ERIC EJ798589: Assessing Potency of High- and Low-Preference Reinforcers with Respect to Response Rate and Response Patterns

2008·0.47 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2008, 41, 177–188 NUMBER2 (SUMMER2008)
ASSESSING POTENCY OF HIGH- AND LOW-PREFERENCE
REINFORCERS WITH RESPECT TO RESPONSE RATE AND
RESPONSE PATTERNS
BECKY PENROD, MICHELE D. WALLACE, AND EDWIN J. DYER
UNIVERSITYOFNEVADA,RENO
Previousresearchhassuggestedthattheavailabilityofhigh-preferencestimulimayoverridethe
reinforcingefficacyofconcurrentlyavailablelow-preferencestimuliunderrelativelylowschedule
requirements (e.g., fixed-ratio 1 schedule). It is unknown if similar effects would be obtained
underhigherschedulerequirements.Thus,thecurrentstudycomparedhigh-preferenceandlow-
preferencereinforcersunderprogressivelyincreasingschedulerequirements.Resultsfor3ofthe
4 participants indicated that high-preference stimuli maintained responding under higher
schedulerequirementsrelativetolow-preferencestimuli.For1participant,high-preferenceand
low-preference stimuli were demonstrated to be equally effective reinforcers under increasing
schedule requirements. Implications with respect to rate of performance and response patterns
arediscussed.
DESCRIPTORS: positivereinforcement,preferenceassessments,progressiveratio,response
effort
_______________________________________________________________________________
A number of researchers have examined reinforcers than those that are identified as less
methods for assessing stimulus preferences and preferred (e.g., chosen less than 20% of
deciding the degree to which such preferences presentations).
are predictive of the reinforcing effects of those Although stimuli identified as highly pre-
stimuli (see Cannella, O’Reilly, & Lancioni, ferred (HP) have been shown to be more
2005, for a review). Various preference assess- effective reinforcers when compared to those
ment methods differ in terms of the manner in that are less preferred (LP; e.g., Fisher et al.,
which the stimuli are presented, with stimuli 1992; Roane et al., 1998), recent investigations
presented individually (Pace, Ivancic, Edwards, have evaluated the reinforcing efficacy of LP
Iwata, & Page, 1985) or in pairs (Fisher et al., stimuli in the absence of HP stimuli more
1992), or with multiple stimuli presented closely. Roscoe, Iwata, and Kahng (1999)
concurrently (DeLeon & Iwata, 1996; Roane, evaluated the reinforcing effectiveness of HP
Vollmer, Ringdahl, & Marcus, 1998; Windsor, and LP stimuli identified through two prefer-
Piche, & Locke, 1994). In general, preference ence assessments (based on the methods
assessments produce a relative ranking of described by Pace et al., 1985, and Fisher et
stimuluspreferencesthatisbasedontherelative al.). The relative reinforcing effects of HP and
amount of time with which the stimuli were
LP stimuli were evaluated in both single- and
manipulated or the number of times one
concurrent-operants arrangements. In the sin-
stimulus was chosen relative to other stimuli.
gle-operant arrangement only the LP stimulus
An implication of such rankings is that stimuli
was available, whereas the HP and LP stimuli
that are more highly preferred (e.g., chosen on
wereavailablesimultaneouslyintheconcurrent-
80% of presentations) are more effective
operantsarrangement.ResultsindicatedthatLP
stimuliwerenotaseffectiveasHPstimuliwhen
Address correspondence to Michele D. Wallace,
California State University, Los Angeles, Charter College both were available (i.e., concurrent-operants
of Education, Division of Special Education and Coun- arrangement). However, when LP stimuli were
seling,KHC1064,LosAngeles,California90032(e-mail:
evaluatedinasingle-operantarrangement,these
[email protected]).
doi:10.1901/jaba.2008.41-177 stimuli produced increases in responding that
177
178 BECKY PENROD et al.
were similar to levels observed with HP stimuli requirement completed; Hodos, 1961). Using
in the concurrent arrangement. That is, when stimuli that were preferred similarly based on
evaluated in a single-operant paradigm, the LP the results of a preference assessment (Fisher et
stimuli were as effective as HP stimuli in terms al., 1992), Roane et al. showed that one
of reinforcement efficacy, which suggested that stimulus was associated with greater response
the concurrent availability of the HP stimuli persistence than the other under increasing
masked the potential reinforcement effects schedule requirements for all participants.
