Training For Siblings Summary

JOURNAL OF APPLIED BEHAVIOR ANALYSIS

BEHAVIORAL TRAINING FOR SIBLINGS OF 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 designed to teach behavior modification procedures to normal siblings of autistic children. Three sibling pairs participated in a multiple-baseline analysis of the effects of training the normal siblings to use behavior modification procedures to teach their autistic brother or sister a variety of learning tasks. Results indicated that the siblings learned to use the behavioral procedures at a high level of proficiency, they used the procedures in a generalization setting, and there were observed improvements in the behavior of the autistic children. In addition, a social validation assessment of the normal siblings' statements about their autistic sibling indicated a decrease in negative statements and an increase in positive statements after training. These results are discussed in terms of the 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 other family members to serve as therapists for their handicapped children are becoming increasingly well 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 therapists or 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 has now been shown to produce positive changes in

This research was supported by U.S. Office of Edu- cation Research Grant G007802084 from the Bureau of Education for the Handicapped and by U.S. Public Health Service Research Grants MH 28210 and MH 28231 from the National Institute of Mental Health. The authors are particularly grateful to Michael Stein- berg for his input. In addition, Gina Richman, John Burke, Kimberly Wheeler, and Paul Graff all pro- vided assistance in this investigation. We are espe- cially appreciative to the children and families who participated in this research. Requests for reprints should be addressed to Laura Schreibman, Psychology Department, 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 that similar techniques could be used profitably to train siblings in families with handicapped chil- dren (Cash & Evans, 1975; Frankel, Tymchuck, & Simmons, 1976; Miller & Cantwell, 1976; Weinrott, 1974). The successful results of the above studies suggest optimism for pursuing this approach with severely handicapped children, such as those diagnosed as autistic. Some initial research has demonstrated the application of these procedures to a limited extent with certain target behaviors exhibited by such children (Col- letti & Harris, 1977). Similarly, research has be- gun to show that interaction with other normal peers outside the family can have a beneficial effect 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 feasibility of attempting to train siblings in more broadly based general behavior modification procedures. This might prove helpful by giving siblings skills that will help them interact more gener- ally with their autistic brothers and sisters (Du- koff, 1980; Sullivan, 1979). The purpose of the present study, therefore, was to investigate the effectiveness of a generalized training program for siblings of autistic children. The present study 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, which they could use in a variety of settings with a variety of target behaviors.

Three main questions were addressed in this investigation: (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 such procedures also take place in nontraining (gen- eralization) environments during unstructured play activities; and (c) would the siblings' im- plementation of behavior modification proce- dures produce measurable improvements in au- tistic children's behavior?

METHOD

Participants The participants in this study were three

sibling pairs, with one child in each pair diag- nosed autistic by at least two outside agencies not associated with our research projects. The three autistic children and their families were participants in a research program conducted jointly by the University of California at Santa Barbara and Claremont McKenna College. All three of the autistic children were currently en- rolled in public school special education pro- grams.

Sibling pair 1 contained an 8-yr-old autistic female and her brother. Her IQ (nonverbal) from the Leiter International Performance Scale was 78. Her social age, from the Vineland Social Maturity Scale was 7.2. She had some very simple speech, and displayed some appropriate social 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 was of normal intelligence, performed well in school, and participated in a variety of extra-curricular activities (e.g., baseball and soccer).

Sibling pair 2 consisted of an 8-yr-old autistic boy and his 11-yr-old sister. The autistic boy had a (nonverbal) Merrill-Palmer IQ of 60. His so- cial age, from the Vineland, was 3.7. He had some simple expressive and receptive labeling skills, with occasional spontaneous speech and appropriate play behaviors. His self-stimulatory behavior was infrequent but relatively intense when it occurred. The sibling was of normal intelligence, and functioned adequately in school and social situations. She was described by her parents (and observed by the experimenters) to be generally quiet and to display some social shyness.

Sibling pair 3 consisted of a 5-yr-old autistic boy 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 had limited expressive and receptive speech, and oc- casionally displayed some echolalia and self- stimulatory behavior. This child's sibling was an 8-yr-old female of above average intelligence who obtained high grades in school. Like the previously 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 any type of clinical pathology or learning disability.

Design A 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 conducted in the children's homes, except for the generali- zation probe sessions which were conducted in a separate building.

