Nina Kurfman SPED 461 IP- part 2 December 2, 2013 Student Description: The student I chose to target for

this assignment, J.B., is seven years old, a female, and as stated on her individual education plan (IEP) has a primary diagnosis of other health impairment and a secondary diagnosis of visually impaired including blindness (VIIB). J.B. is a second grader, but receives most of her instruction in the K-1 cross-categorical classroom. From the practicum placement in this classroom last spring, I know that a team decision was made to progress J.B. to the next grade level but keep in the K-1 classroom. This decision was based on the teachers knowledge and experience with her curriculum interventions and the foreseen low numbers of the K-1 classroom to allow for more 1:1 attention and personalized instruction. As stated, most of her instruction is in the self-contained cross-categorical room, but she is included with her same aged peers during calendar, lunch, and recess. J.B.s curriculum continues to be fostered through the PS-Fall reading system (Project Symbol-Functional Adapted Literacy and Language program) as it did the pervious year. J.B. is able to identify/discriminate 166 words/pictures, which is more than double of where she was last year at this time, while also having the ability to track symbols on two lines independently and answer simple comprehension questions from the text, but has yet to generalize skills to written expression (H. Edwards, personal communication, September 20, 2013). J.B. has proven the ability to identify and discriminate all upper and lower case letters of the alphabet (H. Edwards, personal communication, September 20, 2013). She also receives instruction through the handwriting without tears program to improve motor and handwriting skills (H. Edwards, personal communication, September 20, 2013). Her math skills are emerging, working on number, coin, and time identification (H. Edwards, personal communication, September 20, 2013). In February of 2012, J.B. attended the low vision clinic where she was tested using pictures and it was determined that her acuity to be 4/50 at a distance (H. Edwards, personal communication, September 20, 2013). She has the ability to identify very small objects at a near distance and discriminate visual like symbols, but her left and right sidedness is still inconsistent (H. Edwards, personal communication, September 20, 2013). This past summer the student has begun toilet training, but is still a concern for staff for she has yet to void at school (H. Edwards, personal communication, September 20, 2013). H. Edwards stated J.B. uses her Vantage communication device throughout the day, and believes this use has helped to facilitate more verbalizations the past year and continues to help her verbal communication skills increase, however the minimum verbal skills she does posses are intelligible (personal communication, September 20, 2013). J.B. uses gestures and signs to communicate, but would be unfamiliar to those not familiar with the student.

Student: JB Skill: Requesting a preferred item Initiator: Classroom Teacher Context for Instruction: Instruction of this skill will occur throughout the morning and afternoon routine during small work groups in the natural setting of JBs K-1st crosscategorical classroom. During this time, there are two adults in the classroom (one special education teacher and one para professional) and four students who spend the majority of the day together in the cross cat classroom. Instruction and assessment will be implemented during class time when all students are involved in small work groups, working on independent skills on differentiated skill levels. Skill Sequence: (Class Notes, 2013)

0.&123*(45&.&678,)#$,")5&9&(*+"$,8&

0.&123*(45&.&678,)#$,")5&.&(*+"$,&

0.&123*(45&.&(*+"$,&

!"#$%&'()&(*+"$,&-.&(*+"$,/&

Program Objective: When presented with one symbol, one distractor, two objects, and the prompt What do you want?, JB will independently point to the symbol from the communication broad (one symbol and one distracter) corresponding to the preferred object to communicate her request of the object. Generalization: To promote skill generalization, the instructor will be rotated among the different professionals working directly with the student and will occur across different environments. Rational: Having the ability to request preferred objects is an essential and functional skill within communication (Sigafoos et al., 1996). In addition, this skill gives way to valuable types of reinforcement and provides one with an aspect of control over their environment (Sigafoos et al., 1996). Assessment Procedures: 1. To begin each assessment opportunity, the instructor will hold a preferred object (Ipad) in view of the student. At this time as the item is being offered to the student, the instructor will say, What do you want? and wait ten seconds and record the students response. 2. If the student correctly responds, by pointing the symbol of the preferred item, circle the + on the data sheet of the corresponding opportunity. Wait five seconds before initiating the next opportunity. 3. If the student incorrectly responds, by pointing the distracter, not providing a response within ten seconds, or reaching for the item, circle the on the data sheet of the corresponding opportunity. Wait five seconds before initiating the next opportunity. 4. Repeat until the instructor has provided ten consecutive opportunities for the student to request the item. 5. Calculate the percent of correct responses and graph on skill graph.

Assessment Schedule: During baseline, assessment will occur on each scheduled or naturally occurring opportunity, unit performance is stable. Once instruction begins, assess every third trial until mastery criterion is met. Instructional Procedures: Instruction will occur on every scheduled or naturally occurring opportunity, except on probe trails (every third turn). 1. Instruction will begin by displaying one symbol and one object to the student. Opportunities to request will be initiated by the instructor displaying the preferred item and providing the prompt, What do you want? 2. If the student points to the correct symbol, within ten seconds, present the student with the object. 3. If the student responds incorrectly, by reaching for the object or not responding within ten seconds, physically guide the students finger to the symbol and repeat the verbal prompt. This prompt does not lead to access of the object. 4. Wait five seconds and initiate the next opportunity. 5. If the student provides a correct response, present the student with the object. 6. If the student gives an incorrect response say, No and place the students hand over the incorrect response. Then repeat the verbal prompt while modeling the correct response of placing the students finger on the correct symbol. 7. Provide praise and present the student with the object. 8. Wait five seconds before initiating he next opportunity. 9. When each step of the skill sequence has met mastery, add the next component of the sequence. (Sigafoos et al., 1996) Reinforcement: Specific, direct verbal praise will be provided upon each correct response, in addition to the natural reinforcement of being presented with the object requested. The natural reinforcement will continue, but the verbal praise will be faded accordingly: -every other correct response -every third correct response

-every fourth correct response -no specific verbal praise Maintenance: Follow-up probes will be administered every two weeks. The same procedures used during instruction will be utilized during maintenance. Maintenance will also be facilitated through fading of verbal reinforcement. Research Component: Sigafoos, J., Couzens, D., Roberts, D., Phillips, C., & Goodison, K. (1996). Teaching requests for food and drink to children with multiple disabilities in a graphic communication mode. Journal of Developmental and Physical Disabilities. 3, 247-262. A study done by Sigafoos, Couzens, Roberts, Phillips, and Goodison (1996) demonstrated the success of using time-delay, prompting, reinforcement, and error correction strategies to teach requesting skills to individuals with multiple disabilities. The participants of this study were two young girls with multiple disabilities, including visual impairment. At the time of the study, neither of the two girls used any form of communication to make requests. Before beginning intervention, a preference assessment was conducted to identify preferred food and drinks items for each participant. Once baseline data was collect, intervention began. Intervention stated by selecting a single item and offering ten successive opportunities for the student to request the single item. The instructor presenting the item and saying, If you want this, point to initiated these communicative opportunities. If the participant pointed to the corresponding symbol, the item was given to the participant. If the participant gave an incorrect response or failed to respond within ten seconds, the instructor introduced a physically prompt by guiding the participants finger to the correct symbol while restating the verbal prompt. After five seconds had passes, the instructor initiated another opportunity. If this opportunity were followed by an incorrect response, the instructor would say, No while simultaneously placing the participants hand of the correct symbol. With the instructor still providing physically prompting, the instructor would present the item out of reach and say, If you want this, point to while modeling the correct response. If the participant then responded correctly, praise was given and the item to the participant and the next opportunity began five seconds later. & 17:#'((85&;<5&=>!"74425&?<5&1""42@A()B5&1<5&C&D6)787E%#5&F<&&-9GGH/<&&I"#$%7E:&8,J6"E,8& & K7,%&6"L"4(M3"E,#4&678#*747,7"8&,(&4($#,"&,%"7)&NNF&6"L7$"<&&!"#"$%&'()*( ( +","-./0"*1$-(+)#$2)-)1)"#3(455&OP.@OQO<& In study conducted by Sigafoos, OReilly, Seely-York, and Edrisinha (2003) participants were instructed to use a VOCA device to communicate the request of I want more.

