JOURNAL OF APPLIED BEHAVIOR ANALYSIS

BEHAVIORAL TRAINING FOR SIBLINGSOF AUTISTIC CHILDREN

LAURA SCHREIBMAN, ROBERT E. O'NEILL, AND ROBERT L. KOEGEL

CLAREMONT MCKENNA COLLEGE, CLAREMONT GRADUATE SCHOOL,AND UNIVERSITY OF CALIFORNIA, SANTA BARBARA

The present study was conducted to investigate the effectiveness of a program designedto teach behavior modification procedures to normal siblings of autistic children. Threesibling pairs participated in a multiple-baseline analysis of the effects of training thenormal siblings to use behavior modification procedures to teach their autistic brotheror sister a variety of learning tasks. Results indicated that the siblings learned to usethe behavioral procedures at a high level of proficiency, they used the procedures in ageneralization setting, and there were observed improvements in the behavior of theautistic children. In addition, a social validation assessment of the normal siblings'statements about their autistic sibling indicated a decrease in negative statements and anincrease in positive statements after training. These results are discussed in terms ofthe potential for incorporating siblings into the treatment plan in intervention pro-grams with autistic children.DESCRIPTORS: autistic children, family treatment, generalization, siblings

Procedures for training parents and otherfamily members to serve as therapists for theirhandicapped children are becoming increasinglywell documented in the behavioral literature(see Graziano, 1978; Koegel, Schreibman, John-son, O'Neill, & Dunlap, in press; O'Dell, 1974).Most of this research has concentrated on train-ing the children's parents to serve as therapistsor co-therapists. With respect to autistic chil-dren, several different types of training pro-grams have been developed, and training par-ents in behavior modification procedures hasnow been shown to produce positive changes in

This research was supported by U.S. Office of Edu-cation Research Grant G007802084 from the Bureauof Education for the Handicapped and by U.S. PublicHealth Service Research Grants MH 28210 and MH28231 from the National Institute of Mental Health.The authors are particularly grateful to Michael Stein-berg for his input. In addition, Gina Richman, JohnBurke, Kimberly Wheeler, and Paul Graff all pro-vided assistance in this investigation. We are espe-cially appreciative to the children and families whoparticipated in this research. Requests for reprintsshould be addressed to Laura Schreibman, PsychologyDepartment, Claremont McKenna College, Clare-mont, California 9171 1.

many aspects of autistic children's behavior(e.g., Koegel, Glahn, & Nieminen, 1978;Koegel, Schreibman, Britten, Burke, & O'Neill,1982; Kozloff, 1973; Wolf, Risley, Johnston,Harris, & Allen, 1967; Wolf, Risley, & Mees,1964).

There has also been some suggestion thatsimilar techniques could be used profitably totrain siblings in families with handicapped chil-dren (Cash & Evans, 1975; Frankel, Tymchuck,& Simmons, 1976; Miller & Cantwell, 1976;Weinrott, 1974). The successful results of theabove studies suggest optimism for pursuing thisapproach with severely handicapped children,such as those diagnosed as autistic. Some initialresearch has demonstrated the application ofthese procedures to a limited extent with certaintarget behaviors exhibited by such children (Col-letti & Harris, 1977). Similarly, research has be-gun to show that interaction with other normalpeers outside the family can have a beneficialeffect on certain behaviors of autistic children(Frankel et al., 1976; Ragland, Kerr, & Strain,1978; Strain, Kerr, & Ragland, 1979).The above research shows that siblings can

129

1983, 16,129-138 NUMBER 2 (SUMMER 1983)

LAURA SCHREIBMAN et al.

be trained, and it also supports the feasibilityof attempting to train siblings in more broadlybased general behavior modification procedures.This might prove helpful by giving siblingsskills that will help them interact more gener-ally with their autistic brothers and sisters (Du-koff, 1980; Sullivan, 1979). The purpose of thepresent study, therefore, was to investigate theeffectiveness of a generalized training programfor siblings of autistic children. The presentstudy extends previous work in this area, which,as mentioned above, has focused on teaching sib-lings one aspect of behavior management skills,or on having them teach the handicapped chil-dren only one task or behavior (Cash & Evans,1975; Colletti & Harris, 1977). The present pro-gram focused on giving the siblings a set of gen-eralizable behavior modification skills, whichthey could use in a variety of settings with avariety of target behaviors.

