1-s2.0-002244059290002M-main

0022-44n5/92/165 “0 + “0

0 19’12 TheJournal id School Pyrholngy, Inc

Teaching Socially Valid Social Interaction Responses to Students With Severe Disabilities

in an Integrated School Setting

Edward G. Nientimp and Christine 1. Cole Lehigh University

To evaluate the effects of a procedure designed to teach appropriate social responses to

adolescents with severe disabilities in an integrated school setting this study employed

an ABA withdrawal design, replicated twice with two students, and an AB design with

a third student. Social responses were recorded during daily training sessions and

generalization probes. Treatment involved implementation, by the classroom teacher,

of a constant time delay procedure utilizing nonhandicapped peers and a socially

validated teaching curriculum. The results showed increases in correct responding and

decreases in echolalia following intervention. Generalization of appropriate responding

to nonhandicapped peers was demonstrated to varying degrees for all participants.

Students with severe disabilities such as autism are currently being integrated into educational environments with nonhandicapped peers. In an effort to maximize their potential for successful integration, there is an increasing emphasis on teaching these students appropriate social interaction behaviors. For example, transient social greetings such as “Hi,” “How are you doing?,” or “What’s up?” may promote nonhandicapped peers’ perceptions of students with severe disabilities as sociable rather than socially isolated students (Gay- lord-Ross & Pitts-Conway, 1984). In addition, if efforts to teach appropriate social interaction behaviors are to be maximized, it is critical that the target responses selected be socially valid (Kazdin, 1977) and representative of those used by same-age peers in the natural environment (Nutter & Reid, 1978).

Inasmuch as many individuals with severe disabilities have limited language repertoires and unusual speech patterns, there are inherent problems in teaching functional and meaningful social responses. Problems such as echolalia or a failure to respond when greeted by a peer may extinguish any naturally occurring social interactions initiated by nonhandicapped peers. Additionally, traditional teaching methods designed to strengthen correct responding and reduce or eliminate echolalic responses have been largely unsuccessful in facili- tating generalization of newly acquired responses (e.g., Charlop, Schreibman, & Thibodeau, 1985; Schreibman & Carr, 1978).

Received October 18, 1990; final revision received May 4, 1992.

Address correspondence and reprint requests to Christine L. Cole, School Psychology Pro-

gram, 111 Research Drive, Mountaintop Campus-Bldg A, Lehigh University, Bethlehem, PA

18015-4792.

343

344 Journal of School Psychology

One promising procedure for teaching spontaneous verbalizations and facil-

itating generalization is the time delay technique (Halle, Baer, & Spradlin,

1981; Handen & Zane, 1987; Snell & Gast, 1981). Time delay is a teaching

procedure used to transfer control from a prompt-the controlling discrimina-

tive stimulus- to a more natural stimulus designed to evoke the target behav-

ior (Sulzer-Azaroff & Mayer, 199 1). The procedure involves inserting a period

of time (constant or graduated) between the more natural stimulus and the

prompt until the prompt can be discontinued altogether. Constant time delay

involves holding the interval between the stimulus and the prompt constant

throughout training, whereas graduated, or progressive, time delay involves grad-

ually increasing this interval during training. Time delay has been used suc-

cessfully to teach students with a range of disabilities a variety of behaviors,

including spontaneous language and verbalizations of affection in children

with autism (Charlop et al., 1985; Charlop & Walsh, 1986; Matson, Sevin,

Fridley, & Love, 1990) sight words in preschoolers and elementary students

with developmental disabilities (Doyle, Wolery, Gast, Ault, & Wiley, 1990;

Gast, Ault, Wolery, Doyle, & Belanger, 1988), complex visual discriminations

in children with mild and moderate intellectual handicaps (Smeets, Sriefel, &

Hoogeveen, 1990), check cashing and use of an automatic teller in students

with moderate disabilities (McDonnell & Ferguson, 1989), and use of a vend-

ing machine in children with moderate mental retardation (Browder, Snell, &

Wildonger, 1988).

The present concern was to attempt to promote social interaction with

nonhandicapped peers by students with severe disabilities in a self-contained

special education classroom recently moved from an approved private school

into a local middle school. Although the usefulness of time delay procedures

has been demonstrated with a variety of behaviors and settings, no studies

were found that evaluated time delay in teaching social interaction responses in

an integrated school setting. Thus, the purpose of the present study was to

evaluate the use of constant time delay by the classroom teacher in teaching

students with severe disabilities socially valid social interaction responses to

trainer-initiated and nonhandicapped-peer-initiated social greetings.

