JOURNAL OF APPLED BEHAVIOR ANALYSIS

EFFECTS OF CHOICE MAKING ON THE SERIOUS PROBLEM BEHAVIORS OFSTUDENTS WITH SEVERE HANDICAPS

KATnEE DomTHE MAY INSTITUTE

GLEN DUNLAPFLORIDA MENTAL HEALTH INSTITUTE, UNIVERSITY OF SOUTH FLORIDA

AND

VINCENr WINTERLNGUNIVERSITY OF KENTUCKY

This study assessed the impact of choice making on the serious problem behaviors of 3 studentswith severe autism and/or mental retardation. In the context of within-subject reversal designs, theresults showed consistently reduced levels of problem behaviors (e.g., aggression) when the studentswere given opportunities to make choices among instructional tasks and reinforcers. Additional datashowed no systematic differences in the rate of correct responding between the two conditions. Theresults are discussed in relation to the continuing search for effective, nonintrusive solutions to theoccurrence of serious problem behavior.DESCRIPTORS: severely handicapped, choice behavior, problem behavior, autistic children

A growing emphasis of research and practice hasbeen on the development of effective, nonintrusivetechniques for managing the serious problem be-haviors of persons with developmental disabilities(Homer et al., in press). Numerous authors haveargued in favor of restricting the use of invasiveprocedures and promoting the development ofpos-itive, educative approaches to behavior manage-ment (e.g., Evans & Meyer, 1985; Lovaas & Favell,

This investigation was supported by Cooperative Agree-ment G0087CO234 from the National Institute on Dis-ability and Rehabilitation Research. However, the opinionsexpressed herein do not necessarily reflect the opinions orpolicies ofthe supporting agency, and no official endorsementshould be inferred.

The authors are grateful to JeffAmero and Michele Taylorfor serving as teachers, and to Cynthia Dollman, Mark Ellison,Beth Brookfield, Jennifer Karr, Jenny Turner, Paul Reedy,Kit Hoffman, Jeffrey K. Withstandly, and Lynn FosterJohn-son for assistance in data collection and manuscript and figurepreparation. Appreciation is extended to Steven C. Luce andEdward K. Morris for their helpful comments.

Correspondence and requests for reprints should be sentto Kathleen Dyer, The May Institute, Box 703, 100 SeaView Street, Chatham, Massachusetts 02633, or to GlenDunlap, Department of Child & Family Studies, FloridaMental Health Institute, University ofSouth Florida, Tampa,Florida 33612-3899.

1987). It has been recognized by many that con-tinued progress in this area will rely on appliedresearch designed to delineate and document in-terventions that are both effective and respectful ofa person's dignity (Bannerman, Sheldon, Sherman,& Harchik, 1990; Dunlap, 1985; Homer et al.,in press).One rapidly expanding research emphasis has

been on the effects oflearners' preferences and choice-making opportunities (Guess, Benson, & Siegel-Causey, 1985; Houghton, Bronicki & Guess, 1987;Kishi, Teelucksingh, Zollers, Park-Lee, & Meyer,1988; Parsons & Reid, 1990; Shevin & Klein,1984). The success of procedures using choice andpreference for individuals with severe handicaps hasbeen documented in studies showing reductions insocial avoidance behavior (Koegel, Dyer, & Bell,1987), increases in spontaneous communication(Dyer, 1987; Peck, 1985), and improvements intask performance (Mithaug & Mar, 1980; Parsons,Reid, Reynolds, & Bumgarner, 1990). There hasalso been some suggestion that choice making mayresult in reductions in serious problem behaviors.For example, Dyer (1987) showed that when chil-dren with autism were given choices of preferred

515

1990, 239 515-524 NUMBER 4 (wiNm 1990)

KATHLEEN DYER et a1.

rewards, decreases were shown in stereotyped self-stimulatory behaviors that had been reduced pre-viously only with contingent restraint. This studyalso indicated that there were improvements in oth-er problem behaviors, but these indications camefrom global ratings of on-task behavior rather thanfrom direct observations of specific responses.Therefore, the purpose of the present experimentwas to expand the literature on choice making byfocusing explicitly on the serious problem behaviors(induding aggression and self-injury) of school-agestudents with severe handicaps. In this case, a choice-making package was implemented in which stu-dents were permitted to make selections of rewardsas well as the tasks and materials with which theywould be engaged.