Three main questions were addressed in thisinvestigation: (a) could siblings of autistic chil-dren be taught to conduct correct behavior modi-fication procedures at a high level of proficiency;(b) if so, would their implementation of suchprocedures also take place in nontraining (gen-eralization) environments during unstructuredplay activities; and (c) would the siblings' im-plementation of behavior modification proce-dures produce measurable improvements in au-tistic children's behavior?

METHOD

ParticipantsThe participants in this study were three

sibling pairs, with one child in each pair diag-nosed autistic by at least two outside agenciesnot associated with our research projects. Thethree autistic children and their families wereparticipants in a research program conductedjointly by the University of California at SantaBarbara and Claremont McKenna College. Allthree of the autistic children were currently en-rolled in public school special education pro-grams.

Sibling pair 1 contained an 8-yr-old autisticfemale and her brother. Her IQ (nonverbal)from the Leiter International Performance Scalewas 78. Her social age, from the Vineland SocialMaturity Scale was 7.2. She had some verysimple speech, and displayed some appropriatesocial skills (such as occasional affectionate be-havior). Her main autistic symptoms were de-layed echolalia, a large amount of self-stimula-tory behavior, and occasional intense tantrums.This child's sibling was a 13-yr-old male. He wasof normal intelligence, performed well in school,and participated in a variety of extra-curricularactivities (e.g., baseball and soccer).

Sibling pair 2 consisted of an 8-yr-old autisticboy and his 11-yr-old sister. The autistic boy hada (nonverbal) Merrill-Palmer IQ of 60. His so-cial age, from the Vineland, was 3.7. He hadsome simple expressive and receptive labelingskills, with occasional spontaneous speech andappropriate play behaviors. His self-stimulatorybehavior was infrequent but relatively intensewhen it occurred. The sibling was of normalintelligence, and functioned adequately in schooland social situations. She was described by herparents (and observed by the experimenters) tobe generally quiet and to display some socialshyness.

Sibling pair 3 consisted of a 5-yr-old autisticboy and his sister. The autistic child had a (non-verbal) Merrill-Palmer IQ of 83. His social age,obtained from the Vineland, was 4.2. He hadlimited expressive and receptive speech, and oc-casionally displayed some echolalia and self-stimulatory behavior. This child's sibling was an8-yr-old female of above average intelligencewho obtained high grades in school. Like thepreviously described sibling, she was reported(and observed) to be somewhat "quiet."None of the three normal siblings had ever

been suspected of, or diagnosed as, having anytype of clinical pathology or learning disability.

DesignA multiple-baseline design across sibling pairs

was used to assess the acquisition and generali-

TRAINING SIBLINGS

zation of behavior modification skills by the sib-lings. All experimental sessions were conductedin the children's homes, except for the generali-zation probe sessions which were conducted in aseparate building.

For each sibling pair a predetermined numberof baseline measures were obtained. Duringthese measures the sibling was instructed to at-tempt to teach the autistic child a task that was-determined not to be in the child's repertoire,such as time telling, simple discriminations, orspeech concepts. Table 1 lists the tasks trainedby the sibling pairs during baseline and trainingsessions. Baseline sessions lasted for 15 min each,

Table 1Tasks worked on by sibling-autistic child pairs duringbaseline and training sessions.

Sibling Pair OneDiscrimination between penny, nickel, dime, quar-

ter, and dollarNumber of pennies in a nickelBefore and after concept for numbers (i.e., two is

before three)Number of pennies and nickels in a dimeNumber of pennies, nickels, dimes, in a quarterExpressive labeling of penny, nickel, dime, quarter,

and dollar

Sibling Pair TwoPreposition discrimination (between in, on, under,

behind)Discrimination between pronouns "I" and "you"

(I have vs. you have)Expressive labeling of truck, tractor, and topReceptive labeling of letters (a, c, h, k, r, s, t)Expressive labeling of penny, nickel, dime, quarter,

and dollar

Sibling Pair ThreeBefore and after concepts (for numbers)Receptive identification of money (penny, dime,

nickel)Sorting of picture cards by group (animals, clothes,

etc.)Pronoun discrimination (I have vs. you have)One-to-one correspondence (counting objects)Spelling short words (cat, hat)First/last concepts (What did you touch first?

Last?)Identification of capital lettersShort-term memory after hearing a sentence (Who

went to the store?)

during which the sibling's behavior was recordedin terms of five behavior modification skill areas(see Dependent Measures). The autistic chil-dren's correct responding was also recorded.

