ERIC EJ758911: Further Examination of Factors that Influence Preference for Positive versus Negative Reinforcement

2007·0.44 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2007, 40, 25–44 NUMBER1 (SPRING2007)
FURTHER EXAMINATION OF FACTORS THAT INFLUENCE
PREFERENCE FOR POSITIVE VERSUS NEGATIVE REINFORCEMENT
TIFFANY KODAK
LOUISIANASTATEUNIVERSITY
DOROTHEA C. LERMAN
UNIVERSITYOFHOUSTON–CLEARLAKE
AND
VALERIE M. VOLKERT AND NICOLE TROSCLAIR
LOUISIANASTATEUNIVERSITY
Factors that influence choice between qualitatively different reinforcers (e.g., a food item or
abreakfromwork)areimportanttoconsiderwhenarrangingtreatmentsforproblembehavior.
Previousfindingsindicatethatchildrenwhoengageinproblembehaviormaintainedbyescape
fromdemandsmaychooseafooditemoverthefunctionalreinforcerduringtreatment(DeLeon,
Neidert,Anders,&Rodriguez-Catter,2001;Lallietal.,1999).However,anumberofvariables
may influence choice between concurrently available forms of reinforcement. An analogue for
treatmentsituationsinwhichpositivereinforcementforcomplianceisindirectcompetitionwith
negative reinforcement for problem behavior was used in the current study to evaluate several
variablesthatmayinfluencechoice.Participantswere5childrenwhohadbeendiagnosedwith
developmental disabilities and who engaged in problem behavior maintained by escape from
demands. In the first phase, the effects of task preference and schedule of reinforcement on
choicebetweena30-sbreakandahigh-preferencefooditemwereevaluated.Thefooditemwas
preferred over the break, regardless of the preference level of the task or the reinforcement
schedule,forallbut1participant.Inthesecondphase,thequalityofthebreakwasmanipulated
bycombiningescapewithtoys,attention,orboth.Only1participantshowedpreferenceforthe
enrichedbreak.Inthethirdphase,choiceofamedium-orlow-preferencefooditemversusthe
enrichedbreakwasevaluated.Threeof4participantsshowedpreferenceforthebreakoverthe
lesspreferredfooditem.Resultsextendpreviousresearchbyidentifyingsomeoftheconditions
underwhichindividualswhoengageinescape-maintainedbehaviorwillpreferafoodreinforcer
over thefunctional one.
DESCRIPTORS: choice, differential reinforcement, behavioral economics, positive re-
inforcement, negativereinforcement, escape-maintained behavior
_______________________________________________________________________________
Variables that influence choice between Shade, & Miller, 1994; Tustin, 1994; Vollmer,
reinforcers, including the schedule, delay, and Borrero, Lalli, & Daniel, 1999). Research on
quality of the reinforcer and the effort required choice is important to application because
to gain access to reinforcement, have been multiple reinforcers are often concurrently
examined in a number of studies (e.g., Koehler, available in the natural environment.
Iwata,Roscoe,Rolider,&O’Steen,2005;Neef, Although choice between similar or identical
reinforcers (e.g., two food items) has been
Thisresearchwasbasedonadissertationconductedby arranged in most research in this area, an
the first author in partial fulfillment of the requirements
increasing number of applied studies have
forthe PhDdegree.
Addresscorrespondence toTiffanyKodak,whoisnow examined choice between qualitatively different
at the Munroe-Meyer Institute, UNMC, 985450 Ne- reinforcers (e.g., a break from work or a food
braskaMedicalCenter,Omaha,Nebraska68198(e-mail:
item; DeLeon et al., 2001; Lalli et al., 1999).
