JOURNAL OF APPLIED BEHAVIOR ANALYSIS

TEACHING CHILDREN WITH AUTISM TOUSE PHOTOGRAPHIC ACTIVITY SCHEDULES:MAINTENANCE AND GENERALIZATION OF

COMPLEX RESPONSE CHAINS

GREGORY S. MACDUFF, PATRICIA J. KRANTZ, ANDLYNN E. MCCLANNAHAN

PRINCETON CHILD DEVELOPMENT INSTITUTE

We used a graduated guidance procedure to teach 4 boys with autism to follow photographicactivity schedules to increase on-task and on-schedule behavior. The multiple baseline across par-ticipants design induded baseline, teaching, maintenance, resequencing of photographs, and gen-eralization to novel photographs phases. The results indicated that photographic activity schedules(albums depicting after-school activities) produced sustained engagement, and skills generalized toa new sequence of photographs and to new photographs. The acquisition of schedule-followingskills enabled these children with severe developmental disabilities to display lengthy response chains,independently change activities, and change activities in different group home settings in the absenceof immediate supervision and prompts from others.DESCRIPTORS: autism, prompting, stimulus control, photographic cues, generalization

A goal of behavioral intervention for people withautism is the development of functional skills thatmaximize engagement in appropriate self-care,work, and leisure activities. However, many inter-vention packages rely heavily on verbal instructions,modeling, and gestures. Because these prompts areoften associated with reinforcement during teach-ing, they may acquire stimulus control over targetresponses, with the result that learners may notdisplay target skills in the absence of teachers andprompting procedures. Although complex behav-ioral repertoires may be quickly established in atreatment setting, often they do not generalize orendure over time (see Dunlap & Plienis, 1988;Fowler, 1988; Sailor, Goetz, Anderson, Hunt, &Gee, 1988). Children with autism who have beentaught to play with toys, dust the furniture, andcomplete handwriting worksheets or other func-tional skills often fail to exhibit these responsesspontaneously, perhaps because stimulus controldid not transfer from prompts embedded in thetraining setting to naturally occurring stimuli (Bil-lingsley & Romer, 1983; Snell, 1983).

Reprints may be obtained from the authors, PrincetonChild Development Institute, 300 Cold Soil Road, Princeton,New Jersey 08540.

People with autism have also had difficulty ac-quiring lengthy response chains. Thus, although aboy may have learned to brush his teeth, get dressed,and go to breakfast, he may not complete thissequence without prompts to initiate each activityor each part of an activity. Some investigators haveused visual cues (pictures, photographs, or linedrawings) as sequenced discriminative stimuli thatenable participants to complete the steps in a com-plex task or to change tasks independently. Forexample, Wacker and Berg (1983) used photo-graphs of line drawings to teach adolescents withmoderate and severe retardation to complete com-plex assembly tasks consisting of 18 to 30 steps,andJohnson and Cuvo (1981) used pictorial recipesto help adults with developmental disabilities learncooking skills such as boiling, baking, and broiling.Others have used visual cues to help people withdevelopmental disabilities acquire self-care skills(Thinesen & Bryan, 1981), meal preparation andcooking repertoires (Martin, Rusch, James, Decker,& Trtol, 1982; Robinson-Wilson, 1977), com-puter use (Frank, Wacker, Berg, & McMahon,1985), and derical and laundry tasks (Wacker,Berg, Berrie, & Swatta, 1985). Sowers, Rusch,Connis, and Cummings (1980) taught adults withmental retardation to move independently from

89

1993, 269, 89-97 NUMBEEL 1 (SPRING 1993)

GREGORY S. MAcDUFF et al.

work to lunch or break and back to work; theirresearch is noteworthy because there have been fewexaminations of the use of picture cues to teachpeople to move independently from an activity inone setting to a different activity in another setting.

