort is made to move the child along the continuum from one-to-one tosmall group to large group instruction.

Utilization ofMultiple Instructional Techniques Myth: ABA exclusively uses teacher led,discrete tri­ als methods. Many parents andprofessionals equate ABA with an instructional techniquecalleddiscrete trials. In discretetrialstraining (Anderson, Taras, & O'Malley Cannon, 1996; Koegel, Russo, & Rincover, 1977; Lovaas, 1981;Luyben, 1998), the instructor pro­ vides a concise instruction, question,oractivity when the child is mostattending. If the childanswers cor­ rectly, the teacher praises enthusiastically and may re­ ward the occurrence of the behaviorwith other signs of approval (a toy, pat on the back,food). If the child fails to respond or responds incorrectly, the instructor deliv­ ers feedback andprovides a prompt toensure that the correct ordesired response occurs. The formand level of the prompt (e.g., physical, gesture,verbal,partial physical) is determined individually but shouldbe im­ mediately successful and easilyfaded. Often, the trialis quickly repeated, giving the child an opportunity to practice what has just been shown. Most of us have learned something at one timeor another using this method and it is unlikely byitself to be very controver­ sial. However, discretetrials methods alsohave become closely associated with other aspectsof the instructional context. Initialinstructional sessionscan be very diffi- cult for the learner with autism and as soon as the in­ structor beginsto set limits (e.g., insisting that the child remain seated), many children resist. To limit the child'smobility and to allow more effective prompting, the instructor and child often sit facingeach other in chairs, with the child'slegsresting between the instruc­ tor's. The instructor then presents the lesson in the dis­ crete trials format described above.A well trained in­ structor will use a varietyoftechniques to assist and help the child feel comfortable with the situation(e.g., short sessions,introduction offamiliar activities, physi­ cal and verbal assistance, music or preferred activities interspersed withless preferred tasks, teaching an ap­ propriate escaperesponse). Nevertheless, even with the best laidplans, some children withautism willresist the intervention.

On the positive side, this formofinstruction often results inrapid learning for many children. On the negativeside, critics argue that the selectionoflessons is too teacher directed and that itminimizes individual choice. However, the goal of the behavior analystis to increase alternative responsesso that the childacquires more freedom ofchoice (Alberto & Troutman, 1999).

Children who lack functional communication skills and repeatedly fail toexpress their needs and wants are unlikely to make friends. The discretetrials method had been contrasted with another behavioral approach calledincidental teaching.

This approach has been described and demonstrated to be effective formany children (McGee, Krantz,Mason, & McClannahan, 1983). In this method, the instructor assesses the child's ongoing interests, follows the child's lead, restricts access to high interest items, and con­ structs alesson within the natural context, with apre­ sumably more motivated child. This model produces rapid learning if the childfrequently exhibitsinterests in objects and activities. On the other hand, if the child does not spontaneously show interest inmany things,it is difficult to find enough instructional opportunities. Furthermore, ultimatelyit isdesirable for a child to learn even insituations orcircumstances that are not inherently interesting.

In our opinion, these techniques are both very im­ portant and are not' necessarily incompatible. At any given point in time, each method may have benefits and limitations. It can beargued that adiscrete trialis a naturally embedded part of an incidental teachingop­ portunity (getting the child'sattention, providing clear instruction, offeringpraise and support). For thisrea­ son, we prefer to use the phrase direct teaching rather than discrete trialstolabel the highly teacher directed situation described above.In adirect teaching situation, most instructors use avariety oftechniques tonaturally motivate children(e.g.,incorporating childchoice, re­ inforcing attempts). In our opinionit is important to think ofdiscrete trials and incidental teachingas end pointsalonga continuum ofteaching contexts and tech- at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from 170 AndersonandRomanczyk niques (Table1). Inbetween the two ends is a gradient of situations thatblend teacher directed and child ini­ tiated opportunities forlearning (we have identified one point as activities embedded). Eachpoint along this continuum hasadvantages anddisadvantages and each may be importantforindividuals at any given time in their development. Evenwithin a single session, the best instructors willblend direct teaching andincidental teaching techniques. The decision to teachin a given context is again basedon many individual variablesin­ cluding (but not limitedto) the child's attention,indi­ vidual distractibility, spontaneousinterestin toys and materials, the lesson being taught,and most impor­ tantly, the child's progress.Itshould not be a polarized decision but an individualized decisionbasedon indi­ vidual needsandobjectively measuredoutcomes.