METHOD

SubjectsThree children (Lori, Mary, and George) par-

ticipated in this experiment. All of the childrenexhibited high levels of serious disruptive behaviorsuch as aggression, self-injury, and tantrums. Thebehaviors resulted in placement in a residentialtreatment center for Lori and Mary. George hadalso been institutionalized because of his problembehaviors, but at the time of this study he wasliving at home and receiving assessment and train-ing services from a university-based program. Loriwas 5 years old, and Mary and George were 11.Lori was diagnosed as having a developmental delaywith autistic features, Mary as having pervasivedevelopmental delay with autistic features, andGeorge as having severe mental retardation. Onthe Vineland Social Maturity Scale, Lori and Marywere estimated to have social age scores of 1.8 yearsand 1.0 years, respectively. George received anadaptive behavior composite of 1.5 years on theVineland Adaptive Behavior Scales. Lori was non-verbal and used gestures and a small number ofsigns to communicate. Mary had a vocabulary ofapproximately 10 words that served requestingfunctions. George was primarily echolalic, but heoccasionally used phrases of two to four words toexpress needs and desires.

Before the experimental sessions were conducted,each child received pretraining on how to expresschoices. Lori and Mary were taught to point to apreferred object when presented with two objectsand the statement, "Show me which one you want."For George, the pretraining involved teaching himto discriminate such questions as "What do youwant to work on?" and to limit his selections tothe materials that were available and designated asoptions. Prior to the experiment, the children werealso exposed to sessions in which definitions of thedependent variables were developed and the teach-ers were trained to use the procedures.

SettingAll sessions were conducted with one-to-one

teacher-student ratios in rooms located on the cam-puses of the residential center or the university. Therooms contained at least one table, several chairs,instructional tasks, and videotape equipment. Ses-sions ranged in length from 10 to 20 min, withno more than four sessions per day and 5 daysbetween sessions.

Selection of Tasks and ReinforcersDuring all sessions in each experimental condi-

tion, three to four educational tasks were used foreach student. The students had demonstrated theability to perform each of the tasks during previouseducational activities. The tasks involved prevo-cational and preacademic skills and were selectedbecause they could be handled easily in table worksessions and because they were judged by the stu-dents' teachers to be of approximately equivalentpreference.

Also used in each session were three to five pre-ferred stimuli that were reported by the children'steachers to be functional reinforcers in other con-texts. These preferred stimuli were used as rein-forcers for correct performance on instructional tasks.The schedules of reinforcement, determined by theteachers, were variable-ratio schedules (e.g., VR 3,VR 5) that were maintained throughout all phasesof the experiment. Table 1 lists the tasks and re-inforcers available for each child.

516

CHOICE MAKING AND PROBLEM BEHAVIOR

Table 1Tasks and Reinforcers for Each Student in Each of the Experimental Conditions

Student Tasks Reinforcers

Lori Nine-piece puzzle SodaInserting sticks into a slot CrackersShape box M&M candies

Mary Seven-piece puzzle Teddy bearInserting pennies into a bank CookieInserting buttons into a slot Juice

George Large form puzzle CrackersStacking disks onto a spindle SodaSorting spoons by color JuiceLabeling picture cards- Candy

Potato sticks

Used only in the final choice condition.

Dependent VariablesThe primary dependent variable in this experi-

ment was the percentage of intervals that indudedinstances of serious problem behavior. Problem be-haviors were defined individually for each child andinduded aggression (biting, hitting, kicking, pinch-ing, and scratching) and object misuse (throwing,tearing, banging, and destroying objects) for all 3children, tantrums (screaming, whining, and cry-

ing) and bolting (moving more than 2 ft from thetable) for Lori and George, and self-injury (slappingthe face or body, punching self, head banging,elbow and wrist banging, and forcefully pressingobjects to the face) for Lori. For a self-hit or self-slap to be scored, it had to be forceful and beinitiated from a distance of more than 6 in. Mea-sures of problem behavior were obtained with a

30-s continuous-interval system of data recording.To obtain measures of task performance, correct

responses were counted and then translated intomeasures of rate for each task. To be scored as a

correct response, performance on the educationalmaterials had to conform to the task definitions.Responses that were prompted with physical assis-tance were not counted as correct. Response defi-nitions for each child's tasks are as follows:

Lori. A correct response on the puzzle was scoredwhen Lori placed one puzzle piece into its accom-

panying space on the template. The stick task in-volved inserting sticks into a small hole on top of

a can. A correct response for this task was scoredwhen one stick was put completely into the slot.A correct response on the shape box was scoredwhen Lori placed a block into its accompanyinghole in a container.

