JOURNAL OF APPLIED BEHAVIOR ANALYSIS                       1993,26,121-132                          NUMBER     I (SPRING 1...
122                PATRICIA J. KRANTZ and LYNN E. McCLANNAHANand Frea (1992) demonstrated that when the social        auti...
CHILDREN WITH AUTISM                                                 123    Walt, also age 12, had entered the program 5  ...
124                 PATRICIA J. KRANTZ and LYNN E. McCLANNAHANthat were separated from the speakers previous           ing...
CHILDREN WITH AUTISMincluded the 3 other childrens names. In addition,                to these instructions with his or he...
126                    PATRICIA J. KRANTZ and LYNN E. McCLANNAHAN                                                     Tabl...
CHILDREN WITH AUTISM                                                              127              25                     ...
128                 PATRICIA J. KRANTZ and LYNN E. McCLANNAHAN(11, 17, and 25 initiations during the session).           T...
CHILDREN WITH AUTISM                                                                         129          25              ...
PATRICIA J. KRANTZ and LYNN E. McCLANNAHANpreferred activities were typically followed by pre-      initiations in the las...
CHILDREN WITH AUTISM                                                           131disabilities who were the recipients of ...
132                  PATRICIA J. KRANTZ and LYNN E. M.CLANNAHAN   verely handicapped adolescents following picture prompt ...
Upcoming SlideShare
Loading in …5

Teaching children with autism to initiate to peers effects of a script-fading procedure-krantz


Published on

Published in: Health & Medicine, Education
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Teaching children with autism to initiate to peers effects of a script-fading procedure-krantz

  1. 1. JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1993,26,121-132 NUMBER I (SPRING 1993) TEACHING CHILDREN WITH AUTISM TO INITIATE TO PEERS: EFFECTS OF A SCRIPT-FADING PROCEDURE PATRICIA J. KRANrz AND LYNN E. MCCLANNAHAN PRINCETON CHILD DEVELOPMENT INSTITUTE A script that was systematically faded from end to beginning was used to teach peer initiations about recently completed, current, and future activities. The effectiveness of the script-fading procedure was assessed via a multiple baseline design across 4 children with autism. During baseline, the children seldom initiated to peers, although all had previously acquired some functional expressive language and sometimes spontaneously addressed adults. When the script was introduced, peer initiations increased, and as the script was faded, unscripted initiations increased. With the minimal written prompts available in the final fading steps, initiations generalized to a different setting, time, teacher, and activity; and for 3 of the 4 children, peer initiations were maintained at a 2-month follow-up. After the script was faded, the participants levels of peer initiations were within the same range as a normative sample of 3 nondisabled youngsters. The script-fading procedure enabled children with severe social and verbal deficits to practice context-specific, peer-directed generative language that was not prompted by adults or peer confederates. DESCRIPTORS: autistic children, peer-initiation interventions, antecedent control, prompting, script fading A cardinal feature of autism, observed and in- forcement are apt to produce atypical peer socialvestigated for almost 50 years, is a severe deficit in exchanges marked by brief interaction episodes, toreciprocal social interaction (Kanner, 1943; Rapin, require the ongoing presence of an adult, or to yield1991; Wing & Gould, 1979). Although many decreases in social interaction when adult promptschildren with autism acquire expressive language, are reduced or withdrawn. These issues led Odomthey may infrequently display spontaneous speech and Strain (1986) to conclude that "a clear direc-or may engage in conversation only with significant tion for future research will be establishing a pro-adults and not with other children (Krantz, Rams- cedure in which teacher prompts are not, & McClannahan, 1989). An approach is needed for transferring the stimulus Several authors (e.g., Odom & Strain, 1986; control from the teacher to natural elements in theOke & Schreibman, 1990; Shafer, Egel, & Neef, environment" (p. 69).1984) have reviewed the course of research on peer The development of a technology to enhance theinteraction, noting difficulties encountered and social relatedness and social competence of peopleproblems that remain to be addressed. In brief, with developmental disabilities is receiving increas-when socially competent peers are taught to initiate ing attention (Breen & Haring, 1991), and in theinteractions with children with autism, target chil- case of children with autism, there is some evidencedrens responses increase, but their initiations may that the acquisition of social interaction skills mayremain unchanged or may even be suppressed favorably