[email protected]).
doi:10.1901/jaba.2007.151-05 Such an arrangement may arise when treatment
25
26 TIFFANY KODAK et al.
is implemented with less than perfect integrity. may influence choice between these reinforcers
For example, the teacher of a child with duringtasks.DeLeonetal.(2001),forexample,
developmental disabilities may deliver praise examined choice between food and escape
and food reinforcers for completing tasks but under increasing schedule requirements with
permit escape from the task for disruptive a child who engaged in problem behavior
behavior. maintained by escape. The participant was
In Lalli et al. (1999), the effects of positive required to complete a number of tasks to
and negative reinforcement on problem behav- choose between a food item or a break, while
ior and compliance were examined with 5 problem behavior no longer produced escape.
children who engaged in problem behavior The participant chose thefood item moreoften
maintained by escape from demands. Treat- than the break under low schedule require-
ment consisted of providing either negative ments. However, preference switched to the
reinforcement (i.e., a break from the task) or negative reinforcer when the participant was
food for compliance while problem behavior required to complete 10 tasks before receiving
continued to produce escape. Treatment with the opportunity to choose a reinforcer. The
the food item was associated with greater authors hypothesized that the larger work
reductions in problem behavior and higher requirement functioned as an establishing
levels of compliance than treatment with the operation (EO) by momentarily altering the
functional reinforcer, even though escape was reinforcing value of the break. However, only 1
available for problem behavior. Thus, the individualparticipatedinthestudy.Thus,more
participants chose the food reinforcer over the researchiswarrantedtoevaluatechoicebetween
functional reinforcer when both were concur- a break and food under increasing schedule
rently arranged. requirements.
The use of food reinforcers for appropriate Furtherresearchalsoisneededonfactorsthat
behavior with children who engage in escape- mayinteractwiththereinforcementscheduleto
maintained behavior has some advantages over alter choice between reinforcers. For example,
the use of the functional reinforcer (escape) in in DeLeon et al. (2001), the relative value of
classroom settings. For example, food can be escapeversusfoodunderthethinreinforcement
delivered relatively quickly on rich schedules of schedule may have depended on the preference
reinforcement while the child continues to level of the task. Certain types of tasks may be
engage in academic tasks. Frequent breaks from moreaversivethanothers,eventhoughmultiple
a task may be impractical to implement and tasks may be associated with problem behavior.
limit the child’s participation in ongoing Results of several studies indicate that individ-
classroom activities. Nonetheless, food reinforc- uals with developmental disabilities will engage
ers must compete with escape for problem in higher levels of problem behavior when
behavior when it is difficult for teachers to required to complete less preferred tasks
implement escape extinction with integrity. (Foster-Johnson, Ferro, & Dunlap, 1994;
Examining variables that influence preference Vaughn & Horner, 1997). For example,
for food reinforcers over escape is important to VaughnandHornercomparedratesofproblem
maximize the likelihood of compliance with behavior when participants were required to
academic tasks when problem behavior con- complete high- and low-preference tasks. Al-
tinues to produce the functional reinforcer. though both tasks produced some problem
Although results of Lalli et al. (1999) behavior, the low-preference tasks were associ-
suggested that some children prefer food over ated with much higher rates of problem
escape from demands, a number of variables behavior. Thus, relative to food reinforcement,
REINFORCEMENT SCHEDULE AND QUALITY 27
a break from some tasks (more aversive, less Theextenttowhichchoicebetweenafooditem
preferred, or both) may be more valuable than and a break depends on the preference level of
a break from other tasks as the schedule the food item should be evaluated in further
requirement increases. research.
