Table Of ContentJOURNALOFAPPLIEDBEHAVIORANALYSIS 2008, 41, 423–428 NUMBER3 (FALL2008)
DECREASING PROBLEM BEHAVIOR ASSOCIATED WITH A WALKING
PROGRAM FOR AN INDIVIDUAL WITH DEVELOPMENTAL AND
PHYSICAL DISABILITIES
HENRY S. ROANE
MUNROE-MEYERINSTITUTEANDUNIVERSITYOFNEBRASKA
MEDICALCENTER
AND
MICHAEL E. KELLEY
UNIVERSITYOFSOUTHERNMAINEANDACHIEVE
In the current investigation, a functional analysis suggested that positive reinforcement in the
formofphysicalcontactmaintainedtheself-injuriousbehaviorofagirlwithdevelopmentaland
physicaldisabilities.Weusedtheinformationobtainedfromthefunctionalanalysistodevelopa
treatment for noncompliance with walking in which a therapist removed physical interaction
followinginappropriate behavior during walks.
DESCRIPTORS: ambulation, partial-weight support, physical attention, self-injurious
behavior
_______________________________________________________________________________
Walking is a milestone of typical child the exact behavioral mechanism for the result-
development, and usually occurs around the ing change in behavior is unknown, because
ageof12months.Childrenwithdevelopmental Lalli et al. did not conduct a formal evaluation
and physical disabilities show impairments in of the variables that maintained refusal of
the acquisition of typical developmental mile- walking.
stones such as walking, which may negatively Individuals with developmental disabilities
affect other areas of their development (e.g., may have difficulties with walking, and they
self-sufficiency; Winter & Kiely, 2006). Behav- may exhibit destructive behavior (e.g., self-
ioral treatments, such as differential reinforce- injurious behavior [SIB], flopping, other refusal
ment, have been used to promote walking for behavior) when they are required to walk from
individuals with developmental and physical one location to another. For example, McCord,
disabilities (e.g., Horton & Taylor, 1989; Thomson, and Iwata (2001) hypothesized that
Lancioni et al., 2004, 2005). Lalli, Mauk, terminatinganactivity,initiatinganewactivity,
Goh,andMerlino(1994)reportedthatphysical or movement itself may have maintained the
prompting was successful for increasing com- SIB that occurred for 2 individuals during
pliance with instructions to walk for 2 individ- transitions between activities. Treatments based
uals with developmental disabilities. However, on the outcomes of functional analyses that
assessed these variables were successful in
reducing SIB associated with transitions. How-
This investigation was supported in part by Grant 1
RO1MH069739-01fromtheNationalInstituteofChild ever, the behavior-analytic literature suggests
Health and Human Development. We thank Susan that identifying the exact operant mechanism
Manoglu and Wayne Fisher for their assistance with
forproblembehaviorisnotalwaysnecessaryfor
variousaspects ofthisinvestigation.
AddresscorrespondencetoHenryS.Roane,Centerfor effectivetreatment.Thatis,contingentaccessto
Autism Spectrum Disorders, Munroe-Meyer Institute, arbitrary reinforcers (i.e., those that do not
985450 Nebraska Medical Center, Omaha, Nebraska
maintain problem behavior but are nevertheless
68198(e-mail: [email protected]).
doi:10.1901/jaba.2008.41-423 reinforcers for some other behavior) may
423
424 HENRY S. ROANE and MICHAEL E. KELLEY
effectively reduce problem behavior in some to the next. SIB analysis sessions lasted 10 min,
cases. For example, Fischer, Iwata, and Maza- and walking sessions lasted 5 min. One to four
leski (1997) provided noncontingent access to sessions were conducted daily.