After a preference assessment was completed for each participant, preferred items were identified. Instruction began by placing a tray of a preferred item in front of the participant, as well as their VOCA device. If the participant reached for the tray or touched a blank symbol, this was an incorrect response, and the instructor would use the least amount of physical guidance necessary to get a correct response. The participant was given access to the item after each correct response, prompted or not. Over instructional opportunities, physical guidance as delayed by five seconds to encourage independence. 1 symbol, 1 object 1 + (Ipad) 2 + (Ipad) 3 + (Ipad) 4 + (Ipad) 5 + (Ipad) 6 + (Ipad) 7 + (Ipad) 8 + (Ipad) 9 + (Ipad) 10 + (Ipad) Notes: - (Dist) - (Dist) - (Dist) - (Dist) - (Dist) - (Dist) - (Dist) - (Dist) - (Dist) - (Dist)

.GG& TG& QG& !"#$"%&'()##"$&' PG& SG& RG& HG& OG& 9G& .G& G& .& O& R& P& T& ..& .O& .R& .P& .T& 9.& 9O& 9R& 9P& 9T& *"++,)%+' U#8"47E"& N7347E"9&

Materials: I would have liked to have created a program around the use of the students Vantage device, to help change the student perceptions of the device as work, but was unable to work it out with the classroom staff. Instead of the students Vantage device, this low- tech communication board was utilized. I felt the student would be able to successfully learn to use this communication with her ability to use her other device. Also, with it being a new device, the student may not have negative feelings associated with it and be more open to using it. Even though the students verbal skills are increasing and communicates through sign and gestures, her communication skills would not recognized by those who are not familiar to the student.

Journal of Developmentaland PhysicalDisabilities, Vol. 8, No. 3, 1996

Teaching Requests for Food and Drink to Children with Multiple Disabilities in a Graphic Communication Mode
Jeff Sigafoos, 1 D o n n a Couzens, 1 Donna Roberts, l Cathie Phillips, 2 and Kathy Goodison 2

Graphic communication modes are often considered for children with multiple disabilities as an alternative to speech or gestures. However, there have been few empirical studies demonstrating effective procedures for teaching graphic mode requests to young children with multiple disabilities. In the present study, two children with multiple disabilities were taught to request preferred food and drink items by pointing to corresponding line drawings. Time-delay, prompting, reinforcement, and error correction procedures were implemented in a multiple-probe design to establish discriminated and generalized requests for preferred foods and beverages during morning snack time. The procedures were effective in teaching generalized and discriminated use of the food and drink symbols. Although effective in teaching children to request offered items, the procedures used in the present study need further development to ensure requests are controlled by the more natural conditions of hunger and thirst.
KEY WORDS: requesting; augmentative communication; multiple disabilities.

INTRODUCTION

Requesting preferred objects is an important and functional communication skill. It provides a means to access effective types of reinforcement and to exert some degree of control over the environment (Guess et al., 1974; Reichle et al., 1991). Requests for preferred objects typically emerge 1Fred and Eleanor Schonell Special Educational Research Centre, The University of Queensland, Qld 4072, Australia. 2Xavier Special School, Brisbane, Australia.
247 1056-263X/96/0900-0247509.50/0 9 1996 Plenum

Publishing Corporation

248

Sigafoos ~ a L

early in the developing language of most children and usually without any deliberate or systematic instruction (Bates, 1976). For children with developmental and physical disabilities, however, the natural environment is often not sufficient to ensure the development of functional requesting skills. Instead, deliberate and systematic instruction may be required to establish requests for preferred objects. Systematic instructional procedures for teaching requesting typically combine the use of time-delay, prompting, differential reinforcement, and error correction strategies into a discrete and massed-trial instructional format (Cipani, 1988). For example, a learner might be offered a preferred food item, prompted to make the corresponding request (i.e., "food") and then provided with a small amount of the offered item. Prompts might include the use of verbal, gesture, modelling, or physical assistance. Over successive opportunities, this assistance can be faded using a time-delay procedure until the learner spontaneously requests offered items in the absence of any verbal, gesture, model, or physical prompt (Halle, 1987; Sigafoos and Reichle, 1993). Several studies have demonstrated the effectiveness of these instructional procedures for teaching persons with moderate to profound mental retardation to request preferred objects in both vocal (i.e., speech) and gesture (e.g., manual signs) modes (Duker and van Lent, 1991; Gobbi et al., 1986; McCook et aL, 1988; Sigafoos, 1995). Physical limitations, however, often preclude the use of speech or gestures for requesting among some children with multiple disabilities. In such cases, a graphic mode of communication may represent a viable alternative (Goetz and Hunt, 1994). A graphic mode of communication can begin by teaching the learner to select graphic symbols (e.g., line drawings) to request preferred objects (Reichle et al., 1991). However, there have been relatively few empirical studies demonstrating effective procedures for teaching young children with developmental and physical disabilities to request preferred objects using graphic mode communication (e,g., Sigafoos et aL, 1995; Turnell and Carter, 1994). After teaching graphic mode requests, it may be necessary to implement procedures to bring the selection of symbols under appropriate stimulus control (Reichle et aL, 1991). Developing appropriate stimulus control is often difficult, especially as the number of symbols being taught increases. With two symbols, for example, the learner's response of pointing to a symbol must be controlled not only by the availability of an item to request (e.g., a food item being offered by the teacher), but also by the particular graphics which comprise the corresponding "food" symbol, as opposed to the distinct graphics which might comprise an alternative "drink" symbol (Michael, 1985; Shafer, 1993; Sundberg, 1993). Appropriate stimulus control involves both discrimination and generalization (Albin and

Teaching Requests

249

Horner, 1988). A discrimination would be demonstrated when a child pointed consistently to the "food" symbol when offered a preferred food item, but instead pointed to the "drink" symbol when offered a preferred beverage item. Generalization would be evidenced when the child pointed to the "food" symbol to request a variety of preferred food items and pointed to the "drink" symbol to request a variety of preferred beverage items. Although programming for appropriate stimulus control typically occurs after teaching the response, more recently it has been recommended that appropriate stimulus control can be established as the request is being acquired (O'Neill, 1990). However, when teaching graphic mode requests to young children with multiple disabilities, there are few guidelines on how to program instruction so as to simultaneously establish appropriate stimulus control. Furthermore, difficulties in achieving this dual goal have been reported. In one relevant study, Sigafoos and Reichle (1992) reported difficulty in establishing appropriate stimulus control when teaching an initial repertoire of graphic mode requests to four adults with severe disabilities. In that study, time-delay, prompting, and differential reinforcement procedures were implemented to teach the concurrent use of both explicit (e.g., "apricot") and more generalized (e.g., "fruit") requests. Although the procedures were effective in increasing the percent of correct requests during intervention, each adult continued to display systematic error patterns (e.g., selecting the "apricot" symbol rather than the "fruit" symbol when offered a pear). Analysis of these error patterns suggested that some of the difficulties in obtaining appropriate stimulus control may have resulted from attempting to teach two different types of requests for items within the same general class of objects (e.g., explicit request for apricot, but generalized request for other fruit items). The present study was designed to address these potential difficulties in developing appropriate stimulus control when teaching requests for food and drink in the graphic mode. In the present study, two children with multiple disabilities were taught to request preferred items from two distinct stimulus classes (foods and beverages). In addition, requests for food and drink were taught separately and then intermixed, rather than using a concurrent instructional strategy from the beginning of intervention. These characteristics of instruction were designed to facilitate acquisition of the discrimination between the line drawings designated for food and drink items. In addition, although the children were taught initially to request only one food and beverage item, multiple foods and beverages were offered during a subsequent probe phase to facilitate acquisition of generalized requests for food and drink (Chadsey-Rusch and Halle, 1992; O'Neill, 1990). The study was intended to demonstrate effective procedures

250

Sigafoos etaL

and sequences for teaching generalized and discriminated graphic mode requesting to young children with multiple disabilities that would, in turn, go some way towards the development of effective instructional guidelines.

METHOD Participants
Alison, aged 5, and Emily, aged 6, participated. Both girls attended a school for children with multiple disabilities. A team of professionals at the school nominated Alison and Emily for participation because neither spoke to request nor exhibited any alternative means of requesting food and drink. In addition, occupational and physical therapists from the school reported that both girls had sufficient visual acuity and physical dexterity to discriminate and point to the line drawings for food and drink that were used in the present study. According to records maintained by the school, both girls were diagnosed with cerebral palsy and significant intellectual delay. No formal assessment of intellectual functioning had been conducted, but adaptive behavior was assessed for this study by each child's respective classroom teacher using the TARC Assessment System (Sailor and Mix, 1975). This device was normed on a sample of 283 institutionalized children with severe disabilities from 3 to 16 years of age. It yields an overall standard score with a mean of 50 and a standard deviation of 20, as well as standard scores in several specific domains (e.g., self-care, social skills, receptive and expressive language). Alison received an overall standard score of 60 on the TARC, with standard scores of 68 and 52 for the receptive and expressive language domains, respectively. Emily received an overall standard score of 54, with scores of 52 in the receptive language domain and 26 in the expressive language domain. These assessment results indicate the both children were within one standard deviation of the mean when compared to the standardization sample of 283 severely handicapped children.

Setting
Alison and Emily attended separate but adjoining classrooms. Procedures associated with this study were implemented in a common area between these two classrooms during a morning snack time. The procedures were implemented by each child's respective classroom teacher. These two teachers worked individually with each child at two separate

Teaching Requests

251

tables. At the same time, approximately 8 to 11 children were situated at other tables in the common area for morning snack.

