ERIC EJ750839: Experimental Analysis and Treatment of Multiply Controlled Problem Behavior: A Systematic Replication and Extension

2006·0.14 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2006, 39, 375–379 NUMBER3 (FALL2006)
EXPERIMENTAL ANALYSIS AND TREATMENT OF MULTIPLY
CONTROLLED PROBLEM BEHAVIOR: A SYSTEMATIC REPLICATION
AND EXTENSION
CARRIE S. W. BORRERO AND TIMOTHY R. VOLLMER
UNIVERSITYOFFLORIDA
Weevaluatedinterventionsdesignedtoreducemultiplycontrolledproblembehaviorexhibited
by a young boy with developmental disabilities, using a multiple baseline design. Each
intervention was designed to address a specific social function of problem behavior. Results
showedthattheseparateinterventionswereusefulinreducingproblembehavior,andterminal
schedules were reached by way of schedule thinning (attention condition) and delays to
reinforcement (tangible andescapeconditions).
DESCRIPTORS: aggression, disruption, functional analysis, multiple control
_______________________________________________________________________________
Previous research has shown that problem threesourcesofsociallymediatedreinforcement
behaviormaybesensitivetomultiplesources of identified via functional analysis. We also
reinforcement (Smith, Iwata, Vollmer, & attemptedtothinthescheduleofreinforcement
Zarcone, 1993). Smith et al. evaluated the for each treatment to address the difficulty of
self-injurious behavior (SIB) of 3 individuals implementation.
and determined that SIB was sensitive to both
socially mediated positive reinforcement and
METHOD
reinforcement that was not socially mediated
Participant and Setting
(automatic). Next, interventions were evaluated
Walsh was a 7-year-old boy who had been
in a combined multiple baseline and reversal
diagnosed with mental retardation and who
design for each function of SIB. For 2
engaged in aggression (i.e., hitting and kicking
participants, results of the treatment suggested
others, spitting, and biting) and disruption
that SIB was multiply controlled, whereas
(throwingobjectsandpropertydestruction).He
results for the remaining participant suggested
engaged in unprompted vocalizations and
that SIB was sensitive to only one source of
communicated using full sentences. He was
reinforcement. Despite these results, the gener-
admitted toan inpatienthospitalfacility forthe
ality of this finding may be limited in that all
assessment and treatment of problem behavior.
participants displayed behavior that was main-
Allsessionswereconductedbytrainedresearch-
tained, in part, by automatic reinforcement.
ers on the hospital unit. The unit consisted of
Thus,theapproachusedbySmithetal.hasnot
individual bedrooms as well as a community
yet been applied to multiple sources of socially
living room and kitchen area. All sessions were
mediated reinforcement.
conducted 4 days per week, two to three times
The purpose of the current investigation was
each day, and were 10 min in duration.
to replicate the procedures of Smith et al.
(1993) by evaluating separate interventions for
Response Measurement and Reliability
Frequency data were collected on problem
WethankJohnBorreroforhishelpfulcommentsonan
earlierversionof thismanuscript. behavior (i.e., aggression and disruption),
Address correspondence to Timothy R. Vollmer, De- compliance with instructions (defined as com-
partmentofPsychology,UniversityofFlorida,Gainesville,
pletion of an instruction following a verbal or
Florida32611(e-mail: [email protected]).
doi:10.1901/jaba.2006.170-04 modeled prompt), and appropriate requests
375
376 CARRIE S. W. BORRERO and TIMOTHY R. VOLLMER
(e.g., saying, ‘‘Can I play with that?’’). In- function consisted of noncontingent attention
terobserveragreementwascalculatedfor67%of (NCA) in which attention was initially de-
functional analysis sessions, 30% of baseline liveredcontinuouslywhileproblembehaviorno
sessions, and 49% of treatment sessions. longer resulted in adult attention (i.e., extinc-
Agreement was calculated by dividing each tion). The schedule of reinforcement during
session into 60 intervals (10 s each), dividing NCA was thinned from a continuous schedule
the smaller number of responses by the larger ofattentiontoaterminalscheduleoffixed-time
number of responses within each interval, (FT) 5 min using procedures similar to those
averaging these scores across all intervals, and described by Marcus and Vollmer (1996).