associated with LP stimuli. These findings suggested that although two
It should be noted that in the Roscoe et al. reinforcers may be equally effective when
(1999)investigationthereinforcingeffectsofthe assessed under low schedule requirements
LPstimuliwereassessedusingafixed-ratio(FR) (e.g., FR 1), they may not be equally effective
1 schedule of reinforcement only. Thus, it is when assessed under increasing schedule re-
unknown whetherLP stimuli wouldfunction as quirements.
effective reinforcers under increasing schedule Given the results of previous research, which
requirements. Previous research has shown that suggests that HP and LP stimuli may be equal
two reinforcers may produce similar levels of in terms of reinforcer efficacy when delivered
respondingunderlowschedulerequirementsbut underdenseschedules(e.g.,FR1;Roscoeetal.,
different levels of responding as schedule 1999) and the finding that two stimuli may be
requirements increase (DeLeon, Iwata, Goh, & associatedwithdifferentresponseoutputsunder
Worsdell, 1997; Roane, Lerman, & Vorndran, increasing schedule requirements (e.g., DeLeon
2001; Tustin, 1994). For example, Tustin et al., 1997; Roane et al., 2001; Tustin, 1994),
evaluatedpreferenceforattentionandcombined the purpose of the current study was to
visual and auditory stimulation under various determinetheextenttowhichLPstimuliwould
schedule requirements in a single-operant ar- function as reinforcers under increasing sched-
rangement. Response rates were similar when ule requirements.
reinforcement was delivered continuously (i.e.,
on an FR 1 schedule); however, when schedule
METHOD
requirements increased (e.g., FR 10), one
stimulus was associated with higher levels of Participants and Setting
responding. Similar results were obtained by Four children participated in this study. Sam
DeLeonetal.,whoshowedthatwhentwoedible was a 5-year-old boy who had been diagnosed
stimuli were concurrently available, a greater with autism, Cedar was a 5-year-old girl who
preference for one item over the other was had been diagnosed with autism, Elliot was an
observedastheschedulerequirementsincreased. 8-year-old boy who had been diagnosed with
Roane et al. (2001) evaluated variations in Asperger disorder, and Aden was a 5-year-old
preference between two reinforcers under boy who had been diagnosed with attention
increasing schedule requirements by assessing deficit hyperactivity disorder. All sessions were
two highly ranked stimuli under progressive- conducted in a room (3.7 m by 4.6 m) that
ratio(PR)schedules.InaPRschedule,response contained a table, chairs, and other materials
requirements increase within the course of a that varied depending on the experimental
single session, thus allowing for a relatively condition in effect (described below).
rapid evaluation of reinforcing efficacy under
increasing response requirements. Within PR Response Measurement and
schedules, relative reinforcement effects are Interobserver Agreement
identified by comparing the obtained break The target response for each participant was
point for each stimulus (i.e., the last schedule chosen based on their individual skill deficits.
ASSESSING REINFORCER POTENCY 179
Two different responses were chosen for each using a block-by-block method. Specifically,
participant to evaluate the reinforcing efficacy the number of 10-s intervals during which
oftheLPorHPstimuli.Thetargetresponsefor observers disagreed (on the occurrence or
Sam was visually tracking a moving object (i.e., nonoccurrence of the response) was subtracted
a circular blue card in the HP condition or a from the total number of intervals. For each of
triangulargreencardintheLPcondition).This thedisagreementintervals,thesmallerfrequen-
response was defined as Sam’s hand touching cycountwithinthe intervalwas dividedbythe
any part of the card after the experimenter larger frequency count within the interval to
changed its location. The target response for yield a partial agreement measure for each
Cedar consisted of picking up either one red interval. These values then were added to the
bean (HP condition) or one kidney bean (LP numberofintervalsduringwhichtheobservers
condition)withherthumbandindexfingerand agreed. This number was divided by the total
placing the bean into a small hole (approxi- number of intervals in each session and
mately the size of a dime) cut into the top of a multiplied by 100% to yield a percentage of
container. For Elliot the target response con- agreement.