Thequalityofthebreakandfooditeminthe The purpose of this study was to extend
DeLeon et al. (2001) and Lalli et al. (1999) previous research on reinforcer choice (DeLeon
studiesalsomayhaveinfluencedchoicebetween et al., 2001; Lalli et al., 1999) by examining
reinforcers. In both studies, a highly preferred howpreferenceforeitherafooditemorabreak
food reinforcer was available concurrently with would be influenced by the schedule require-
a break alone (i.e., no attention or toys were ments, preference level of the task, and
available during the break). Previous research variations in the quality of the reinforcer. The
hasshownthatchildrenwithescape-maintained methodology was arranged as an analogue of
behavior may prefer escape to an enriched treatment situations in which both task com-
environment(i.e.,abreakwithtoys)overescape pliance and problem behavior produce re-
alone(Golonkaetal.,2000;Zarcone,Fisher,& inforcement.Theprimaryinterestofthisbridge
Piazza, 1996). In Golonka et al., for example, study was on reinforcer choice rather than on
treatment for negatively reinforced problem the treatment of problem behavior. In the first
behavior was more effective when compliance phase, DeLeon et al. was replicated and
produced an enriched break rather than a break extended by evaluating choice between food
alone. Both participants also chose the enriched and escape when reinforcement was delivered
break more often when the breaks were avail- contingent on compliance to either high-
able concurrently for task compliance. Results preference or low-preference tasks. In the
of Pizza et al. (1997) showed that treatment second and third phases, the quality of the
withmultiplefunctionalreinforcers(i.e.,abreak break and food item was manipulated to
combined with tangible items or attention) was examinehowthisfactorinfluenceschoiceunder
more effective than treatment with a single relatively thin schedules of reinforcement.
reinforcer (i.e., a break alone) for children who
engaged in multiply controlled problem behav- METHOD
ior. In the natural environment, a break from Participants, Settings, and Materials
demands may be combined with access to toys
Five children, aged 4 to 8 years, participated
or attention from adults or peers (i.e., a higher
in the study. Participants had been diagnosed
quality break). Therefore, further research
with developmental disabilities, autism, or both
should evaluate whether a more natural, higher
and had been referred for the treatment of
quality form of escape alters preference for
inappropriate behavior that interfered with task
a food reinforcer over a break.
completion. Table 1 displays each participant’s
Alternatively, the quality of the food re-
problem behavior. All of the children were
inforcer may influence choice between reinforc-
reported to have some visual discrimination
ers. In Lalli et al. (1999) and DeLeon et al. skills. Larry, Sam, Mary, and Scott communi-
(2001), highly preferred food items were cated vocally using complete sentences and
identified via systematic preference assessments. followed three-step instructions. Casey had
These food reinforcers may have been of higher been diagnosed with moderate mental retarda-
quality than those typically used in the natural tion, communicated through gestures or by
environment because teachers and parents are guiding people towards objects, and followed
less likely to conduct systematic preference some one-step instructions (e.g., ‘‘sit down’’).
assessments when selecting food reinforcers. None of the participants had any sensory or
28 TIFFANY KODAK et al.
Table 1
Problem Behaviorfor Each Participant andPreferenceAssessment Results
High-and High-and
Participant Problembehavior low-preferencefoods High-preferencetoys low-preferencetasks
Casey Aggression,SIB GingercookieandRiceChexH BumbleballHand Puttingpiecesinapuzzleand
discoball matchingletters
Larry Aggression,disruption, Carobchipandraisin Children’sbook Receptiveidentificationof
inappropriate andavideo coloredbearsand
vocalizations stringingbeads
Sam Aggression,disruption SourcreamandonionPringlesH MardiGrasbeads Receptiveidentificationof
andGummyBearsHand andBumble animalsandletterpuzzle
regularPringlesHanddried BallH
apples
Mary SIB,inappropriate Fruitsnack,apricot(medium- MardiGrasbeads Stringingbeadsandshape
vocalizations preferencefooditem),and andsquishyball puzzle
strawberryjuice
Scott Aggression,disruption, M&MsHanddriedapple Light-upsnakeand Pegboardandreceptive
inappropriate playtools identificationofopposites
vocalizations
physical deficits except Mary, who had been hitting, pushing, scratching, pinching, and
diagnosed with visual impairments. Sam began pulling hair), disruption (i.e., throwing materi-
taking stimulant medication during the third als, flopping, crying, and spitting), self-injury
phase (i.e., at Session 113). A functional (i.e., hand biting, arm biting, and head and
analysis was conducted prior to the study to body hitting), and inappropriate vocalizations
identify the variables that maintained problem (i.e., whining, screaming, and saying ‘‘no’’).