For each sibling pair a predetermined number of baseline measures were obtained. During these 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, or speech concepts. Table 1 lists the tasks trained by the sibling pairs during baseline and training sessions. Baseline sessions lasted for 15 min each,

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

Sibling Pair One Discrimination between penny, nickel, dime, quar-

ter, and dollar Number of pennies in a nickel Before and after concept for numbers (i.e., two is

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

and dollar

Sibling Pair Two Preposition discrimination (between in, on, under,

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

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

and dollar

Sibling Pair Three Before 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 letters Short-term memory after hearing a sentence (Who

went to the store?)

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

Sibling Training Procedures After baseline, the siblings were trained in

the following manner. During the first training session the sibling and trainer viewed a video- tape which presented examples of behavior therapy with autistic children. Reinforcement, shaping, chaining, and discrete trial techniques were discussed (cf. Koegel & Schreibman, 1982). These techniques were explained in reference to both appropriate and inappropriate behaviors (e.g., reinforcement, extinction). For example, with respect to reinforcement the siblings were taught how to provide positive consequences contingent on appropriate task responding and good behavior; with respect to extinction the siblings were taught to withhold attention or other positive consequences contingent on inap- propriate behavior. (Also, the siblings were told that if the autistic child displayed disruptive be- havior, such as tantrums, they were to ignore it and continue to try and work with the child [paced instructions). None of the children in this study exhibited disruptive behavior severe enough to warrant outside intervention by the trainers during the training sessions.) Frequent stops were made to explain and clarify points presented in the videotape, and to simplify tech- nical terms with which the siblings would be unfamiliar. The trainer also gave the sibling examples of how the techniques and procedures could be applied to more general target behav- iors and situations besides those shown on the tape.

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

131

LAURA SCHREIBMAN et a,.

brother or sister. A specific target behavior was chosen and worked on for approximately 30 min. The trainer would periodically interrupt briefly with corrective or positive feedback on the sibling's progress. If the sibling had trouble applying a particular procedure, the trainer modeled the procedure and asked the sibling to try 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 an additional 15 min for the purpose of data col- lection. Since this was a probe to determine the sibling's proficiency, there was no intervention by the trainer. Eight such sessions with training and data collection were conducted for each child.

Dependent Measures Primary measures. These were the siblings'

correct use of behavior modification procedures and the autistic children's correct responding. Data were collected during each baseline and training session. The siblings' behavior was scored during every other 30-sec interval for correct 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 discrete trials. The sibling had to use a procedure cor- rectly for the entire 30-sec interval to be scored correct (+) for that interval. The scoring defi- nitions for the five skill areas were taken from a scoring system used in a previous study (on parent training) by Koegel et al. (1978), and are presented in Table 2. An overall percent cor- rect score across the five categories was calcu- lated for each baseline and training session by dividing the number of intervals scored cor- rect (+) by the total number of intervals, and multiplying by 100.

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

as a (+) for that interval. A percent correct responding was calculated for each baseline and posttraining session using the same formula as was presented above with the sibling percent correct scores.

Generalization measures. Along with the above measures, we gathered additional probe data related to the generalization of the proce- dures carried out with the autistic children by their siblings, both during training and after training. These probes were conducted in a room located at the research project site and decorated to resemble a living room. A variety of toys was placed in the room, and the sibling was instructed to interact and play with his or her autistic brother or sister while they were alone in the room. The sibling was not asked to conduct any training. These sessions were covertly videotaped for scoring. The siblings' in- teractions were characterized in terms of discrete trials and were scored for the same teaching cate- gories as the home data sessions. For example, if the sibling said, "Get the ball," the autistic child brought the ball, and the sibling delivered a consequence ("thanks, good job"), these would be 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, the correct responding by the autistic children was also recorded. These generalization data were collected to assess both stimulus and response generalization in that they were in a novel and unstructured setting and involved different be- haviors.

Social validation assessment. As mentioned in the introduction, much of the motivation for this research stemmed from parental concern about the 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 by randomly asking the parents for examples of comments the siblings were making about their brothers and sisters both before and after train- ing. To minimize demand characteristics and to

TRAINING SIBLINGS

Table 2 Instructions for Scorers

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

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

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

3. The SD should be consistent with that given on the 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-task or overly disruptive behavior. Prompts

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

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

1. Shaping involves the correct reinforcement of successive approximations. In order to have a good shaping procedure, each reinforced response should be at least as good as the last one. (Score shaping with reference to responses actually observed during the scoring period.) Consequences

1. Consequences should be delivered immediately after the response. Immediately is defined as within 3 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 with a frown are ambiguous.