affect inappropriate or socially unaccept-(Odom, Hoyson, Jamieson, & Strain, 1985); fur- able responses. For example, Oke and Schreibmanther, maintenance and cross-setting generalization (1990) reported that the disruptive behavior of aof social interaction skills have been difficult to boy with autism remained unchanged when non-achieve. Other complications arise because teaching disabled peers were trained to initiate interactionstrategies that indude adult prompting and rein- with him, but decreased when he was taught to initiate interaction. Lee and Odom (1991) found Reprints may be obtained from the authors, Princeton that when 2 children with autism increased theirChild Development Institute, 300 Cold Soil Road, Princeton, engagement in social interaction with peers, stereo-New Jersey 08540. typic behavior decreased. Koegel, Koegel, Hurley 121
  2. 2. 122 PATRICIA J. KRANTZ and LYNN E. McCLANNAHANand Frea (1992) demonstrated that when the social autism often appear to be unable to discern ap-responsivity of 4 children with autism increased, propriate conversational topics, the script identifiedthere were reductions in disruptive behavior. These recently completed, present, and future activities asrecent studies provide additional grounds for con- topics for discussion. After children were observedtinued exploration of procedures that will amplify to dependably use the script, it was faded from endthe social interaction repertoires of children with to beginning in five fading steps. We ajso assessedautism. unscripted peer initiations, responses, cross-setting Our study was designed to assess the effects of generalization, and maintenance over a 2-montha new procedure script fading-on the social ini- period.tiations of children with autism. Loveland and Tun-ali (1991) discussed social scripts as "sets of ex- METHODpectations for human behavioral events" (p. 178),and suggested that identifying and applying scripts Participants and Settingto relevant social situations might be especially dif- The 4 participants-Kate, Mike, Walt, andficult for people with autism, because to do so Ross-attended the Princeton Child Developmentrequires social information and understanding of Institutes day school and intervention program forcultural norms. They found that people with au- 5.5 hr per day, 5 days per week. Kate and Walttism, when compared to language-matched people resided in a community-based group-home treat-with Down syndrome, made fewer responses to a ment program; Mike and Ross lived with their ownconversational "social script" about a stolen wallet. families. All 4 children met the DSM-III-R (APA,But what if the script had been specifically taught? 1987) criteria for autism and, in each case, autism Goldstein, Wickstrom, Hoyson, Jamieson, and had been diagnosed by one or more outside agen-Odom (1988) taught typical and language-delayed cies. The parents of all 4 children gave informedpreschoolers to display motor and verbal scripted consent for the youngsters participation in thisresponses pertaining to two dramatic play situa- research.tions-a hamburger stand and a barber shop. Dur- Kate, age 12, had been enrolled in the programing training, the teacher prompted children to do for 8 years; at admission she was completely non-or say the components of the script (e.g., say "I verbal and displayed tantrums and high-rate self-want a hamburger"). Although some teacher injurious head hitting. A Peabody Picture Vocab-prompting was necessary in the generalization set- ulary Test-R (PPVT), administered just prior toting (a free-play period), the investigators noted the study, yielded a receptive language age-equiv-that script training reduced the need for intrusive alent score of 5-1. The Woodcock Reading Masteryprompts, and the teacher did not become an in- Tests revealed a reading grade-equivalent score ofteraction partner. In addition, following training, 1.9. On the McCarthy Scales of Childrens Abili-the participants exhibited theme-related social re- ties, she received a score of <50, and her age-sponses that were not directly trained. equivalent score on the Vineland Adaptive Behavior The social uses of language represent major dif- Scales was 5-2.ficulties for many people with autism (Volkmar, Mike, age 12, had been in treatment for 5 years; 1987), and we have noted that some children who when first seen, he had no expressive language,experience considerable difficulty with vocal instruc- frequently engaged in inappropriate laughing andtions and prompts are consistently able to respond crying, and systematically attempted to isolate him-to pictorial or written stimuli that remain available self from others. His PPVT age-equivalent scorefor review (MacDuff, Krantz, & McClannahan, was 4-5, his Woodcock reading grade-equivalent 1993). Thus, in the present study, we examined score was 1.8, his score on the McCarthy Scalesthe effects of a written script and a script-fading was <50, and his Vineland age-equivalent scoreprocedure on peer initiations. Because children with was 5-1.