In summary, it is ouropinionthatcontinuum based models use many differentinstructional techniquesto achieve the desiredresults moving up and down the continuum ofteaching techniques (Table1),matching the approach to the child's level and most effective style of learning. Opportunities for Social Integration Myth: ABA cannot be applied in socially integrated school settings.

Some behavioral researchers have madecompelling argumentsfor thebenefits ofteaching children withdevelopmental disabilities inintegrated settings (Jenkins, Odom, &Speltz, 1989). The underly­ ing assumption isreasonable, thatis, young children with disabilities willimprove byobserving the social and language modelsof their typically developing peers. In fact, many behavioralprogramshavereported the use of a social integrationmodel(Anderson et al., 1994; McGee, Daly, &Jacobs, 1994; Strain &Cordisco, 1994). Although mostbehavior analystsrecognize the importance of socialintegration for manychildren,one can find considerable variabilityas to when the integra­ tion should occur.

A few programs havehypothesized thatfor some children, aperiod ofindividual andsegregated small group instruction helps inpreparing thechildren for later inclusionary experiences(Andersonet aI., 1994; Handleman & Harris, 1994).Theseprograms typically provide agradual progression fromindividual (homeor center based) to small group (oftensegregated) to large group instruction (integrated). Many oftheseprograms also report success in transitioning childrentokinder­ garten placements withinregular schools.

When inclusion is the choice of parentsand profes­ sionals, it is importanttoremember thatsimple expo­ sure to typically developingchildrenmay not be suffi­ cient to produce measurable gains.Furthermore, close proximity does notprecludethatchildren might dem­ onstrate highlyidiosyncratic patternsofparticipation (Kohler, Strain, & Shearer,1996).There must be an individual focus for the evaluationofinclusion thatin­ cludes parents andteachers. As part of the decision making progress, theseindividuals mustmake decisions about thesequence of thecurriculum and thebalance between individualization and inclusion.

The behavioral strategiesandtechniques thatare ap­ plied in segregated settings also can be appliedin inte­ grated settings. Anderson et al. (1994) describedan in­ tegrated settingin which teachersorganized learning centers and the children were free to move among areas as they chose. Specific educationalobjectivesfor each child were embeddedintothese play activities, often using task analysis so thatteaching occurred at the child's current level and graduallyincreased in com­ plexity. While theseactivities were occurring,students were removed individually or as part of a small group for brief periods ofinstruction onother educational objectives thatwere less easy to embedorthat required significant repetitionfor learning. Methodsfor ongoing assessment anddata driven decision makingwere di­ rectly built into the system. The programemployed be­ havioral practices commonto most quality programs:

(1) clearly definedskills to be learned;(2)baseline and ongoing assessment; (3) systematicintroduction of be­ havior intervention methods;and (4) a broadcontext for learning thatincluded directteaching, embedded trials, and incidentalteaching, In summary, thecontinuum-based models typically employ a strategyof acontinuum of services from in­ tense individualized services in aspecializedsetting(or home) tomore typical classrooms withrequired sup­ port. This allows the rapidacquisition of skills and a systematic andplanned approach tointegration. Table 1 A Brief Description of Three DifferentPoints Along a ContinuumofContexts ForLearning Direct teaching Activitiesembedded Incidentalteaching • Usually one-to-one • Highly structured • Teacher directed • Many instructional opportunities • Skills acquired quickly • Distractions minimized • Easier tomanage behavior problems • Generalization may berestricted • Usually small group • Shared teacher/child control • Some naturaldistractions • Typically few instructionalopportunities • Moderate level ofstructure • Challenging to findreinforcing activities • Generalization may beenhanced • Small and large group • Most natural • Child directed • Natural distractions • Challenging to findreinforcing activities • Typically fewer instructional opportunities • Generalization may beenhanced at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 171 Individualized InstructionalCurriculum Myth: ABA does not develop functional skills. A common characteristic ofchildren withautism is the failure togeneralize skills across conditionsand re­ sponses dissimilar from theoriginaltraining. For ex­ ample, children wholearn toidentify severalcommon body parts (nose, ear, eye) when appliedtothemselves, may fail to generalizewhenasked toidentify thesame parts on a doll. Similarly, a child who learnstoimitate simple motor movements (e.g., pat stomach, touch handtohead) could fail to learnthe more general rule (i.e., "Do as I do"). Many of the skills targetedfor initial instruction (e.g., motor imitation,matchingcom­ mon objects,following instructions) are considered functional in that they provide the foundation for teaching many other things(communication, academic, self-care) .