4. Consequences should be consistent. Reinforce- ment should follow each correct response, unless the child has reached an advanced level, in which case the reinforcement may be scheduled. Once the sibling begins to punish or say "No," he or she should do it consistently.

5. Consequences should be effective. They should be 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 trials 1. Within the natural flow of the sibling/autistic

child interactions, observable (although usually very brief) intertrial intervals should be discriminable. That is, the interactions should not be run together in a manner that makes it difficult to discriminate where one 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 a child who systematically uses delayed consequences or sophisticated reinforcement schedules, these should be 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 table may be modifiable slightly after future research is conducted in this area.

try to prevent bias from affecting these data, we simply asked the parents for comments the children made, rather than directly asking about "attitudinal" shifts. Also, whenever possible, we had two experimenters obtain samples of com- ments from the same parents at different times so that comments that were reported on multiple occasions could be emphasized in the data sample.

Reliability Reliability 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 sibling behavior and correct responding by the autistic children. For sibling behavior and correct re- sponding, an agreement was scored if both ob- servers recorded a plus (+) or both observers recorded a minus (-) for a particular sibling skill area or child correct responding during an interval. Reliability was calculated by dividing the number of agreements by the number of agreements 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, with a mean of 94%; from 71% to 100% for the prompts category, with a mean of 93%; from 67% to 100% for the consequences category, with a mean of 93%; from 86% to 100% for the discrete trial category, with a mean of 96%; and from 87% to 100% for the autistic child correct responding category, with a mean of 97%.

RESULTS

Siblings' Correct Use of Behavior Modification Procedures

Figure 1 shows (solid line) the siblings' aver- age percent correct use of behavior modification procedures during baseline and training ses- sions. As can be seen, prior to training the siblings showed no consistently high levels of correct behavior modification usage (although sibling 3 inconsistently exhibited high levels in three scattered pretraining sessions). The mean correct performance levels for siblings 1, 2, and 3 during baseline were 53%, 42%, and 73%, respectively. The introduction of the training procedures produced consistent increases above baseline levels for all three siblings. Further, all three siblings exhibited consistent and stable levels 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 were acquiring and using the techniques at a steady high level of proficiency. A more detailed analysis of the siblings' be-

havior is presented in Table 3, which shows the average percent correct use of each teaching component, both pre- and posttraining (except shaping, which was never used for these tasks). These data show that, prior to training, the sib- lings in this study were able to provide relatively clear 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 the components to high levels of correct perfor- mance, which produced positive changes in the autistic children's behavior.

Autistic Children's Correct Responding The percent correct responding by the autistic

children is also shown in Figure 1, represented by the dashed line. None of the autistic children showed steady improvements prior to the intro- duction of the training procedures. The mean percent correct responding for autistic children 1, 2, and 3 was 9%, 16%, and 15%, respec- tively. However, after training began, and as the normal siblings became more proficient and con- sistent in their teaching skills, the percentage of correct responding showed increases for all three children (means of 45, 58, and 45). The correct responding for autistic child 1 shows some de- crease during the latter stages of treatment, but the level of correct responding remained above baseline levels.

Generalization Probes Figure 1 also shows the normal siblings' per-

cent correct use of behavior modification, as well as the autistic children's correct responding, dur- ing the generalization sessions. Looking at per- cent correct usage of behavior modification across 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 responding for the autistic children across the two probes

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

Conse- Discrete SD Prompt quences Trials

Pre Post Pre Post Pre Post Pre Post

Sib 1 79 89 31 71 44 88 51 90 Sib 2 84 100 33 100 37 97 8 97 Sib 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

,,- . .\ .

N 1*

Training I

&--_

+ T 100 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

SESSIONS Fig. 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

80 W 60 0 Z 40 2 20 0o ol

F- 100

c 80

0 60 ° 40

o 0 w I- z IL 100 0 F 80 z w 60 0 a: 40 w L- 20

0

135

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

LAURA SCHREIBMAN et al.

was child 1, 68% and 65%; child 2, 83% and 92%; and for child 3, 47% and 45 %.