Materials

Reinforcers
Teachers identified food and drink items to which each child showed a preference. Preference was based initially on observation of the food and drink items the children had brought from home and consumed for morning snack in the past. Once a list of potentially preferred food and drink items had been generated, a reinforcer sampling procedure was implemented to validate these preferences. Specifically, small amounts of each food and drink item were offered one at a time. If the child selected and consumed the item within 10 sec of it being offered, that item was retained for use in subsequent phases of the study. Using these procedures, 10 food items and 10 drink items were identified and retained for each child. Examples of the food items that were identified and retained for Alison included cookies, cheese, and raisins. Examples of drink items for Alison included fruit-flavored cordial, chocolate milk, and water. Emily's food items included bananas, crackers, and potato chips and her drink items included fruit-flavored cordial, milk, and lemonade.

Communication Symbols
Black and white COMPIC symbols (Compic Development Association, 1986) were selected initially to represent food and drink items. Both symbols measured 10 7.5 cm and showed the outline of a person from the waist up. In the "food" symbol, this person was holding an apple to the mouth, whereas in the "drink" symbol, the person was holding a cup. The words "eat" and "drink" were printed in lower case letters above each respective line drawing. At a later point in the study, line drawings were colored to facilitate discrimination training (see Procedures).

Design and Procedural Overview

Procedures associated with this study were designed to teach Alison and Emily to request a variety of food and drink items by pointing to corresponding line drawings. Baseline observations were implemented initially

252

SigafoosetaL

to verify that neither child pointed to the correct symbol consistently when offered a variety of food and drink items. Intervention procedures were then implemented across food and drink items in a multiple-probe design (Homer and Baer, 1978). Sessions were implemented three to four times per week. During each session, the teacher provided 10 or 20 opportunities for the child to request food and/or drink items depending on the phase of the study. During sessions when opportunities to request both food and drink items were scheduled, the 10 food items were offered in succession and then after a short break (e.g., 2-3 min), the drink items were offered in 10 successive opportunities. However, at a later point in the study (see Concurrent Instruction), opportunities to request both food and drink were intermixed. When providing opportunities, the teachers sat directly across from the child with the communication symbols placed on the table approximately 15-20 cm in front of the child and centered approximately 10 cm to the left and right of midline. The left/right placement of symbols was altered randomly across opportunities. In all phases of the study, a correct request was recorded when the child pointed to the corresponding line drawing, and only that drawing, within 10 s of being offered the item. An incorrect request was recorded when the child pointed to the symbol for food when offered a drink item and vice versa. A "no response" occurred if the child did not point to a symbol within 10 s of being offered a food or drink item.

Procedures

Baseline
To initiate each baseline opportunity, the teacher held a plate containing a small amount of one of the food items or a cup containing a small amount of one of the drink items in view of the child. As the item was offered, the teacher said "If you want this, point to eat (drink)." During the ensuing 10 s, the teacher recorded if the child was correct, incorrect, or made no response. At the end of the 10 s interval, the child was given the item that had been offered, regardless of whether the correct symbol had been selected. After the child had consumed the item, the next opportunity was initiated by the teacher with a different food or drink item. Opportunities continued until all 10 food (drink) items had been offered once, in a predetermined random order.

Teaching Requests

253

Single-Instance Training
Intervention began with single-instance training. During these sessions, a single item was selected randomly from the array of preferred food and drink items and only this item was offered on 10 successive opportunities by the teacher. For Alison, the food item was a cookie and the drink item was fruit cordial. Banana and fruit cordial were selected for Emily. Requesting opportunities were again initiated by the teacher who displayed the item and said "If you want this, point to eat (drink)." When the learner pointed to the correct symbol and only that symbol, the teacher gave the child the offered item. If the child did not point to the correct symbol within 10 s, the teacher physically guided the child's finger to the correct symbol while repeating the question "If you want this, point to eat (drink)." These physically prompted requests did not result in access to the offered item and were followed approximately 5 s later by initiation of the next opportunity. If the child pointed to the incorrect symbol, the teacher said "No" and simultaneously placed her hand over the incorrect symbol. Keeping her hand in this position, the teacher put the cup (plate) on the table out of the child's reach and then repeated "If you want this, point to eat (drink)" while modeling simultaneously the correct response by placing her index finger at the top edge of the appropriate symbol. If the child responded correctly to this level of prompting, praise was given (e.g., "Right. That's eat [drink].") and approximately 5 s later the next requesting opportunity was initiated. If the child did not point to the correct symbol within 10 s, the teacher physically guided the child's finger to the c o r r e c t symbol and again initiated the next o p p o r t u n i t y approximately 5 s later. After several responses had been prompted in this manner, subsequent error correction no longer involved covering the incorrect symbol. Instead the teacher said "No" and modeled the correct response. Towards the end of single-instance training, the verbal "No" was often sufficient to prompt correct responses. Single-instance training was designed to establish the requesting responses in the presence of a single exemplar from each class of objects. Once established, generalization of the requesting responses to the remaining food and drink items was assessed during multiple-instance training sessions.

Multiple-Instance Training
During multiple-instance training for food items, the teacher provided one opportunity for the child to request each of the 10 food items within

254

Sigafoos et

aL

each session. However, only one type of beverage was offered within a multiple-instance training session for drink items, The exact beverage item that was offered varied across sessions to ensure the child was offered the full array of preferred beverages over the course of the study. For example, during one session the child might be offered a cup containing water. In the next session, the child might be offered a small amount of milk on each of the 10 requesting opportunities. Because beverage items were always offered in the same opaque cup, it was reasoned that the cup itself would become the discriminative stimulus for selecting the "drink" symbol. In addition, only a single beverage item was used within a session because of the undesirability of providing the child with a small amount of 10 different beverages within a relatively brief period of time. Procedures for initiating requesting opportunities and reinforcing and prompting correct responses were identical to those used during single-instance training. This phase of the study was designed to test for generalization of the food and drink requests to the remaining exemplars from these two respective stimulus classes that were not the direct focus of intervention during singleinstance training.

Concurrent Instruction
In this phase, the child was provided with five opportunities to request food items and then five opportunities to request the selected drink item for that session. This sequence was then repeated until a total of 20 opportunities (i.e., 10 food, 10 drink) had been implemented. These concurrent opportunities were implemented to determine if the discriminations established under the previous instructional sequence (i.e., serial instruction) would be maintained when opportunities to request food and drink items were intermixed. Except for this change from serial to concurrent sequencing, the procedures used were identical to those described for multiple-instance training.

Serial Instruction
Because Alison's performance under concurrent instruction diminished somewhat (see Results), requesting opportunities were again arranged in a serial sequence. Specifically, Alison was provided with one opportunity to request each of the 10 preferred food items and then, following a short break (2-3 min), 10 opportunities to request the selected beverage item

Teaching Requests

255

were provided. Procedures replicated those implemented during multipleinstance training.

Add Color
In an effort to increase the difference between symbols, color was added to the line drawings and then removed at various points in the study. For Alison, the drawing of the apple in the "food" symbol was colored red and the cup in the "drink" symbol was colored green. These colors were added to the symbols during the four sessions prior to the follow-up phase because of difficulty in maintaining acceptable levels of accuracy during concurrent instruction (see Results). The symbols were returned to their original black and white configuration for the three follow-up sessions. For Emily, the symbols were enlarged to 12 cm x 15 cm and the apple in the "food" symbol was colored green. These changes to Emily's symbols occurred during single-instance training for food because of deteriorating performance. Her symbols were returned to their original size and color in the five sessions prior to follow-up. It was hypothesized that the difficulties noted above for Alison and Emily were due in part to the fact that the food and drink symbols looked quite similar. Adding color to each symbol was considered the easiest way to make the symbols look different. This modification was made after the initial intervention phase when the data suggested that the children were having difficulty discriminating between the two symbols. Although it may have been better to have used colored symbols from the onset, we started with the standard food and drink symbols because we wanted to establish a discrimination on the basis of the differing configurations of the drawing. If color had been maintained as the basis for symbol discrimination, the total number of symbols these children could acquire in the future may have been limited by the need to associate each symbol with a distinct color.

Follow-up
Follow-up sessions occurred at two-week intervals. The procedures used during follow-up sessions replicated those implemented during concurrent instruction. Opportunities to request food and drink at morning snack were not provided in the interim between follow-up sessions. Instead, Alison and Emily became involved in another study designed to increase their repertoire of graphic mode requests.

256

Sigafoos et aL

Interobserver Agreement
One of two independent observers collected agreement data on the children's responses during all phases of the study. For Alison, 37% of the baseline opportunities and 42% of the intervention opportunities were scored for agreement. With Emily, 37% and 32% of the respective baseline and intervention opportunities were observed. Agreement was calculated on a trial-by-trial basis by comparing the type of response (correct, incorrect, no response) recorded by the teacher with that recorded by the independent observer. Interobserver agreement was calculated using the formula: Agreements/(Agreements + Disagreements) x 100%. Agreement ranged from 80% to 100%. Mean agreement on Alison's responses was 100% in baseline and 99% during intervention. The means for Emily were 98% and 97% during baseline and intervention, respectively.

