Você está na página 1de 13

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2017, 9999, n/a–n/a NUMBER 9999 ()

USING PROGRESSIVE RATIO SCHEDULES TO EVALUATE TOKENS AS


GENERALIZED CONDITIONED REINFORCERS
DANIELLE RUSSELL, EINAR T. INGVARSSON AND JENNIFER L. HAGGAR
UNIVERSITY OF NORTH TEXAS AND CHILD STUDY CENTER

AND

JOSHUA JESSEL
UNIVERSITY OF NORTH TEXAS

The properties of operant reinforcers are dynamic and dependent on a number of variables, such
as schedule and effort. There has been sparse research on the generalized conditioned properties
of token reinforcement. We evaluated leisure items, edible items, and tokens using a progressive
ratio schedule with three children with diagnoses of ASD and developmental delays. The highest
break points occurred during the token reinforcement condition for two out of three partici-
pants, but response rates tended to be higher with edibles. We then evaluated the effects of pre-
session access to edibles on the break points of edible items and tokens with two participants.
Break points decreased only in the edible reinforcement condition, and the participants chose to
work for leisure items rather than edibles when presession access to edibles was in place. These
findings suggest that the tokens functioned as generalized conditioned reinforcers.
Key words: generalized conditioned reinforcement, progressive ratio schedule, motivating
operations, tokens

Efficient identification of reinforcers and eval- Mazur, 1997; Piazza, Fisher, Hagopian, Bow-
uation of their ongoing effectiveness is a critical man, & Toole, 1996). However, reinforcer effec-
aspect of behavioral interventions. Applied tiveness may fluctuate dependent on various
researchers typically conduct preference assess- parameters, including fixed ratio (FR) require-
ments to determine a relative ranking of stimulus ments (DeLeon, Iwata, Goh, & Worsdell, 1997;
value, under the assumption that the ranking Tustin, 1994). DeLeon et al. (1997) compared
predicts reinforcer effectiveness (e.g., DeLeon & the reinforcing potency of two similar stimuli
Iwata, 1996; Fisher et al., 1992; Pace, Ivancic, previously found to be roughly equivalent in
Edwards, Iwata, & Page, 1985; Roane, Vollmer, preference ranking. When these stimuli were
Ringdahl, & Marcus, 1998; Windsor, Piche, & available under concurrent FR 1 schedules, both
Locke, 1994). After preference is determined, a participants—adults diagnosed with develop-
reinforcer assessment with mastered responses mental disabilities—showed little or no differ-
might be conducted, during which each response ences in preference for either stimulus. However,
produces access to a preferred stimulus (Fisher & when the schedule requirements increased to FR
5, a clear preference for one of the stimuli
This study was conducted in partial fulfillment of the emerged for both participants. This difference
requirements for the master’s degree of the first author at maintained when the schedule requirements
the University of North Texas. We thank Kimberly Fair-
man and Melinda Robison for assistance with data collec-
increased to FR 10. Thus, reinforcer potency is
tion. Joshua Jessel is now affiliated with Queens College. not an absolute value for a given stimulus.
Address correspondence to: Einar Ingvarsson, who is Given that schedule requirements influence
now at the Virginia Institute of Autism, 943 Glenwood reinforcer potency and the terminal goal of
Station LN, STE 201, Charlottesville, VA 22901.
Email: eingvarsson@viaschool.org most applied efforts is to identify preferred
doi: 10.1002/jaba.424 stimuli that will maintain and increase
© 2017 Society for the Experimental Analysis of Behavior
1
2 DANIELLE RUSSELL et al.