arbitrary reinforcers (i.e., preferred items) to
reduce attention-maintained problem behavior Response Measurement and
exhibited by 2 individuals. In the current Interobserver Agreement
investigation, physical attention was identified During the SIB analysis, data were collected
as a reinforcer for 1 individual’s SIB. This on self-hitting (defined as open or closed
reinforcerwas thenappliedto decreaseproblem contact of the hand to the head or body) and
behavior during walking. head banging (defined as forceful contact
between Gail’s head and a surface). During
the walking analysis, the primary dependent
METHOD
measure was foot withdrawals, which were
Participant and setting defined as Gail lifting the bases of both feet
Gail was a 16-year-old girl who had been off of the ground and not supporting her own
diagnosed with moderate mental retardation, weight for a period of at least 3 s. Data also
generalized anxiety disorder, and cerebral de- were collected on correct steps, which were
generative chorea; she was enrolled in a day defined as lifting one foot off of the ground for
program for the treatment of SIB and for the less than 3 s and subsequently placing her foot
development of academic and vocational pro- in front of her body such that the base of her
grams. She communicated through brief (e.g., foot made complete contact with the floor.
one- to two-word) vocalizations and exhibited Frequency data were used to record SIB and
limited self-help skills, although she was able to correct steps and were converted to responses
complete some vocational tasks (e.g., preparing per minute by dividing the observed frequency
foods) with minimal physical guidance. Due to of these responses by the length of the
her physical disabilities, she required assistance observation period (5 or 10 min). Duration
from several individuals when completing data were collected on foot withdrawals by
complex tasks (e.g., toileting, walking). How- recordingthenumberofsecondsthatGail’sfeet
ever, she often became noncompliant during were lifted off of the ground. These data were
such tasks (e.g., lifting her feet off the ground converted to a percentage-of-session measure
during transitions), such that she required for the purpose of data analysis by dividing the
complete physical guidance to complete the total length of the response by the duration of
task. Consequently, she was confined to a theobservation(300 sor600 s)andconverting
wheelchair for the majority of her day, even this ratio to a percentage.
though she possessed sufficient musculature to Duringallsessions,anobservercollecteddata
support herself during standing and walking. on a laptop computer while following Gail on
Based on her physical disabilities and her the walk. A second observer collected data
noncompliance with physical tasks, one of her during 71% of all SIB analysis sessions and
educational goals was the development of a 42% of all walking sessions. Interobserver
daily walking routine. agreement coefficients were calculated by par-
Sessions for the SIB analysis were conducted titioning each session into 10-s intervals and
in a fully padded room (3 m by 3 m) equipped dividingthenumberof10-sintervalswithexact
with a one-way observation window. The agreement (i.e., both observers recording the
walking analysis was conducted in empty same duration or frequency of occurrences
hallways approximately 10 m in length, and within a given interval) by the number of 10-s
each session consisted of walking from one hall intervals with agreements plus those with
INAPPROPRIATE WALKING 425
disagreements and converting this ratio to a support her full weight, issued prompts every
percentage. Mean interobserver agreement was 30 s,andcontinuedalongtheroutebutdidnot
98% for SIB, 99% for foot withdrawals, and deliver praise.
98% for correct steps. Accessto physicalattention was manipulated
during the treatment analysis for walking.
Procedure
Specifically, during the treatment condition
SIB analysis. Prior to the walking analyses, for the walking analysis, physical attention was
the function of Gail’s SIB had been assessed on removed contingenton the occurrenceof a foot
several occasions and had suggested that withdrawal by the therapists crouching down
physical attention (i.e., holding on to a andmovingapproximately1 mawayfromGail
therapist) was one variable suspected to main- for a period of 10 s, such that Gail could sit on
tain SIB. Thus, a multielement comparison of her knees on the floor but could not physically
continuous and contingent physical attention interact with the therapists. Thus, the therapists
was conducted. In the continuous contact no longer supported Gail’s weight if she
condition, a therapist sat next to Gail on the withdrew her feet, and she was either allowed
floor such that Gail was able to wrap her arms to sit on the ground or stand unassisted. At the
aroundthetherapist,orthetherapistheldhands end of the 10-s interval, the therapists either
with Gail throughout the session. No other picked Gail up or reimplemented the support
formofattentionwasprovided.Thecontingent procedure (if she was standing on her own).