RESULTS Figures 1 and 2 show the percent of correct requests for Alison and Emily. In each figure, requests for food are plotted in the upper panel and requests for drink are plotted in the lower panel. The percent of correct requests for food and drink are shown on a session-by-session basis across all phases of the study. During initial baseline sessions, Alison's performance (Fig. 1) was characterized by a position-bias. Specifically, Alison always pointed to the symbol placed to the left of her midline, except on three occasions during which no response was made. This position-bias accounts for the ascending trends in baseline as the "correct" symbol was placed to the left of midline on 6 of the 10 opportunities during the second and third baseline sessions for drink and food, respectively. With the introduction of single-instance training involving cookies, the percent of correct "food" requests increased. Baseline sessions with drink items that were implemented at the end of single-instance training for cookies produced no correct responses as Alison now exhibited a systematic error pattern consisting of always pointing to the "food" symbol. When single-instance training was extended to "drink," the percent of correct requests increased. Correct selection of the "food" symbol continued under the multiple-instance training phase, except during one of the three sessions when performance dropped to 20% correct. When opportunities to request both food and drink were intermixed (i.e., concurrent instruction), the percent of correct requests for both food and drink items decreased. A return to the serial instructional sequence produced a recovery in terms of percent correct. A second attempt to intermix food

T e a c h i n g Requests

257

Intervention
too

I M~It .ICo:~w~la
i i

Ii*ri~
i ) i

Co~r
i i

A&t Colar

Follow-up

9o 80 70 60 50 40 30. 20. g= 0 D II0, o, too 9 90. 80. 70, 60. 50. 403020I0o-

'//'
i i i

i i

i i i I i i

p!

i i

i i i

'
~.at I i I i i

I i
i i i i

FOOD

. . . .

'1

.II

i i i I i i i

/
o-o l0
20 30

IiRINK

246

SESSIONS

W~EKS

Fig. 1. Percentage of correct requests for food (upper panel) and drink (lower panel) across sessions for Alison.

Interven~n
I00 90" 80" 70 60 50 40 i" 30 LI~I4I A ~ Color Ij M u l l

Co~un'~

IR e m o ~

Follow-up

aE'g \f i=g

FOOD

~
~ 0 ~ 9

2o
lO o IOO 80 70 60 50' 40. 30" 20" 10OI

~
~ ~

9o

DRINK

T
10

20 SESSIONS

30

2468 WEEKS

Fig. 2. Percentage of correct requests for food (upper panel) and drink (lower panel) across sessions for Emily.

258

Sigafoos et aL

and drink opportunities resulted again in a gradual deterioration. At this point, it was decided to continue concurrent instruction and to color parts of the symbols to increase the visible difference between the two line drawings. Accurate levels of requesting associated with these symbol modifications were maintained during follow-up sessions when the original black and white line drawings were used. Emily (Fig. 2) typically pointed to both line drawings in rapid succession during initial baseline opportunities. Single-instance training with banana was associated with an increase in the correct selection of the "food" symbol and only that symbol. However, a descending trend emerged over the last four sessions within this phase owing to an increase in the number of incorrect and no responses (i.e., pointing to the "drink" symbol when offered a food item and failing to point to the "food" symbol within 10 s). Because of this trend, the symbols were enlarged and color was added to the "food" symbol in an effort to facilitate the discrimination between the two line drawings. After these changes were made, the percent of correct requests for banana stabilized above 80%. Implementation of baseline opportunities for drink items at this point (Sessions 13-15 and 20) revealed an increase in correct selection of the "drink" symbol when Emily was offered the cup. Although this result could suggest that intervention to teach requests for food had generalized to offers of beverages, the results from Emily's final four baseline sessions compromise the integrity of the multipie-probe design. However, when opportunities to request both food and drink items were then arranged with the introduction of concurrent instruction, correct requesting was maintained, indicating that Emily was in fact using her two symbols in a discriminated manner. It is therefore possible that teaching Emily to point to the food symbol when offered something to eat was sufficient to ensure that she would come to select the other [drink] symbol when offered the cup. In the next phase (Remove color), the original sized black and white line drawings were used with no detriment to performance. Follow-up sessions revealed that Emily continued to make requests for food and drink items accurately.

DISCUSSION Results of the present study suggest that the procedures associated with intervention were effective in developing appropriate stimulus control over requests for food and drink. Over several intervention phases, a discrimination between the two line drawings appeared to develop as the children came to select the "food" symbol when offered a food item and, alternatively, the "drink" symbol when offered a beverage. These requests

Teaching Requests

259

appeared appropriately generalized as well, given that correct responses were maintained when the children were offered a variety of food and drink items. These results are consistent with previous studies which have demonstrated the effectiveness of behavior analysis procedures for teaching children with developmental disabilities to request preferred objects in vocal and gesture modes (Duker and Remington, 1991; Duker and van Lent, 1991; Gobbi et al., 1986; McCook et al., 1988; Sigafoos, 1995). In addition, these results are consistent with a small, but growing literature demonstrating the potential viability of beginning communication intervention for some individuals by teaching requests for preferred objects in a graphic communication mode (e.g., Glennen and Calculator, 1985; Reichle and Brown, 1986; Romski et al., 1988; Sigafoos et al., 1995; Turnell and Carter, 1994). When beginning intervention to teach requesting, it is important to select communication modes which best match individual needs and abilities (Goetz and Hunt, 1994; Reichle et aL, 1991). For some, a graphic mode may represent an appropriate means of communicating requests, but for others, gesture, vocal, or some combination of these three modes may be indicated (Butterfield et al., 1995). Despite the generally favorable effects of intervention, there was some difficulty in obtaining appropriate stimulus control over requesting. Alison, for example, exhibited an increase in the number of incorrect responses as requests for food and drink were intermixed. Similar problems in teaching discriminations to children with developmental disabilities have been reported (e.g., Lovaas, 1977). Sigafoos and Reichle (1992), for example, noted significant difficulties in obtaining discriminations among line drawings when intervention to teach requesting began with concurrent instruction. In comparison, somewhat fewer difficulties were evident in the present study, which may have resulted from first establishing requests for food and drink separately and only then adopting a concurrent instructional strategy. Saunders and Spradlin (1989) also found a gradual introduction of concurrent instruction was effective for two adults with moderate mental retardation who experienced initial difficulty on a discrimination task. Because opportunities to request food and drink were intermixed, but never randomized, it is possible that the children performed accurately by adopting a win-stay, lose-shift strategy. If this were the case, performance during concurrent instruction could have stabilized, as it generally did, near 80%-90%, even if the child continued to make errors each time the item offered was switched from food to drink and vice versa. A review of the raw data indicated that incorrect responses during concurrent instruction did occur at these shift points, but not exclusively. In addition, if such a strategy had been adopted it may have been used only temporarily as both

260

Sigafoos et al.

children made relatively few errors during follow-up sessions. Alison, however, made two errors during her last follow-up session and both of these occurred as the item offered was switched from food to drink. Given the plausibility that Alison and Emily may have adopted a win-stay, lose-shift strategy, the present data must be viewed with caution. Further limitations also necessitate a cautious interpretation of the data. Specifically, Emily began to request drink items before intervention, making it difficult to attribute improvement to the intervention procedures. Nonetheless, in three of the four cases, the percent of correct requests increased only when intervention procedures were introduced. Another limitation stems from the numerous phases that comprised intervention. Most of these phases were not introduced and withdrawn in a systematic manner and thus their effects on requesting were not demonstrated equivocally. For example, it appeared that adding color facilitated discrimination among the line drawings, but it is possible that the percentage of correct requests would have increased with continued intervention and without any modification to the symbols. Sequential introduction of various intervention phases represented a formative or "radical" methodology (Skinner, 1956; Sundberg, 1980). Changes were made to the independent variable (i.e., the intervention procedures), based on changes in the children's behavior. While responsive to the learner's ongoing behavior, it is often difficult to implement such formative modifications in a manner that conforms precisely to a specific research design. Opportunities to request food and drink were arranged differently in some phases of the study. During multiple-instance training and concurrent instruction, for example, a variety of food items were used, whereas only one drink item was offered within a given session. In addition, the drink item was always offered in the same cup. As a result, "food" and "drink" may have represented different types of requests. Pointing to the "food" symbol could be interpreted as a rather generalized request, in that it was eventually emitted in the presence of a variety of food items. In contrast, pointing to the "drink" symbol may have been taught as a more explicit request, under the control of a particular item (i.e., the cup). The extent to which such differences influence stimulus control remains an area for further study. In both paradigms, however, opportunities to request were initiated by the teacher who offered an item and said "If you want this, point to eat (drink)." Because requests for food and drink would be of greater benefit when controlled by hunger and thirst, respectively, further research is needed to develop procedures for bringing requests under the control of such private events (Schnaitter, 1978; Skinner, 1945). One method by which to achieve this desired outcome might be to use transfer of stimulus control procedures to bring requests for food and drink under

Teaching Requests

261

the control of certain public accompaniments to hunger and thirst (Sigafoos and Reichle, 1993). For example, if the teacher were to hear the child's stomach growling, then an opportunity to request food could be provided. Alternatively, opportunities for requesting might be provided in the presence of certain collateral behaviors. For example, a child sitting in the shade fanning herself might assumed to be thirsty and this could set the occasion for requesting a drink. In the present study, requests were taught during the regular morning snack time when the children were likely to be hungry and thirsty due to a natural schedule of mild deprivation. It remains to be determined, however, if requests taught under these latter conditions would occur at other times when the child was hungry or thirsty.