During initial training situations, some teachers main­ tain unusually tightcontrol over the instructions pro­ vided, the materialspresented, thechild's positioning, and other setting conditions. This level of controlis often needed to helpchildren attendto the task, mini­ mize distractions, and break the activity into a series of small steps for learning.Teaching in this manner may lead to rapidacquisition of skills but sometimesmini­ mizes generalization. However,aschildren becomein­ creasingly more responsive toinstruction, it ispossible to simultaneously teachunder more natural conditions or to initiate instruction within anembedded or inci­ dental teaching format.We again referto the impor­ tance of using the full contextcontinuum forlearning (Table 1). For manychildren, instructional targetsare quickly learned andgeneralized tosettings, materials, and con­ ditions different from the learningcontext. For other children, verylittle generalization occurs andmostev­ erything the child learnswill have to be taughtdirectly.

The failure togeneralize (ordevelop rulegoverned be­ havior) may be a resultof many variables: (1)poorly selected goals for learning(learned behaviors did not meet anatural community ofreinforcement), (2) fail­ ure of the instructor to actively programgeneralization, and (3) limitations in our currenttechnology toteach and promote generalization for themostchallenging children. It is fair to say that mostchildren willbenefit, to a greater or lessextent, fromintensive behavioral interventions. However,therates oflearning and the range ofoutcomes forchildren may be as broadas the disorder itself (Weiss, 1999). Some childrenfail to dis­ play newly acquiredskillsexcept under highly teacher directed conditions (someonegives them averbal in­ struction, modelsthedesired behavior, orgestures to­ ward materials). Clearly,practitioners of ABA cannot be satisfied withproducing simplebehavior changein highly controlled conditions withoutfullyconsidering how the skill will be used in more natural situations. However, thereis little evidence tosupport analterna­ tive hypothesis that learning in all cases is likely to be more successful whenprovided in thenatural context.

Fortunately, there is an emerging set oftechniques (train diversity, incorporate functionalmediators, in­ clude natural communities) that whenapplied system­ atically oftenproduce thedesired results(Stokes &Os­ nes, 1988). It is importanttoprovide anindividual child focus for the use and evaluationofthese instructional techniques. In summary, thecontinuum basedmodels consider many variables whendetermining skills to betaught and the context forlearning to occur. As we stated earlier, acurriculum forchildren withautism cannot be seen as a linearsequence of skills to be learnedinorder. For the very young child, many skill targetsarechosen because they are believed to befunctional indevelop­ ing a foundation forteaching other skills. In the older child, skills are functionalbecausetheyenable the child to live a more independent lifestyle.Onceatarget skill has been selected, other factors may influencethe con­ text in which teachingwill occur (e.g., child'sattention, need forone-to-one). Again,weargue that the indi­ vidual child focus be maintainedand the full context continuum forlearning beavailable.

Development of aBehavioral Support Plan Myth: ABA usesaversive techniques tomanage be­ havior problems. Childrenwithautism oftendisplay challenging behaviors that interferewithlearning. It is our sense that allcontinuum basedmodels emphasize methods toteach new skills rather than asingular focus on managing challenging behavior.Initially,problem behaviors arehandled througha richschedule of rein­ forcement that builds adaptive skills. For example, in these programs, much ofearlylearning ishelping the child to acquireandgeneralize the ability to follow simple instructions and to useappropriate behaviorto gain attention orescape fromundesired requests.Many behavior problems improveas the child'sabilityto communicate andrespond to simple instructions im­ proves (Luiselli, 1990). As it was indicatedearlier,the first job of the behavioranalystis to make the sessions fun for the child (e.g., short,familiar, andmotivating).