Although some investigations have shown thatthe use of visual cues can decrease dependence onteachers and enhance generalization and mainte-nance ofnewly acquired skills, most of these studieshave used intervention packages that combine pic-tures, line drawings, or photographs with otherprocedures. Research using visual cueing procedureshas not yet clarified which components of treatmentpackages are responsible for the reported results,or whether entire treatment packages are essential.Our investigation was designed to assess the effectsof a two-component teaching package (photo-graphs and graduated guidance) on participants'on-task and on-schedule behavior. Our goal is toextend the literature with 4 youths whose group-home intervention programs relied on verbal con-tacts (instructions, questions, and praise statements)by caregivers. Each youth received an average ofone or two verbal contacts per minute. Althoughthis rate of staff contacts resulted in very high levelsof appropriate engagement and low levels of ste-reotypic behavior, most responses were prompted,and previously taught skills were seldom sponta-neously displayed. Several attempts to fade verbalprompts had been abandoned because decreases inprompts were associated with increases in off-taskand disruptive behavior. In this context, we ex-amined the effects of photographic activity sched-ules, taught with graduated guidance, on the ac-quisition, maintenance, and generalization ofcomplex response chains that required the partici-pants to remain engaged and to move into differentsettings in their group home without prompts.

METHOD

ParticipantsThe 4 participants in this study were Mike and

Walter, age 9; Steve, age 11; and Roy, age 14.They met the DSM-III-R (APA, 1987) criteria forautism, and an independent diagnosis of autismhad been conferred by outside agencies before their

program enrollment. The boys' scores on the Pea-body Picture Vocabulary Test ranged from 2.1 to3.9 years (M = 3.2); Steve was unable to obtaina basal score. Age-equivalent scores on the VinelandAdaptive Behavior Scale were 5.5 for Mike andWalter, 5.3 for Roy, and 3.3 for Steve. Informedparental consent was obtained for each participant.

All participants had long histories of disruptivebehavior, including aggression, tantrums, and run-ning away, and all displayed high rates of stereo-typic behavior in the absence of structured pro-gramming. In addition, all had severe languagedeficits; they exhibited echolalia, vocal noise, non-contextual speech, and lack of spontaneous lan-guage. The boys were dependent on ongoing su-pervision and verbal prompts to complete self-help,housekeeping, and leisure activities. Prior to thisinvestigation, incidental teaching and discrete-trialteaching procedures were used to teach home-livingskills such as vacuuming, dusting, and table setting,as well as leisure skills such as using manipulativetoys, biking, and rollerskating. All participants hadacquired picture-object correspondence skills beforethe study began, and each had a limited experiencewith a photographic activity schedule that either(a) displayed the steps in preparing a bag lunchfor the next school day or (b) depicted the stepsnecessary to obtain a preferred beverage. None hadever used photographic cues that prompted leisureskills or that depicted a sequence of different ac-tivities.

SettingThe study was conducted in a community-based

Teaching-Family Model group home where theparticipants had resided for 1.1 to 4.2 years (M =

2.1 years). The home was staffed by live-in teachingparents (a married couple) and two other full-timetherapists. Like other Teaching-Family programs,the home was family-style and consumer-evaluated(McClannahan, Krantz, McGee, & MacDuff, 1984).Sessions were conducted in the living room, familyroom, and participants' bedrooms.

Photographic Activity SchedulesEach participant's photographic activity schedule

was displayed in a three-ring binder. Each binder

90

PHOTOGRAPHIC ACTIVITY SCHEDULES

contained six photographs (35 mm) depicting lei-sure and homework activities; each photo (7 cm

by 1 1.5 cm) was mounted in the center of a singlepage of white paper (21.5 cm by 28 cm) that was

inserted in a plastic page protector. Photographsdisplayed materials against a plain backgroundwithout distractors; for example, a picture of a

snack showed only a plate with two cookies.Initially, the first three photographs in Mike's

schedule were Colorforms ®, handwriting work-sheets, and Tinker Toys®0; Walt's first, second, andthird photos were handwriting worksheets, LincolnLogs®, and Colorforms®0; Roy's were Memory®game, Lego® blocks, and Cootie® game; andSteve's were Lego blocks, Colorforms ®, and Per-fection® . The three remaining activities in the boys'schedules were snack, puzzle, and TV in the fourth,fifth, and sixth positions, respectively. Some of theleisure materials depicted in the photographic ac-

tivity schedules were displayed on shelves abovethe youths' desks, some were located on their dress-ers, and others (e.g., the TV) were located severalrooms away from the boys' bedrooms in the familyroom.