convertingtheresultingproducttoapercentage. During the tangible baseline condition, brief
Agreement for problem behavior averaged 90% access to preferred tangible items was presented
(range, 67% to 100%) during the functional for 30 s contingent on problem behavior. The
analysis, 88% (range, 73% to 95%) during the treatmentforthetangibleconditionconsistedof
baseline analysis, and 99% (range, 81% to differentialreinforcementofalternativebehavior
100%) during the treatment analysis. Agree- (DRA), in which problem behavior no longer
ment averaged 98% (range, 95% to 100%) for produced access to items (extinction) and all
appropriaterequestsfortangibleitemsand98% appropriate requests resulted in 30-s access to
(range, 89% to 100%) for compliance during preferred items. During the tangible treatment,
baseline, and 98% (range, 75% to 100%) for thedelaytoreinforcementwasincreasedfromno
appropriaterequestsfortangibleitemsand98% delay (0 s) following instances of appropriate
(range, 78% to 100%) for compliance during verbal behavior to a 5-min delay, using proce-
the treatment analysis. dures similar to those described by Vollmer,
Borrero, Lalli, and Daniel (1999).
Procedure
Finally, in the escape baseline condition,
Functional analysis. Prior to the functional escape from instructional demands was pro-
analysis, a preference assessment was conducted vided for 30 s contingent on problem behavior.
to identify preferred stimuli for inclusion in During the treatment for the escape function,
subsequentsessions(Roane,Vollmer,Ringdahl, problem behavior no longer resulted in a 30-s
& Marcus, 1998). Next, a functional analysis break from instructional demands (extinction),
wasconductedusingproceduressimilartothose and a DRA component was implemented in
describedbyIwata,Dorsey,Slifer,Bauman,and which each instance of compliance resulted in
Richman (1982/1994). During the functional a 30-s break. During the 30-s break, Walsh
analysis, four test conditions (attention, escape, typicallyattemptedtoinitiateconversationwith
alone,andtangible)werecomparedtoacontrol the therapist and to request preferred tangible
condition in a multielement design. items. Thus, he also received access to tangible
Treatment. Following the functional analysis, items and attention during the 30-s break from
separateinterventionsweredevelopedtoaddress demands. During the escape treatment, the
each maintaining reinforcement contingency. schedule of reinforcement for compliance was
The treatments were introduced sequentially subsequently thinned from a fixed-ratio (FR) 1
using a multiple baseline design. All baseline to a variable-ratio (VR) 3 using procedures
conditions were identical to the relevant test similar to those described by Lalli et al. (1999).
conditions from the functional analysis. During
the attention baseline condition, a brief repri-
RESULTS AND DISCUSSION
mand (e.g., ‘‘Don’t do that; you’re hurting
me.’’) was delivered contingent on problem During the functional analysis, high rates of
behavior. The treatment for the attention problem behavior were observed during the
EXPERIMENTAL ANALYSES AND MULTIPLY CONTROLLED BEHAVIOR 377
attention (M 5 3.9 responses per minute), 0.7 responses per minute) compared to baseline
escape (M 5 6.2), and tangible (M 5 3.1) levels (M 5 3.7), and compliance increased
conditions, compared to the alone (M 5 0.9) from baseline (M 5 20%) to treatment (M 5
and control (M 5 0.05) conditions, suggesting 81%). Although problem behavior was more
behavioral sensitivity to adult attention, escape variable in the escape-based treatment than in
from demands, and access to tangible items the other treatment conditions, the schedule
(data not presented in Figure 1; these were requirement for escape was successfully in-
reported previously in Borrero, Vollmer, Bor- creased from an FR 1 (Session 42) to a VR 3
rero, & Bourret, 2005). (Session 140).
Figure 1 shows the results of the treatment Hanley, Iwata, and McCord (2003) reported
evaluation for the attention, tangible, and that as many as 15% of functional analyses
escape conditions. The baseline phases include revealed multiple control of problem behavior,
sessions from the pretreatment functional which suggests that interventions designed to
analysis. During the attention analysis, NCA address multiply controlled problem behavior
plus extinction resulted in an immediate de- represent an important area for investigation.
crease in rates of problem behavior (M 5 0.2 From a practical perspective, treatment of
responses per minute), which remained low as multiply controlled problem behavior may
thereinforcer–reinforcerintervalincreasedfrom be difficult for caregivers to implement
2 s (Session 61) to 5 min (Session 120). because separate interventions for each re-
Walsh’s mother served as therapist during one inforcer must be implemented with regard to
session(Session122),inwhichratesofproblem the specific establishing operation in place.
behavior remained low. Therefore, in the current investigation, separate
Prior to the tangible treatment analysis, two interventions were developed for each function
additional baseline sessions were conducted while the parameters of each intervention (e.g.,
after the initial functional analysis sessions. delays and response requirement) were thinned
The intervention resulted in a decrease in to enhance the overall practicality of their
problem behavior (M 5 0.2 responses per implementation.