REFERENCES
Albin, R. W., and Horner, R. H. (1988). Generalization with precision. In Horner, R. H., Dunlap, G., and Koegel R. L. (eds.), Generalization and Maintenance: Life-style Changes in Applied Settings, Paul H. Brookes Publishing Co., Baltimore, pp. 99-120. Bates, E. (1976). Language and Context: The Acquisition of Pragmatics, Academic Press, New York. Butterfield, N., Arthur, M., and Sigafoos, J. (1995). Partners in Everyday Communicative Exchanges, MacLennan and Petty, Sydney. Chadsey-Rusch, J., and Halle, J. (1992). The application of general-case instruction to the requesting repertoires of learners with severe disabilities. J. Assoc. Pers. Sev. Hand. 17: 121-132. Cipani, E. (1988). Behavior Analysis Language Program: Theory, Assessment, and Training Practices for Personnel Working with People with Severe Handicaps, Edmark Corporation, Bellevue, WA. Compic Development Association (1986). Compic Pictographs, Compic Development Association, North Balwyn,Victoria. Duker, P. C., and Remington, B. (1991). Manual sign-based communication for individuals with severe or profound mental handicap. In Remington, B. (ed.), The Challenge of Severe Mental Handicap: A Behaviour Analytic Approach, John Wiley & Sons, Chichester, pp. 167-187. Duker, P. C., and van Lent, C. (1991). Inducing variability in communicative gestures used by severely retarded individuals. J. Appl. Behav. Anal. 24: 379-386. Goetz, L., and Hunt, P. (1994). Augmentative and alternative communication. In E. Cipani, E., and Spooner, F. (eds.), Curricular and Instructional Approaches for Persons with Severe Disabilities, Allyn & Bacon, Boston, pp. 263-288. Glennen, S. L., and Calculator, S. N. (1985). Training functional communication board use: A pragmatic approach. Augment. Alter. Comm. 1: 134-142. Gobbi, L., Cipani, E., Hudson, C., and Lapenta-Neudeck, R. (1986). Developing spontaneous requesting among children with severe mental retardation. Ment. Retard. 24: 357-363. Guess, D., Sailor, W., and Baer, D. M. (1974). To teach language to retarded children. In Schiefelbusch, R. L., and Lloyd, L. L., (eds.), Language Perspectives: Acquisition, Retardation, and Intervention, University Park Press, Baltimore, pp. 529-563. Halle, J. W. (1987). Teaching language in the natural environment: An analysis of spontaneity. J. Assoc. Pers. Sev. Hand. 12: 28-37. Horner, R. D., and Baer, D. M. (1978). Multiple-probe technique: A variation of the multiple-baseline. J. Appl. Behav. Anal. 11: 189-196.

262

Sigafoos e t a L

Lovaas, O. I. (1977). The Autistic Child: Language Development Through Behavior Modification, Irvington Publishers, New York. McCook, B., Cipani, E., Madigan, K., and LaCampagne, J. (1988). Developing requesting behavior: Acquisition, fluency, and generality. Ment. Retard. 26: 137-143. Michael, J. (1985), Two kinds of verbal behavior plus a possible third. AnaL Verbal Behav. 3: 1-4. O'Neill, R. E. (1990). Establishing verbal repertoires: Toward the application of general case analysis and programming. AnaL Verbal Behav. 8: 113-126. Reichle, J., and Brown, L. (1986). Teaching the use of a multi-page direct selection communication board to an adult with autism. J. Assoc. Pers. Sev. Hand. 11: 68-73. Reichle, J., York, J., and Sigafoos, J. (1991). Implementing Augmentative and Alternative Communication: Strategies for Learners with Severe Disabilities, Paul H. Brookes Publishing Co., Baltimore. Romski, M. A., Sevcik, R. A., and Pate, J. L. (1988). Establishment of symbolic communication in persons with severe mental retardation. J. Speech Hear. Disorders 53: 94-107. Sailor, W., and Mix, B. (1975). The Topeka Association for Retarded Citizen's Assessment System, Pro-ed, Austin, TX. Saunders, K. J., and Spradlin, J. E. (1989). Conditional discrimination in mentally retarded adults: The effects of training the component simple discriminations. J. Exp. AnaL Behav. 52: 1-12. Schnaitter, R. (1978). Private causes. Behaviorism 6: 1-12. Shafer, E. (1993). Teaching topography-based and selection-based verbal behavior to developmentally disabled individuals: Some considerations. Anal. Verbal Behav. 11: 117-133. Sigafoos, J. (1995). Testing for spontaneous use of requests after sign language training with two severely handicapped adults. Behav. Interven. i0: 1-16. Sigafoos, J., Couzens, D., Pennell, D., Shaw, D., and Dudfield, G. (1995). Discrimination of picture requests for missing items among young children with developmental disabilities. J. Behav. Educ. 5: 295-317. Sigafoos, J., and Reichle, J. (1992). Comparing explicit to generalized requesting in an augmentative communication mode. J. Develop. Phys. Disabil. 4: 167-188. Sigafoos, J., and Reichle, J. (1993). Establishing spontaneous verbal behavior. In Gable, R. A., and Warren, S, F. (eds.), Advances in Mental Retardation and Developmental Disabilities, VoL 5: Strategies for Teaching Students with Mild to Severe Mental Retardation, Jessica Kingsley, London., pp. 191-230. Skinner, B. F. (1945). The operational analysis of psychological terms. Psych. Review 52: 270-277. Skinner, B. F. (1956). A case history in scientific method. Amer. Psych. 11: 221-233. Sundberg, M. L. (1980). Developing a Verbal Repertoire Using Sign Language and Skinner's Analysis of Verbal Behavior, Western Michigan University, Kalamazoo, MI, Unpublished doctoral dissertation. Sundberg, M. L. (1993). Selecting a response form for nonverbal persons: Facilitated communication, pointing systems, or sign language? Anal Verb. Behav. 11: 99-116. Turnell, R., and Carter, M. (1994). Establishing a repertoire of requesting for a student with severe and multiple disabilities using tangible symbols and naturalistic time delay. Aust. New Zealand J. Develop. Disabil. 19: 193-207.

Research in Developmental Disabilities 25 (2004) 371383

Teaching students with developmental disabilities to locate their AAC device

Jeff Sigafoos*, Mark OReilly, Sue Seely-York, Chaturi Edrisinha
Department of Special Education, The University of Texas at Austin, Austin, TX 78712-1290, USA Received 18 March 2003; received in revised form 14 July 2003; accepted 15 July 2003

Abstract Students with autism and related developmental disabilities who do not speak are often taught to use some type of augmentative and alternative communication (AAC) system, such as a voice output communication aid (VOCA). One problem with such devices is that the person may be unable to communicate when the device is not readily accessible. We rst taught three nonverbal students with autism to use a VOCA to request access to preferred items. Following this initial acquisition phase, however, none of the students would locate their VOCA when it was not within reach. A least-to-most prompting procedure was implemented to teach the students to locate their AAC device. The effectiveness of this procedure for teaching VOCA location skills was evaluated in a delayed multiple-baseline across subjects design. The results showed that the intervention was effective in teaching the students to locate their AAC device when they needed it to request access to preferred objects. Teaching VOCA location skills may be a useful and necessary component in AAC interventions for some people with developmental disabilities. # 2004 Elsevier Ltd. All rights reserved.
Keywords: voice output communication aids; AAC intervention; location training; least-to-most prompting; autism; developmental disability

Individuals with developmental disabilities who have limited or no speech are candidates for augmentative and alternative communication (AAC; Reichle, Beukelman, & Light, 2002). While there are many different AAC systems, the various options can be classied into two modes: aided and unaided (Beukelman & Mirenda, 1998). Aided systems involve the use of external equipment or
*

Corresponding author. Tel.: 1-512-475-8572; fax: 1-512-471-2471. E-mail address: j.sigafoos@mail.utexas.edu (J. Sigafoos).