sistedofpickingupeitheronekidneybean(HP A second observer simultaneously but inde-
condition) or one red bean (LP condition) with pendently collected data during 50% of all
a pair of tweezers, using his thumb and index sessions. Mean interobserver agreement for
finger as if holding a pencil, and placing the target responses was 94% for Sam (range,
bean into a plastic container. Aden’s target 82% to 100%), 95% for Cedar (range, 88%
response was tracing either the first (HP to 100%), 95% for Aden (range, 78% to
condition) or the second half (LP condition) 100%), and 94% for Elliot (range, 83% to
of the alphabet by moving a pencil along the 100%). Mean agreement for reinforcer delivery
dotted lines of each letter, such that when he was 91% for Sam (range, 76% to 100%), 97%
was finished, the traced letter correctly resem- forCedar(range,88%to100%),94%forAden
bled the appropriate form of the letter. Sam, (range, 85% to 100%), and 94% for Elliot
Cedar, and Elliot had no history with the (range, 77% to 100%).
responses that we selected. Consequently, these
3 participants received preexperimental training Preference Assessment
using either a least-to-most prompting hierar- Two preference assessments were conducted
chy (Sam and Elliot) or stimulus fading to determine each participant’s preference for
(Cedar). Aden had practiced tracing the 10fooditems.Beforeconductingthepreference
alphabet in his regular schooling and demon- assessments, each food item was sampled bythe
strated similar precision with the first and participants. For all participants, the preference
second halves of the alphabet during preexperi- assessments were conducted in a specific order
mental observations. (described below), and all preference assess-
Inadditiontotheabovementionedresponses, ments were conducted at least 2 hr before or
data also were collected on reinforcer delivery, after meals. Foods assessed during the prefer-
which was defined as the experimenter placing ence assessments were available to the partici-
an edible reinforcer directly in front of the pants only during the experimental sessions.
participant(Elliot,Cedar,andAden)ordirectly Single-stimulus method. For the single-stimu-
into the participant’s mouth (Sam). lus (SS) preference assessment, the procedures
Handheld computers equipped with Observe described by Pace et al.(1985) were followed in
software were used to collect data on the which each item was presented singly, 10 times
frequency of the target responses and reinforcer each,inacounterbalancedorderoverthecourse
delivery.Interobserveragreementwascalculated of five sessions. The stimulus was placed on a
180 BECKY PENROD et al.
plate approximately 0.7 m in front of the Reinforcer Assessment
participant. If the participant reached for the Baseline. During baseline, materials needed
item,heorshewasallowedtoconsumetheitem. forengagingintheresponseineithertheHPor
If 5 s elapsed without the participant reaching the LP condition, depending on which condi-
for the item, the experimenter prompted the tion followed, were placed in front of the
participant to pick up the food and consume it. participant, and there were no programmed
Thetrialwasthenrepeated.Ifagaintherewasno consequences for the emission of the target
response,theexperimenterremovedtheitemand response. Throughout this and all subsequent
presented the next food. The SS assessment was conditions, the therapist told Aden that he
conducted first for all participants. could respond if he wanted to but that he did
Paired-choice method. A paired-choice (PC) nothavetorespond.Sam,Cedar,andElliotdid
preference assessment also was conducted for not receive any verbal prompts throughout the
each participant following the procedures study. All baseline sessions were 10 min in
described by Fisher et al. (1992). During this duration, and two to four sessions were
assessment, each of the same 10 foods used in conducted per day.
the SS assessment was paired once with every FR 1 conditions. During the FR 1/HP
other food in a counterbalanced order. Specif- condition, the therapist delivered the HP item
ically, two food items were placed on separate to the participant on an FR 1 schedule
plates next to each other approximately 0.7 m following the emission of the target response.
in front of the participant. The participant was The FR 1/LP condition was identical to the FR
allowed access to whichever food he or she 1/HP condition, with the exception that the
reached for, and the food that was not selected therapist delivered the LP item to the partici-
was removed. Attempts to reach for both foods pant following each response emitted on an FR
at once were blocked, and the plates were 1 schedule. All FR 1 sessions were 10 min in
withdrawn briefly and re-presented. If the duration. During the FR 1 conditions, two to
participant did not reach for either food, the four sessions were conducted daily.