behavior. Only participants whose problem Compliance was defined as completing a de-
behavior was maintained by escape from mand within 5 s of a verbal or model prompt.
demands were included in the study. Asecondindependentobservercollecteddata
All assessment and treatment sessions were during 36% to 54% of sessions for each
conductedinanunusedroomattheparticipant’s participant. Interobserver agreement was calcu-
school or in therapy rooms at a university-based lated for reinforcer choice, problem behavior,
early intervention program for children with andcompliancebydividingthetotalnumberof
autism. The rooms contained a desk and chair, occurrence agreements across consecutive 10-s
chairsfordatacollectors,andanyrelevantsession intervals by the total number of occurrence
materials.The therapistand datacollectorswere agreements plus disagreements and multiplying
present during all sessions. Sessions were con- by 100%. Mean interobserver agreement for
ducted the same time each day with each reinforcer choice, problem behavior, and com-
participant. All sessions were conducted prior pliance across participants was 98% (range,
to lunch or at least 1 hr after lunch. 60% to 100%), 94% (range, 50% to 100%),
and 99% (range, 90% to 100%), respectively.
Response Measurement and Reliability
Frequencydataonreinforcerchoice,problem General Procedure
behavior, and compliance were collected via Preference assessments. Assessments were con-
laptop computers by trained observers during ductedtoidentifypreferencerankingsoftangible
all assessment and treatment sessions. Reinforc- items,fooditems,andtasksbasedonprocedures
erchoicewasdefinedaspointingtoortouching described by Fisher et al. (1992) and Lattimore,
oneoftwocouponsoritemsassociatedwiththe Parsons, and Reid (2002). Tangible items, food
reinforcer after a verbal prompt. Inappropriate items,and taskswere assessedseparately.Highly
behavior (see Table 1) included aggression (i.e., preferredtangibleitemswereincludedinspecific
REINFORCEMENT SCHEDULE AND QUALITY 29
conditions of the functional analysis and in access to food was a functional reinforcer for
Phases 2 and 3. The highest ranked food items problem behavior. During the food condition,
wereusedincertainconditionsofthefunctional participants were given presession access to
analysis and during all phases of the study. The a highly preferred food. Food was removed at
lowestrankedfooditemswereusedinPhase3(a thebeginningofthesessionandreturnedfor20 s
medium-preferencefooditemalsowasevaluated contingent on the target behavior.
for Mary in Phase 3). The task preference Discriminationtraining.Priorto baseline,the
assessment included tasks that were delivered participant was taught to discriminate between
during the demand condition of the functional two coupons (Sam, Scott, and Larry) or two
analysis. During each trial of the task preference three-dimensional items (Mary and Casey) and
assessment,twotaskswereplacedinfrontofthe to touch the coupons or items to obtain the
participant,andtheparticipantwasinstructedto designated reinforcer. Coupons were used for
‘‘pick one.’’ When the participant pointed to or participants whose teachers reported that they
touched one task, the nonchosen task was could discriminate among different pictures.