Mary. A correct response on the puzzle wasscored when Mary placed one puzzle piece into itsaccompanying space on the template. The buttontask involved inserting buttons into a small holeon top of a can, with a correct response being scoredwhen one button was put completely through theslot. A correct response on the penny task was scoredwhen Mary put a penny into a slot in a bank.

George. A correct response on the puzzle taskwas scored when a puzzle piece was inserted intoits appropriate place. A correct response on thestacking spindles was counted when a donut-shapeddisk was placed successfully on a spindle. Plasticspoons were sorted by color and a correct responsewas recorded when a spoon was placed in its ap-propriate receptade. Responses to the picture cardswere scored as correct when George accurately la-beled a picture upon request.

Reliability of Dependent VariablesReliability measures were obtained for each of

the dependent variables during the experiment orfrom videotape obtained throughout each experi-mental condition. Interobserver agreement on theoccurrence of problem behaviors was assessed for21% of the experimental sessions, and reliability

KATHLEEN DYER et al.

of the task performance data was assessed for 61%of the sessions. An agreement in the recording ofproblem behavior was counted when two indepen-dent observers scored a 30-s interval in an identicalmanner.

Interobserver agreement for problem behaviorwas calculated by dividing the number of agree-ments by the sum ofagreements plus disagreementsand multiplying the obtained quotient by 100.Percentage agreement for problem behavior was92% (range, 75% to 100%); occurrence reliabilitywas 90% (range, 71% to 100%); nonoccurrencereliability was 91% (range, 75% to 100%). Reli-ability for rate of correct responding was calculatedby dividing the smallest frequency count (obtainedby one observer) by the largest frequency count(obtained by the other observer) and multiplyingthe quotient by 100. Percentage agreement for rateof correct responding was 96% (range, 80% to100%).

Design and Experimental ConditionsTo demonstrate replicability of treatment effects,

reversal designs (Barlow & Hersen, 1984) wereused, with the order of conditions alternated acrosschildren. For each child, two conditions were pre-sented, a choice condition and a no-choice condi-tion. In the choice condition, the child was providedwith opportunities to choose from the availableselection of tasks and reinforcers. George indicatedhis selections verbally, and Mary and Lori indicatedtheir selections by pointing. Ifa task was completedduring a session, the child was asked to choose newmaterials. The children were also permitted to con-tinue work on the same materials if they chose.Also, if the child voluntarily requested a change inmaterials, such a change was permitted. Reinforcerswere also selected, but the opportunity to select areinforcer was provided in accordance with the pre-vailing reinforcement schedule.

In the no-choice condition, the same tasks andreinforcers were provided, but always according toan independent schedule and always at the teacher'sinitiation. To keep the teaching sessions as naturalas possible, the specific scheduling of tasks in someof the no-choice conditions was left to the discretionof the teacher. In other no-choice conditions, teach-

ers were given explicit instructions to maintain anequal balance of tasks within the sessions. Anyexpression of choice that occurred during this con-dition was followed by a brief explanation that theschedule must be followed (e.g., "We need to dothis other work now.").

In all experimental conditions, problem behav-iors were addressed according to guidelines set forthin each child's regular program. In general, problembehaviors were ignored whenever possible and ag-gressive responses were blocked in a protective man-ner. For all 3 children, physical prompts were usedoccasionally to continue instruction. For example,if the child engaged in excessive motor activity thatprevented attending to task-related instructions, heor she was prompted with verbal and occasionalphysical guidance to sit quiedy. In accordance withher ongoing habilitation plan, some ofMary's prob-lem behaviors were managed with additional con-tingencies. Specifically, Mary's aggression was fol-lowed by a brief regime of contingent exercise, andher instances of object misuse were followed by 5s of corrective positive practice. All procedures wereused in the same manner across all conditions inthe experiment.