Sibling Training ProceduresAfter baseline, the siblings were trained in

the following manner. During the first trainingsession the sibling and trainer viewed a video-tape which presented examples of behaviortherapy with autistic children. Reinforcement,shaping, chaining, and discrete trial techniqueswere discussed (cf. Koegel & Schreibman, 1982).These techniques were explained in referenceto both appropriate and inappropriate behaviors(e.g., reinforcement, extinction). For example,with respect to reinforcement the siblings weretaught how to provide positive consequencescontingent on appropriate task responding andgood behavior; with respect to extinction thesiblings were taught to withhold attention orother positive consequences contingent on inap-propriate behavior. (Also, the siblings were toldthat if the autistic child displayed disruptive be-havior, such as tantrums, they were to ignoreit and continue to try and work with the child[paced instructions). None of the children inthis study exhibited disruptive behavior severeenough to warrant outside intervention by thetrainers during the training sessions.) Frequentstops were made to explain and clarify pointspresented in the videotape, and to simplify tech-nical terms with which the siblings would beunfamiliar. The trainer also gave the siblingexamples of how the techniques and procedurescould be applied to more general target behav-iors and situations besides those shown on thetape.

Next, the trainer and sibling discussed how be-havior modification procedures could be appliedto everyday situations involving problem be-haviors, using examples from a training manualdesigned for use with exceptional children(Baker, Brightman, Heifetz, & Murphy, 1976).The next training step involved instruction tothe siblings while they worked with their autistic

131

LAURA SCHREIBMAN et a,.

brother or sister. A specific target behavior waschosen and worked on for approximately 30min. The trainer would periodically interruptbriefly with corrective or positive feedback onthe sibling's progress. If the sibling had troubleapplying a particular procedure, the trainermodeled the procedure and asked the sibling totry again until he or she could perform it cor-rectly.

After each training period a new (untrained)task was chosen and the sibling worked for anadditional 15 min for the purpose of data col-lection. Since this was a probe to determine thesibling's proficiency, there was no interventionby the trainer. Eight such sessions with trainingand data collection were conducted for eachchild.

Dependent MeasuresPrimary measures. These were the siblings'

correct use of behavior modification proceduresand the autistic children's correct responding.Data were collected during each baseline andtraining session. The siblings' behavior wasscored during every other 30-sec interval forcorrect use of five behavior modification proce-dures. These areas were the use of SD's (instruc-tions and questions), use of prompts, use of shap-ing, use of consequences, and the use of discretetrials. The sibling had to use a procedure cor-rectly for the entire 30-sec interval to be scoredcorrect (+) for that interval. The scoring defi-nitions for the five skill areas were taken froma scoring system used in a previous study (onparent training) by Koegel et al. (1978), andare presented in Table 2. An overall percent cor-rect score across the five categories was calcu-lated for each baseline and training session bydividing the number of intervals scored cor-rect (+) by the total number of intervals, andmultiplying by 100.

During these same intervals, the correct re-sponding by the autistic children was also re-corded. The child had to respond correctly (orattempt to respond correctly) to the sibling'sinstructions for the entire interval to be scored

as a (+) for that interval. A percent correctresponding was calculated for each baseline andposttraining session using the same formulaas was presented above with the sibling percentcorrect scores.

Generalization measures. Along with theabove measures, we gathered additional probedata related to the generalization of the proce-dures carried out with the autistic children bytheir siblings, both during training and aftertraining. These probes were conducted in aroom located at the research project site anddecorated to resemble a living room. A varietyof toys was placed in the room, and the siblingwas instructed to interact and play with his orher autistic brother or sister while they werealone in the room. The sibling was not askedto conduct any training. These sessions werecovertly videotaped for scoring. The siblings' in-teractions were characterized in terms of discretetrials and were scored for the same teaching cate-gories as the home data sessions. For example,if the sibling said, "Get the ball," the autisticchild brought the ball, and the sibling delivereda consequence ("thanks, good job"), these wouldbe scored as the components of one discrete trial(i.e., "Get the ball" an SD, bringing the ball-the response, "thank you" the consequence).As with the baseline and training sessions, thecorrect responding by the autistic children wasalso recorded. These generalization data werecollected to assess both stimulus and responsegeneralization in that they were in a novel andunstructured setting and involved different be-haviors.