experimenter prompted the participant to PR conditions. During the PR-HP condition,
sample both items, and the trial was repeated. thetherapistdeliveredtheHPitemfollowingthe
If again the participant did not reach for either completionofaprogressivelyincreasingresponse
food,theexperimenterremovedbothitemsand requirement. Specifically, the therapist imple-
presented the next set of foods. mented a PR schedule based on the procedures
Based on the results of the preference described by Roane et al. (2001) in which the
assessments, HP and LP stimuli were identified therapist delivered the HP item on an FR 1
for each participant. The HP stimulus was the schedule,whichthenincreasedtoFR2,andthen
food item approached most frequently during to FR 3, and continued in this fashion until no
both assessments. Specifically, stimuli chosen responding occurred for 3 min in any given
between 80% and 100% of trials during both session. The therapist conducted two trials for
the SS and PC assessments were identified as eachschedulerequirementtopreventrapidratio
the HP items. The LP stimuli were those items strain (as described by Roane et al.).
for which there was the largest difference The PR-LP condition was identical to the
between the two assessments. Specifically, PR-HP condition, with the exception that the
stimuli chosen between 80% and 100% of therapist presented the LP item rather than the
trials during the SS assessment but less than HP item contingent on responding using a PR
40% of trials during the PC assessment were schedule. The same PR schedule was in effect
selected as LP items. for both the HP and LP stimuli.
ASSESSING REINFORCER POTENCY 181
Basedonthe3-minterminationcriterion,the stimuli. The cumulative number of responses
duration of sessions during the PR conditions across the PR-HP and PR-LP conditions was
varied as a function of response persistence. determined by adding the total number of
There was no limit imposed on the duration of responses from the first session to the total
each session; however, no session continued for number of responses from the second session
more than 1 hr for any participant. Across and so on.
participants, the mean session durations were
24 min and 17 min during the PR-HP condi-
RESULTS
tion and PR-LP condition, respectively. Based
Preference Assessments
on time constraints, one PR session was
conducted per day for each participant. Results from the preference assessments are
depicted in Figure 1. HP stimuli (i.e., stimuli
Experimental Design chosen frequently during both assessments)
Theexperimental conditionsdescribed above were peanut butter balls for Sam, cheese
were arranged in a multiple baseline design crackers for Cedar, rice crispy treats for Elliot,
across participants. The HP and LP conditions andchocolatechipcookiesforAden.LPstimuli
were introduced in a counterbalanced order (i.e., stimuli chosen frequently during the SS
across participants. That is, 2 participants were preference assessment but rarely during the PC
exposed to the HP conditions first followed by preference assessment) were pretzels for Sam,
the LP conditions, and the other 2 participants licoriceforCedar,cheesecrackersforElliot,and
were exposed to the LP conditions followed by raisins for Aden.
the HP conditions.
Reinforcer Assessment
Data Analysis Figure 2 depicts the results of the reinforcer
The relative reinforcement effects of HP and assessments. During baseline, Sam rarely en-
LP stimuli were evaluated by comparing gagedinthetargetresponse(M50.4responses
response rates associated with each stimulus perminute).DuringtheFR1/LPandtheFR1/
across conditions and the average break point HP conditions, rates of responding increased
associated with each stimulus under the PR relative to baseline. Rates of responding were
schedules. A response rate for each session was slightly higher in the FR 1/LP condition (M 5
determined by dividing the total number of 5.0)thanintheFR1/HPcondition(M53.1).
responses by the session duration (in minutes) During the PR-LP condition, rates of respond-
to yield the number of responses per minute. ing were lower (M 5 1.9) than during the PR-
Based on differences between the types of HP condition (M 5 5.7). During the PR-LP
schedules in effect, the response rates in the condition, the obtained break points for each
two FR conditions and those obtained in the sessionwereasfollows:FR5,FR7,FR8,FR6,
two PR conditions were compared separately. FR 3, FR 6, FR 6; with an average break point
During the PR conditions, the average break of FR 6. The obtained break point for each
point for each stimulus was determined by session of the PR-HP condition was as follows:
adding the obtained break points (i.e., the last FR 13, FR 18, FR 16, FR 6, and FR 13, and
schedulerequirementcompleted)acrosssessions the average break point was FR 13.