Teachers and ObserversAll teachers had extensive experience in the use

of behavioral techniques with severely handicappedchildren, including a minimum of 1 year supervisedpracticum in clinic settings. The observers had ex-tensive backgrounds in the recording of operation-ally defined behavior of children with disabilities.Before any experimental data were recorded, eachobserver was trained to record each ofthe dependentvariables until interobserver agreement reached atleast 80% for three consecutive practice sessions.To control for the potential effects of experimenterbias, the teachers were naive with respect to theexperimental hypothesis for 59% of the experi-mental sessions.

RESULTS

The data points in Figure 1 show the percentageof intervals with problem behavior during eachsession for each experimental condition in the re-

CHOICE MAKING AND PROBLEM BEHAVIOR

versal analyses. Each child exhibited lower levels ofproblem behavior during the choice condition.When choices were first presented to Mary, herproblem behavior decreased to a low of 5% duringthe last session in the condition. A reversal to theno-choice condition resulted in an immediate in-crease in problem behavior, with an average of78%across the condition. During the subsequent choicecondition, Mary's problem behavior decreased to alow of 0%. The remaining two graphs reveal es-sentially the same effects for the other 2 children.That is, the choice condition always produced lowerlevels of problem behavior than did the no-choicecondition.

For George and Mary, a subset of their problembehavior (i.e., aggression) was considered to be amost urgent and severe problem (Lori did not dis-play aggression during the experiment). To assessthe potential effects ofchoice on George and Mary'saggressive behaviors, the videotapes were reviewed,and the data on aggressive behavior were separatedfrom the pool of problem behaviors for these 2children. The shaded portions of Figure 1 show theresults of this analysis. Both Mary and Georgeshowed higher levels of aggression in the no-choicecondition than in the choice condition. Mary's ag-gressive behavior decreased to 0 during the last twosessions of both choice conditions, and George'sdid not display any aggressive behavior during anyof the sessions in the choice condition.

Data for rate of correct unprompted respondingacross all experimental conditions are shown in Ta-ble 2. These data show no consistent differences inrate of responding across the conditions.

Data were also collected on the relative propor-tions of tasks and reinforcers that were presentedby the teachers in the no-choice condition and thosethat were presented (i.e., selected by the children)in the choice condition. These data are shown inTable 3. All of the children selected all of the tasksin the choice condition that were presented in theno-choice condition. Although there was some ev-idence of preference (e.g., Mary appeared to favorthe pennies and buttons over the puzzle), in generalthere were few consistent differences across tasksand conditions.

The data for reinforcer presentation and selection

are also presented in Table 3. All of the childrenselected all of the reinforcers in the choice conditionthat were presented in the no-choice condition.During the choice condition, however, a somewhathigher proportion of crackers and candy were se-lected by George and Lori, respectively.

In summary, the principal findings from thisinvestigation are that (a) the choice conditions al-ways produced lower levels of problem behaviorthan did the no-choice conditions, (b) during thechoice conditions, levels of serious aggressive be-haviors were lower for Mary and nonexistent forGeorge, and (c) there were no systematic differencesin the rate ofunprompted correct responding acrossboth conditions. In addition, it is important to notethat all of the children selected all of the tasks andreinforcers during the choice condition that werepresented by the teachers during the no-choice con-dition.

DISCUSSION

The results of this experiment add to the liter-ature by providing direct evidence that choice mak-ing can produce reductions in specific, objectivelymeasured problem behaviors. Although the mag-nitude of the differences was not always great, adear distinction was demonstrated by the reversaldesign. Additionally, whereas other studies usedpreference assessments that were conducted beforeeach session (Dyer, 1987; Koegel et al., 1987),this study showed that simply providing choices ofpreferred reinforcers and maintenance tasks duringthe session reduced problem behavior. Thus, it maynot be necessary to conduct daily preference as-sessments to achieve positive results. Indeed, suchassessments can be time consuming and, thus, im-practical for many practitioners in applied settings.The present procedure of making choices availableon a continuing basis was demonstrated to be ef-fective and may be more efficient than previouslyreported strategies.

It is important to note that although the pro-vision of choice-making opportunities systemati-cally influenced problem behavior, there were noeffects on the rate of responding on the instructionaltasks. Similar results were found by Cox (1988),

KATHLEEN DYER et al.