Positive resultscan be enhanced byteaching the child an appropriate communication skill that replacesthe problem behaviors. The directteaching approach de­ scribed aboveoftenis the context forteaching the child the earliest skills to be learned.Gradually, expectations are increased andemerging skills arereinforced during natural school,home,andcommunity contexts.

Some behavior problems continuetopersist during or outside the direct teaching contextregardless of well intentioned efforts. At thispoint, welltrained behavior analysts develop anintervention planafter obtaining information throughadetailed functional assessment (O'Neill et al., 1997). In this approach,thebehavior at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from 172 AndersonandRomanczyk analyst conducts careful assessments to identify events that reliably predictandmaintain thebehavior. Subse­ quently, a behaviorsupportplan isdeveloped thatis linked to the assessment andmodifies antecedent events, teaches replacement behaviors,and changes maintaining consequences.

In summary, the continuumbased models primarily use a skills orientedapproach toimproving the child's behavior. Whenserious behavior problems do not im­ prove with skill reinforcementandredirection proce­ dures, a careful analysis is conductedand abehavior support plan isdeveloped. Conclusions At this time, controlledresearchindicates thatcom­ prehensive, wellconducted use of ABA offers the most hope for childrenwith autism and theirfamilies. In this article, we first providedthereader with adescription of programmatic andmethodologic commonelements of ABA. It is our opinionthatlittle disagreement exists within the behavioral community regardingthegeneral elements thatshould be presentin a quality program.

However, as iscommon with anyapproach, oftenthere are differences in the specific interventiontechniques that are chosen, the proportionof onetechnique versus another, and thesequencing and timing of theiruse.

The diversity of these techniqueshascreated confusion regarding thecombination ofbehavioral techniques that is most likely to be effective. It is our opinionthat ABA includes a large numberofconceptually consis­ tent techniques thatcan, and should, be used in various combinations across manydifferentcontexts. References Alberto, P. A., &Troutman, A. C. (1999). Applied behavior analysisfor teachers. Upper SaddleRiver, NJ: Prentice-Hall. Anderson, S. R.,Avery, D. L.,DiPietro, E. K.,Edwards, G. L., &Christian, W. P. (1987). Intensivehome-based early intervention withautistic children. Education and Treatment of Children, 10,352-366. Anderson, S. R.,Campbell, S., &O'MalleyCannon,B. (1994).

The May CenterforEarly Childhood Education. In S. L.

Harris &J. S. Handleman (Eds), Preschooleducation pro­ grams for children with autism. (pp. 15-36) Austin, TX: Pro-Ed.

Anderson, S. R.,Taras, M., &O'Malley Cannon,B. (1996).

Teaching new skills to young childrenwith autism. In C.

Maurice, G. Green, &S. C. Luce (Eds.), Behavioralinter­ vention for young children withautism: A manual forparents and professionals. (pp. 181-184) Austin,TX:Pro-Ed.

Baer, D. M., Wolf, M. M., &Risley, T. R. (1968). Some cur­ rent dimensions ofapplied behavior analysis. Journal ofAp­ plied Behavior Analysis, 1, 91-97.

Barlow, D. H. (1988). Anxiety and its disorders. New York:

Guilford.

Barlow, D., Hayes, C., &Nelson, R. (1984). The scientist prac­ titioner. Boston: Allyn and Bacon.

Berkell Zager,D. (Ed.). (1999). Autism: Identification, educa­ tion, and treatment (2nd ed.). Hillsdale,NJ:Erlbaum.

Birnbrauer, J. S., &Leach, D. J. (1993). The MurdockEarly Intervention Programafter 2 years. Behaviour Change, 10, 63-74. California Department ofDevelopmental Services. (1999). A reportto the legislature: Changesin the population ofpersons withautism and pervasive developmental disordersin Cali­ fornia's Developmental Services system: 1987 through1998. Sacramento, CA:California Healthand Human Services Agency.