Dependent VariablesOn-task. On-task was recorded if participants

were (a) visually attending to any appropriate playor work materials, (b) looking at their photographicschedules, (c) manipulating play or work materialsappropriately (i.e., as they were designed to beused), or (d) in transition from one scheduled ac-

tivity to another. Off-task was scored if they (a)used materials in a manner other than that for whichthey were designed, (b) manipulated but did not

visually attend to the materials, (c) engaged ininappropriate behavior (e.g., aggression, tantrums,

stereotypies), or (d) did not engage in activities or

use materials.On-schedule. On-schedule was scored if, at the

moment of observation, a participant was engagedin the activity depicted on the page to which hisactivity schedule was open. For example, if a boywas building with Lego® blocks and his notebookwas open to a page displaying a photograph ofLego blocks, on-schedule was scored. However,if he was using the blocks and the photographic

activity schedule showed handwriting worksheets,the observers scored on-task but off-schedule. Fur-ther, off-schedule was recorded if on-task criteriawere not met.

Independent VariablesVerbal contacts. These were defined as verbal

instructions, questions, or praise statements (e.g.,"Steve, look at your puzzle," "What color will youuse, Roy?" or "Good, Mike, you're looking at yourblocks!").

Gestures and gestural prompts. These promptsinduded all pointing, motioning, or nodding to-ward children or materials, as well as pointing tospecific toys, materials, or photographs that rep-resented the next tasks in a sequence. Thus, bothnonspecific gestures (e.g., pointing toward a child)and gestural prompts (e.g., pointing to the lastpiece to be placed in a puzzle) were scored. Thisbroad definition was used to identify any trainerbehavior that could potentially influence partici-pants' performance.

Manualprompts. Manual prompts were definedas orienting a youth's head toward materials, hand-over-hand prompts, and light touches such as thosethat occur when manual guidance is faded.

Measurement ProceduresDuring all sessions, independent observers used

a 60-s momentary time-sampling procedure to scoreon-task and on-schedule. Additional observers re-corded verbal contacts, gestures and gesturalprompts, and manual prompts with a 60-s partial-interval procedure.

Experimental DesignA multiple baseline design across participants

was used to assess the effects of photographic ac-tivity schedules on on-task and on-schedule be-havior during baseline, teaching, maintenance, re-sequencing of pictorial schedules, and generalizationto novel photographs.

Experimental ConditionsSessions were 60 min in duration. Prior to all

sessions, participants were seated on a bench in theliving room; their photographic activity schedules

GREGORY S. MAcDUFF et al.

were located on a table approximately 1 m fromthe bench and directly in front of them. Activityschedules and the depicted materials were presentduring all conditions. Sessions began when the pri-mary data collector gave the instruction, "Everyonelook at me; please find something to do." Thisstandard instruction was used throughout all phasesof the research, and no rewards were delivered bythe teacher.

Baseline. After the standard instruction was giv-en, no additional manual, gestural, or verbalprompts were delivered, and inappropriate behaviorwas ignored. The teacher (the first author) wasnever present during baseline.

Teaching use of pictorial schedules. Partici-pants successively entered the teaching condition.At first, the teacher stood next to the bench wherethe boys were seated, and after the primary observergave the initial instruction, the teacher waited 10s for a target child to stand up. If the participantdid not get up during this interval, the teacherplaced his hand on the boy's shoulder and manuallyguided him to his photographic activity schedule.If the child got up but did not move toward thephotographic schedule, the same prompt was de-livered. Graduated guidance, delivered from be-hind the youth, was used to help him complete thesequence of activities pictured in his activity sched-ule, in the order in which photographs were pre-sented. During initial sessions, the boy was man-ually prompted to pick up his notebook, carry itto his bedroom, open it, point to the first picture,gather the necessary materials, complete the activ-ity, put materials away, and turn the page to thenext activity. A child was manually prompted toput materials away and move on to the next activitywhen he (a) used all the available materials (e.g.,if 30 Lego® blocks were provided, he used allpieces) or (b) completed all items on his worksheets.If a participant progressed to the last scheduledactivity (TV) before 60 min had elapsed, he con-tinued this activity until the session ended. Theteaching procedure specifically prohibited verbalcontacts, gestures, and gestural prompts, and calledfor the preventive use of manual prompts to de-crease errors.