METHOD

Participants and Setting

Two boys, aged 12-O and 13-4, and one girl, aged 12-8, participated in the

study (see Table 1). All had a psychiatric diagnosis of autism or pervasive

developmental disorder and a history of aggressive, self-injurious, or autistic-

like (ritualistic and/or self-stimulatory) behavior that had necessitated inten-

sive educational programming. All three students were verbal, although ver-

balizations were often prompt-dependent and/or echolalic.

Student records indicated that all participants had attended “special” schools

Nientimp and Cole 345

Table 1 Subject Characteristics

Psychiatric Participant Age Sex IQ” Instrument diagnosi?

1 13-4 M 38 Stanford-Binet Early infantile autism

2 12-8 F Autism 3 12-O M 32 Stanford-Bmet Pervasive

developmental disorder

“IQscore for Participant 2 was not available. ‘Provided by psychiatrists using clinical interviews.

until the present year, when they began receiving services in a self-contained

special education classroom at an integrated middle school in eastern Pennsyl-

vania. Middle school enrollment was made up of approximately 700 sixth-

through eighth-grade students, a majority of whom were nonhandicapped. All

training for the study was conducted in the self-contained classroom; three

other students with similar handicaps were present in the room. Two certified

special education teachers served as primary trainers and data collectors.

Social Validation Procedures

Social Validation 1 (SVl). During SVl, a sample of nonhandicapped same-

age peers (20% of the student body) was surveyed to identify the most com-

monly used transient social greetings. The survey was distributed voluntarily

by homeroom teachers. Students were asked to list what they most commonly

say to classmates when they pass each other in the hallway. A total of 197

completed surveys were returned. The five most frequently reported greetings

are presented in Table 2.

Social Validation 2 (SV2). During SV2, nonhandicapped peers were again

surveyed to determine appropriate responses to the most commonly used tran-

sient social greetings previously identified. The SV2 survey listed the five most

commonly used greetings identified in SVl. Students were asked to fill in what

Table 2 Five Most Common Greetings and Associated Responses

Identified by Nonhandicapped Peers

Greetings

Hi What’s up H&l Yo Hey

Hello Not much Hi What What


they would most often say in response to each particular greeting from a peer.

A total of 144 completed surveys were returned. The five most frequently

reported responses are presented in Table 2.

Dependent Variables

The dependent variables were the participants’ responses to social greetings,

which were categorized as follows:

1. Correct responses were verbal responses given within 5 seconds of the peer/

teacher-initiated greeting that had been identified by nonhandicapped peers as

the most common response to that particular greeting (Table 2). For example,

the correct response to the greeting “Hi” was “Hello.” whereas a correct re-

sponse to “Yo” was “What .”

2. Error responses were defined as no verbal response within 5 seconds of the

peer/teacher-initiated greeting or any verbal response other than the one iden-

tified by nonhandicapped peers as the most common response to that particu-

lar greeting (excluding echolalic responses).

3. Echolalic responses were verbal responses in which all or part of the peer/

teacher-initiated greeting was repeated.

4. Prompted correct responses were correct verbal responses, as defined above,

given after the delay prompt (recorded during intervention only).

Observational Procedures

Data were collected during daily training sessions in which the first author

initiated socially \,alid greetings and recorded students’ responses as correct,

error, echolalic, or prompted correct as defined above. Instructional trials

involved (a) having all three participants present and (b) having the trainer

greet students one at a time in random order. Each student was presented with

five different greetings, each ofwhich was assessed twice. Thus, every training

session provided each participant 10 opportunities to respond. Daily training

sessions lasted approximately 10 minutes.

Data were also recorded in a similar manner by the first author following

generalization probes in which live same-age nonhandicapped peers were

trained to serve as initiators. Generalization probes were conducted during

each phase of the study for a total of five probes. Nonhandicapped peers were

selected for participation as a result of their expressed interest in the special

education program, which was new to the school.