  3. 3. CHILDREN WITH AUTISM 123 Walt, also age 12, had entered the program 5 when schedule components were resequenced andyears earlier with a small verbal-imitation reper- when new schedule components (representing pre-toire. He had a history of aggression (e.g., hitting, viously acquired skills) were added, and all usedbiting, pinching, and kicking) and self-injury, as these schedules during the school day. Kate, forwell as stereotyped body movements and a poten- example, followed a written schedule that includedtially life-threatening disregard for danger (he at- components such as "book bag," "mirror,"tempted to jump out of moving vehicles and throw "homework," and "computer" prompts for herflammable objects on stove burners). He achieved to get her book bag at the beginning of the schoola PPVT age-equivalent score of 4-9, a Woodcock day, look in the mirror and correct her appearancereading grade-equivalent score of 1. 5, and a Vine- if necessary, check her homework, and practiceland age-equivalent score of 5-4. arithmetic facts on the classroom computer. Ross, age 9, had been in treatment for 5 years; The participants had also developed some func-at program enrollment, he was echolalic and had tional expressive language, albeit atypically appropriate spontaneous speech. He actively All responded verbally when addressed by adults,avoided social contact and frequently displayed ste- and they sometimes directed spontaneous requestsreotypies. His PPVT age-equivalent score was 4-4, to teachers, therapists, and parents (e.g., "I wanthis Woodcock reading grade-equivalent score was a cookie" or "I want to ride the bike"), but they1.6, his score on the McCarthy Scales was <50, rarely initiated conversation with caregivers otherand his Vineland age-equivalent score was 4-6. than to request items or activities. In the year preceding the study, the WISC-R For 9 months prior to the study, the 4 partici-was administered to each child by representatives pants were dassmates; their class schedule was de-of his or her referring agency. Kate, Mike, Walt, signed specifically for them and was not shared byand Ross achieved verbal IQ scores of 50, <45, other children. Although their teachers modeled<45, and 49, respectively, and full-scale IQ scores and verbally prompted social initiations and pro-of 49, 49, 46, and 48. vided behavior-specific praise and preferred snacks, In summary, all 4 participants had histories of play materials, and activities contingent upon im-impaired imitation skills, severe communication def- itative or prompted initiations, these proceduresicits, and minimal or absent academic, social, lei- were not effective in producing spontaneous peersure, and home-living skills. After 5 to 8 years of initiations; the youngsters failed to initiate in theintervention, they were seldom aggressive or self- absence of adult prompts, although it was notedinjurious, exhibited fewer ritualistic and stereotyped that they sometimes responded to other childrensresponses, and had acquired some early academic prompted initiations.skills. The setting was a school and research center for Before the study, the children had learned to children with autism. Presession activities occurredindependently follow photographic activity sched- in an outdoor play area. All sessions (except gen-ules (Wacker & Berg, 1983, 1984; Wacker, Berg, eralization sessions) were conducted in a typicalBerrie, & Swatta, 1985) and, later, written activity classroom furnished with desks, chairs, and book-schedules that specified lengthy chains of academic, cases; generalization sessions were held in a largeself-care, and leisure responses. Schedule-following conference room furnished with upholstered chairsskills were taught without verbal or gestural and rectangular tables.prompts; manual prompts were delivered from be-hind the youths in a sequence of most to least Dependent Variablesintrusive, and graduated guidance was followed by Initiation to peers was defined as understand-spatial fading (cf. MacDuff et al., 1993). As a able statements or questions that were unpromptedresult of these relatively errorless procedures, all 4 by an adult, that were directed to another child bychildren had learned to follow written schedules using his or her name or by facing him or her, and
  4. 4. 124 PATRICIA J. KRANTZ and LYNN E. McCLANNAHANthat were separated from the speakers previous ing, and painting-were systematically rotatedvocalizations by a change in topic or a change in across sessions. Upon entering the classroom, eachrecipient of interaction. If a child initiated and then child found art materials at his or her place, as wellimmediately repeated the same sentence or ques- as a single sheet of paper that presented the writtention, the repetition was not counted as an initiation. instructions, "Do your art" and "Talk a lot." ToScripted initiations were those that matched the ensure that participants were aware of these instruc-written script, with the exception that conjunctions, tions, the teacher stood behind each child and man-artides, prepositions, or pronouns could be altered ually guided him or her to point to each instructionor deleted (e.g., substituting "and" for "or," or with a pencil and to move the pencil along beneathdropping "the"), and verb tense could be changed. the text. Because of the childrens prior experienceFor example, "Ross, you like my picture," was with written activity schedules and manual prompts,coded as a scripted initiation, although the script it was anticipated that this procedure would promptread "Ross, I like your picture." Unscripted