0891-4222/$ see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ridd.2003.07.002

372

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

auxiliary materials such as communication boards with graphic symbols (Reichle, York, & Sigafoos, 1991), the picture-exchange communication system (PECS; Bondy & Frost, 2001), and voice output communication aids (VOCAs; Schepis, Reid, & Behrman, 1996). Unaided modes do not involve external materials and include gestures and manual signs (Duker, 1988). VOCAs may represent a promising option for some individuals with autism and related developmental disabilities (Schlosser & Blischak, 2001). Schepis et al. (1996) delineated several potential advantages of VOCAs. Specically, the speech output feature may provide a more natural and understandable signal to listeners. In addition, voice output combines attention-gaining with the communicative act and this may increase the probability of listeners attending to the persons communicative attempts. Furthermore, VOCA may be programmed to produce messages so precise (e.g., I need help opening this jar.) that misunderstandings are reduced. Despite these potential advantages, data-based reports on the use of VOCAs in AAC interventions for individuals with autism and related developmental disabilities are comparatively rare. Schlosser and Blischak (2001) considered some of the possible reasons for this relatively limited use of VOCAs in AAC interventions for people with autism. First, there may be a tendency to view people with autism as primarily visual students and thus conclude that voice output would be less salient for these individuals and perhaps even function as a distraction during intervention. There is little evidence to support this position, however, and in one relevant study voice output improved the efcacy of communication intervention for a student with autism (Schlosser, Blischak, Belore, Bartley, & Barnett, 1998). Second, VOCAs are generally more expensive than other AAC systems and the mere provision of a VOCA is no guarantee that the person will be capable of or interested in using that device for communicating with others. In the absence of substantial evidence that individuals with autism can be taught to use VOCAs for functional communication, clinicians may be justiably reluctant to prescribe such devices. To counter this concern, there is an emerging, albeit modest, literature demonstrating effective procedures for teaching individuals with autism and related developmental disabilities to use VOCAs for functional communication (Schlosser & Blischak, 2001; Schlosser et al., 1998; Sigafoos & Drasgow, 2001) and some individuals prefer to use VOCAs over other low-tech AAC systems (Soto, Belore, Schlosser, & Haynes, 1993). A third reason for this apparent reluctance to use VOCAs in AAC interventions for people with autism may arise from the obvious fact that VOCAs are less portable than unaided modes of communication and generally more cumbersome than other low-tech options (Sigafoos & Iacono, 1993). The portability issue is less of a concern for individuals who use wheelchairs because the VOCA can be afxed to the laptray or wheelchair frame, thus making it readily accessible. For individuals who are ambulatory, however, the concern is that the device may be too difcult to carry around and hence not always available when needed. This concern might be addressed during

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

373

intervention by teaching individuals to locate their VOCA when it is not in their immediate vicinity. Teaching individuals to locate their VOCA has been neglected in interventions to establish AAC skills in people with autism and related developmental disabilities. Instead, VOCA training often begins by teaching the person to request a preferred object by pointing to a line drawing (e.g., WANT) that is afxed to the VOCA. Touching the WANT symbol produces a relevant voice output (e.g., I want more.). During the initial acquisition phase, the VOCA is always placed within reach of the individual. While this conguration can lead to rapid acquisition (Sigafoos, Didden, & OReilly, 2003), it is unclear if, following acquisition, the student would seek out the VOCA to make a request if it was not within reach. The present study investigated this issue in three students with autism. We sought to determine if the students would locate their VOCA when it was not in their immediate vicinity. This assessment was completed following acquisition training during which the VOCA was always within reach. If VOCA location skills were not demonstrated, then we further aimed to evaluate the effectiveness of a least-to-most prompting procedure to teach the students to search for and locate their VOCA when it was needed to communicate. The goal of intervention was to establish VOCA location skills because these were considered highly relevant to the functional use of such devices by students who are ambulatory.

1. Method 1.1. Students and settings The three individuals who participated in this study were selected because they had autism and no speech. Owing to the severe nature of their disability and their lack of speech, intervention to teach the use of a VOCA was identied as an instructional priority. The intervention described in the present study was part of a larger program focused on teaching various aspects of VOCA use including device operation and symbol discrimination. Megan was a 20-year-old female. She was diagnosed with autism, mental retardation, and severe bilateral hearing loss. She was considered untestable on standardized intelligence tests. Megan did not speak, but made occasional whining and humming sounds. She communicated mainly through informal gestures (reaching for objects, pushing items away) and by guiding an adults hand to an object. Megan was also able to follow some simple gestured commands such as Stand up, Sit down, and Come here. She had limited social skills and displayed frequent and severe self-injurious head banging that had left a permanent welt on her forehead. She attended a vocational training program for adults with developmental disabilities. Sessions associated with this study were conducted during morning break-time in an ofce at the training center.

374

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

Jason was 16 years old. He was diagnosed with pervasive developmental disability not otherwise specied (PDD/NOS) and severe mental retardation. His IQ was estimated at less than 30, although he had not been formally assessed. He was nonverbal and communicated mainly through facial expressions, gestures, a few manual signs (i.e., MORE, EAT, DRINK), screaming, and by guiding an adults hand to objects. Jason was toilet trained and was able to dress himself and brush his teeth with assistance. He had limited social skills and displayed occasional aggression, self-injury, and property destruction. Jason attended a private residential school program for students with autism. He was educated in a self-contained classroom with ve other students. The classroom was staffed by two teachers. The sessions associated with this study were conducted at a table in his classroom during a morning snack activity. Ryan was 12 years old and was diagnosed with autism. His IQ could not be estimated, but he was considered untestable on standardized measures of intelligence. His vision and hearing were within the normal range and he did not appear to have any limitations with respect to ne and gross motor skills. He did not speak, but could vocalize and make speech-like sounds. Ryan communicated mainly by pointing to objects that he wanted. He also used a few manual signs (e.g., WANT, MORE) that appeared to function as requests. Ryan usually responded appropriately to simple spoken and gestured commands from adults such as Come here, Wash your hands, and Sit down. He could feed, dress, and toilet himself, but needed considerable assistance with other self-care tasks (e.g., tying his shoes, bathing, brushing his teeth). He initiated interactions with adults by approaching, making eye contact, and touching them, but tended to ignore his peers. He showed interest and ability in using reactive toys and could play independently for 1020 min when provided with a preferred toy. His problems behaviors consisted of frequent tantrums that included screaming, dropping to the oor, head banging, and hand biting. He would also occasionally hit and head-butt others and sniff the hair of strangers. Ryan attended the same school program as Jason and his training was conducted in his classroom during a morning snack activity. 1.2. Prior intervention history In the 2 weeks prior to the start of the present study, each student was taught to use a Tech/Talk 6X81 voice output communication aid (Advanced Multimedia Devices, Inc.). The Tech/Talk consists of eight panels, each of which can accommodate a 7 cm 7 cm graphic symbol and a digitized recorded message. For this study, only one randomly selected panel had a graphic symbol. The symbol was a black and white line drawing representing WANT from the MayerJohnson Picture Communication Symbols Combination Book (Mayer-Johnson Co., 1994). Pressing this panel activated the recorded message (I want more.). The other eight panels were blank. This prior intervention began with a preference assessment to identify a set of preferred edible objects that each student would be taught to request during the

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

375

morning snack/break-time. Each student was on a special diet unrelated to the present study that restricted the types of foods that they could eat. From the list of acceptable snack foods, a set of potential items was assembled (e.g., raisins, popcorn, peanuts, carob chips). The preference assessment consisted of offering a small sample of each item 12 times. Results showed that the students never failed to select and consume each item when offered. These snack foods were therefore used as reinforcers in the present study. To teach the initial general request WANT, the trainer and student sat at a table. The Tech/Talk device was placed within the students reach and several portions of each preferred edible were placed on a tray. The tray was placed on the table and in view, but it was out of the students reach. During the preference assessment it was determined that each student would reliably reach for the tray of items when it was visible. We made use of this tendency to reach in teaching the requesting response. Specically, as the student reached for the tray of items, his or her hand was physically guided to press the WANT symbol on the Tech/Talk using the least amount of physical guidance necessary. Reaching for the items or touching a blank panel were considered incorrect responses. If the student made an incorrect response he or she was prompted to touch the correct panel on the Tech/Talk, again using the least amount of physical guidance necessary. Each correct response, whether prompted or not, was followed by allowing the student to select a small portion of one of the available items from the tray. Over successive opportunities, physical guidance was delayed by 5 s to promote independent use of the device for requesting. Training was conducted in 5min sessions and continued until the student made 10 independent and correct requests (i.e., pressing the correct panel without rst reaching for the tray of preferred edibles) within a single session. All three students reached criterion within 20 min of training. After acquisition, Jason and Ryan were also taught to turn the volume on the Tech/Talk to the on position when the voice output function was turned off. Megan was not taught this step of VOCA operation because of her hearing impairment. To teach this skill, the device was turned off so that a correct panel press did not result in activating the recorded message. When the volume was turned off and when Jason and Ryan attempted to make a request by pressing the WANT symbol, they were prompted to turn the volume knob to the on position. The prompt consisted of rst pointing to the volume knob and then physically assisting the student. Prompting was delayed 5 s to promote independent performance of this step. Jason and Ryan learned to turn the device on after approximately 10 min of training. The baseline and intervention phase of the present study began one week after this initial acquisition training. The baseline phase provided an assessment of VOCA use under two conditions: (a) when the device was within reach and (b) when the VOCA was placed out or reach. Baseline was followed by an intervention phase to evaluate the effectiveness of a least-to-most prompting procedure to teach the students to locate their VOCA when it was placed out of reach and not located within their immediate vicinity.