Social validation assessment. As mentioned inthe introduction, much of the motivation for thisresearch stemmed from parental concern aboutthe interactions between their normal and au-tistic children. Therefore, we conducted an in-formal social validation assessment (cf. Wolf,1978) of the effects of the training program byrandomly asking the parents for examples ofcomments the siblings were making about theirbrothers and sisters both before and after train-ing. To minimize demand characteristics and to

TRAINING SIBLINGS

Table 2Instructions for Scorers

Observers were told to mark a "+," a "-," or"N.A." (not applicable) for each category for each30-sec interval. Observers watched a session for 30sec, and scored while the session continued for thenext 30 sec.SD's

1. The SD should be clear and discriminable; thatis, it should stand apart from anything else that thesibling says. A good SD has a distinct beginning and adistinct end.

2. The SD should be appropriate to the task. If thesibling is teaching the child to point to a red card, heshould not mistakenly say, "point to the blue card."

3. The SD should be consistent with that given onthe previous trial for that task.

4. The SD should be uninterrupted.5. When the SD is presented, the child should be

attending. The child should not be engaged in off-taskor overly disruptive behavior.Prompts

1. The prompt must be effective; that is, it mustevoke a correct response.

2. The prompt should be faded (as the teachingsituation advances). This should be scored with re-spect to the overall teaching session.Shaping

1. Shaping involves the correct reinforcement ofsuccessive approximations. In order to have a goodshaping procedure, each reinforced response shouldbe at least as good as the last one. (Score shaping withreference to responses actually observed during thescoring period.)Consequences

1. Consequences should be delivered immediatelyafter the response. Immediately is defined as within3 sec.

2. Consequences should be contingent; that is,reinforcement only for correct responses, and non-reinforcement only for incorrect responses (or pun-ishment for inappropriate responses).

3. The consequences should be unambiguous. A"No" said with a smile or a "Good boy" given witha frown are ambiguous.

4. Consequences should be consistent. Reinforce-ment should follow each correct response, unless thechild has reached an advanced level, in which casethe reinforcement may be scheduled. Once the siblingbegins to punish or say "No," he or she should do itconsistently.

5. Consequences should be effective. They shouldbe tailored to each child. Reinforcers should be some-thing the child is eager for, and punishments some-thing the child dislikes or attempts to avoid.

Discrete trials1. Within the natural flow of the sibling/autistic

child interactions, observable (although usually verybrief) intertrial intervals should be discriminable.That is, the interactions should not be run together ina manner that makes it difficult to discriminate whereone interaction ends and the other begins.

N.B. 1. In working with siblings it is rare to en-counter a child who uses very sophisticated program-ming techniques. However, should one encounter achild who systematically uses delayed consequencesor sophisticated reinforcement schedules, these shouldbe taken into consideration in the scoring.

2. In the absence of research specifically ad-dressing the need for very stringent shaping criteria,it is conceivable that the definition in the above tablemay be modifiable slightly after future research isconducted in this area.

try to prevent bias from affecting these data, wesimply asked the parents for comments thechildren made, rather than directly asking about"attitudinal" shifts. Also, whenever possible, wehad two experimenters obtain samples of com-ments from the same parents at different times sothat comments that were reported on multipleoccasions could be emphasized in the datasample.

ReliabilityReliability data were obtained by a second

observer for at least one fourth of all the base-

line, training, and generalization sessions. Ob-server agreement was calculated for both siblingbehavior and correct responding by the autisticchildren. For sibling behavior and correct re-sponding, an agreement was scored if both ob-servers recorded a plus (+) or both observersrecorded a minus (-) for a particular siblingskill area or child correct responding during aninterval. Reliability was calculated by dividingthe number of agreements by the number ofagreements plus the number of disagreements,and multiplying by 100.

Observer agreement for the sessions ranged

133

LAURA SCHREIBMAN et al.

from 67% to 100% for the SD category, witha mean of 94%; from 71% to 100% for theprompts category, with a mean of 93%; from67% to 100% for the consequences category,with a mean of 93%; from 86% to 100% forthe discrete trial category, with a mean of 96%;and from 87% to 100% for the autistic childcorrect responding category, with a mean of97%.

