83

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2004, 37, 8387 NUMBER 1 (SPRING 2004)

AN EVALUATION OF RESPONSE COST INTHE TREATMENT OF INAPPROPRIATE VOCALIZATIONS

MAINTAINED BY AUTOMATIC REINFORCEMENT

TERRY S. FALCOMATA, HENRY S. ROANE,ALYSON N. HOVANETZ, AND TRACY L. KETTERING

MARCUS INSTITUTE AND

EMORY UNIVERSITY SCHOOL OF MEDICINE

AND

KRIS M. KEENEYGWINNETT COUNTY PUBLIC SCHOOLS, GEORGIA

In the current study, we examined the utility of a procedure consisting of noncontingentreinforcement with and without response cost in the treatment of inappropriate vocali-zations maintained by automatic reinforcement. Results are discussed in terms of exam-ining the variables that contribute to the effectiveness of response cost as treatment forproblem behavior maintained by automatic reinforcement.

DESCRIPTORS: autism, automatic reinforcement, inappropriate vocalizations, non-contingent reinforcement, response cost

Automatic reinforcement, a reinforcementcontingency that exists independent of thesocial environment, presents numerous dif-ficulties to behavior analysts (see Vollmer,1994, for a discussion). Recent research onthe treatment of problem behavior main-tained by automatic reinforcement has fo-cused on the use of noncontingent access topreferred items (NCR; e.g. Shore, Iwata,DeLeon, Kahng, & Smith, 1997). WhenNCR is used in the treatment of automati-cally reinforced problem behavior, there aretwo competing sources of reinforcementavailable. As such, NCR may not always beeffective as a treatment for problem behaviormaintained by automatic reinforcement.However, if the reinforcer that is used in theNCR procedures is removed contingent on

We thank Michael Kelley for his helpful commentson an earlier version of this article.

Requests for reprints should be sent to Henry S.Roane, Marcus Institute, 1920 Briarcliff Road, Atlan-ta, Georgia 30329 (e-mail: roane@marcus.org).

the occurrence of the target behavior (i.e.,response cost), an increase in appropriate be-havior and a decrease in problem behaviormay be observed.

Keeney, Fisher, Adelinis, and Wilder(2000) recently evaluated a procedure inwhich noncontingent access to positive re-inforcers was used to treat problem behaviormaintained by negative reinforcement. Re-sults showed that NCR alone was not effec-tive as a treatment. However, when a re-sponse-cost procedure was implementedcontingent on problem behavior, inappro-priate responding decreased. Other studies(e.g., Mason & Iwata, 1990) have shownsimilar results for destructive behavior main-tained by attention (positive reinforcement).In the current investigation, we attempted toreplicate the results of the prior studies byevaluating the effects of NCR alone andNCR with response cost in the treatment ofproblem behavior maintained by automaticreinforcement.

Evan Schneiderman

Evan Schneiderman

Evan Schneiderman

Evan Schneiderman

Evan Schneiderman

Evan Schneiderman

84 TERRY S. FALCOMATA et al.

METHOD

Participant, Setting, and Data Collection

Derek, an 18-year-old man who had beendiagnosed with autism, was enrolled in afull-day (6-hr) program for the assessmentand treatment of multiple problem behav-iors including inappropriate vocalizations.All sessions were conducted in two roomsmeasuring 4 m by 5 m and were 10 min induration. The rooms contained only stimulithat were necessary for the condition in ef-fect (e.g., toys, chairs).

Observers who had been previouslytrained in behavioral observation collecteddata on laptop computers from behind aone-way mirror. During all analyses, dura-tion data were collected on inappropriate vo-calizations (defined as any vocalization thatwas not appropriate to the context; e.g.,singing, imitation of drumbeat or guitar)and on choice behavior (defined as Derekhaving his full body inside one of tworooms) during the choice analysis. Agree-ment coefficients for inappropriate vocaliza-tions and choice responding were calculatedby dividing the smaller duration (in seconds)by the larger duration and multiplying by100%. Reliability data for inappropriate vo-calizations were collected on 39% of func-tional analysis sessions and averaged 93.5%(range, 80% to 100%), 67% of treatmentanalysis sessions and averaged 94.5% (range,53.3% to 100%), and 100% of choice anal-ysis sessions and averaged 98.5% (range,96.9% to 100%). Reliability data for choiceresponding were collected during all choiceanalysis sessions and were always 100%.

Procedure

Functional analysis. A functional analysiswas conducted using procedures similar tothose described by Iwata, Dorsey, Slifer,Bauman, and Richman (1982/1994) toidentify the reinforcer that maintained De-reks inappropriate vocalizations. Following

the functional analysis, a phase of consecu-tive alone sessions was implemented to eval-uate further the occurrence of the target be-havior in the absence of programmed con-tingencies (Vollmer, Marcus, Ringdahl, &Roane, 1995).

Treatment analysis. The treatment analysisconsisted of a comparison of baseline andtwo treatment conditions, NCR and NCRplus response cost, in a reversal (ABCACBC)design. Baseline sessions were similar to thealone condition of the functional analysis,except that the therapist was present and apreferred item (a Walkman! radio; identifiedvia a preference assessment) was placed inthe room to control for its presence in latertreatment conditions. During baseline, De-rek was not allowed to use the radio eventhough it was present. During the NCRcondition, Derek was given continuous ac-cess to the radio, and the therapist providedno consequences for occurrences of the tar-get behavior.