Manual prompts were always delivered from be-hind the youth. Graduated guidance was initiallyavailable for all tasks specified by the photographicschedule, but prompts were faded in frequency andintensity as rapidly as possible. Fading began bymoving from graduated guidance to spatial fading(i.e., the teacher changed the location of manualprompts). Subsequently, the teacher moved toshadowing-he followed the youth's movementswith his hands near the boy, but without makingphysical contact (Cooper, 1987; Foxx & Azrin,1973). However, if a boy engaged in inappropriatebehavior, or if he paused for an extended periodof time, prompts were reinstated.When a youth was scored as on-task and on-

schedule during at least 80% of time samples withshadowing, the teacher began to fade his physicalproximity. The teaching condition ended for a youthwhen he remained on-task and on-schedule duringat least 80% of time samples for five consecutivesessions after the teacher's physical proximity hadbeen faded. (Because Mike and Walter shared onebedroom and Roy and Steve shared another, theteacher was present in each bedroom until his prox-imity was faded for the second boy in each dyad.The teacher was present in the living room andfamily room until his proximity was faded for Steve.)

Maintenance. During maintenance, the teacherwas present to prompt the youth entering the teach-ing condition, but boys in maintenance received noprompts. For example, although the teacher stoodnext to the bench at the beginning of the sessionor was present in the room of a child who hadrecently completed training, he did not provide anyprompts to participants who had completed train-ing.

Resequencing photographic activity schedules.At the beginning of this condition, four of the sixoriginal activities in Mike's, Walter's, and Roy'sschedules were resequenced (i.e., all activities exceptsnack and TV were randomly assigned to new po-sitions in their schedules). All other aspects of theschedules remained constant, no prompts were de-livered, and the teacher was absent. This manip-ulation was performed to assess whether the boyswere using their schedules or following now-fa-

92

PHOTOGRAPHIC ACTIVITY SCHEDULES

miliar routines. Because of time constraints, Stevedid not participate in the resequencing condition.

Generalization. The teacher was absentthroughout this condition. During the generaliza-tion phase, two of the six original activities werereplaced with two similar but novel leisure activitiesfor each boy. In Mike's schedule, Colorforms ® andTinker Toys® were replaced by a marble game andcoloring; in Walt's, Lincoln Logs®0 and Color-forms® were replaced by Pipeworks®0 and Lego®0blocks; in Roy's, Memory® and Cootie® gameswere replaced by Perfection® and a peg game; andin Steve's schedule, Legions of Power®0 and Ra-magon® replaced Lego® blocks and Perfection ®.Although these new activities called on previouslyacquired skills, none had ever been the topic ofinstruction, nor had any of them been componentsof photographic activity schedules.

Interobserver AgreementInterobserver agreement data were obtained for

the dependent and the independent variables in atleast 30% of sessions across all conditions. Interval-by-interval percentage interobserver agreement wascalculated by dividing the number of agreementsby the number of agreements plus disagreementsand multiplying by 100%. Mean interobserveragreement on the occurrence of on-task was 96%(range, 0% to 100%), and mean agreement onnonoccurrence of on-task was 95% (range, 0% to100%). For on-schedule, mean agreement was 99%for both occurrence and nonoccurrence, and rangesfor both were 98% to 100%. No verbal or gesturalprompts were ever scored; interobserver agreementon nonoccurrence was 100%. Mean agreement onthe occurrence of manual prompts was 99% (range,50% to 100%), and mean agreement on nonoc-currence of manual prompts was 99% (range, 99%to 100%).

RESULTS

On-TaskFigure 1 shows levels of on-task for the 4 par-

ticipants across all conditions. During baseline, theboys, with the exception of Steve, displayed con-

siderable variability across sessions; on-task was al-most never scored for Steve.

With each application of teaching pictorialschedules, on-task immediately increased for allyouths. Mean on-task performance was 99% forMike, Walt, and Roy, and 97% for Steve. Duringmaintenance, resequencing of photographs, andgeneralization to novel photographs, all youths dis-played high and stable on-task performances. Meansfor Mike and Walt across these three conditionswere 99%; Roy's on-task means were 97% in main-tenance, 96% in resequencing, and 97% in gen-eralization. Finally, Steve's on-task averaged 91%in maintenance and 96% during generalization; hedid not participate in the picture-resequencing phase.