RESULTS

Siblings' Correct Use ofBehavior Modification Procedures

Figure 1 shows (solid line) the siblings' aver-age percent correct use of behavior modificationprocedures during baseline and training ses-sions. As can be seen, prior to training thesiblings showed no consistently high levels ofcorrect behavior modification usage (althoughsibling 3 inconsistently exhibited high levels inthree scattered pretraining sessions). The meancorrect performance levels for siblings 1, 2, and3 during baseline were 53%, 42%, and 73%,respectively. The introduction of the trainingprocedures produced consistent increases abovebaseline levels for all three siblings. Further,all three siblings exhibited consistent and stablelevels of correct procedures after training began.The mean correct performances during the in-tervention for siblings 1, 2, and 3 were 87%,98%, and 100%, indicating that they wereacquiring and using the techniques at a steadyhigh level of proficiency.A more detailed analysis of the siblings' be-

havior is presented in Table 3, which shows theaverage percent correct use of each teachingcomponent, both pre- and posttraining (exceptshaping, which was never used for these tasks).These data show that, prior to training, the sib-lings in this study were able to provide relativelyclear instructions and questions, but had diffi-culty using effective prompts and consequences,and clear discrete trials. The training was ef-

fective in improving the siblings' use of all thecomponents to high levels of correct perfor-mance, which produced positive changes in theautistic children's behavior.

Autistic Children's Correct RespondingThe percent correct responding by the autistic

children is also shown in Figure 1, representedby the dashed line. None of the autistic childrenshowed steady improvements prior to the intro-duction of the training procedures. The meanpercent correct responding for autistic children1, 2, and 3 was 9%, 16%, and 15%, respec-tively. However, after training began, and as thenormal siblings became more proficient and con-sistent in their teaching skills, the percentage ofcorrect responding showed increases for all threechildren (means of 45, 58, and 45). The correctresponding for autistic child 1 shows some de-crease during the latter stages of treatment, butthe level of correct responding remained abovebaseline levels.

Generalization ProbesFigure 1 also shows the normal siblings' per-

cent correct use of behavior modification, as wellas the autistic children's correct responding, dur-ing the generalization sessions. Looking at per-cent correct usage of behavior modificationacross the two probes for each normal sibling,the results indicate for sibling 1, 74% and 72%;for sibling 2, 92% and 68%; and for sibling 3,98% and 100%. The percent correct respondingfor the autistic children across the two probes

Table 3Average percent correct use of each teaching tech-nique (pre- and posttraining) by the siblings.

Conse- DiscreteSD Prompt quences Trials

Pre Post Pre Post Pre Post Pre Post

Sib 1 79 89 31 71 44 88 51 90Sib 2 84 100 33 100 37 97 8 97Sib 3 98 100 73 100 67 100 61 100

TRAINING SIBLINGS

Baseline

Pair #1

go,* . .*0

I I I I

1 2 3 4

Pair #2

1 2 3 4 5 6

Pair #3

,,-. . .

N1*

Training I

&–_

+ T100 w%0

%01

I I I I I I I I

5 6 7 8 9 10 11 12

+ a+

O..

/

/4

I~~~~~~\ ,,

ftft 1%-0

I I I I 1 I I I I

17 8 9 10 11 12 13 14

+ tl _ ,X9

ow . . . .

/10

I/

+

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SESSIONSFig. 1. Percent correct usage by sibling of correct behavior modification procedures and autistic children's

appropriate responding. Data are presented from both experimental and generalization settings.

-~~Sib. (Exp. Setting)— Aut. Child

(Exp. Setting)o) Sib. (Gen. Setting)+ Aut. Child

(Gen. Setting)

,l

100

80W 600Z 402 200o ol

F- 100

c 80

0 60° 40

o 0wI-zIL 1000F 80zw 600a: 40wL- 20

0

135

_ . .~~~~~~~~~~~

LAURA SCHREIBMAN et al.

was child 1, 68% and 65%; child 2, 83% and92%; and for child 3, 47% and 45 %.

Socid ValidaotuTable 4 shows that the parents reported that,

prior to training, the normal children most oftenexpressed comments that could be characterizedas being neutral or tolerant, with occasionalnegative elements. For example, the siblingwould say things such as "He looks weird andhe acts funny, but he's my brother and I guesswe have to put up with him." The table alsoshows that the training appeared to produce apositive shift in the siblings' verbalizations asreflected in comments such as "He does okay ifwe ask him right."

DISCUSSION

Overall, the results are quite encouraging insupport of the idea of training autistic children'ssiblings in behavior modification. The siblingsbecame proficient in behavioral teaching skillsand their increased skills enabled them to pro-duce improvements in the autistic children's cor-rect responding. In addition, the parents werepleased with the changes in the siblings andthey all felt the training had been very beneficial.