ini- the child to read the instructions aloud; however,tiations were verbal productions that differed from for Session 1 only, the teacher was asked to use athe script by more than conjunctions, articles, prep- single verbal prompt (reading the first word of anositions, pronouns, or changes in verb tense; the instruction) if a child failed to respond to manualquestion, "Would you like some more paper?" was prompts. During Session 1, the teacher providedscored as unscripted because the noun "paper" did one verbal prompt ("Do .. .") to 1 participantnot occur in the script. A response was defined as (Mike). In all subsequent sessions, verbal promptsany contextual utterance (word, phrase, or sentence) were exduded from the protocol. After promptingthat was not prompted by the teacher and that reading of the written instructions, the teacher movedoccurred within 5 s of a statement or question to the periphery of the dassroom; she interacteddirected to the target child (as indicated by use of with the participants only if they directed questionsthe childs name, or by facing the child who was to her.initiating). Examples of responses were "what?" Script. The same three art activities were rotated"okay," and "yes, I do." across sessions, and the two written instructions presented during baseline continued to be displayedExperimental Conditions but were followed by scripts consisting of 10 state- Vocal initiations and responses to peers were ments and questions (e.g., "[Name], did you likemeasured during baseline and script conditions in to [swing/rollerskate/ride the bike] outside to-a multiple baseline design across the 4 children. day?" "Hey [name], would you like some candyBecause this study focused on the use of written or chips?" "[Name], do you want to use one ofinstructions and scripts, each child was tested on my [pencils/crayons/brushes]?" "[Name], wonttarget words before the study began, and teaching it be fun to go to the [park/store/farm] on Funwas provided until all participants achieved 100% Friday?"). Scripts were constructed so that theiraccuracy on oral reading of the individual words content reflected (a) activities the children had re-used during baseline and script conditions. All ses- cently completed (e.g., using playground equip-sions during subsequent experimental conditions ment), (b) activities the children were planning (i.e.,were preceded by an outdoor recess and by a brief "Fun Friday" activities that included parties anddiscussion, guided by a teacher, during which the trips to a park, a nearby farm, or a local conveniencechildren were assisted in planning "Fun Friday," store), and (c) objects in the school environmenta weekly outing or special event. Subsequently, they (e.g., candy, bowls of potato chips, and materialsentered a classroom where they were seated in a on their desks). Blank sections of the scripts wererectangular arrangement created by pushing their completed by the teacher immediately before eachdesks together, or a conference room where they session, so that scripts reflected activities that thewere seated at a rectangular table. children had recently completed or discussed. Scripts Baseline. Three art activities-drawing, color- were individualized, so that each participants script
  5. 5. CHILDREN WITH AUTISMincluded the 3 other childrens names. In addition, to these instructions with his or her pencil and tothree different versions of the script were created move the pencil from left to right along the bottomby randomly assigning the 10 questions and state- of the text. Subsequently, the teacher moved to thements to Positions 1 through 10; these versions edge of the activity area, and provided no furtherwere systematically rotated across children. prompts. Standing behind a participant, the teacher man- During Generalization Sessions 4 through 6, theually guided him or her to pick up a pencil, point children found not only the written instructions butto an instruction or a scripted statement or question, also the faded script as it was then being presentedand move the pencil along below the text. If nec- to them during the script condition. Thus, Kate,essary, the teacher also manually guided the childs Mike, and Walt encountered Fading Step 5 (quo-head to face another child to whom a statement or tation marks), and Ross encountered Fading Stepquestion was addressed. If the child did not ver- 4 (quotation marks and one upper-case letter). Inbalize the statement or question within 5 s, the these sessions, no prompts were delivered by themanual guidance procedure was repeated. If the teacher.child read or said a statement or read or asked a On the same day as Session 9, 3 children withoutquestion, the teacher used the same type of manual disabilities, ages 9, 10, and 12, were invited toguidance to ensure that the child placed a check visit the school. These children were previouslymark to the left of that portion of the script. acquainted, having met at social gatherings at- Manual prompts were faded as quickly as pos- tended by their parents. Before this visit, their par-sible; no prompts were delivered to Kate, Mike, ents provided consent for observation and mea-Walt, and Ross after Sessions 15, 18, 23, and 27, surement of their interactions with one another.respectively, and the teacher remained at the pe- During their visit, they were scheduled to spendriphery of the classroom throughout subsequent some time together in the same classroom used bysessions. After manual guidance had been faded the 4 participants while the latter pursued theirfor a target child, fading of the script began. Scripts activities in a different area. On their desks, thewere faded from end to beginning, in five phases. nondisabled youngsters found art materials and sin-For example, the fading steps for the question gle sheets of paper bearing the instructions "Do"Mike, what do you like to do best on Fun Fri- your art" and "Talk a lot." The teacher was presentdays?" were (a) "Mike, what do you like to do at the periphery of the classroom but provided nobest, (b) "Mike, what do you, (c) "Mike, what, verbal or manual prompts.