On-ScheduleFigure 2 reports the participants' on-schedule

data during all experimental conditions. Duringbaseline, on-schedule was never scored for any ofthe youths. Teaching pictorial schedules producedon-schedule means of 99% for Mike, Walter, andRoy, and 96% for Steve. In maintenance, on-sched-ule means were 98%, 99%, 97%, and 91% forMike, Walter, Roy, and Steve, respectively. Fur-ther, the boys continued to use their schedules afterthe photographs were resequenced, obtaining meanon-schedule scores of 97% (Mike), 99% (Walter),and 95% (Roy). The presentation of novel pho-tographs and materials during the generalizationphase produced on-schedule means of 99% forMike and Walter, 97% for Roy, and 96% for Steve.

Independent VariablesDuring baseline, no verbal contacts, gestures or

gestural prompts, or manual prompts were record-ed. In the first five sessions of the teaching phase,the percentage of intervals scored for the occurrenceof manual prompts ranged from 3% to 8% forMike (M = 4%), 0% to 22% for Walt (M =8%), 20% to 40% for Roy (M = 23%), and 8%to 37% (M = 19%) for Steve. The last five trainingsessions for each participant contained no manualprompts, and no verbal contacts, gestures, or ges-tural prompts were scored throughout this condi-tion. No prompts of any type were scored during

GREGORY S. MAcDUFF et al.

Un

a) u -,a)

cn (v4-

r

CUa u_co 4) 2-

e

I .. ... ... I ....

I- IMW

I II II II II II II II II II I

a)

C:

C:4C

r-

C

U) U =

2) V)A& - -0-146-40" %.a- - - %

On-Task

ManualPrompts

84

36

-_ -_ _ _ __-_-

96

148 60 72 84 i 96 10I- - - - - - - - --I I I

~~~~~~I I I

I I I II I I I

I I I I

I I I I

I_ _ _ _ _ _ I I

lX l l l~~~~~~~~I Il[

* j j : - it j j z s ^: I

I ~~~~I148 60 72 84 1 96 1 10

I I

I _ _ _ _ _ _ _

____ _ _ _ III

I I I~~~~~~~~~~~~I I II I I

I II I I

I.I

IA-.. A....

0 12 244

36 48 60 72 84 96

SESSIONS

Figure 1. Percentage of 60-s momentary time samples scored as on-task, and percentage of 60-s intervals scored formanual prompts during 1-hr sessions.

94

c

4Ji

N)C-a)C

80

60

40

20

80

60

40

20

108<H

I -

zo0oc(1)0.

<Ix< cr0LL

U.

c 0o LLwuc(1)0

u

LI

0L-J

LLQF

Hz

Ucr

0 L

'8

~~~~~~~~~~~I

II

100-

80-

60-

40 -

20

)8

108

r- I

o I-I I

. 1L..... .. ....

-Ir - - 111IIIIIIIIII

it -5IIII

r-IIIIII

-I Ar - iazr

IIIIIII

IIIIIII

IIIIII

-A as---D&A-

PHOTOGRAPHIC ACTIVITY SCHEDULES

C__ _E L.a) U4)_, en a)-

I-_

cna)

a)

U-)

CY) 0

4)~ 4)

C _ r0 m~~~

4-,D , aL)

_-n r cm 4c - u WT- c c- (.n C.D

0 1 2 241 36 48 60 72 84 1 96 g 108L I

O - 1A~10".jl I ..JaL". A

D , , , p~~~~~~~~~~~ALTER

O- _ _ _ _ _ _ _ __._ _ _ __._ _

0 1 2 241 36 1 48 ----60 72 84 96 108

O-6 AAA !, II I I,I

00-

o ~~~~I I Im

*

0

0 1 2 24 36 48 60 72 84 96 108

SESSIONSFigure 2. Percentage of 60-s time samples scored as on-schedule during 1-hr sessions.

95

I I

I I

I I

I I

I I

I I

lOC

80

60

40

20LLJ

I

Z 8C0(O~ 6C

Ar

ILz 2(0

,4)

(nLL

Q 104

< 8(

(I-)

iili 64

LL 210

LiiIn I

zLiU 10

W 8C

0~

6'

44

21

n-

IIIIIIIII

AAAAAAAAAAAA.A

IIIIIIII

IhAdw AAM "A 11 664 ri-di -aia 4 A 4 bm" A 4 AA 04" p Mfi&AA AAr - - V - la A- i to -

l II I I I

GREGORY S. MAcDUFF et al.

maintenance, resequencing, or generalization phases.The boys completed these sessions without assis-tance or guidance, and from Session 90 on, theteacher was no longer present.