Cohen, D. J., &Volkmar, F. R. (Eds.). (1997). Handbook of autism and pervasive developmental disorders (2nd Ed.). New York: Wiley.

Dawson, G., &Osterling, J. (1997).Earlyintervention in au­ tism: Effectiveness andcommon elements ofcurrent ap­ proaches. In M. J.Guralnick (Eds.). Theeffectiveness of early intervention: Second generation research. (pp. 307-326) Baltimore: Paul H.Brookes.

Delmolino, L. M., & Romanczyk, R. G. (1995). Facilitated communication: A critical review. The Behavior Therapist, 18(2),27-30.

Fenske, E. C.,Zalenski, S.,Krantz, P. J., & McClannahan, L. E. (1985). Age at interventionandtreatment outcomefor autistic children in acomprehensive interventionprogram. Analysis and Intervention in Developmental Disabilities,5, 49-58.

Green, G. (1996a). Evaluatingclaimsabouttreatments for autism. In C. Maurice,G.Greene, &S. C. Luce (Eds.), Be­ havioral intervention for young childrenwithautism. (pp.

15-28) Austin, TX:Pro-Ed.

Green, G. (1996b). Earlybehavioral intervention for autism:

What doesresearch tell us? In C. Maurice, G. Greene, & S. C. Luce (Eds.), Behavioral interventionfor young children with autism. (pp. 29~4) Austin, TX:Pro-Ed.

Handleman, J. S., &Harris,S. L. (1994). The Douglass Devel­ opmental Disabilities Center.In S. L.Harris &J. S.Handle­ man (Ecls.), Preschool education programs forchildren with autism. (pp. 71-86) Austin,TX:Pro-Ed.

Harris, S. L., &Handleman, J. S. (1994). Preschooleducation programs forchildren withautism. Austin, TX:Pro-Ed.

Harris, S. L.,Handleman, J. S.,Gordon, R., Kristoff, B., & Fuentes,F. (1991). Changesin cognitive and languagefunc­ tioning of preschool childrenwith autism. Journal ofAutism and Developmental Disabilities, 21, 281-290.

Hersen, M., & Barlow, D. (1981). Singlecaseexperimental designs. New York: Pergamon.

Iwata, B. A., Vollmer,T. R.,Zarcone, J. R., &Rodgers, T. A.

(1993). Treatment classification and selectionbased on be­ havioral function. In R. Van Houten &S.Axelrod (Eds.), Behavior analysis and treatment (pp. 101-168). New York:

Plenum.

Jenkins, J. R.,Odom, S. L., &Speltz, M. L. (1989). Effects of social integration onpreschool childrenwithhandicaps. Ex­ ceptional Children, 55, 420~28. Kamps, D. M., Walker,D.,Dugan, E. P.,Leonard, B. R., Thi­ badeau, S.F., Marshall, K.,Grossnickle, L., &Boland,B.

(1991). Small groupinstruction forschool-aged students with autism and developmentaldisabilities. Focuson Autis­ ticBehavior, 6,1-18.

Koegel, R. L., Russo, D. C., &Rincover, A. (1977). Assessing and training teachers in thegeneralized use ofbehavior modification with autistic children. Journal ofApplied Be­ havior Analysis,10,197-205. Kohler, F. W., Strain, P. S., &Shearer, D. D. (1996). Examin­ ing levels of social inclusion within an integratedpreschool for children with autism. In L. K. Koegel, R. L. Koegel, & G.Dunlap (Eds.), Positive behavioral support: Including people withdifficult behavior in the community. (pp. 305­ 332) Baltimore: Paul H.Brookes.

Lovaas, O. I.(1981). Teaching developmentally disabledchil­ dren: The ME book. Baltimore: UniversityParkPress.

Lovaas, O. I. (1987). Behavioraltreatmentandnormal educa- at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from Continuum-BasedBehavioralModels 173 tionalandintellectual functioning in youngautisticchildren.