responding over time, Roane, Lerman, and reinforcers, the tokens may acquire conditioned
Vorndran (2001) suggested using progressive reinforcing properties (Cooper et al., 2007;
ratio (i.e., PR) schedules to evaluate reinforcer DeFulio, Yankelevitz, Bullock, & Hackenberg,
potency. In a PR schedule, the number of 2014; Hackenberg, 2009). Early laboratory
responses required to access reinforcement sys- research evaluated token systems and the extent
tematically increases within-session. With PR to which tokens functioned as conditioned rein-
schedules, relative reinforcement effects can be forcers (Cowles, 1937; Kelleher, 1956, 1958;
identified by comparing the obtained break Wolfe, 1936). These studies demonstrated that
point for each stimulus (Hodos, 1961). One poker chips could be established as conditioned
advantage of this method is that it allows for an reinforcers for chimpanzees.
evaluation of reinforcing efficacy under increas- There are a number of advantages to using
ing response requirements. generalized conditioned reinforcers (Kazdin &
In two experiments, Roane et al. (2001) evalu- Bootzin, 1972). Conditioned reinforcers are less
ated the effectiveness of preferred stimuli using susceptible to the effects of satiation and can be
PR schedules. Four individuals who were diag- used to bridge the delay between initial
nosed with developmental disabilities participated responses and access to preferred items or activi-
in the first study and three of those individuals ties. Additionally, conditioned reinforcers main-
participated in the second study. Roane tain performance and responding over extended
et al. found that whereas response patterns of periods of time, involve minimal interruptions,
higher and lower preferred stimuli failed to differ- and allow for sequences of responses to be rein-
entiate with FR 1 schedules, PR schedules pro- forced easily. They can also be more potent than
duced differences in break points between various any single primary reinforcer.
stimuli. They also discovered that stimuli with Ayllon and Azrin (1968) described several
higher break points were overall more effective in potential advantages of tokens over other gener-
the treatment of destructive behavior. alized conditioned reinforcers (such as praise
Even following more extensive reinforcer statements). First, the number of tokens quan-
assessments using PR schedules, reinforcer effec- titatively represents the amount of reinforce-
tiveness can fluctuate as a function of motivat- ment earned. Second, tokens are portable, can
ing operations (MO; Laraway, Snycerski, remain with the client, and can be delivered
Michael, & Poling, 2003). A common concern easily across a variety of contexts and environ-
with edible and leisure reinforcers is the deleteri- ments. Third, there is no limit to the potential
ous effects of abolishing operations on the rein- number of tokens one can earn. Fourth, tokens
forcing potency of these stimuli (McAdam are durable and allow for consistency and stan-
et al., 2005; Murphy, McSweeney, Smith, & dardization. Fifth, tokens offer a visual repre-
McComas, 2003). One major advantage of sentation of improvement and progress (see
tokens is that they should not be subject to MO also Kazdin & Bootzin, 1972).
effects specific to each backup reinforcer, if they Whereas a considerable number of basic
function as generalized conditioned reinforcers research studies with nonhuman subjects have
(Cooper, Heron, & Heward, 2007). explored the reinforcing properties of tokens, little
Token economies are reinforcement systems is known about the extent to which tokens func-
that involve the contingent delivery of tokens tion as generalized conditioned reinforcers in the
following the occurrence of the targeted behav- context of application (Hackenberg, 2009). An
ior. Accumulated tokens can later be exchanged exception is a study by Moher, Gould, Hegg, and
for back-up reinforcers (e.g., food and toys). Mahoney (2008). The participants were five chil-
Through their association with back-up dren with autism and other developmental
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 3

disabilities. Moher et al. conducted reinforcer spectrum disorder. According to the


assessments with tokens and manipulated the Woodcock-Johnson III Test of Achievement
MO (i.e., presession free access) for the backup (WJ-III; Woodcock, McGrew, & Mather,
edible reinforcers. In a parametric analysis, tokens 2001), he functioned on a late-first grade/early-
were exchangeable with one, two, or three backup second grade level. Prior to the study, Damien
reinforcers. Moher et al. found that tokens exclu- had experience using tokens for about
sively exchangeable for one backup reinforcer lost 18 months. On average, Damien completed
the ability to maintain responding on dense FR two to three responses before earning each
schedules when preceded by sessions of free access token, and he earned about 10 tokens before
to that reinforcer. However, tokens continued to accessing reinforcers. According to his teacher’s
maintain responding when exchangeable for mul- report, Damien typically chose to exchange his
tiple backup reinforcers, even when the value of tokens for video games, board games, and occa-
one of the reinforcers was previously abolished via sionally more lessons.
satiation. These findings suggest that the tokens Carmen was an 8-year-old girl who was diag-
functioned as generalized conditioned reinforcers. nosed with developmental delays and functioned
We conducted the current study as an exten- on a kindergarten grade level, according to the
sion of Moher et al. (2008) by including PR WJ-III. She had experience using tokens for about
schedules to determine reinforcer effectiveness. 14 months prior to the study. On average, she
In addition, we included qualitatively different earned about 10 tokens before accessing reinforcers
reinforcers (leisure items and edible items) in in the classroom. With difficult tasks, only one
an evaluation of the effects of different levels of response was required per token, but with mas-
presession exposure on the effectiveness of tered tasks, two to three responses were required
tokens (exchangeable for both edibles and lei- before earning a token. According to her teacher’s
sure items) versus edibles only. report, Carmen typically chose to exchange her
tokens for small toys (i.e., Polly Pockets, LEGO®
bricks, etc.), board games, and candy.
METHOD
Zane was a 7-year-old boy who had a diagnosis
Participants of attention deficit hyperactivity disorder and
Three students attending a school for children autism spectrum disorder. He functioned on a
with developmental disabilities served as partici- first grade level according to the WJ-III. Zane
pants. The school director identified participants had experience using tokens for about 2 years.
based on whether they had experience with token The number of responses required to earn a single
economies, had low levels of problem behavior token varied from one to five, depending on task
and noncompliance, and had acquired a sufficient difficulty. According to his teacher’s report, Zane
number of math facts to be used as tasks during typically earned about 10 tokens before accessing
the study. All three children, Damien, Carmen, reinforcers and chose to exchange his tokens for
and Zane, participated in Phase 1 of the study. LEGO® bricks, Wii®, iPad®, and jelly beans.
However, only Damien and Carmen participated
in Phase 2. All three participants had received
standard token training, based largely on the pro- Setting
cedures described by Cooper et al. (2007). The All conditions and sessions were conducted in
participants had extensive experience (at least a room (approximately 2 m x 3 m) containing a
14 months) with token economies at the school. bookshelf, two desks, four chairs, a tripod with
Damien was a 7-year-old boy who was diag- video camera, and a three-tier plastic storage bin.
nosed with developmental delays and autism Other children were not present in the room
4 DANIELLE RUSSELL et al.