(d) "M, and (e)". Follow-up. At a 2-month follow-up, all partic- Measurement Proceduresipants used fading Step 5 (one pair of quotation Each observation period began immediately aftermarks). The follow-up session was separated from the last child was seated and continued for 10Sessions 1 through 42 by 1 month of vacation and consecutive minutes. Two observers (or pairs of 1 month of a new school year. The teacher did not observers) were stationed at the periphery of thedeliver prompts during this session. classroom; each observer or observer pair collected Generalization. Generalization sessions (which data on 2 participants, and observers assignmentsoccurred on the same days as Sessions 34 through to target children were rotated across sessions. A39 of the script condition) were conducted in a continuous event recording system (in 1 -min in-different setting (a conference room), with a dif- tervals) was used to score sequentially scripted andferent teacher, at a different time of day, and with unscripted initiations and responses to peers; andifferent materials (puzzles). During the first three audible signal from an Apple Ile® classroom com-generalization sessions, a single piece of paper placed puter marked the end of each 1-min interval.on each childs work surface contained the instruc- In three sessions for each participant during thetions "Do your puzzle" and "Talk a lot," and the script condition and before the first fading step,teacher manually guided each participant to point additional observers scored whether the target child
  6. 6. 126 PATRICIA J. KRANTZ and LYNN E. McCLANNAHAN Table 1 Interobserver Agreement Percentages on Total Number of Initiations, Number of Scripted and Unscripted Initiations, and Number of Responses by Child during Classroom and Generalization Sessions Total Scripted Unscripted initiations initiations initiations Responses Child Mean Range Mean Range Mean Range Mean RangeClassroom sessionsKate 97 86-100 97 83-100 98 75-100 90 89-100Mike 98 84-100 98 75-100 98 82-100 98 86-100Walt 96 88-100 98 86-100 96 92-100 98 83-100Ross 91 0-100 99 75-100 91 0-100 90 33-100Generalization sessionsKate 99 94-100 100 94 88-100 100Mike 100 100 100 100Walt 100 100 - 100 100Ross 100 97 86-100 93 67-100 100 Note. Data on classroom sessions do not indude follow-up.said or read each statement or question induded in For the follow-up session conducted in the class-his or her script. The order of questions and state- room, interobserver agreement on total peer initi-ments was not assessed; observers measured only ations was 100%, 94%, 100%, and 100%; agree-whether the components of the script were used by ment on scripted initiations was 75%, 100%, 100%,the child during the observation period. These data, and 100%; and agreement on unscripted initiationsas well as other observations by the experimenters was 88%, 90%, 100%, and 100% for Kate, Mike,(e.g., whether Kate was self-restraining), were used Walt, and Ross, respectively. Agreement on re-in determining when to begin script fading. sponses was 100% for each of the 4 participants. For the 3 nondisabled children, interobserver agree-Interobserver Agreement ment on initiations was 85% for Child 1, 100% Observers sequentially recorded the initial of each for Child 2, and 100% for Child 3 (M = 95%).child who initiated or responded; scripted and un- The primary independent variable, use of thescripted initiations were coded minus and plus, script, was assessed three times for each participant,respectively. Agreements were scored only if both during dassroom sessions that occurred after theobservers coded the same child initial, in the same script was introduced but before fading began. Twocolumn, and in the same sequence, and used the observers scored whether a target child read or saidsame symbol to designate a scripted or unscripted each sentence or question induded in the script.initiation. Percentage agreement on each of the de- During these sessions, both observers scored everypendent variables was calculated by dividing total scripted sentence or question as present for all 4number of agreements by total number of agree- children. Thus, agreement on childrens use of thements plus disagreements, and multiplying by script was 100%.100%. Interobserver agreement was obtained dur-ing 80% of dassroom sessions, in the 2-month RESULTSfollow-up session, in five of six generalization ses-sions, and in the session conducted with nondis- Figure 1 displays the number of peer initiationsabled children. Mean agreement and ranges of and responses exhibited by each child during base-interobserver agreement on the occurrence of total line and script conditions in the dassroom. In base-initiations, number of scripted and unscripted in- line, neither Kate nor Mike initiated, but duringitiations, and number ofresponses during classroom the script condition their mean initiations were 15and generalization sessions are shown in Table 1. and 13, respectively (not induding follow-up). The