Journal of Consulting and Clinical Psychology, 55, 3-9.

Luiselli, J. K. (1990). Recent developments innonaversive treatment: A review of rationale,methods,andrecommen­ dations. In A. C. Repp &N. N. Singh (Eds.),Perspectives on the use ofnonaversive and aversive interventions forpersons with developmental disabilities (pp. 73-86).Sycamore, IL:

Sycamore Publishing.

Luyben, P. (1998). Discretetrial training: Harmfulorhelpful?

Editorial comment in theNewsletter ofthe New York State Association forBehavior Analysis, 10, 4-5.

Matson, J. L.(Ed.). (1994). Autism in children and adults. Pa­ cific Grove, CA:Brookes Cole.

Matson, J. L.,Benavidez, D. A.,Compton, L. S., Paclawskyj, T., & Baglio, C. (1996). Behavioraltreatmentofautistic persons: A review of researchfrom 1980 to the present.

Research in Developmental Disabilities,17, 433-465.

Matson, J. L., Sevin, J. A., Box, M. L., Francis, K. L., &Sevin, B. M. (1993). An evaluationof twomethods forincreasing self-initiated verbalizations inautistic children. Journal of Applied Behavior Analysis, 26(3), 389-398.

Maurice, C. (1993). Let me hearyourvoice. New York: Faw­ cett Columbine.

McClamahon, L. E.

& Krantz,P. J. (1994). The Princeton Child Development Institute.In S. L.Harrs &J. S.Handle­ man (Eds.), Preschool educationprograms for children with authis. (pp.107-126) Austin,TX:Pro-Ed.

McEachin, J. J., Smith, T., &Lovaas, O. I. (1993). Long-term outcome forchildren withautism whoreceived earlyinten­ sive behavioral treatment.AmericanJournal on Mental Re­ tardation, 97(4), 359-372.

McGee, G. G., Daly, T., &Jacobs, H. A. (1994). The Walden preschool. In S. L.Harris&J. S.Handleman (Eds.),Pre­ school education programs for children with autism. Austin, TX: Pro-Ed.

McGee, G. G.,Krantz, P. J.,Mason, D., & McClannahan, L. E. (1983). A modifiedincidental-teaching procedurefor autistic youth: Acquisitionandgeneralization ofreceptive object labels. Journal ofApplied Behavior Analysis, 16, 329­ 338.

Miltenberger, R. G. (1998).Methodsfor assessing antecedent influences onchallenging behaviors.In J. K. Luiselli &M. J.

Cameron (Eds.),Antecedent controlprocedures for the be­ havioral support ofpersonswithdevelopmental disabilities.

(pp. 47-66) Baltimore:Paul H.Brookes.

New York State Department ofHealth, EarlyIntervention Program. (1999a).Clinical practice guideline: Report of the recommendations. AutismlP D D, assessment and interven­ tion in young children (age 0-3 years) (No. 4215).Albany, NY: New York State Department of Health.

New York State Department ofHealth, EarlyIntervention Program. (1999b).Clinicalpractice guideline: Guideline technical report. Autism/pervasive developmentaldisorders, assessment and interventionforyoung children (ages 0-3 years) (No. 4217). Albany,NY: New York StateDepart­ ment ofHealth.

Olley, G.,&Gutentag, S. S. (1999). Autism: Historicalover­ view, definition, andcharacteristics. In D.Berkell Zager (Ed.),Autism: Identification, education, and treatment (2nd ed.). (pp.3-22) Hillsdale, NJ:Erlbaum.

O'Neill, R. E.,Horner, R. H.,Albin, R. W., Sprague, J. R., Storey, K., & Newton, J. S. (1997). Functionalassessment and program development forproblem behavior: A practical handbook. PacificGrove, CA:Brookes Cole.

Palmieri, L. A.,Valluripalli, L.,Arnstein, R., &Romanczyk, R. G. (1998, May).Use of theinternet forinformation searches: ABA, autism, and the web. Paper presented at the 24th annual convention of theAssociation forBehavior Analysis, Orlando,FL. Pelham, W. E. (1989). Drugsandbehavior therapy.In M. Hersen (Ed.),Innovations in childbehavior therapy.(pp.