Significantly, the data show that the siblingsused their skills in a different environment andalso in a much less structured type of interactionthan had occurred during the training sessions.Generalization of the acquired skills is very im-portant if the training is to be expected to havea generalized positive effect on the siblings'interactions. Although these data must be in-terpreted with some caution since no baselineassessments were conducted in the generalizationsettings, the results are encouraging.The data have implications for several areas

of the functioning of the children's families. Weobserved several different types of improved in-teractions among the siblings with respect bothto their general comments and to their abilityto control and evoke appropriate behavior fromthe autistic children. We also realize, however,

Table 4Sample comments made by siblings about their au-tistic brothers and sisters prior to and after training.

Prior to Training After Training

"He looks weird and acts "He does okay if we askfunny, but he's my him right."brother and I guess we "I don't get so mad at herhave to put up with when I ask her righthim." and she does okay."

"She breaks things and "She behaves a lot bettermesses up my stuff, but when we work withI guess we have to be her right."patient." "It's nicer to be around

"He never does anything him when he playsfun with me, because good with us."he doesn't know how." "We get along easier

"We have to pay atten- when I make him paytion to her a lot, even attention."though she never "My friends like himknows what we're better when he listensdoing." to us."

"Sometimes he acts funny "He gets along better if Iaround my friends, but know how to ask him."I don't always get madbecause he's mybrother."

"She causes problems alot, but I guess wehave to put up withher."

"I know she's part of thefamily, but she's hardto get along with."

that with such increased control there could alsobe a potential for problems, such as the siblingsmisusing them or "exploiting" their brothers andsisters. We did not observe any of this type ofproblem in this study, but we recognize thatit may arise. Thus, we feel that careful monitor-ing, especially by trained parents, may be animportant component of this approach. This andother specific questions concerning the relativeroles of trained and untrained parents in con-junction with trained versus untrained siblingspresent a number of interesting questions forfuture research.

In general, the positive possibilities of theseprocedures seem quite broad. The data suggestthat siblings may be useful adjuncts to trainedparents in teaching and maintaining behaviorsin many areas in addition to those appropriate

136

TRAINING SIBLINGS 137

social and play behaviors measured in the pres-ent study. For example, along with parents, sib-lings could play an important role in providingcontinuity between the home and school forvarious educational programs, thus alleviatingpossible contrast problems and promoting gen-eralization (cf. Koegel, Egel, & Williams, 1980).Siblings also are in an ideal position to serve asfacilitation agents for the children's appropriatesocial interactions with other children in theneighborhood, thus increasing the potential fora more normalized development for the autisticchild; and conversely, helping to prevent thedevelopment of maladaptive interactions, lead-ing to progressively more and more abnormalsituations as the autistic child matures (cf. Pat-terson, 1975; Patterson & Reid, 1970).

In relation to this latter point, recent researchby Wellen (Note 1) and Wellen and Broen,1982) has suggested that in some situations oldersiblings of normal and language delayed chil-dren may hinder the children's linguistic devel-opment by interfering with their language in-teractions with adults. Training programs forsiblings may alter such potentially problematicsituations by teaching the siblings how topromote and facilitate appropriate languageproduction by their handicapped brothers andsisters.

All these types of sibling participation willperhaps function to lessen the burden that fami-lies with autistic children often have to bear.The positive results presented here and the po-tential beneficial results from future work makethis seem like an especially important area forfurther research and teaching. The results of thepresent study provide encouraging evidencefor the continuing development of interventionstrategies that involve the family members ofautistic and other severely handicapped childrenas primary sources of treatment delivery.

REFERENCE NOTE1. Wellen, C. J. Older siblings' effects on the speech

young children hear and produce. Unpublished

doctoral dissertation, University of Minnesota,1981.

REFERENCES

Baker, B. L., Brightman, A. J., Heifetz, L. J., &Murphy, D. M. Behavior problems. Champaign,Ill.: Research Press, 1976.

Cash, W. M., & Evans, I. M. Training preschoolchildren to modify their retarded siblings' be-havior. Journal of Behavior Therapy and Experi-mental Psychiatry, 1975, 6, 13-16.

Colletti, G., & Harris, S. L. Behavior modificationin the home: Siblings as behavior modifiers, par-ents as observers. Journal of Abnormal ChildPsychology, 1977, 5, 21-20.