physicalattentionconditionwasidenticaltothe With the exception of the brief praise, there
continuous attention condition, except that the were no differential consequences in place for
therapist moved close to Gail such that she correct steps. The length of the treatment
could wrap her arms around the therapist or sessions was corrected for the amount of time
hold hands with the therapist for 20 s contin- that the therapists were not supporting Gail,
gent on SIB. The results of the SIB analysis such that all baseline and treatment observa-
suggestedthatphysicalattentionfunctionedasa tions consisted of 5 min of therapist support.
positive reinforcement for SIB. This informa-
tion was used to develop a treatment for
RESULTS AND DISCUSSION
walking.
Walking analysis. During the initial walking Figure 1 (top) shows the results of the SIB
analysis, two conditions were compared in a analysis. Higher rates of SIB were observed in
reversal (ABAB) design. Both conditions con- the contingent contact condition (M 5 28.4
sisted of a modified partial-weight support responses per minute) than in the continuous
program (e.g., Hesse, 2001) in which two contact condition (M 5 9.5), suggesting that
therapists walked along either side of Gail to access to physical attention functioned as a
provide assistance and stability while also positive reinforcer for SIB.
supporting her weight by holding her at the Figure 1 (middle) shows the occurrence of
wrists,underherarmpits,andacrossherchestif foot withdrawal. In baseline, relatively high
she withdrew her feet. If she did not withdraw levels of foot withdrawals were observed (M 5
her feet, the therapist held her under her 66% of the session). When physical attention
armpits and held her wrists. During baseline, was removed contingent on foot withdrawal, a
shewaspromptedevery30 stotakeasteporto decrease in that behavior was observed (M 5
continue walking, and she received brief praise 7%ofthesession).Footwithdrawalsreemerged
(e.g., ‘‘good job walking Gail, keep going’’) if duringthereversaltobaseline(M588%ofthe
she was walking appropriately. If she withdrew session)anddecreasedagainwhenthetreatment
her feet at any time, the therapists continued to was reintroduced (M 5 15% of the session).
426 HENRY S. ROANE and MICHAEL E. KELLEY
Figure 1. Responses per minute of SIB (top) during the SIB analysis, percentage of session with foot withdrawals
(middle), andresponses per minuteof correct steps (bottom)during the walkinganalysis.
Figure 1 (bottom) shows the occurrence of for foot withdrawals, a corresponding increase
correct steps. During baseline, near-zero rates in correct steps was observed (M 5 56.6).
were observed(M5 0.2 responses perminute). Correct steps decreased during the reversal to
Bycontrast,whenthetreatmentwasintroduced baseline (M 5 7.4) and gradually increased
INAPPROPRIATE WALKING 427
duringthefinalphaseoftreatment(M551.4). to that applied in previous research (e.g.,
Relatively low rates of SIB (M 5 1.1) occurred Fischer et al., 1997) in which an arbitrary
throughout the walking analysis (data not reinforcer (i.e., one that has not been demon-
shown). strated to maintain a specific aberrant response)
In the current investigation, contingent was applied to the treatment of destructive
removal of physical attention decreased prob- behavior. A second limitation of the current
lem behavior (i.e., foot withdrawals) that analyses is that data were not collected on
occurred when walking. Results of the SIB holding onto the therapist during the walking
analysissuggestedthatphysicalattention(inthe analysis.Thesedatamayhavefurthersupported
form of Gail intertwining her arms with those the anecdotal observation that Gail engaged in
of the therapists or holding onto the therapists’ high rates of this response in the walking
torsos) functioned as a reinforcer. Although analysis.Third,itispossiblethathigherratesof
physicalattentionwasnotdirectlyevaluatedasa SIB were observed in the contingent contact
reinforcer during the walking analysis, it is condition than in the continuous contact
possible that the removal of physical attention condition because the type of physical attention
for 10 s contingent on a foot withdrawal delivered interfered with the occurrence of SIB
functioned as punishment (i.e., contingent (i.e., less SIB occurred in the continuous
cessation of a preferred activity). Likewise, it is contact condition because Gail’s hands were
possiblethatcorrect stepsincreasedbecause this on the therapist). However, SIB did occur in
responseresultedinprolongedaccesstophysical the continuous contact condition, suggesting
interaction and increased the probability that that SIB was possible during this condition. It
Gail would contact praise; however, the praise should be noted that SIB was treated through
contingency was identical across both condi- additional analyses that were not included in
tions. Alternatively, foot withdrawals and the current investigation.