for the HP and LP items independently and Similar results were obtained for Cedar.
then dividing this value by the total number of During baseline, Cedar rarely engaged in the
sessions. In addition, cumulative records were response (M 5 0.2 responses per minute).
generated to determine if different patterns of Similar levels of responding were observed in
responding were associated with HP and LP boththeFR1/LPandFR1/HPconditions(Ms
182 BECKY PENROD et al.
Figure1. PercentageofedibleitemschosenduringtheSSandPCpreferenceassessmentsforSam,Cedar,Elliot,and
Aden.The single asterisk denotes theHPstimuli, andthe double asterisk denotes the LPstimuli.
ASSESSING REINFORCER POTENCY 183
Figure2. ResponsesperminuteofthetargetresponseforSam,Cedar,Aden,andElliotduringbaseline,FR1,and
PRsessions withHPandLPreinforcers.
184 BECKY PENROD et al.
5 3.3 and 2.5 responses per minute, respec- varied greatly across the PR-HP and PR-LP
tively). Lower rates of responding occurred in conditions. To illustrate, for Sam the total
thePR-LPcondition(M51.2)thaninthePR- session duration was approximately 219 min,
HP condition (M 5 2.3). The average break with a total of 377 responses across seven
point during the PR-LP condition was FR 2, sessions during the PR-LP condition, as
and the break point for each session was FR 2, opposed to a total duration of approximately
FR4,FR1,FR5,FR1,andFR1.Theaverage 183 min, with a total of 1,083 responses across
break point during the PR-HP condition for five sessions during the PR-HP condition.
Cedar was FR 4, and the break points for each Similar patterns were observed for Cedar (i.e.,
sessionwereFR5,FR2,FR6,FR4,andFR4. 48 min of total session time and 73 total
During baseline, Aden rarely engaged in the responses in the PR-LP condition, 57 min of
target response (M 5 0.8 responses per total session time and 141 responses in the PR-
minute). During the FR 1/HP and the FR 1/ HPcondition),Aden(i.e.,38totalminutesand
LP conditions, rates of responding increased 118 total responses in the PR-LP condition,
relative to baseline and were lower in the FR 1/ 114totalminutesand259responsesinthePR-
HP condition (M 5 3.9) than in the FR 1/LP HP condition), and Elliot (i.e., 157 total
condition (M 5 5.0). Rates of responding were minutes and 546 total responses in the PR-LP
slightly lower in the PR-HP condition (M 5 condition, 264 total minutesand 643 responses
2.2) than in the PR-LP condition (M 5 3.1). in the PR-HP condition). Thus, across partic-
The average break point was FR 6 for the PR- ipants the means for the PR-LP condition were
HP condition, and the obtained break points 116 minand279responses(M52.4responses
for each session were FR 7, FR 6, FR 4, FR 8, per minute) and were 155 min and 532
FR 4, and FR 4. The obtained break points for responses (M 5 3.4) for the PR-HP condition.
each session during the PR-LP condition were
FR 4, FR 6, FR 3, and FR 3, and the average Cumulative Responses
break point was FR 4. Cumulative records allow a direct inspection
For Elliot, baseline rates of responding were of rate and changes in rate otherwise not
low (M 5 0.9 responses per minute). During possiblewithothertypes ofgraphs.Specifically,
theFR1/HPandtheFR1/LPconditions,rates relative rates of responding can be determined
increased relative to baseline and were higher in by visually comparing the slopes of two
the FR 1/HP condition (M 5 5.6) than in the performances; if the slope of one performance
FR 1/LP condition (M 5 1.7). During the PR- is steeper than the other, this indicates a higher
HP condition, rates of responding were some- response rate. In addition, cumulative records
what lower (M 5 2.5) than those observed in reveal patterns of responding (i.e., break and
the PR-LP condition (M 5 3.0). The break run) not shown by a typical line graph.