Although levels of inappropriate vocali-zations decreased below baseline levels dur-ing NCR, the observed reduction was notclinically significant. Therefore, a response-cost component was added to the NCR con-dition. This condition was similar to theNCR condition except that Derek lost accessto the radio for 5 s contingent on inappro-priate vocalizations. At the end of the 5-sinterval, the radio was returned.

Choice analysis. Following the treatmentanalysis, an analysis was conducted to testthe hypothesis that Derek preferred the ra-dio relative to engaging in inappropriate vo-calizations. During the choice analysis, De-rek could enter one of two rooms. In theradio room, Derek had free access to theradio (the response-cost procedure was notin place), whereas in the sing room he wasable to sing (similar to the alone conditionof the functional analysis). Prior to each ses-sion, the therapist directed Derek to entereach room to expose him to the contingen-

85AUTOMATIC REINFORCEMENT

Figure 1. Average percentage of engagement in inappropriate vocalizations during the functional analysis(top panel) and treatment analysis (middle panel) and choice responding during the choice analysis (bottompanel). BL ! baseline.

86 TERRY S. FALCOMATA et al.

cies associated with each room. No therapistwas present in either room.

RESULTS AND DISCUSSION

The results of the functional analysis sug-gested that inappropriate vocalizations weremaintained by automatic reinforcement.This hypothesis was further supported be-cause the behavior persisted during the alonephase (Figure 1, top).

The middle panel of Figure 1 shows theresults of the treatment analysis. Dereks in-appropriate vocalizations averaged 99.1%during baseline and 55.6% during the sub-sequent NCR condition. The addition ofthe response-cost component produced animmediate reduction in inappropriate vocal-izations (M ! 1.2%) relative to NCR alone.Inappropriate vocalizations reemerged dur-ing the reversal to baseline (M ! 98.9%),decreased during NCR plus response cost(M ! 0.3%), and returned to prior levelsduring the reversal to the NCR condition(M ! 60.9%). During the final phase ofNCR plus response cost, inappropriate vo-calizations again decreased (M ! 1.6%).

During the choice analysis (Figure 1, bot-tom), Derek allocated the majority of his re-sponding to the radio room (M ! 98.2%)and did not enter the sing room. Thesedata suggest that Derek preferred listeningto the radio relative to singing. It should benoted that while Derek was in the radioroom he displayed low levels of inappropri-ate vocalizations (M ! 1.1%; data notshown).

In the current investigation, the use of re-sponse cost reduced the occurrence of au-tomatically reinforced problem behavior.These results are similar to the results of pre-vious investigations (e.g., Keeney et al.,2000; Mason & Iwata, 1990) and suggestthat for some individuals, the loss of high-preference stimuli may compete with en-

gagement in automatically reinforced prob-lem behavior.

Although the provision of an alternativereinforcer decreased problem behavior some-what, the removal of the alternative rein-forcer reduced problem behavior to near-zero levels. Apparently Derek preferred thestimulation derived from the radio relativeto that derived from inappropriate vocaliza-tions. Results of the choice analysis support-ed this hypothesis.

It was unexpected that Derek displayedlow levels of vocalizations during the choiceanalysis. Although these sessions were similarto the NCR condition of the treatment anal-ysis, rates of vocalizations were near zero inthe choice analysis. It is possible that the his-tory of exposure to the response-cost pro-cedure during the treatment analysis mayhave contributed to this effect.

A potential limitation of the current re-sults is that the specific reinforcing proper-ties of the inappropriate vocalizations werenot identified. For example, it is possiblethat the behavior was maintained by certainresponse products (e.g., auditory or physical)or a combination of products. In addition,the current investigation involved only 1participant and only one schedule of re-sponse cost (continuous). Future researchshould examine the effects of various re-sponse-cost schedules.

REFERENCES

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K.E., & Richman, G. S. (1994). Toward a func-tional analysis of self-injury. Journal of Applied Be-havior Analysis, 27, 197209. (Reprinted fromAnalysis and Intervention in Developmental Dis-abilities, 2, 320, 1982)

Keeney, K. M., Fisher, W. W., Adelinis, J. D., & Wil-der, D. A. (2000). The effects of response cost inthe treatment of aberrant behavior maintained bynegative reinforcement. Journal of Applied Behav-ior Analysis, 33, 255258.

Mason, S. A., & Iwata, B. A. (1990). Artificial effects

87AUTOMATIC REINFORCEMENT

of sensory-integrative therapy on self-injurious be-havior. Journal of Applied Behavior Analysis, 23,361370.

Shore, B. A., Iwata, B. A., DeLeon, I. G., Kahng, S.W., & Smith, R. G. (1997). An analysis of re-inforcing substitutability using object manipula-tion and self-injury as competing responses. Jour-nal of Applied Behavior Analysis, 30, 2141.

Vollmer, T. R. (1994). The concept of automatic re-inforcement: Implications for behavioral research

in developmental disabilities. Research in Devel-opmental Disabilities, 15, 187207.

Vollmer, T. R., Marcus, B. A., Ringdahl, J. E., &Roane, H. S. (1995). Progressing from brief as-sessments to extended experimental analyses in theevaluation of aberrant behavior. Journal of AppliedBehavior Analysis, 28, 561576.

Received March 20, 2003Final acceptance November 12, 2003Action Editor, Timothy Vollmer