during sessions. One session was conducted per collection app on an iPod. The experimenters
day for each participant, 3 to 5 days per week (first and third author) and trained observers
(depending on participant availability). collected data on correct and incorrect
responses, total session time, and the progres-
sive ratio (PR) break point. Observers scored a
Tasks and Materials
response each time the child placed the com-
During sessions, participants worked on pre-
pleted task (math problem) in the pan. When
viously mastered academic skills. The first
the child had written the correct answer on the
author interviewed their teachers to identify
sheet, observers scored that response as correct.
skills that had previously been mastered and
Observers counted a response as incorrect when
selected tasks on the basis of each participant’s
either (a) the answer was not written clearly
individual level of development and academic
enough for the therapist to read or (b) the child
ability. Addition math facts were chosen for all
wrote the wrong answer on the sheet. The
participants. Carmen worked on addition math
number of correct responses was then divided
facts 0 and 1 + 0 through 11. Both Zane and
by the duration of the session to calculate rate
Damien worked on addition math facts
of responses per min (rpm). All participants
0 through 5 + 0 through 11.
responded at near 100% accuracy throughout
In all sessions and conditions, one participant
the study (M = 96.4%, SD = 2); therefore, we
was seated at a desk and the experimenter
report only correct responses.
(either the first or third author) sat to his or her
Reinforcer access time was subtracted from
right side. Each math fact was printed on a
session duration prior to calculating rate. Session
10.8 cm x 14.0 cm laminated paper. An alumi-
time began after delivering the initial instruc-
num pan (22.9 cm x 27.9 cm) with a card
tions and ended when the child asked to be fin-
attached with the printed phrase “I’m done”
ished or when the break point criterion was
was located on the far left corner of the desk.
reached. Observers circled the PR break point
The math fact sheets were stacked in the far
when the child was off-task for 2 min or when
right corner of the desk. After taking a math fact
the child said, “I’m done.” If the session ended
sheet from the stack and writing the answer on
after 2 min of off-task behavior, this duration
the sheet under the equals sign, the participant
was also subtracted prior to calculating the rate.
placed the completed math fact in the pan. The
In the token condition, the experimenter also
participant was given a dry-erase marker to write
noted whether the participants selected an edible
the answers and a tissue to erase mistakes.
or leisure item as a backup reinforcer.
The token board consisted of a laminated file
folder with the label “token board” at the top.
It had three long strips of Velcro centered Interobserver Agreement (IOA)
beneath the label. The tokens were brown
Interobserver agreement (IOA) was assessed
paper circles, about the size of a quarter, lami-
for at least 30% of sessions by having a second
nated with the word “token” printed across the
observer independently collect data from videos
center and Velcro affixed to the back. In the
using pencil and paper. We then counted point
token condition, the therapist placed the token
by point agreements versus disagreements
board in between the pan and the sheets.
across the two observers for incorrect and cor-
rect responses and PR break point, and calcu-
Measurement lated IOA for each session by dividing the total
All sessions were video recorded and the pri- number of agreements by the total number of
mary observers scored responses using a data agreements plus disagreements and multiplying
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 5