  7. 7. CHILDREN WITH AUTISM 127 25 2 Month 20 Follow-up 15 10 5. a1) 0 Co 25 0 20 - C) 15 n _0 10 - 5- 0. () 0 25 - . 0 20 - 15 - 4--) Co (I3 10 - A I A 5- ___ ~ ~ ~ ~ 4-) t t tt 0- 12 34 5 0 l - - I 25- -0 at 20- A :3 15- z7 10- 5- 0 t f ft O0 I 1 2 3 45 - I I - 0 5 10 15 20 25 30 35 40 Sessions Figure 1. Number of initiations and number of responses by Kate, Mike, Walt, and Ross during baseline, script, andfollow-up sessions in the classroom. Arrows indicate the introduction of fading steps.mean number of initiations for Walt increased from At the 2-month follow-up, when all 4 partici-0.1 during baseline to 17 during the script con- pants were on the fifth fading step (quotationdition. Rosss means were two during baseline and marks), Kate, Mike, and Walt continued to initiate14 during the script condition. to their peers at about the same levels as previously
  8. 8. 128 PATRICIA J. KRANTZ and LYNN E. McCLANNAHAN(11, 17, and 25 initiations during the session). Thus, for all of the participants except Ross, mostAlthough Ross displayed fewer initiations at follow- initiations observed during Sessions 4 through 6up, his score of 6 was above his baseline mean. were unscripted. Because there were no initiations in Sessions 1 During Session 42 of the script condition in theto 5, none of the participants responded. But in classroom, observers scored a total of 14 initiationsSession 6, when Kate entered the script condition, for Kate, 17 for Mike, 17 for Walt, and 21 forall 3 of the other youngsters made some responses. Ross. It is interesting to note that the participantsIn general, childrens responses increased as each scores are representative of the range of peer ini-successive youngster began to use the script; after tiations exhibited by the nondisabled children- 13Kate left baseline, mean responses to peers were for Child 1, 20 for Child 2, and 24 for Child 3.five for Mike, six for Walt, and nine for Ross;during the script condition, mean responses were10, 14, 11, and 14 for Kate, Mike, Walt, and DISCUSSIONRoss, respectively. People with autism characteristically display def- Figure 2 shows the number of scripted and un- icits in social competence, and that was true of thescripted initiations during dassroom sessions. After participants in this study. During baseline in theFading Step 5 was introduced, mean scripted and classroom, the children followed the written in-unscripted initiations were two and 13 for Kate, struction "Do your art," pursuing coloring, draw-seven and seven for Mike, six and 12 for Walt, ing, and painting activities; they did not, however,and six and 11 for Ross. With the exception of follow the written instruction to "Talk a lot." Fur-Ross, who was the last to encounter the fifth phase ther, Mike did not display peer initiations afterof script fading, scripted initiations decreased as Kate entered the script condition, nor did Waltsunscripted initiations increased. During follow-up, initiations to peers increase when both Kate andthe number of unscripted initiations was main- Mike had begun to use the script. Videotapes oftained at about the same level for Kate, Mike, and Sessions 13 and 21 showed that Rosss initiationsWalt, but decreased for Ross. during baseline were efforts to prevent peer inter- During the first three generalization sessions, action, e.g., "I dont like to talk," and "I dontwhen the children received only the instructions want to talk." Thus, none of the participants ap-"Do your puzzle" and "Talk a lot," Kate made peared to imitate his or her peers initiations. 10 initiations in Session 1 (three scripted and seven The introduction and systematic fading of theunscripted) but subsequently ceased to initiate. Mike script significantly increased social initiations; it maymade three unscripted initiations in Session 1 but be noted that, although there were only 10 scripteddid not initiate in Sessions 2 and 3. Walt did not sentences and questions, the children typically madeinitiate during the first three generalization sessions, far more than 10 initiations per session. Sentencesand Ross made only one unscripted initiation in and questions were deliberately not numbered inSession 3. But when the faded script was reintro- order not to convey any numeric criterion. As fading duced (Sessions 4 through 6), the number of peer proceeded, the participants continued to "check initiations increased. During these sessions, Kate off scripted initiations, as had been manually made 16, 20, and nine initiations (all unscripted) prompted before fading began; by Fading Step 5, and Mike made nine, eight, and nine initiations, they placed multiple check marks beside the quo- of which three, seven, and five were unscripted. tation marks or started new columns of check marks. Walt made three, 11, and six initiations; two, 10, Placing check marks adjacent to completed tasks and five of these were unscripted. Ross initiated was a familiar response-the children also made two, zero, and 14 times during Sessions 4 through check marks beside items on their written activity 6, making one, zero, and six unscripted initiations. schedules. The schedules were arranged so that less
  9. 9. CHILDREN WITH AUTISM 129 25 2 Month Follow up 20 - Co 15.Z 0 10.I S. 0. 0 m 25 I - - - .1 1 2 3 4 5.I I 204-) I 15 _0 I 10 I U C) 5, I 0.r D L _ . 25 - I _0 20 - I I U 15 - I 10 - I 0 (a) 5- U 0 m 7C 0- 0 re A . __j a I I I I I I I a 9 I I L.I I. 25 1 2 3 4 5 I - _ . 0 20 15 I I z 10 0 5 10 15 20 25 30 35 40 Sessions Figure 2. Number of scripted and unscripted initiations by each participant during baseline, script, and follow-upsessions in the classroom. Arrows 1 through 5 indicate the introduction of fading steps.