Socid Validaotu Table 4 shows that the parents reported that,

prior to training, the normal children most often expressed comments that could be characterized as being neutral or tolerant, with occasional negative elements. For example, the sibling would say things such as "He looks weird and he acts funny, but he's my brother and I guess we have to put up with him." The table also shows that the training appeared to produce a positive shift in the siblings' verbalizations as reflected in comments such as "He does okay if we ask him right."

DISCUSSION

Overall, the results are quite encouraging in support of the idea of training autistic children's siblings in behavior modification. The siblings became proficient in behavioral teaching skills and their increased skills enabled them to pro- duce improvements in the autistic children's cor- rect responding. In addition, the parents were pleased with the changes in the siblings and they all felt the training had been very beneficial.

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

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

Table 4 Sample 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 ask funny, but he's my him right." brother and I guess we "I don't get so mad at her have to put up with when I ask her right him." and she does okay."

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

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

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

"Sometimes he acts funny "He gets along better if I around my friends, but know how to ask him." I don't always get mad because he's my brother."

"She causes problems a lot, but I guess we have to put up with her."

"I know she's part of the family, but she's hard to get along with."

that with such increased control there could also be a potential for problems, such as the siblings misusing them or "exploiting" their brothers and sisters. We did not observe any of this type of problem in this study, but we recognize that it may arise. Thus, we feel that careful monitor- ing, especially by trained parents, may be an important component of this approach. This and other specific questions concerning the relative roles of trained and untrained parents in con- junction with trained versus untrained siblings present a number of interesting questions for future research.

In general, the positive possibilities of these procedures seem quite broad. The data suggest that siblings may be useful adjuncts to trained parents in teaching and maintaining behaviors in 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 providing continuity between the home and school for various educational programs, thus alleviating possible contrast problems and promoting gen- eralization (cf. Koegel, Egel, & Williams, 1980). Siblings also are in an ideal position to serve as facilitation agents for the children's appropriate social interactions with other children in the neighborhood, thus increasing the potential for a more normalized development for the autistic child; and conversely, helping to prevent the development of maladaptive interactions, lead- ing to progressively more and more abnormal situations as the autistic child matures (cf. Pat- terson, 1975; Patterson & Reid, 1970).

In relation to this latter point, recent research by Wellen (Note 1) and Wellen and Broen, 1982) has suggested that in some situations older siblings 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 for siblings may alter such potentially problematic situations by teaching the siblings how to promote and facilitate appropriate language production by their handicapped brothers and sisters.

All these types of sibling participation will perhaps 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 make this seem like an especially important area for further research and teaching. The results of the present study provide encouraging evidence for the continuing development of intervention strategies that involve the family members of autistic and other severely handicapped children as primary sources of treatment delivery.

REFERENCE NOTE 1. 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 preschool children 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 modification in the home: Siblings as behavior modifiers, par- ents as observers. Journal of Abnormal Child Psychology, 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 autistic children: Implications of current research. In E. R. Ritvo (Ed.), Autism: Diagnosis, current re- search, and management. New York: Spectrum Publications, 1976.

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

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

Koegel, R. L., Glahn, T. J., & Nieminen, G. S. Gen- eralization of parent training results. Journal of Applied 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 and understanding autistic children. Houston: College- Hill Press, 1982.

Koegel, R. L., Schreibman, L., Johnson, J., O'Neill, R. E., & Dunlap, G. Collateral effects of parent training on families with autistic children. In R. F. Dangel & R. A. Polster (Eds.), Behavioral parent 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 Journal of 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 and architect of a coercive system. In E. J. Mash, L. A. Hammerlynck, & L. C. Handy (Eds.), Behavior modification and families. New York: Brunner/ Mazel, 1975.

Patterson, G. R., & Reid, J. B. Reciprocity and coercion: 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. Effects of 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. Effects of peer-mediated social initiations and prompt- ing/reinforcement procedures on the social be- havior of autistic children. Journal of Autism and Development Disorders, 1979, 9, 41-54.

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

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

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

Wolf, M. M. Social validity: The case for subjective measurement or how applied behavior analysis is finding its heart. Journal of Applied Behavior Analysis, 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 of an autistic child: A follow-up and extension. Be- haviour Research and Therapy, 1967, 5, 103-111.

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

Received April 12, 1982 Final acceptance December 2, 1982