Dukoff, S. A sibling's view. In C. D. Webster, M.M. Konstantareas, J. Oxman, & J. E. Mack (Eds.),Autism: New directions in research and educa-tion. New York: Pergamon Press, 1980.

Frankel, F., Tymchuck, A. J., & Simmons, J. Q.Operant analysis and intervention with autisticchildren: Implications of current research. InE. R. Ritvo (Ed.), Autism: Diagnosis, current re-search, and management. New York: SpectrumPublications, 1976.

Graziano, A. M. Parents as behavior therapists. InM. Hersen, R. M. Eisler, & P. M. Miller (Eds.),Progress in behavior modification: Vol. 4. NewYork: Academic Press, 1978.

Koegel, R. L., Egel, A. L., & Williams, J. A. Behav-ioral contrast and generalization across settingsin the treatment of autistic children. Journal ofExperimental Child Psychology, 1980, 30, 422-437.

Koegel, R. L., Glahn, T. J., & Nieminen, G. S. Gen-eralization of parent training results. Journal ofApplied Behavior Analysis, 1978, 11, 95-109.

Koegel, R. L., & Schreibman, L. How to teach au-tistic and other severely handicapped children.Lawrence, Kansas: H & H Enterprises, 1982.

Koegel, R. L., Schreibman, L., Britten, K., Burke, J. C.,& O'Neill, R. E. A comparison of parent train-ing to direct child treatment. In R. L. Koegel, A.Rincover, & A. L. Egel (Eds.), Educating andunderstanding autistic children. Houston: College-Hill Press, 1982.

Koegel, R. L., Schreibman, L., Johnson, J., O'Neill,R. E., & Dunlap, G. Collateral effects of parenttraining on families with autistic children. InR. F. Dangel & R. A. Polster (Eds.), Behavioralparent training: Issues in research and practice.New York: Guilford Press, in press.

Kozloff, M. Reaching the autistic child. Champaign,Ill.: Research Press, 1973.

Miller, N. B., & Cantwell, D. P. Siblings as thera-pists: A behavioral approach. American Journalof Psychiatry, 1976, 133, 447-450.

138 LAURA SCHREIBMAN et al.

O'Dell, S. Training parents in behavior modifica-tion: A review. Psychological Bulletin, 1974, 81,418-433.

Patterson, G. R. The aggressive child: Victim andarchitect of a coercive system. In E. J. Mash, L. A.Hammerlynck, & L. C. Handy (Eds.), Behaviormodification and families. New York: Brunner/Mazel, 1975.

Patterson, G. R., & Reid, J. B. Reciprocity andcoercion: Two facets of social systems. In C. Neu-ringer & J. D. Michael (Eds.), Behavior modifica-tion in clinical psychology. New York: Appleton-Century-Crofts, 1970.

Ragland, E. U., Kerr, M. M., & Strain, P. S. Effectsof peer social initiation on the behavior of with-drawn autistic children. Behavior Modification,1978, 2, 565-578.

Strain, P. S., Kerr, M. M., & Ragland, E. U. Effectsof peer-mediated social initiations and prompt-ing/reinforcement procedures on the social be-havior of autistic children. Journal of Autism andDevelopment Disorders, 1979, 9, 41-54.

Sullivan, R. C. Siblings of autistic children. Journalof Autism and Developmental Disorders, 1979,9, 287-298.

Weinrott, M. R. A training program in behaviormodification for siblings of the retarded. Ameri-can Journal of Orthopsychiatry, 1974, 44, 362-375.

Wellen, C. J., & Broen, P. A. The interruption ofyoung children's responses by older siblings.Journal of Speech and Hearing Disorders, 1982,47, 204-210.

Wolf, M. M. Social validity: The case for subjectivemeasurement or how applied behavior analysisis finding its heart. Journal of Applied BehaviorAnalysis, 1978, 11, 203-214.

Wolf, M. M., Risley, T. R., Johnston, M., Harris,F., & Allen, E. Application of operant condi-tioning procedures to the behavior problems ofan autistic child: A follow-up and extension. Be-haviour Research and Therapy, 1967, 5, 103-111.

Wolf, M. M., Risley, T. R., & Mees, H. Applicationof operant conditioning procedures to the be-havior problems of an autistic child. BehaviourResearch and Therapy, 1964, 1, 305-312.

Received April 12, 1982Final acceptance December 2, 1982