376

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

1.3. Response denitions and measurement Two types of responses were dened and recorded in the present study. First, correct use of the Tech/Talk for requesting was recorded when a student pressed the WANT symbol on the Tech/Talk with sufcient force to activate the recorded message I want more. This rst dependent measure was relevant only during baseline and only during the condition when the Tech/Talk was placed within reach of the individual. A request was counted as correct only of the response occurred within 30 s of the beginning of a trialsignaled by the trainer giving the verbal cueor was initiated by the student without cueing at least once every 30 s. In addition, a request was correct only if the student did not rst reach for the available snacks. The second response was relevant during the second condition when the Tech/ Talk was placed out of reach, which occurred during baseline and intervention. A correct response under this condition required the individual to locate the VOCA in the classroom, bring it back to the table, turn it on (Jason and Ryan only), and then press the WANT symbol to activate the recorded message I want more. Varying amounts of time were allowed for this sequence to be completed because the VOCA was placed in different locations throughout the classroom. However, a correct response under this condition was only counted if the students initiated the rst step in the sequence (i.e., getting up from their chair at the table to begin the search for the VOCA) within 30 s of the beginning of an opportunity and then completed the entire sequence without prompting. The trainer recorded the presence/absence of correct responding on a trial-by-trial basis on a data sheet. 1.4. Procedures All sessions began by seating the student at the table with the trainer. Three different people assumed the role of the trainer as a way of programming for generalization. The student was then given one portion of each of the snack items from the tray. The tray containing additional portions of each item was then moved out of reach and the trainer provided an initial verbal cue by saying Let me know if you want more. The verbal cue signaled the start of a requesting opportunity. This cue was repeated every 30 s if necessary. In addition, however, the student could self-initiate the beginning of a opportunity as well. This occurred whenever the student made a correct request during an inter-trial interval. That is, if the student made a correct request at least once every 30 s, then the verbal cue was not given as it was considered unnecessary to signal an opportunity. In this scenario, an opportunity began when the student self-initiated a correct response. Sessions were conducted once or twice per week. Each session typically lasted about 10 min and included at least three opportunities, although the length of sessions and number of opportunities per session varied. This variation was due to the fact that sessions were to some extent paced by the students behavior. That is, the number of opportunities that could be completed during the snack activity

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

377

depended to some extent on how quickly each student initiated requests or waited to be prompted, what level of prompting was required, and how long it took the student to consume a snack item once it had been requested. 1.5. Baseline During baseline, the VOCA was either placed within reach or moved out of the students reach. When the VOCA was within reach, access to the tray of items was contingent upon a correct VOCA response (i.e., pressing the WANT symbol with sufcient force to activate the recorded message I want more.). After each correct response, the VOCA was removed from the table by the trainer or reliability observer and placed in a location elsewhere in the classroom (Jason and Ryan) or ofce area (Megan). The location of the VOCA varied randomly across opportunities. Sometimes it was in view of the student, but at other times it was placed in a location where it could not be seen by the student. Sometimes it was placed on a different table, desk, or chair, but at other times it might be placed on a shelf or held by the trainer or reliability observer. In all cases, the VOCA was placed at least 2 m from the student making it necessary for him or her to get up from the table to obtain the VOCA. The intent of altering the location of the VOCA was to simulate what might typically happen in a classroom or vocational training center. In such settings it would not be unusual for a persons VOCA to be moved around and placed in various locations throughout the day. After the VOCA was removed from the table and placed in a location, the trainer waited up to 30 s for the student to initiate a correct response. If the student had not initiated a correct response within 30 s, the device was returned to the table by the reliability observer for the beginning of the next opportunity in which the VOCA was placed within reach of the student. 1.6. Intervention For all opportunities during this condition, the VOCA was placed out of reach in a random location in the classroom (Jason and Ryan) or the ofce area (Megan). Once the VOCA was placed out of reach, the trainer waited 30 s for initiation of a correct response, which involved the student leaving his or her chair and beginning to search the room for the VOCA. If the student did not initiate a correct response within 30 s, a least-to-most prompting hierarchy was implemented (Duker, Didden, & Sigafoos, 2004). Specically, the trainer rst used a gesture prompt, which consisted of pointing to where the VOCA had been placed. If the student did not initiate the correct response within 10 s of this prompt, the pointing prompt was paired with a verbal instruction (e.g., Go get it. Can you nd it? Its over there). If the student still did not initiate a correct response within 10 s, physical guidance was used to prompt initiation of the response. Physical guidance involved using the least amount of physical assistance that was necessary to prompt the person to get up from the chair and walk over to the device. Once the device was located and returned to the table, access to the tray of

378

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

items occurred when the student pressed the WANT symbol to activate the message I want more. After this, the VOCA was again removed to start the next opportunity. A modication was made to these intervention procedures for Megan beginning with the seventh block of opportunities (see Fig. 1). This was necessary
Baseline
100 90 80 70 60 50 40 30 20 10 0

Intervention

Prompt to return

Megan

100 90 80

Percentage Correct

70 60 50 40 30 20 10 0

Jason

100 90 80 70 60 50 40 30 20 10 0

Ryan

Blocks of 3 Opportunities
Fig. 1. Percentage of correct responses when the VOCA was within reach (~) and when it was located out of reach and placed in various locations in the room (*). Data are graphed in blocks of three opportunities during baseline and intervention for Megan, Jason, and Ryan.

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

379

because during the previous four blocks of intervention opportunities, Megan would locate the VOCA, press the correct panel, and then return to the table to select a snack item. That is, she found the device and used it, but she did not bring it back to the table with her. These distant requests were unanticipated and initially they were considered appropriate and reinforced, but after 12 such opportunities, physical guidance was used to prompt her to return to the table with the device before pressing the panel to activate the message. 1.7. Experimental design Baseline and intervention were introduced sequentially across students in a delayed multiple-baseline design (Richards, Taylor, Ramasamy, & Richards, 1999). The start of baseline had to be delayed for Jason and Ryan because they could not be scheduled to begin the program at same time as Megan. In addition, varying amounts of intervention data were collected for each student due to changes in their respective daily schedules, which made it impossible to continue data collection for the same length of time. 1.8. Inter-observer agreement The trainer and a second staff person who served as the reliability observer independently recorded correct use of the VOCA during each opportunity. An agreement was scored if the two agreed on whether or not correct VOCA use had occurred during the opportunity. Any discrepancy between the two observers was counted as a disagreement. Inter-observer agreement was calculated for each block of three opportunities using the formula: agreements=agreement disagreements 100%. Agreement during baseline and intervention was always 100%.

2. Results Fig. 1 shows the percentage of correct responses across blocks of three opportunities during baseline and intervention. During baseline when the VOCA was placed within reach (~), all three students would consistently press the WANT symbol so as to activate the recorded message I want more. In most cases, these responses were self-initiated and occurred as soon as the student had nished consuming a previous snack item. However, when the VOCA was placed out of reach (*), none of the students ever made a correct response, which in this situation required retrieving the VOCA from its displaced location. The percentage of correct responses when the VOCA was placed out of reach increased with the onset of intervention and stabilized at a high level (i.e., two out of three correct or better) for all three students. Although Megan did not return with the VOCA to the table during the rst 12 intervention opportunities, she did consistently locate the VOCA and press the correct panel to activate the recorded

380

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

message I want more. This was considered sufcient and functional at the time and so her performance during these initial intervention opportunities was recorded as correct. However, more stringent criteria were established for her with the onset of the seventh block of opportunities. At this point, she was prompted to return to the table with the VOCA if she did not do so on her own. With this modication, she began to return with the VOCA 100% of the time over her nal nine intervention opportunities.