by 100. Mean IOA for both Carmen and in the upcoming progressive ratio conditions.
Damien was 100% for correct responses and For Damien, the top three edibles were 3 Mus-
PR break point. Mean interobserver agreement keteers Bar®, M&M®, and chocolate chip
for Zane was 99% (range, 98% - 100%) for cookie. iPad®, LEGO® bricks, and an Angry
correct responses and 100% for the PR break Birds Game were the top three leisure items.
point. For Carmen, the top three edibles were jelly
bean, Oreo®, and KitKat®. The top three lei-
Procedural Fidelity sure items were iPad®, bowling, and an Angry
Birds game. For Zane, the top three edibles
To assess procedural fidelity, we created a
were Doritos®, Oreo®, and KitKat®. The top
checklist specifying important aspects of the pro-
three leisure items were iPad®, bowling, and
cedures. These included giving instructions, pre-
LEGO® bricks.
senting the task, delivering consequences
Progressive ratio assessment. There were three
according to the PR schedule, implementing the
conditions in the progressive ratio assessment:
break point, managing the token system, and
edible, leisure, and token reinforcement. The
implementing presession access procedures. An
top three edibles from the preference assess-
independent observer observed randomly selected
ment were used in the edible condition, the
video-recorded sessions (with at least one session
top three leisure items were used in the leisure
from each phase and condition for each partici-
condition, and tokens exchangeable for any of
pant) and scored experimenter behavior using the
those six items were used in the token condi-
checklist. This observer had not previously been a
tion. For the duration of the study (both
part of running the study or collecting data on
Phases 1 and 2), the participants did not have
participant behavior. We collected procedural
access to these edibles and leisure items at the
fidelity data for 25.8% of sessions for Damien,
school. Additionally, the experimenters asked
23.8% of sessions for Carmen, and 25% of ses-
the participants’ parents to ensure that they did
sions for Zane. Procedural fidelity was 100% for
not have access to these items at home.
Damien, 99.5% (range, 89.5%-100%) for Car-
Prior to each session, the experimenter placed
men, and 100% for Zane. We also obtained IOA
the items relevant to the current experimental
for the procedural fidelity measures for 12.5% of
condition on the table and the child selected
scored sessions for Damien, 20% of sessions for
what he or she wanted to earn during that ses-
Carmen, and 20% of sessions for Zane. IOA for
sion. Immediately after the child selected an
the procedural fidelity measures was 100% for all
item, the experimenter removed all other items.
three participants.
Session materials were then placed on the child’s
desk (marker, tissue for erasing, math facts, and
Procedures: Reinforcer Assessment (Phase 1) “finished” pan). The experimenter then issued
Preference assessment. Two paired-stimulus the following instruction: “Today you will be
preference assessments were conducted with working on math facts. You will receive
each child (Fisher et al., 1992). One assessment _______ for completing math problems. Just
included edible items and the other included grab a sheet from the stack, answer the problem
leisure items. Five to eight items were paired and then put it in the finished bin. If at any
once with every other item in a counterba- point you want to be all done working, you can
lanced sequence. Following a selection tell me, ‘I’m done’ (experimenter points to the
response, the participant had access to the edi- ‘I’m done’ card). What do you say when you
ble/leisure item for 1 min. The top three edible want to be all done?” The session began after
and leisure items were then used as reinforcers the experimenter delivered the instructions and
6 DANIELLE RUSSELL et al.