DISCUSSION

Clinical data indicated that the 4 participantshad learned to do a variety of recreational andhome-living tasks, but baseline data showed thatthe presence of leisure and homework materials didnot prompt sustained engagement. During theteaching condition, however, participants met cri-terion in only 13 to 27 sessions; subsequently, theycontinued to be on-task and on-schedule withoutprompts from the teacher, and their schedule-fol-lowing skills generalized to new sequences of ac-tivities and to novel photographs and materials withno additional training. The photographic schedulesenabled the boys to display lengthy and complexchains of previously mastered, functional behavior.Anecdotally, it also appeared that when followingtheir schedules, they engaged in fewer aberrant be-haviors; this may be a topic for future research.

The present study extends previous investiga-tions of pictorial and photographic cues in severalways. First, prompts were specified and measured.Most earlier studies used training packages thatincluded multiple prompting procedures, makingan analysis of the effects of specific classes of trainingstimuli difficult or impossible. Only one other studyof visual cues measured all of the prompts used intraining (Johnson & Cuvo, 1981); our researchisolated the effects of graduated guidance. Our in-vestigation also measured another variable, on-schedule, that has been frequently discussed butseldom assessed. Several other researchers have re-ported that participants eventually discontinued ormodified their use of photographs or pictures. Forexample, Thinesen and Bryan (1981) reported thatafter 1 week of training, all 15 participants stoppedpaging through their albums, using only the firstor a randomly selected photograph to complete anentire grooming sequence. Connis (1979) notedthat subjects eventually stopped checking each pic-ture before changing work tasks. In our study, the

resequencing and generalization conditions estab-lished that the photographs were relevant discrim-inative stimuli, and that participants were not mere-ly engaging in familiar routines.

The participants' consistent use of schedules maybe related to the training procedures. The trainerensured that the boys turned single pages, pointedto photographs before beginning tasks, obtainedthe depicted materials, and completed scheduledactivities in sequence. Graduated guidance was usedto prevent errors, lengthy delays that might haveinterrupted response chains, and stereotypies thatmight otherwise have become embedded in re-sponse chains. The boys continued to demonstratelook-then-do sequences throughout the study.

The nature of the target tasks may also influencethe posttraining use of visual cues. People whoengage in one repetitive task (e.g., assembly orpackaging tasks) may cease to use visual promptsbecause tasks are familiar and unchanging. Butactivities that are frequently resequenced (e.g., gro-cery shopping or work assignments in an industrialkitchen) may require continued reference to pic-torial cues.

It is interesting that in the current investigation,boys not yet in training were never observed toimitate the schedule-following behavior of those inthe teaching phase. Further, no resistance to manualguidance was noted. After training, off-task wasoccasionally scored, but children were not observedto engage in functional activities other than thosecued by their photographic schedules. The relativelysmall number of manual prompts required to teachthe participants to follow schedules may have beenrelated to (a) their prior mastery of picture-objectcorrespondence skills, (b) their previous acquisitionof the types of skills needed to complete depictedtasks, and (c) the fading of graduated guidance.The manual prompting procedure was selected be-cause it was hoped that nonverbal prompts deliv-ered from behind a youth would prevent perfor-mance from becoming dependent on the teacher'spresence. Although least-to-most prompt sequencesare frequently used with participants such as these,the most-to-least sequence described above was ef-fective in preventing errors that might otherwise

96

PHOTOGRAPHIC ACTIVITY SCHEDULES 97

have become embedded in schedule-following re-sponse chains.

Before this investigation began, the 4 partici-pants were dependent upon verbal prompts to re-main constructively engaged in appropriate activ-ities. In the absence of verbal prompts fromsupervising adults, it appeared that stimulus controltransferred to photographs and materials that wereavailable in the group home. When the study end-ed, all 4 boys were able to display complex home-living and recreational repertoires for an hour, dur-ing which time they frequently changed tasks andmoved to different areas of their group home with-out adults' prompts. Photographic activity sched-ules, taught with graduated guidance, became func-tional discriminative stimuli that promoted sustainedengagement after training ceased and fostered gen-eralized responding to new activity sequences andnovel leisure materials.

REFERENCES

American Psychiatric Association. (1987). Diagnostic andstatistical manual of mental disorders (3rd ed. rev.).Washington, DC: Author.