removed, and the participant was prompted to Items were used for participants who did not
complete three responses from the task that was reportedly possess such discrimination skills. If
chosen. For example, if stringing beads was three-dimensional items were used, one item
selected, the participant was required to string represented the break (e.g., a timer) and the
three beads. Problem behavior was exposed to other item represented the food (e.g., a bag of
extinction during preference assessment trials. food). During discrimination training, the
The task that the participant chose most often therapist physically guided the participant to
wasdesignatedthehigh-preferencetask;thetask choose one coupon or item and delivered the
that was chosen least often was designated the consequenceassociatedwiththecouponoritem
low-preference task. High- and low-preference (i.e., either a 30-s break from discrimination
tasks were used in Phase 1. Table 1 displays the training or a small piece of food). After
results of the preference assessments conducted physically guiding the participant to pick each
with each participant. Additional preference coupon or item a minimum of five times, the
assessments (i.e., multiple stimulus without therapist required the participant to comply
replacement) were conducted throughout the with one instruction (e.g., string one bead) and
study with various participants(i.e., Sam, Larry, then permitted the participant to choose
andScott;seefurtherdiscussionbelow)basedon between the two coupons or items. This
procedures described by DeLeon and Iwata procedure was conducted a minimum of three
(1996). times. The therapist then asked the child to
Functional analysis. A functional analysis of point to or say the coupon or item associated
problem behavior was conducted based on with either a break or food item. If the child
procedures described by Iwata et al. (1994). could accurately perform this activity twice for
Functionalanalysisconditionsincludedattention, each coupon or item, discrimination training
demand, toy play, tangible, and no interaction was completed. If the participant did not
(Casey and Mary only). Sessions (10 min each) accurately discriminate between the coupons,
were randomly alternated in a multielement de- discrimination training was conducted with
sign. Additional assessments (i.e., pairwise com- items in place of the coupons, and training
parisons) were conducted with Sam and Scott to continued until the criteria above were met.
clarify functional analysis results. A pairwise Following discrimination training, probes were
comparison of food and toy play conditions was conducted daily prior to sessions to ensure that
conducted with all participants to determine if the participant’s choice remained under dis-
30 TIFFANY KODAK et al.
criminative control of the two coupons or schedule beginning with FR 1. Each session
items. During daily probes, the therapist forced ended when the participant had received five
a choice for each coupon or item once and opportunities to choose between reinforcers.
repeatedthelaststepofdiscriminationtraining. Whentheparticipanthadcompliedwiththe
requirednumberofdemands(dependingonthe
Experimental Design
schedule), the therapist placed both coupons or
High-preference and low-preference tasks
items on the table at equal distances from the
werealternatedinamultielementdesignduring
participant. The therapist said, ‘‘pick one.’’ If
Phase 1. The effects of the schedule on
the participant chose the snack coupon or item,
reinforcerchoicealsowereevaluatedinareversal
the participant was given a small piece of
design in Phase 1. During Phase 2, the
a highly preferred food. The next demand
parametersofthebreakweremanipulatedusing
began immediately after delivery of the food
a reversal design. In Phase 3, the effects of
item so that the positive reinforcer was not
reinforcer quality on choice were evaluated
confounded with a break from the task. If the
using a reversal design.
participant chose the break coupon or item, the
therapist turned away from the participant and
PHASE 1
provided a 30-s break from task demands. All
Procedure
participants chose a coupon or item within 5 s
The purpose of Phase 1 was to evaluate
of the initial verbal prompt to pick one.
preference for reinforcers under increasing
When reinforcer choice remained stable for
schedule requirements with high- versus low-
at least three sessions under FR 1, the schedule
preferencetasks.High-andlow-preferencetasks
was increased to FR 2. Reinforcer choice was
were alternated in a multielement design.
considered stable if it varied by 20% or less
During all sessions, the experimenter presented
from one session to the next. The schedule
instructional trials using a graduated three-step
continued to be increased (FR 2, FR 5, FR 10,
prompting procedure (verbal, model, physical
FR 20, FR 40) if choice remained stable for at
prompts). No programmed consequences were
least three consecutive sessions under each
provided for problem behavior (i.e., problem
schedule value until preference appeared to
behavior was exposed to extinction).
change(e.g.,switchedfromthefooditemtothe
Baseline (no reinforcement). Sessions with the
break) or until the schedule reached FR 40. If
most and least preferred tasks were conducted
preference changed at or before the schedule
aminimum offivetimeseach.Noprogrammed
reached FR 40, FR 2 and the highest schedule
consequences were provided for compliance.
requirement for that participant was replicated.
Five trials were conducted during each session.