Interobserver Agreement

Interobserver agreement checks were conducted by the second classroom

teacher at least once during each phase of the study for approximately 21% of

the daily observation periods. The second observer was accustomed to the

format of data collection and instructional method. Procedures were described


by the first author and written on a systematic instructional plan. Since agreement was consistently at or near 100%) no further training was required. Percentage agreement between the two observers was defined as number of agreements divided by the total number of agreements plus number of dis- agreements multiplied by 100. The level of interobserver agreement averaged 97% (range 90%-100%) during all baseline and intervention phases.

Experimental Design

A combined treatment withdrawal and replication across participants procedure was used to evaluate treatment effects. Owing to time constraints (i.e., end of the school year), the final treatment withdrawal phase was not implemented for Participant 3.

Experimental Conditions

Baseline 1. During Baseline 1, the participants’ responses to the live socially valid greetings were assessed. The trainer initiated each greeting twice to each student. Thus, there was a total of 10 discrete trials in which each participant had an opportunity to respond. The order in which the greetings were presented was randomized and participants were not provided the same greeting on two consecutive trials. Following each greeting, the trainer coded the student’s response. Correct responses were praised; no prompting or correction was provided for incorrect responses.

Prior to the generalization probes, the first author met briefly with each nonhandicapped peer and explained the procedure. For each generalization probe, peer trainers stopped at the participants’ classroom on their way to lunch when participants typically were engaged in free-time leisure activities. As the nonhandicapped peer entered the room, he or she selected an index card that was pinned to a bulletin board near the door. Each card listed a student’s name, the greeting to be given, and the response to be praised. Nonhandicapped peers were told to deliver the specified greeting, “give the student a chance to respond,” and praise the correct response. As with the teacher trainer, no prompting or correction was provided for incorrect responses. Following each greeting, the teacher coded the participant’s response. Each nonhandicapped peer initiated one greeting to one participant per day. Generalization data were reported for a participant as soon as all live target responses had been probed.

Constant Time Delay. At the beginning of each training session, participants were provided two warm-up trials for each of the five target responses. The trainer presented greetings in the same manner as during Baseline 1, but immediately prompted the correct response. If the participant responded correctly, labeled verbal praise was provided (e.g., “That’s good, saying hello!“). If the student engaged in an error or echolalic response or provided no response


for 5 seconds, the trainer responded by saying “No, you should say (specific

mjmzse)” and averting his gaze for 3-5 seconds. No data were recorded during warm-up trials and these trials were conducted only by the teachers.

Following the warm-up trials, a constant delay of 5 seconds was implemented for the remainder of the training session. The trainer initiated a social greeting to a student and paused for 5 seconds. If the participant responded correctly within 5 seconds, he/she received labeled verbal praise. If an error or echolalic response was given, the trainer firmly stated “No” and provided the correct response. If the student did not respond for 5 seconds, the trainer presented the correct response (e.g., “Hello”). If the participant then correctly repeated the response, the trainer provided labeled verbal praise and scored the response as prompted correct. If the participant engaged in an error or echolalic response following the model prompt, a correction procedure identical to that used during the warm-up trials was implemented. Each student was provided 10 discrete trials in which to respond to the five different greetings.

Generalization probes were conducted by nonhandicapped peers in the same manner as during Baseline 1.

Baseline 2. Conditions during this phase were identical to those in Baseline 1.

RESULTS

Participant 1

As shown in Figure 1, Participant 1 averaged 5% independent correct responses and 85% echolalic responses per session during Baseline 1. During Constant Time Delay, correct responses increased to an average of 70% (100% correct during the last 3 days of intervention). Echolalic responses decreased to an average of 2 1% during this phase. On the first day of treatment withdrawal (Baseline 2), correct responding dropped to 60% and echolalic responses increased to 40%. However, correct responding rebounded to 100% correct during the final 3 days of the study.

During generalization probes, percentage of correct responses by Partici- pant 1 increased steadily throughout all phases from 20 % on Day 3 and Day 9, to 60% on Days 17 and 19, to 100% on Day 28.

Participant 2

During Baseline 1, Participant 2 averaged 8 % independent correct responses per session (Fig. 2). Other responses during this phase included echolalic (average 22%) and error (average 70%) responses. On Day 11 during Con- stant Time Delay, Participant 2 reached 100% independent correct responding and echolalia decreased to 0% for the remainder of the study (with the excep- tion of 10% on Day 13).