Billingsley, F. F., & Romer, L. T. (1983). Responseprompting and the transfer of stimulus control: Methods,research, and a conceptual framework. Journal of theAssociation for the Severely Handicapped, 8, 3-12.

Connis, R. T. (1979). The effects of sequential pictorialcues, self-recording, and praise on the job task sequencingof retarded adults. Journal of Applied Behavior Anal-ysis, 12, 353-362.

Cooper, J. 0. (1987). Stimulus control. In J. 0. Cooper,T. E. Heron, & W. L. Heward (Eds.), Applied behavioranalysis (pp. 298-326). Columbus, OH: Merrill.

Dunlap, G., & Plienis, A. J. (1988). Generalization andmaintenance of unsupervised responding via remote con-tingencies. In R. H. Homer, G. Dunlap, & R. L. Koegel(Eds.), Generalization and maintenance: Life-stylechanges in applied settings (pp. 121-142). Baltimore:Paul H. Brookes.

Fowler, S. A. (1988). The effects of peer-mediated inter-ventions on establishing, maintaining, and generalizingchildren's behavior changes. In R. H. Horner, G. Dunlap,& R. L. Koegel (Eds.), Generalization and mainte-nance: Life-style changes in applied settings (pp. 143-170). Baltimore: Paul H. Brookes.

Foxx, R. M., & Azrin, N. H. (1973). Toilet training theretarded: A rapid program for day and nighttime in-dependent toileting. Champaign, IL: Research Press.

Frank, A. R., Wacker, D. P., Berg, W. K., & McMahon,C. M. (1985). Teaching selected microcomputer skillsto retarded students via picture prompts. Journal ofApplied Behavior Analysis, 18, 179-185.

Johnson, B., & Cuvo, A. (1981). Teaching mentally re-tarded adults to cook. Behavior Modification, 5, 187-202.

Martin, J., Rusch, F., James, V., Decker, P., & Trtol, K.(1982). The use of picture cues to establish self-controlin the preparation of complex meals by mentally retardedadults. Applied Research in Mental Retardation, 3,105-119.

McClannahan, L. E., Krantz, P. J., McGee, G. G., & MacDuff,G. S. (1984). Teaching-Family Model for autistic chil-dren. In W. P. Christian, G. T. Hannah, & T. J. Glahn(Eds.), Programming efective human services: Strategiesfor institutional change and client transition (pp. 383-406). New York: Plenum.

Robinson-Wilson, M. A. (1977). Picture recipe cards asan approach to teaching severely and profoundly retardedadults to cook. In M. E. Snell (Ed.), Systematic instruc-tion of the moderately and severely handicapped (pp.69-73). Columbus, OH: Merrill.

Sailor, W., Goetz, L., Anderson, J., Hunt, P., & Gee, K.(1988). Research on community intensive instructionas a model for building functional, generalized skills. InR. H. Homer, G. Dunlap, & R. L. Koegel (Eds.), Gen-eralization and maintenance: Life-style changes in ap-plied settings (pp. 67-98). Baltimore: Paul H. Brookes.

Snell, M. E. (1983). Implementing and monitoring theIEP: Intervention strategies. In M. E. Snell (Ed.), Sys-tematic instruction ofthe moderately and severely hand-icapped (pp. 113-145). Columbus, OH: Merrill.

Sowers,J., Rusch, F., Connis, R., & Cummings, L. (1980).Teaching mentally retarded adults to time-manage in avocational setting. Journal of Applied Behavior Anal-ysis, 13, 119-128.

Thinesen, P., & Bryan, A. (1981). The use of sequentialpicture cues in the initiation and maintenance of groom-ing behaviors with mentally retarded adults. Mental Re-tardation, 19, 246-250.

Wacker, D. P., & Berg, W. K. (1983). Effects of pictureprompts on the acquisition of complex vocational tasksby mentally retarded adolescents. Journal of AppliedBehavior Analysis, 16, 417-433.

Wacker, D. P., Berg, W. K., Berrie, P., & Swatta, P. (1985).Generalization and maintenance of complex skills by se-verely handicapped adolescents following picture prompttraining. Journal of Applied Behavior Analysis, 18,329-336.

Received July 15, 1991Initial editorial decision September 23, 1991Revisions received December 23, 1991; March 3, 1992;

October 20, 1992Final acceptance November 9, 1992Action Editor, F. Charles Mace