Thepurposeofbaselinewastoevaluatelevelsof
complianceandproblembehaviorintheabsence PHASE 2
of reinforcement for appropriate behavior. Procedure
Reinforcer choice. All procedures were the All of the children participated in Phase 2
same as in baseline, but reinforcement was because results of Phase 1 indicateda preference
provided for task compliance. The participant for the food item under relatively thin re-
had the opportunity to choose between two inforcement schedules (e.g., FR 20 or FR 40).
reinforcers contingent on compliance following The purpose of Phase 2 was to examine how
averbalormodelprompt.Thenumberoftimes preference would change if the break contained
the participant was required to complete the access to other positive reinforcers. Because the
task to gain access to reinforcement was parameters of the break were manipulated, the
gradually increased on a fixed-ratio (FR) conditionsmostlikelytoincreasethevalueofthe
REINFORCEMENT SCHEDULE AND QUALITY 31
breakwereineffectduringthisphase.Thus,the high-preference food item and the enriched
low-preference task and the thinnest schedule break) served as the baseline data for Phase 3,
underwhichthefoodreinforcerwasconsistently withtheexceptionofMary,whowasthefirstto
preferred over the break were used. All other participate in Phase 3. We hypothesized that
procedures were identical to those in Phase 1. a lower preference food item would lead to
Baseline. Data from the last phase in Phase 1 a complete switch in preference to the enriched
(i.e., when the highest schedule requirement break even when a denser schedule of re-
reachedbytheparticipantwasreplicated)served inforcement was in place (e.g., FR 10).
as the baseline data for Phase 2, with two Therefore, we began Phase 3 under the FR 10
exceptions. First, an additional baseline was schedule with Mary only. When her preference
conducted with Sam when a change in setting did not change, the same reinforcement
occurred following Phase 1. The baseline for schedules that had been implemented in Phase
Phase 2 was conducted in the new setting. 2 were used for the remaining participants.
Second, a relatively rich schedule (FR 5) was Reinforcer choice. Participants could choose
chosen as the baseline schedule for Scott between a low-preference food item or an
because reinforcer choice in Phase 1 was
enriched break (i.e., a break with access to two
somewhat variable even under rich schedules
highly preferred toys and adult attention). A
of reinforcement (e.g., the replication of FR 2).
medium-preference food item (i.e., food item
Thus, a baseline phase with this schedule was
ranked in the middle of the items in the food
implemented during Phase 2.
preference assessment) also was evaluated with
Reinforcer choice. Highly preferred tangible
Mary only. If participants chose the enriched
items (i.e., toys), therapist attention, or both
break more often than in baseline, a reversal to
were systematically combined with the break.
the high-preference food item was implemen-
When the break was selected, the therapist
ted. Following the reversal, the low-preference
removed the task materials, provided the two
food item was reinstated to replicate the change
most highly preferred toys, or delivered atten-
in preference.
tion (i.e., conversation) for 30 s. The separate
The procedures conducted in Phase 3 were
effects of attention and tangible items during
modified for Sam after a change in preference
the break were evaluated for 1 participant to
from the low-preference food to the break was
identify the particular variables (attention,
not replicated (i.e., Sam chose the low-
tangible items, or both) that were responsible
preference food over the break). An additional
for the change in preference. This evaluation
preference assessment was conducted to iden-
was followed by a reversal to the break only
tify a food item that was less preferred than
(i.e., baseline) and replication of the break plus
the current food item. The least preferred food
the relevant variable (attention).
item that Sam would consume was evaluated
PHASE 3 as the second low-preference food item. We
Procedure also hypothesized that the enriched break may
Four of the 5 children (Sam, Scott, Casey, have acquired some aversive properties during
and Mary) participated in Phase 3 because they the course of the study (e.g., Sam began to
continued to show preference for the food item push the toys away and wouldn’t talk with the
overtheenrichedbreakinPhase2.Thepurpose therapist during the break). Therefore, choice
wasto examinehow preference would change if between the food items and a nonenriched
the quality of the food item was manipulated. break (i.e., a 30-s break with no toys and
Baseline. Data from the last phase in Phase 2 attention) was evaluated at the end of the
(whentheparticipantcouldchoosebetweenthe phase.