Generalization probes for Participant 2 showed an increase from 0 % correct

Boseline 1 Conatmt Time Delay Baseline 2

100

80

; E 8 60

z -z :: 40 u

E

20

0 100

80

f

b 60

”

B x 40 E Ii

20

0 100

80

60

4

r” y 40

20

0 100

80

60

1

40

20

0

Figure 1. Daily percentage of independent correct, prompted correct, echolalic, and error responses for Participant 1 during Baseline 1, Constant Time Delay, and Baseline 2 conditions. Asterisks indicate percentage independent correct responses during nonhandicapped-peer-initiated generalization probes.

349

Figure 2. Daily percentage of independent correct, prompted correct, echolalic, and errm responses for Participant 2 during Baseline 1, Constant Time Delay, and Baseline 2 conditions. Asterisks indicate percentage independent correct responses during nonhandicapped-peer-initiated generalization probes.

350


responses on Days 3 and 9 to 40% on Days 14 and 17 (intervention). There was a subsequent decrease to 20% during Baseline 2.

Participant 3

As noted in Figure 3, Participant 3 made no correct responses during Baseline 1 (97% echolalic and 3% error responses). However, during Constant Time Delay, correct responding increased to an overall average of 39%) while echolalic responses decreased to an average of 17 % Other responses were prompted corrects (average 17 %) and errors (average 25 %). Participant 3 scored at or above 50% correct responding during the last 7 days of intervention, with echolalia consistently at zero during this period.

For all initial generalization probes for Participant 3 correct responding stood at 0% but it increased during Constant Time Delay to 20% on Day 19 and 60% on Day 22.

DISCUSSION

The results of this study showed that independent appropriate social interaction responses increased and echolalic responding decreased for all participants. As the behavior of Participants 1 and 2 did not return to baseline levels after withdrawal of treatment, it is difficult to state with confidence that these results were due solely to the time delay strategy. At least three alternative explanations should also be considered: (a) Peer modeling may have been responsible for initial as well as maintenance effects; (b) the continued use of praise for correct responses throughout Baseline 2 conditions may have been a factor in the maintenance of correct responding; and/or (c) some other influ- ential event may have coincided with the introduction of time delay. Future research should address these alternative explanations by attempting to control for the effect of observing peers and by using a design better suited to skills acquisition research (e.g., multiple baseline design). However, in this case the low initial levels of correct responding, the fairly rapid and dramatic increases in independent correct responses during intervention, and the fact that these results were replicated across three similar participants lends support to the notion that the treatment was in fact responsible for the observed changes. Additionally, from an educational perspective, students’ continued positive responding following treatment withdrawal suggests the constant time delay procedure may also facilitate maintenance of correct response.

Correct responses to teacher-initiated greetings were also generalized to greetings from nonhandicapped peers for Participant 1. However, Participant 2, who acquired the target responses in the shortest amount of time during discrete trial training, failed to generalize this behavior to nonhandicapped peers at a level comparable with that of her discrete trial performance. Partici- pant 3 showed generalization effects only during the final two generalization

Boseline Cocstont Time Dcl~y

100

OL X Xkd

60

40

20

0

Figure 3. Daily percentage of independent correct, prompted correct, echolalic, and error responses for Participant 3 during Baseline 1 and Constant Time Delay conditions. Asterisks indicate percentage independent correct responses during nonhandicapped-peer-initiated generalization probes.

352


probes. Thus, although one student readily generalized the newly acquired responses, the others did not. One hypothesis is that specific characteristics of the individual subjects may have contributed to some degree to the variability in responding. Participant 1, for example, was observed to be much more friendly and relaxed (e.g., smiling, approaching, verbalizing) when peer tutors entered the room than were Participants 2 and 3, who often averted gaze. This variability across students suggests that training procedures (e.g., training in vivo versus in simulation) may need to be individualized.

Although Participant 3 did not learn the target responses as readily as his peers, his data provide some interesting results. Figure 3 illustrates that, as independent correct responses increased, echolalic responses decreased but error rate increased. This suggests that although this student learned that echolalic responses were not correct, he replaced them with nonecholalic guesses that resulted in increased errors.