191-206) New York: Springer.

Pelham, W. E., & Murphy, H. A. (1986). Attention deficits and conduct disorders. In M. Hersen (Ed.),Pharmacological and behavioral treatments: An integrative approach. (pp.

108-148) New York: Wiley.

Powers, M. D., &Handleman, J. S. (1984).Behavioralassess­ ment of severedevelopmental disabilities.Rockville,MD: Aspen Systems.

Romanczyk, R. G. (1994). Autism.In V. S. Ramachandran (Ed.), Theencyclopedia ofhuman behavior (Vol. 1, pp. 327­ 336). San Diego: AcademicPress.

Romanczyk, R. G. (1996). Behavioralanalysis and assessment:

The cornerstonetoeffectiveness. In C.Maurice, S. Luce, & G. Green (Eds.),Behavioral intervention foryoung children with autism. (pp.195-217) Austin,TX:Pro-Ed.

Romanczyk, R. G.,Lockshin, S., &Navalta,C. (1994). Differ­ ential diagnosis of autism. In J. Matson(Ed.), Autism in children and adults: Etiology, assessment, and intervention. (pp. 99-125) Pacific Grove,CA:Brookes Cole.

Romanczyk, R. G., & Matthews,A. L. (1998). Physiological state asantecedent: Utilizationinfunctional analysis. In J. K. Luiselli & M. J. Cameron (Eds.),Antecedent control procedures for thebehavioral support ofpersonswith devel­ opmental disabilities. (pp.93-121) NewYork: Paul H.

Brookes.

Romanczyk, R. G.,Matey, L., &Lockshin, S. B. (1994). The children's unit fortreatment andevaluation. In S.L.Harris & J. S. Handleman (Eds.),Preschool educationprograms for children with autism. (pp.181-223) Austin,TX:Pro-Ed.

Schopler, E., &Mesibov, G. M. (Eds.).(1988). Diagnosis and assessment in autism.New York: Plenum.

Sheinkopf, S. J., &Siegel, B. (1998). Home-basedbehavioral treatment of youngchildrenwith autism. Journal ofAutism andDevelopmental Disorders,28, 15-23.

Sidman, M. (1988). Tactics ofscientific research. Boston:Au­ thors Cooperative.

Smith, T. (1996). Are other treatments effective?In C. Mau­ rice, G. Greene, &S. C. Luce (Eds.),Behavioral interven­ tion for youngchildren with autism. (pp.45-59) Austin, TX:

Pro-Ed.

Smith, T., Eikeseth,S.,Levstrand, M., &Lovaas, O. I. (1997).

Intensive behavioral treatment forpreschoolers withsevere mental retardation andpervasive developmental disorder.

American Journalon Mental Retardation, 102, 238-249.

Spiegler, M. D., &Guevremont, D. C. (1993).Contemporary behavior therapy.PacificGrove, CA:Brookes Cole.

Stokes, T. F., &Osnes, P. G. (1988). The developingapplied technology ofgeneralization andmaintenance. In R.Hor­ ner, G. Dunlap, &R. L.Koegel (Eds.),Generalization and maintenance: Life-stylechangesinapplied settings. (pp.

5-19) Baltimore: Paul H.Brookes.

Strain, P. S., & Cordisco, L. K. (1994). LEAP Preschool. In S. L. Harris &J. S. Handleman (Eds.), Preschool education programs for children with autism. (pp. 225-244)Austin, TX: Pro-Ed.

Sulzar-Azaroff, B., &Mayer,G. R. (1991). Behavioranalysis for lasting change. Chicago:Holt,Rinehart, andWinston.

Weiss, M. J. (1999). Differentialratesof skill acquisition and outcomes of earlyintensive behavioral intervention for au­ tism. Behavioral interventions: Theory and practice in resi­ dential and community-based clinical programs, 14, 3-22.

Article received: April29, 1999 Final acceptance: August2, 1999 Editor in charge: Fredda Brown at FLORIDA INTERNATIONAL UNIV on June 17, 2015 rps.sagepub.com Downloaded from