  10. 10. PATRICIA J. KRANTZ and LYNN E. McCLANNAHANpreferred activities were typically followed by pre- initiations in the last phase of script fading in theferred activities, and rewards were delivered by adults classroom. Only Ross, who had attempted to pre-for schedule completion. This prior experience with vent or terminate peer interactions throughout thewritten activity schedules may have accounted for study, displayed a decrement in unscripted initia-their use of scripts in the absence of rewards from tions at follow-up.the teacher. As the children practiced peer initiations during As the script was faded, unscripted initiations script fading, some recurrent interaction patternsincreased. Videotapes of Sessions 37, 39, and 41 took on the characteristics of teasing: Ross, whoin the classroom indicated that initiations scored as repeatedly tried to end social interaction, becameunscripted included use of adjectives (e.g., "Walt, the recipient of a large number of questions; Mike,would you like some salty vinegar chips?"), intro- who was often given a treat that he did not prefer,duction of new nouns (e.g., "Mike, would you like became increasingly specific about his candy pref-more Tootsie Rolls®?" and "Wont it be fun to erences, sometimes to no avail. The experimentersgo to the seashore?"), and combination of scripted viewed these recurrent interactions as reminiscentsentences and questions to form new productions of the teasing behavior of nondisabled children.(e.g., "Did you like to swing on Fun Friday?" and The numbers of peer initiations made by the"I like to talk to you on Fun Fridays"). Still other participants during Session 42 (14 to 21) were notunscripted initiations were completely novel (e.g., outside the range of initiations (13 to 24) made"Kate, can I use your water?" [for a paint brush]). by the 3 nondisabled children when they engagedAfter Session 37, Kate exhibited only unscripted in the same art activity. The participants script wasinitiations. designed to indude content relevant to current, The videotapes also documented that Mike, recently completed, and future activities. ContentWalt, and Ross appropriately used scripted state- analysis of the nondisabled childrens videotapedments and questions after Fading Step 5 (quotation session indicated that they also discussed recentlymarks); perhaps the use of three different versions completed, current, and future activities (one, five,of the script contributed to their varied selections and nine peer initiations, respectively). However,of content and recipients of interaction and pre- the majority of their initiations were either requestsvented repetitive or stereotyped initiations. for personal information (e.g., "When is your birth- Although peer initiations did not generalize in day?" "What school do you go to?" "Whats yourthe absence of the script, the prompts provided in favorite TV show?") or elaborations on a topicFading Steps 4 (Ross) and 5 (Kate, Mike, and introduced by a previous speaker (e.g., "My fa-Walt) were sufficient to promote transfer to a new vorite subject is science," "I met my best friend atsetting, time, teacher, and activity. These minimal Scouts"). Analysis of their initiations offers a basiswritten prompts not only produced cross-setting for designing future scripts.generalization (which is often difficult to achieve It should be noted that, although peer initiationsfor children with autism; Oke & Schreibman, 1990), were very infrequent during baseline in the dass-but did so in the absence of teacher intervention, room, the participants were not silent-they wereso that teachers did not become key interaction responding to other childrens initiations. Illustra-partners. Further, the children adapted the content tive responses were "Yes it will," "No," "Yes, Iof their initiations to the new situation; they did want a potato chip," "Thank you," and "Dontnot say "I like your picture," but instead said "I talk."like your puzzle." Before Kate entered the script condition, there Peer initiations taught with the script-fading pro- were no peer initiations and no responses. As eachcedure proved quite durable for 3 of the 4 partic- successive child began to use the script, responsesipants. After 1 month of summer vacation and 1 by children still in baseline increased. Other inves-month of a new school year, Kates, Mikes, and tigators (cf. Odom et al., 1985) have noted in-Walts performances were quite similar to their creases in the social responding of children with