NO;CHOICE: CHOICEI 11!t_

MARY]

pd, - AGGRESSIVE BEHAVIOR

CHOICE NO CHOICE CHOICE

1I

Ia|0

520

CHOICE

'L\Ai

NO CHOICE

--1-

IC

oUU-I

0

Ed

aIL

le

*100.

80-

60

40 --

200

100

80

40

20

0

100

80

60

40

20

0 +

CHOICE MAKING AND PROBLEM BEHAVIOR

Table 2Rate of Unprompted Correct Responding Across Each Experimental Condition

Child Condition Task

Mary Pennies Buttons Puzzle

Choice 3.61 7.58 1.00No choice 5.30 6.00 0.20Choice 7.00 6.00

George Sorting Stacking Rubber Picturespoons spindles puzzles cards

No choice 6.36 10.49 7.89Choice 8.60 9.00 8.00No choice 6.70 10.47Choice 6.75 10.65 3.81

Lori Sticks Blocks Puzzle

Choice 5.49 4.13 1.67No choice 7.06 6.89 7.90Choice 8.07 6.40 8.60No choice 7.00 7.02 6.84Choice 9.19 6.75 9.55

who suggested that systematic differences in per- peted with the reinforcers provided by engaging informance across choice and no-choice conditions problem behavior (Dyer, 1987; Lovaas & New-might not occur in tasks that have been previously som, 1976).acquired by the study participants. Thus, it is im- It might be argued that similar results could beportant to evaluate the effects of this procedure on achieved by having the teachers deliver optimalnew skills. proportions of preferred tasks and reinforcers rather

Because the independent variable in this study than by providing choices. However, a number ofwas a package that permitted choices of both re- studies (e.g., Dyer, 1987; Green et al., 1988; Par-inforcers and tasks, the effects of each element on sons et al., 1990) have shown that teachers' selec-problem behavior cannot be separated. However, tions are not as effective in identifying reinforcersthe data on the children's selection of tasks and as is a procedure of systematic assessment of pre-reinforcers offer some suggestions. Specifically, be- ferred stimuli. Also, in the present study, Mary'scause Mary selected the puzzle at a relatively low teacher was naive with respect to the experimentalrate during the choice condition, it is possible that hypothesis during all but the last session, and Lori'sthis task was less preferred and that the problem teacher was naive throughout the entire experiment.behavior served an escape function during the no- Further, during Lori's sessions in the no-choice con-choice condition (Carr & Durand, 1985). In con- ditions, the proportions of tasks and reinforcers weretrast, George and Lori selected a particular rein- nearly equivalent. These controls reduce the poten-forcer at a relatively high rate during the choice tial for the beneficial effects of choice to be influ-condition. This suggests that for these children, the enced by teacher selection of preferred tasks andreinforcers provided by engaging in the task com- reinforcers. Additional studies should address this

Figure 1. Results of the reversal analyses measuring the influence of choice and no-choice conditions on the amount ofproblem behavior exhibited by 3 students with severe handicaps. The data points reflect the percentage of intervals withany problem behavior, and the shaded portions of the graphs for Mary and George depict those intervals with aggressiveresponding.

521

KATHLEEN DYER et al.

O-oO

. O. .1

*8 ON 0 WIN Ml%.e n -i 0

r- - C Nt O b 000-O - _n- n<,

CN V-* 3IU

.L\D0 e- (N 00 -- e(, -- N(N

0.

. . .

I VI v - VI% e(.

4 nen,: ,"

- %.Oef.' -q r~-r- qt

-0a0 r-'.t f.' -

. . . .

- -4 00 (N -,.zm lq n n n(A

4)4)

0000 0000 0

.0

ZuZu uZuzuI

522

o *5

0n

rc

0

* .

IA '.0

64)

.5V

4)

-0

0A

"0C:"0

UI

0H *

4)"" 0

0

4)

r.bof.44.

t

r.

4)

4)0

P4

H

0.

"0

"0

u

0* (U .

bo V

bo 'A

V0

4)

V).8u

4)

4)0

\0 cf% \0 0 'O VN cf) 00 V,. 0\" 00 C41% '-p 3 "o I% r- et-, .,q

. . . . . . . .

-M ('14 0 0N '..4 V., "

ce .. .