3. Discussion Despite correct and self-initiated use of the VOCA when it was within reach, the baseline phase of the present study demonstrated that none of the three students would go to nd their VOCA when it was located out of reach. Instead, the students would typically attempt to gain access to the preferred items in the absence of the VOCA by reaching for items and leading the trainers hand. These responses were ignored and occasionally the students escalated to tantrums and self-injury, which were also ignored. As intervention progressed, all three students learned to quickly and consistently locate their VOCA when it had been placed in various locations throughout the classroom (Jason and Ryan) or ofce area (Megan) and to do so without prompting, although Ryan needed a pointing prompt to locate the VOCA on two out of his last six opportunities. These results suggest that the intervention was effective in teaching the students to locate their VOCA. As they learned to retrieve the VOCA from its various locations, reaching, leading, tantrums, and self-injury ceased to occur. During baseline and prior to acquisition of the VOCA location skill, reaching, leading, tantrums, and self-injury could be viewed as communicative repair strategies that emerged when a previously effective form of requesting (i.e., using the VOCA) was no longer possible because it was out of reach and also often out of sight. When the VOCA was present, however, it was used instead of these other forms, probably because it was always effective in enabling the student to gain access to the preferred items. As a result there was no need for the student to repair by reaching or leading, for example. Although the data on these collateral effects are anecdotal, these observations are consistent with research on functional communication training, which has shown that existing prelinguistic and problematic forms of communication can be replaced by teaching functionally equivalent AAC skills (Durand, 1999; Keen, Sigafoos, & Woodyatt, 2001). With intervention, the students learned to get up from the table, search the room until they located the VOCA, bring it back to the table, and use it to make the required request. The fact that the students searched the room for the VOCA even when it had been placed out of sight suggests that these for students, the VOCA existed even when it could not be seen (i.e., object permanence) (Piaget, 1929). This aspect of cognitive development could inuence efforts to teach VOCA location and related AAC skills.

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

381

The motivation for searching the room to nd the VOCA was no doubt established by the fact that the VOCA had a history of being needed to request access to the tray of preferred edibles. Because intervention occurred at the usual morning snack time and the available items were highly preferred, we assumed that the students were motivated to request these items. This assumption was supported by the fact that Megan, Jason, and Ryan never failed to select and consume an item once they gained access to the tray. As training progressed, the students latency in getting up from the table to begin the search decreased, although this observation is again anecdotal. For example, at the beginning of intervention, Jason would typically wait 2030 s before getting up from the table to search for his VOCA. By the end of intervention, however, he would get up from the table to retrieve the VOCA within 1 or 2 s of it being removed from the table by the trainer. The fact that all three students learned to locate the VOCA and did so with increasing speed as intervention progressed suggests that the VOCA had become a secondary [conditioned] reinforcer. This is a likely explanation given that in the context of the present study, VOCA use was the only way these students could gain access to the primary [edible] reinforcers that were offered by the trainer in the snack activity. Curiously, Ryan and Jason would often smile and giggle as they were searching the classroom for their VOCA. Megan, on other hand, appeared annoyed. She occasionally made negative vocalizations (e.g., whining sounds) as she got up from her chair to retrieve the device. This suggests individual differences in the extent to which students will tolerate being made to search for their AAC device. Although Megan also appeared motivated to request the food items, it is possible that her overall level of motivation was somewhat less than it was for Ryan and Jason. In any event, future research should examine the inuence of motivation to communicate in relation to the presence or absence of the required AAC device. One might predict that higher levels of motivation are required when the device must rst be located because nding the device takes additional time and effort. The ability to locate the VOCA when it was needed generalized in the sense that the students learned to search the room until they found the VOCA device, which was placed in numerous and varied locations throughout the intervention phase. Although we did not formally assess generalization to other trainers, settings, communication devices, and settings, the interventions incorporated a number of procedures to program generalization (Stokes & Baer, 1977). Specically, three different people were used as trainers and the VOCA had to be retrieved from many different locations. Still, future researchers could extend the present study by including a more systematic analysis of generalization, especially generalization to other types of aided AAC devices and across settings. For Ryan and Jason, a correct response during intervention involved not only locating the device and returning it to the table, but also turning it on so that it would produce the prerecorded message, I want more. when the WANT symbol was pressed. Their rapid acquisition and continued use of this skill suggests that

382

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

they were attentive to the voice output. In fact, both Ryan and Jason would typically rst test whether the device was on by pressing the WANT symbol and only if no voice output was forthcoming did they then adjust the on/off volume control on the back of the device. This suggests a conditional discrimination based on the presence/absence of voice output. Far from being a distraction, the voice output feature may have provided an important source of feedback for Ryan and Jason. It is unclear if Megan could hear the voice output, but even in such cases a VOCA might be indicated because the voice output would likely provide a readily understood signal to listeners. Despite the demonstrated effectiveness of the least-to-most prompting procedure in teaching VOCA location skills, the present study has limitations that necessitate a cautious interpretation of the results. First, the study involved only three students and so systematic replication across more individuals is necessary to validate the procedure for teaching VOCA location skills. Second, we did not obtain follow-up data to determine if VOCA location skills would be maintained over time. Maintenance data could not be collected for Megan because she moved to another program and maintenance probes for Ryan and Jason would have been confounded by the fact that they continue to receive VOCA training that included locating the device when necessary. Additional research is necessary to assess maintenance when VOCA locations skills are taught using a least-to-most prompt hierarchy. In summary, procedures to teach VOCA location skills would seem highly relevant for AAC interventions involving students who are ambulatory. The present study demonstrated the effectiveness of a least-to-most prompting procedure for teaching this skill to three students with developmental disabilities who could walk. Functional and continued use of such devices by students who are ambulatory may depend on their ability to locate their VOCA, because it is unlikely that they will have the device with them at all times.

Acknowledgments Appreciation is extended to Alonzo Andrews and staff at the Autism Treatment Center in San Antonio, Texas for their generous support and cooperation during the conduct of this study.

References
Beukelman, D. R., & Mirenda, P. (1998). Augmentative and alternative communication: Management of severe communication disabilities in children and adults (2nd ed.). Baltimore: Paul H. Brookes Publishing Co. Bondy, A., & Frost, L. (2001). The picture exchange communication system. Behavior Modification, 25, 725744. Duker, P. (1988). Teaching the developmentally handicapped communicative gesturing: A how-to-dobook. Berwyn, The Netherlands: Swets.

J. Sigafoos et al. / Research in Developmental Disabilities 25 (2004) 371383

383

Duker, P., Didden, R., & Sigafoos, J. (2004). One-to-one training: Instructional strategies for learners with developmental disabilities. Austin, TX: Pro-ed. Durand, V. M. (1999). Functional communication training using assistive devices: Recruiting verbal communities of reinforcement. Journal of Applied Behavior Analysis, 32, 247267. Keen, D., Sigafoos, J., & Woodyatt, G. (2001). Replacing prelinguistic behaviors with functional communication. Journal of Autism and Developmental Disabilities, 31, 385398. Mayer-Johnson Co. (1994). Picture Communication Symbols Combination Book. Solano Beach, CA: Author. Piaget, J. (1929). The childs conception of the world. New York: Harcourt Brace. Reichle, J., Beukelman, D. R., & Light, J. C. (Eds.). (2002). Exemplary practices for beginning communicators: Implications for AAC. Baltimore: Paul H. Brookes Publishing Co. Reichle, J., York, J., & Sigafoos, J. (1991). Implementing augmentative and alternative communication: Strategies for students with severe disabilities. Baltimore: Paul H. Brookes Publishing Co. Richards, S. B., Taylor, R. L., Ramasamy, R., & Richards, R. Y. (1999). Single subject research: Applications in educational and clinical settings. San Diego: Singular Publishing. Schepis, M. M., Reid, D. H., & Behrman, M. M. (1996). Acquisition and functional use of voice output communication by persons with profound multiple disabilities. Behavior Modification, 20, 451468. Schlosser, R. W., & Blischak, D. M. (2001). Is there a role for speech output in interventions for persons with autism? A review. Focus on Autism and Other Developmental Disabilities, 16, 170178. Schlosser, R. W., Blischak, D. M., Belfiore, P. J., Bartley, C., & Barnett, N. (1998). The effects of synthetic speech output and orthographic feedback on spelling in a student with autism: A preliminary study. Journal of Autism and Developmental Disabilities, 28, 319329. Sigafoos, J., Didden, R., & OReilly, M. (2003). Effects of speech output on maintenance of requesting and frequency of vocalizations in three children with developmental disabilities. Augmentative and Alternative Communication, 19, 3747. Sigafoos, J., & Drasgow, E. (2001). Conditional use of aided and unaided AAC: A review and clinical case demonstration. Focus on Autism and Other Developmental Disabilities, 16, 152161. Sigafoos, J., & Iacono, T. (1993). Selecting augmentative communication devices for persons with severe disabilities: Some factors for educational teams to consider. Australia and New Zealand Journal of Developmental Disabilities, 18, 133146. Soto, G., Belfiore, P. J., Schlosser, R. W., & Haynes, C. (1993). Teaching specific requests: A comparative analysis of skill acquisition and preference using two augmentative and alternative communication aids. Education and Training in Mental Retardation, 28, 169178. Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10, 349367.