  11. 11. CHILDREN WITH AUTISM 131disabilities who were the recipients of initiations training on social and communicative interaction. Edu-from nondisabled peers, but increases in responses cation and Treatment of Children, 11, 97-117. Kanner, L. (1943). Autistic disturbances of affective con-have not been dependably associated with in creases tact. Nervous Child, 2, social initiations. In the present study, accelerated Koegel, L. K., Koegel, R. L., Hurley, C., & Frea, W. D.peer initiations resulted in gains in responses to (1992). Improving social skills and disruptive behavior in children with autism through self-management. Jour-peers, although responses were never rewarded by nal of Applied Behavior Analysis, 25, 341-353.the teacher. It may be noted, however, that some Krantz, P. J., Ramsland, S. E., & McClannahan, L. E.responses were followed by snacks; an affirmative (1989). Conversational skills for autistic adolescents: Anresponse to the scripted question "Would you like autistic peer as prompter. Behavioral Residential Treat- ment, 4, 171-189.some candy or chips?" usually resulted in delivery Lee, S., & Odom, S. L. (1991). The relationship betweenof these items. stereotypic behavior and peer social interaction for chil- After learning the script, the participants recom- dren with severe disabilities. Manuscript submitted forbined elements of scripted sentences and questions publication. Loveland, K. A., & Tunali, B. (1991). Social scripts forand drew on their existing verbal repertoires to add conversational interactions in autism and Down syn-new words and phrases (i.e., they exhibited gen- drome. Journal of Autism and Developmental Disor-erative language). Additional research is needed to ders, 21, 177-186.determine whether script-fading procedures can MacDuff, G. S., Krantz, P. J., & McClannahan, L. E. (1993). Teaching children with autism to use photo-achieve increases in the length and complexity of graphic activity schedules: Maintenance and generaliza-childrens verbal productions, whether the use of tion of complex response chains. Journal of Appliedmultiple scripts will promote conversational variety, Behavior Analysis, 26, 89-97. Odom, S. L., Hoyson, M., Jamieson, B., & Strain, P. S.whether initiations taught with a script-fading pro- (1985). Increasing handicapped preschoolers peer socialcedure will generalize to more remote settings (e.g., interactions: Cross-setting and component analysis.Jour-home and community), and whether written nal of Applied Behavior Analysis, 18, 3-16.prompts can ultimately be faded to the single writ- Odom, S. L., & Strain, P. S. (1986). A comparison of peer-initiation and teacher-antecedent interventions forten instruction, "Talk a lot." promoting reciprocal social interaction of autistic pre- A particular strength of the script-fading pro- schoolers. Journal of Applied Behavior Analysis, 19,cedure is that it diminishes the involvement of 59-71.adults and confederates in the social interactions of Oke, N. J., & Schreibman, L. (1990). Training social initiations to a high-functioning autistic child: Assess-children with autism. Adult prompts have often ment of collateral behavior change and generalization inbeen seen as intrusive, resulting in atypical social a case study. Journal of Autism and Developmentalexchanges, and initiations by peer confederates have Disorders, 20, 479-497.been difficult to maintain. In this study, a script Rapin, I. (1991). Autistic children: Diagnosis and dinical features. Pediatrics, 87, 751-760.that was faded to minimal written prompts enabled Shafer, M. S., Egel, A. L., & Neef, N. A. (1984). Trainingthe participants to engage in contextual social ex- mildly handicapped peers to facilitate changes in thechanges; peer initiations generalized to a different social interaction skills of autistic children. Journal ofsetting, time, and activity, and were maintained Applied Behavior Analysis, 17, 461-476. Volkmar, F. R. (1987). Social development. In D. J. Co-over a 2-month period. hen, A. M. Donnellan, & R. Paul (Eds.), Handbook of autism and developmental disorders (pp. 41-60). New York: Wiley. Wacker, D. P., & Berg, W. K. (1983). Effects of picture REFERENCES prompts on the acquisition of complex vocational tasksAmerican Psychiatric Association. (1987). Diagnostic and by mentally retarded adolescents. Journal of Applied statistical manual of mental disorders (3rd ed. rev.). Behavior Analysis, 16, 417-433. Washington, DC: Author. Wacker, D. P., & Berg, W. K. (1984). Training adoles-Breen, C. G., & Haring, T. G. (1991). Effects of contextual cents with severe handicaps to set up job tasks indepen- competence on social initiations.Journal ofApplied Be- dently using picture prompts. Analysis and Intervention havior Analysis, 24, 337-347. in Developmental Disabilities, 4, 353-365.Goldstein, H., Wickstrom, S., Hoyson, M., Jamieson, B., Wacker, D. P., Berg, W. K., Berrie, P., & Swatta, P. (1985). & Odom, S. (1988). Effects of sociodramatic script Generalization and maintenance of complex skills by se-
  12. 12. 132 PATRICIA J. KRANTZ and LYNN E. M.CLANNAHAN verely handicapped adolescents following picture prompt Received November 1, 1991 training. Journal of Applied Behavior Analysis, 18, Initial editorial decision January 3, 1992 329-336. Revisions received August 29, 1992; October 19, 1992;Wing, L., & Gould, J. (1979). Severe impairments of November 13, 1992 social interaction and associated abnormalities in children: Final acceptance November 13, 1992 Epidemiology and classification. Journal of Autism and Action Editor, Scott McConnell Developmental Disorders, 9, 11-30.