64

-L-d.nzP4 CL

CHOICE MAKING AND PROBLEM BEHAVIOR 523

issue by implementing a yoking procedure in whichequivalent proportions of tasks and reinforcers arepresented in the no-choice and choice conditions.

Another possible explanation for the results isthat the opportunity to choose tasks as well asreinforcers may have been a reinforcer in itself. Thishypothesis is supported by a study conducted byBrigham and Sherman (1973) in which childrenof normal development responded at higher ratesduring a condition in which they were allowed tochoose their reinforcers, as opposed to a conditionin which the experimenter selected the reinforcers.Because the reinforcers were the same in both con-ditions, the authors suggested that the opportunityto choose may have been as important as the re-inforcers provided. Along these lines, Monty, Gel-ler, Savage, and Perlmuter (1979) found that col-lege students exhibited improved performance in alearning task when they were offered an attractivechoice compared to a condition in which they wereoffered an unattractive choice. These authors sug-gested that the positive effects of choice-makingopportunities may be attributed to the extent towhich a subject perceives control in the situation.To control for possible sequence effects in this

investigation, the order of choice and no-choiceconditions was counterbalanced across children.However, it is noteworthy that George's aggressionincreased in the second no-choice condition. Thiseffect was also reported anecdotally by teachers fromdassrooms in which choice-making opportunitieswere introduced and then taken away (Dyer, Wil-liams, Santarcangelo, & Luce, 1987). Given theseobservations, it is important to investigate the po-tentially deleterious effects of withholding oppor-tunities for control (i.e., choice) after a period inwhich choice-making opportunities are provided.

The children who participated in this study ap-peared to have preference for specific tangible re-inforcers (George and Lori) or tasks (Mary). It isimportant to determine whether similar resultswould be found by students who were motivatedprimarily by social reinforcement. In addition, thereis a need for controlled studies examining the effectof choice making on problem behavior in naturalsettings. In this regard, pilot data collected by the

authors suggest that choice-making strategies canbe integrated into a variety of everyday activities(Dunlap, Dunlap, Clarke, & Robbins, 1990; Gar-ling, Carroll, Luce, & Dyer, 1987) and that theseoptions can effectively reduce levels of problembehavior.

In summary, this study suggests that choice-making options provide a simple strategy that canbe used to reduce serious problems exhibited bystudents with severe handicaps. The fact that theseresults contribute to a growing body of literaturethat stresses the importance of increasing personalautonomy for persons with severe handicaps indi-cates that this is an important area of future in-vestigation.

REFERENCES

Bannerman, D. J., Sheldon,J. B., Sherman,J. A., & Harchik,A. E. (1990). Balancing the right to habilitation withthe right to personal liberties: The rights of people withdevelopmental disabilities to eat too many doughnutsand take a nap. Journal ofApplied Behavior Analysis,23, 79-89.

Barlow, D. H., & Hersen, M. (1984). Single-case exper-imental designs. New York: Pergamon Press.

Brigham, T. A., & ShermanJ. A. (1973). Effects ofchoiceand immediac of reinforcement on single response andswitching behavior of children. Journal of the Experi-mental Analysis of Behavior, 19, 425-435.

Carr, E. G., & Durand, V. M. (1985). Reducing behaviorproblems through functional communication training.Journal of Applied Behavior Analysis, 18, 111-126.

Cox, S. (1988, December). The effects of choice-makingon the production rate of three tasks performed by fourstudents with severe handicaps. In E. Siegel-Causey(Chair), Choice-making as a means of empowermentsession: The role ofchoice-making in the lives ofpersonswith severe handicaps. Implications for teachers, re-searchers, and personnel preparation programs. Paperpresented at the meeting of the Association for Personswith Severe Handicaps, Washington, DC.

Dunlap, G. (1985). Review of An educative approach tobehaviorproblems.Journal ofthe Associationfor Personswith Severe Handicaps, 10, 237-238.

Dunlap, G., Dunlap, L. K., Clarke, M., & Robbins, F. R.(1990). Functional assessment and curricular revisionin solving the serious behavior challenges of a studentwith multiple disabilities. Manuscript submitted forpublication.

Dyer, K. (1987). The competition of autistic stereotypedbehavior with usual and specially assessed reinforcers.Research in Developmental Disabilities, 8, 607-626.