minute) relative to baseline (M 5 2.8). Results of this analysis are similar to those
Appropriate requests for items increased in the reported by Smith et al. (1993) in that
treatment condition (M 5 0.9) relative to systematic decreases in problem behavior were
baseline (M 5 0.2), and problem behavior observed when treatments were introduced for
remained low as the delay to reinforcement was the separate reinforcers that maintained prob-
increased from 5 s (Session 80) to 5 min lem behavior. In addition, the current in-
(Session 139). As expected, the rate of requests vestigation extends the work of Smith et al. by
decreased slightly throughout the analysis, suggesting the effectiveness of the assessment
because there were fewer opportunities to model when applied to multiple sources of
request items during delay thinning. Walsh’s socially mediated reinforcement.
mother conducted one session (Session 122) The multiple baseline evaluation conducted
with no delay to reinforcement, and problem in the current study was somewhat unique in
behavior did not occur. that the independent variable (treatment) was
In the escape treatment evaluation, nine introduced sequentially across functions rather
baseline sessions were conducted in addition than participants, settings, or response topog-
to the initial functional analysis sessions. The raphy. Typically, multiple baseline designs
implementation of DRA plus extinction re- include the same independent variable across
sulted in a decrease in problem behavior (M 5 participants, settings, responses, or functions.
378 CARRIE S. W. BORRERO and TIMOTHY R. VOLLMER
Figure1. Resultsofthetreatmentanalysisfortheattentioncondition(top),tangiblecondition(middle),andescape
condition(bottom).Responsesperminuteofproblembehaviorareshownontheyaxisonallpanels,andthepercentage
ofcompliance is shown onthe secondary yaxis for the escapecondition.
EXPERIMENTAL ANALYSES AND MULTIPLY CONTROLLED BEHAVIOR 379
By contrast, in the current investigation, Iwata,B.A.,Dorsey,M.F.,Slifer,K.J.,Bauman,K.E.,
& Richman, G. S. (1994). Toward a functional
different interventions were implemented
analysis of self-injury. Journal of Applied Behavior
(NCA and two types of DRA contingencies) Analysis, 27, 197–209. (Reprinted from Analysis and
across the respective baselines. Although these Intervention in Developmental Disabilities, 2, 3–20,
1982)
interventionscanbeconsideredsimilarinterms
Lalli, J. S., Vollmer, T. R., Progar, P. R., Wright, C.,
of being based on the results of a functional
Borrero,J.,&Daniel,D.,etal.(1999).Competition
analysis, they involved the delivery of different between positive and negative reinforcement in the
reinforcers andcontingencies and thusmay also treatment of escape behavior. Journal of Applied
Behavior Analysis,32,285–296.
be conceptualized as three separate indepen-
Marcus, B. A., & Vollmer, T. R. (1996). Combining
dent-variable manipulations. Future investiga-
noncontingent reinforcement and differential rein-
tions in the treatment of multiply controlled forcement schedules astreatment for aberrant behav-
behavior may evaluate the effectiveness of ior. JournalofApplied BehaviorAnalysis,29, 43–51.
Roane,H.S.,Vollmer,T.R.,Ringdahl,J.E.,&Marcus,
separate and identical interventions across all
B.A.(1998).Evaluationofabriefstimuluspreference
baselines to determine the optimal procedures assessment. Journal of Applied Behavior Analysis, 31,
for evaluating treatments for multiply con- 605–620.
Smith,R.G.,Iwata,B.A.,Vollmer,T.R.,&Zarcone,J.
trolled problem behavior.
R. (1993). Experimental analysis and treatment of
multiply controlled self-injury. Journal of Applied
Behavior Analysis,26,183–196.
REFERENCES
Vollmer, T. R., Borrero, J. C., Lalli, J. S., & Daniel, D.
Borrero, C. S. W., Vollmer, T. R., Borrero, J. C., & (1999). Evaluating self-control and impulsivity in
Bourret, J. (2005). A method for evaluating param- children with severe behavior disorders. Journal of
eters of reinforcement during parent-child interac- Applied BehaviorAnalysis,32, 451–466.
tions. Research in Developmental Disabilities, 26,
503–600.
Hanley, G. P., Iwata, B. A., & McCord, B. E. (2003). Received December 13, 2004
Functional analysis of problem behavior: A review. Final acceptanceMarch 26, 2006
JournalofAppliedBehavior Analysis,36,147–185. Action Editor,Henry S.Roane