correct steps may have been affected by the The current study adds to the literature
avoidance of some aversive properties of the regarding behavioral approaches to promoting
treatment procedures. For example, foot with- walkinginatleasttwoways.First,itprovidesan
drawalsmay have decreased as a function of the example of a likely reductive procedure (e.g.,
discomfort associated with sitting on the floor, punishment).Second,itdemonstratestheuseof
or correct steps may have increased due to the alternative assessments prior to treatment de-
avoidance of losing bodily support. It is also velopment.Inthecurrentinvestigation,theSIB
possible that the loss of physical support may analysis essentially functioned as a reinforcer
have affected the potency of praise as a assessment in which a potent positive reinforcer
reinforcer for correct steps, which may have was identified and subsequently manipulated
increased Gail’s motivation to engage in correct during treatment. In sum, the current results
steps. suggest that reinforcers functionally related to
One limitation of the current investigation is other(i.e.,nontarget)responsescanbearranged
that we did not directly evaluate the specific to construct an effective program thatpromotes
sources of reinforcement that maintained prob- appropriate behavior during walking.
lembehavior during walks (i.e., viaa functional
analysis). However, the treatment for inappro-
REFERENCES
priate behavior during walks was developed
Fischer, S. M., Iwata, B. A., & Mazaleski, J. L. (1997).
based on the results of the SIB analysis, which
Noncontingent delivery of arbitrary reinforcers as
suggestedthatphysicalattentionfunctionedasa treatment for self-injurious behavior. Journal of
reinforcer. Thus, the current method is similar Applied BehaviorAnalysis,30, 239–249.
428 HENRY S. ROANE and MICHAEL E. KELLEY
Hesse,S.(2001).Locomotiontherapyinneurorehabilita- Lancioni,G.E.,Singh,N.N.,O’Reilly,M.F.,Oliva,D.,
tion. Neurorehabilitation,16,133–139. Campodonico, F., & Groeneweg, J. (2004). Impact
Horton, S. V., & Taylor, D. C. (1989). The use of of favorite stimuli on the behavior of persons with
behavior therapy and physical therapy to promote multipledisabilitieswhileusingatreadmill.Journalof
independentambulationinapreschoolerwithmental VisualImpairment and Blindness, 98,304–309.
retardation and cerebral palsy. Research in Develop- McCord, B. E., Thomson, R. J., & Iwata, B. A. (2001).
mentalDisabilities,10,363–375. Functional analysis and treatment of self- injury
Lalli, J. S., Mauk, J. E., Goh, H., & Merlino, J. (1994). associatedwithtransitions.JournalofAppliedBehavior
Successful behavioral intervention to treat children Analysis, 34, 195–210.
who are reluctant to ambulate. Developmental Med- Winter, S., & Kiely, M. (2006). Cerebral palsy. In I. L.
icine andChildNeurology,36,625–629. Rubin & A. C. Crocker (Eds.), Developmental
Lancioni, G. E., Singh, N. N., O’Reilly, M. F., disabilities: Delivery of medical care for children and
Campodonico, F., Piazzolla, G., Scalini, L., et al. adults(pp. 233–248). Baltimore:Brookes.
(2005). Impact of favorite stimuli automatically
delivered on step responses of persons with multiple Received February 21,2007
disabilitiesduringtheiruseofwalkerdevices.Research Final acceptance July4,2007
inDevelopmental Disabilities,26,71–76. Action Editor,Joel Ringdahl