32 TIFFANY KODAK et al.
RESULTS tions indicated that problem behavior was
maintained by escape from demands and not
Functional Analysis
by access to food items. In the first phase of
Results of the functional analysis and pair-
Scott’s functional analysis, the highest rates of
wise comparisons are presented in Figure 1.
problem behavior occurred in the tangible
Casey’s functional analysis suggested that prob-
condition, suggesting that problem behavior
lem behavior was maintained by negative
was maintained by access to toys. Problem
reinforcement in the form of escape from
behavior wasalsosomewhatelevatedduring the
demands. Initially, rates of problem behavior
attention condition, increasing to high levels
were highest in the play and demand condi-
during the final attention session. Due to the
tions. However, results were more clearly
nature of the problem behavior during this
differentiated during the last 14 sessions of the
session (i.e., severe aggression), additional
assessment, with the highest rates of problem
attention sessions were not conducted. Because
behavior occurring during the demand condi-
parent reports and previous observations in the
tion. During the pairwise comparison of food
classroom suggestedthat problembehaviormay
andtoyplayconditions,higherratesofproblem
have been maintained by escape from demands,
behavior occurred in the toy play condition
additional sessions with demand and toy play
than in the food condition, suggesting that the
conditions were conducted to further evaluate
behavior was not maintained by access to food.
this potential function. Results of the pairwise
Nonetheless, food cannot be excluded as
comparisons suggested that problem behavior
a possible maintaining reinforcer because prob-
was sensitive to negative reinforcement and
lem behavior was observed in the food
access to food items.
condition and no food was provided in the
toy play condition. Larry exhibited the highest Phases 1, 2, and 3
rates of problem behavior during the tangible The primary dependent variable was re-
and demand conditions of the functional inforcer choice, which is depicted in Figures 2
analysis, suggesting that his problem behavior through 6. Due to the lengthy nature of the
was maintained by access to tangible items and evaluation, data from Phase 1 are graphed
escape from demands. Results of the pairwise separatelyfromthedatacollectedduringPhases
comparison of food and toy play conditions 2 and 3 for each participant. However, data
indicated that problem behavior was not from Phase 1 that served as the initial baseline
maintained by access to food items. Sam’s condition for Phase 2 are reproduced in the
results suggested that his problem behavior was second graph for some participants. Results for
maintained by adult attention and escape from the two secondary dependent variables (i.e.,
demands. However, the rate of problem problem behavior and compliance) are briefly
behavior began to increase during the last two summarizedbelow(sessiondataareavailableby
sessions of the tangible condition, so a pairwise contacting the first author).
comparison of the tangible and toy play Choice. Overall, Casey showed a strong
conditions then was conducted. Results in- preference for the food over the break during
dicated that problem behavior was also main- Phase 1, even under the highest schedule
tained by access to toys. The pairwise compar- requirement (Figure 2). Although responding
ison of toy play and food conditions suggested was variable under his first exposure to FR 40,
that problem behavior was not maintained by Casey showed a clear preference for the food
access to food items. Results of Mary’s item when FR 40 was replicated. Choice
multielement functional analysis and the pair- between reinforcers was similar across the high-
wise comparison of food and toy play condi- and low-preference tasks. When attention and
REINFORCEMENT SCHEDULE AND QUALITY 33
Figure1. Responsesperminuteofproblembehavioracrossconditionsofthemultielementfunctionalanalysesand
pairwise comparisons for Casey, Larry,Sam,Mary, andScott.
34 TIFFANY KODAK et al.
Figure2. PercentageoftrialsinwhichCaseychosethefoodversusthebreakacrossincreasingschedulerequirements
inPhase1(top);low5low-preferencetask;high5high-preferencetask.PercentageoftrialsinwhichCaseychoseeither
thehigh-preference foodorthelow-preferencefoodversusthebreakduringPhases2and3(bottom).EB5enriched
break(breakwithtangibleitems andattention).