A unique aspect of this study was the use of a social validation procedure to determine the target greeting responses to be taught. It was noted at the onset of this project that few, if any, spontaneous nonhandicapped peer-initiated greetings were directed toward the participants. It was hoped that by using socially validated responses, involving nonhandicapped peer tutors in the training, and increasing participants’ competence, more spontaneous interactions would occur. Although no formal data were collected, informal observa- tions indicate this did not occur. This suggests that although the so&l signify-

cunce of the target behaviors was established, this did not automatically produce behavioral gains that were socially important to the person, in that ongoing peer relationships may not have been influenced. Thus, it may be more beneficial to focus on actively training students with handicaps to inittate and sustain social interactions with nonhandicapped peers (Gaylord-Ross & Pitts-Conway, 1984).

In summary, constant time delay has been used successfully in several pre- vious studies to teach a variety of communication as well as other responses. The current study extended this method to teach socially valid social interaction responses to students with severe disabilities in an integrated school setting. Although alternative explanations for the results were not ruled out, the results are dramatic and socially significant enough to warrant further investi- gation and consideration for use by classroom teachers of students with severe disabilities.

REFERENCES

Browder, D. M., Snell, M. E., & Wildonger, B. A. (1988). Simulation and community- based instruction of vending machines with time-delay. Education and lhining in Mental Retardation, 23, 175-185.

Charlop, M. M., Schreibman, L., & Thibodeau, M. G. (1985). Increasing spontaneous verbal responding in autistic children using a time delay procedure. Journal of Applied Behauior An&is, 18, 155-166.


Charlop, M. H., & Walsh, M. E. (1986). I ncreasing autistic children’s spontaneous verbalizations of affection: An assessment of time delay and peer modeling procedures. Journal ojApp&d Behavior Analysis, 19, 307-314.

Doyle, P. M., Wolery, M., Gast, D. L., Ault, M. J., & Wiley, K. (1990). Comparison of constant time delay and the system of least prompts in teaching preschoolers with developmental delays. Research in Developmental Disabilities, 1 I, l-2 2.

Gast, D. L., Ault, M. J., Wolery, M., Doyle, P. M., & Belanger, S. (1988). Compari- son of constant time delay and the system of least prompts in teaching sight word reading to students with moderate retardation. Education and ?;aininE in Mental Retar- &ion, 23, 117-128.

Gaylord-Ross, R. J., & Pitts-Conway, V. (1984). Social behavior development in integrated secondary autistic programs. In N. Certo, N. Haring, & R. York (Eds.), Public school intepation of severely handicapped students: Rational issues and progressive alterna- tives (pp. 197-219). Baltimore: Paul H. Brookes.

Halle, J. W., Baer, B. M., & Spradlin, J. E. (1981). Teacher’s generalized use of delay as a stimulus control procedure to increase language use in handicapped children. Journal of Applied Behavior Analysis, 14, 389-409.

Handen, B. L., & Zane, T. (1987). Delayed prompting: A review of procedural variations and results. Research in Developmental Disabilities, 8, 307-330.

Kazdin, A. (1977). Assessing the clinical or applied importance of behavior change through social validation. Behavior Modtjiiation, 1, 427-452.

Matson, J. L., Sevin, J. A., Fridley, D., & Love, S. R. (1990). Increasing spontaneous language in three autistic children. Journal ofApplied Behavior Analysis, 23, 227-233.

McDonnell, J., & Ferguson, B. (1989). A comparison of time delay and decreasing prompt hierarchy strategies in teaching banking skills to students with moderate handicaps. Journal ofApplied Behavior Analysis, 22, 85-91.

Nutter, D., & Reid, D. H. (1978). Teaching retarded women a clothing selection skill using community norms. Journal ofApplied Behavior Analysis, 11, 475-487.

Schreibman, L., & Carr, E. G. (1978). Elimination of echolalic responding to ques- tions through the training of a generalized verbal response. Journal ofApplied Behavior Analysis, 11, 453-463.

Smeets, P. M., Scriefel, S., & Hoogeveen, F. R. (1990). Time-delay discrimination training: Replication with different stimuli and different populations. Research in Developmental Disabilities, 11, 2 17-240.

Snell, M. E., & Gast, D. L. (1981). Applying the delay procedure to the instruction of the severely handicapped. Journal ofthe Assoctationfor the Severely Handicapped, 6, 3-14.

Sulzer-Azaroff, B., & Mayer, G. R. (1991). B ha e vtor analysis&r lasting change. Fort Worth, TX: Holt, Rinehart and Winston.