524 KATHLEEN DYER et al.

Dyer, K., Williams, L., Santarcangelo, S., & Luce, S. C.(1987, November). Training teachers to use "moti-vational" techniques: Effects on spontaneous commu-nication and severe disruptive behavior for studentswith severe handicaps. Poster presented at the meetingofthe Association for the Advancement ofBehavior Ther-apy, Boston.

Evans, I. M., & Meyer, L. H. (1985). An educative ap-proach to behaviorproblems. Baltimore: Paul H. Brookes.

Garling, S. C., Carroll, H., Luce, S. C., & Dyer, K. (1987,October). Reduction ofdisruptive mealtime behaviors:A comparison of choice-making options versus more re-strictive time-out contingencies. Poster presented at theannual Berkshire Conference on Behavior Analysis andBehavior Therapy, Amherst, MA.

Green, C. W., Reid, D. H., White, L. K., Halford, R. C.,Brittain, D. P., & Gardner, S. M. (1988). Identifyingreinforcers for persons with profound handicaps: Staffopinion versus systematic assessment ofpreferences.Jour-nal of Applied Behavior Analysis, 21, 31-43.

Guess, D., Benson, H. S., & Siegel-Causey, E. (1985).Concepts and issues related to choice-making and au-tonomy among persons with severe disabilities. journalof the Association for Persons with Severe Handicaps,10, 79-86.

Homer, R. H., Dunlap, G., Koegel, R. L., Carr, E. G.,Sailor, W., Anderson, J., Albin, R. W., & O'Neill, R.E. (in press). Toward a technology of "nonaversive"behavioral support. Journal of the Association for Per-sons with Severe Handicaps.

Houghton, J., Bronicki, G. J., & Guess, D. (1987). Op-portunities to express preferences and make choices amongstudents with severe disabilities in classroom settings.Journal of the Association for Persons with SevereHandicaps, 12, 18-27.

Kishi, G., Teelucksingh, B., Zollers, N., Park-Lee, S., &Meyer, L. H. (1988). Daily decision-making in com-munity residences: A social comparison of adults withand without mental retardation. American Journal onMental Retardation, 92, 430-435.

Koegel, R. L., Dyer, D., & Bell, L. K. (1987). The influence

of child-preferred activities on autistic children's socialbehavior. Journal of Applied Behavior Analysis, 20,243-252.

Lovaas, 0. I., & Favell, J. E. (1987). Protection for clientsundergoing aversive/restrictive interventions. Educationand Treatment of Children, 10, 311-325.

Lovaas, 0. I., & Newsom, C. D. (1976). Behavior mod-ification with psychotic children. In H. Leitenberg (Ed.),Handbook ofbehavior modification and behavior ther-apy (pp. 303-360). Englewood Cliffs, NJ: Prentice Hall.

Mithaug, D. E., & Mar, D. K. (1980). The relation be-tween choosing and working prevocational tasks in twoseverely retarded young adults. Journal of Applied Be-havior Analysis, 13, 177-182.

Monty, R. A., Geller, E. S., Savage, R. E., & Perlmuter, L.C. (1979). The freedom to choose is not always sochoice. Journal of Experimental Psychology, 37, 170-178.

Parsons, M. B., & Reid, D. H. (1990). Assessing foodpreferences among persons with profound mental retar-dation: Providing opportunities to make choices.Journalof Applied Behavior Analysis, 23, 183-195.

Parsons, M. B., Reid, D. H., Reynolds, J., & Bumgarner,M. (1990). Effects of chosen versus assigned jobs onthe work performance of persons with severe handicaps.Journal of Applied Behavior Analysis, 23, 253-258.

Peck, C. A. (1985). Increasing opportunities for socialcontrol by children with autism and severe handicaps:Effects on student behavior and perceived classroom cli-mate.Journal ofthe Association for Persons with SevereHandicaps, 10, 183-193.

Shevin, M., & Klein, N. K. (1984). The importance ofchoice-making skills for students with severe disabilities.Journal of the Association for Persons with SevereHandicaps, 9, 159-166.

Received June 6, 1989Initial editorial decision December 6, 1989Revision received February 19, 1990Final acceptance August 8, 1990Action Editor, Terry J. Page