points for each session during the PR-HP The top panel of Figure 3 depicts the
condition were FR 8, FR 7, FR 12, FR 6, FR cumulative number of responses across all
5,FR9,FR4,FR5,andFR8,andtheaverage sessions during the PR-LP and PR-HP condi-
break point was FR 7. The average break point tions for Sam. During the PR-HP condition,
during the PR-LP condition was FR 7, and the Sam consistently responded across the sessions
break points for each session were FR 6, FR 4, and the slope of the line was steep, but during
FR 12, FR 10, FR 11, FR 2, FR 5, and FR 5. the PR-LP condition, responding was accom-
It should be noted that although similar paniedbymorebreaksandthustheslopeofthe
response rates and break points were observed line was not as steep. Even though the pattern
under the PR conditions for most participants, of responding was fairly stable during both the
session length and overall response output PR-HP and the PR-LP conditions, it should be
ASSESSING REINFORCER POTENCY 185
Figure3. CumulativenumberofresponsesacrosssessionsduringthePR-HPandPR-LPconditionsforSam,Cedar,
Aden,andElliot.
186 BECKY PENROD et al.
noted that more responding occurred in a tion than under the PR-LP condition for 3 of
shorter period of time and in fewer sessions the 4 participants (Sam, Cedar, and Aden). For
during the PR-HP condition. these participants, the average break point was
The second panel of Figure 3 depicts the lower in the LP condition than in the HP
same data for Cedar. During the PR-HP condition;however,ingeneral,thedifferencein
condition, Cedar consistently responded across break points was minimal. Specifically, for 2
the PR-HP sessions and responding was participants (Cedar and Aden) the difference in
accompanied by more breaks in the PR-LP break points was within two responses of one
condition.Again,agreaternumberofresponses another (e.g., FR4 in the PR-LP condition and
was observed in fewer sessions during the PR- FR 6 in the PR-HP condition for Aden), and
HP condition. for 1 participant (Elliot), the average break
The third panel of Figure 3 depicts the same point was identical in both PR conditions.
data for Aden. During the PR-HP condition,
Aden consistently responded and the slope of
DISCUSSION
the line was steep. Although initially the slope
Results of the current study replicated those
of the line during the PR-LP condition
of Roscoe et al. (1999) by demonstrating that
resembled the slope during the PR-HP condi-
LP stimuli maintained responding during FR 1
tion, pausing was observed after Session 3,
schedules of reinforcement for all participants.
which is denoted by divergence in the slopes of
Moreover, the current results extended those
the two lines thereafter.
obtainedbyRoscoeetal.,inthattheLPstimuli
The bottom panel of Figure 3 depicts the
used in the current investigation also main-
same data for Elliot. Similar patterns of
tained responding under PR schedules of
responding were observed during both the
reinforcement. It is noteworthy that, although
PR-LP and PR-HP conditions, in that the
LP stimuli functioned as reinforcers under both
slopes of the lines were similar, suggesting that
FR 1 and PR schedules, the LP stimuli were
the HP and LP stimuli were equally effective in
generally not as effective as the HP stimuli in
maintainingrespondingunderincreasingsched-
ule requirements. terms of response persistence under increasing
PR schedules. Nevertheless, the current results
In summary, results indicated that baseline
suggestthatLPstimulimayfunctionaseffective
rates of responding were low for all 4
participants. When an FR 1 schedule was in reinforcers in a variety of contexts.
effect, responding increased relative to baseline Roaneetal.(2001)demonstratedthathigher
in both the LP and HP conditions for all rates of responding were associated with one of
participants. For Sam and Aden, higher rates of two stimuli under increasing schedule require-
responding were observed under the FR-LP ments, suggesting that even though two stimuli
conditionthanundertheFR-HPcondition;for may be equally effective reinforcers for low-
Elliot, higher rates of responding were observed effort responses, one stimulus may be more
undertheFR-HPconditionthanundertheFR- effective than the other for high-effort respons-
LP condition. Cedar was the only participant es.ThecurrentresultssupportthoseofRoaneet
for whom similar rates of responding were al., in that both LP and HP stimuli were
observed under boththe LP and HP conditions similarly effective under low schedule require-
when an FR 1 schedule was in effect. When a ments, yet under higher schedule requirements
PR schedule was implemented, responding one stimulus (the HP item) was generally the
maintained under both the LP and HP more effective reinforcer. The current results
conditions for all participants; however, more also highlight the use of PR schedules when
responding occurred under the PR-HP condi- conductingreinforcerassessments.PRschedules