the child answered the question by saying, finished or when no target responses occurred
“I’m done.” for 2 min. When this occurred in the edible
During sessions, the children worked on and leisure conditions, the experimenter imme-
math fact tasks by writing the solution to each diately walked the child back to his or her class-
problem in the designated area on the lami- room. In the token condition, the experimenter
nated sheet. The participant then placed the and child counted the number of tokens and
completed sheet in the “finished” pan, took the the child was given access to the item he or she
next sheet from the stack, and continued. selected prior to the session. Each token was
When the participant completed the current worth one edible or 1-min access to the leisure
ratio schedule requirement, the experimenter item. For example, if the break point was PR
placed a hand on the math sheets and delivered 10, the child had earned 10 tokens that were
the programmed consequence. The ratio sched- exchangeable for 10 min of play time with the
ule requirements increased arithmetically (step leisure item or 10 pieces of the edible.
size of 1) beginning with one response to access Design. The three reinforcement conditions
reinforcement, then 2, 3, and so on. The exper- were arranged sequentially according to an
imenter delivered an edible item in the edible ABC design with the order of conditions coun-
condition, a leisure item in the leisure condi- terbalanced across participants. Sessions were
tion, and a token in the token condition. The conducted until the experimenters (first and
child’s selection at the beginning of session second authors) judged that the data were suffi-
determined which edible or leisure item was ciently stable. A minimum of five sessions was
used. In the edible and leisure conditions, the conducted in each condition. Stability was
participant was given 1-min access to consume judged by first looking at the most recent three
the edible or play with the leisure item after data points. The condition ended if a trend was
completing each schedule requirement. If the absent. If a possible trend was present, the
participant made an error, the therapist said, experimenters examined the most recent five
“That’s not right” and allowed the child a few data points and ended the condition if a trend
seconds to correct the error. If the participant was absent. The number of sessions in each
did not know the answer, the experimenter condition for all participants ranged from 5 to
vocally prompted the correct answer. Any 13 (M = 8.9, SD = 2.6).
incorrect or prompted answers were counted
toward the FR schedule requirement after the
child wrote the correct answer on the sheet and Procedures: Presession Access Evaluation
placed it in the pan. Thus, an incorrect (Phase 2)
response followed by a prompted response was Baseline. There were two conditions in this
counted only once towards the schedule phase: edible and token. The leisure items con-
requirement. If the participant tried to engage dition was not included because only the moti-
in conversation during the task or reinforcer vating operation (MO) for edibles was
time, the experimenter responded by saying, manipulated. However, the participants could
“We can talk about that later if you want” and still choose to earn leisure items in the token
ignored all additional initiations within 30 s of condition. The procedures were similar to the
the verbal redirection. The experimenter ver- progressive ratio assessment. Unlike the pro-
bally redirected the child again if needed after gressive ratio assessment, the participants did
the 30 s had passed. not choose backup reinforcers at the beginning
The PR break point was reached and the ses- of the session, but rather at the end of session
sion was terminated when the child asked to be after all the tokens had been earned. This
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 7

change was made to better approximate the (i.e., unlimited time). Again, this was because
procedures in place in their classroom. she tended to eat very slowly. Presession access
Presession access comparison. During this lasted on average 23 min and 40 s for Carmen
phase, each session in both conditions was pre- during this phase.
ceded by a presession access period. During Return to baseline. The second baseline was
presession access, the child was seated at the identical to the initial baseline, except the par-
desk and given access to the top three edibles ticipants had access to a broader array of leisure
from the preference assessment. The instruc- items as backup reinforcers in the token condi-
tions given were as follows: “You can eat as tion, as in the preceding presession-access
much of the snacks as you want. If at any point condition.
you are full, you can just tell me, ‘I’m full’.” Experimental design. During the presession
The experimenter then started a timer. During access comparison, the progressive ratio condi-
the first presession access condition the maxi- tions (token vs. edible) were conducted in a
mum time was set to 5 min. When the timer multielement design. The order of the condi-
beeped or when the participant stated that he tions was determined in a quasirandom fashion
or she was full or wanted to stop eating, the PR such that no condition was conducted more
session for that day was initiated. than two times in a row. The experimental
The initial duration of presession access did evaluation of the effects of presession access was
not result in a change in behavior. Because the conducted in a reversal design.
participants consumed the edibles rather slowly,
we concluded that 5 min might not be enough
to allow for satiation. Thus, we increased the
RESULTS
presession access period to 10 min. At the same
time, we observed that the participants had Phase 1: Reinforcer Assessment
rarely picked a leisure item as a backup rein- Figure 1 displays the response rate and break
forcer in the token condition in baseline. In point performance across the three reinforce-
fact, Carmen had picked leisure items only ment conditions during the progressive ratio
once, and Damien had never picked the leisure assessment. Break points for each condition
items. We suspected that the leisure items were relatively similar for Damien, with the
might have lost their reinforcing value. Addi- highest mean break point in the edible condi-
tionally, Damien refused to exchange his tokens tion (M = 10.2, SD = 2.1), followed by the lei-
for either leisure or edible items in the token sure items (M = 9.4, SD = 3.6), and token
condition during the 5-min presession access, reinforcement (M = 7.6, SD = 3.7). The
opting instead to go back to his classroom. response rates were also highest in the edible
Therefore, we also increased the range of leisure condition (M = 6.5 rpm, SD = .8), followed
items that the participants could choose as by the leisure items (M = 5.9 rpm, SD = .9),
backup reinforcers in the token condition. and then tokens (M = 5.6 rpm, SD = 1.1).
From that point on, the participants could For Carmen, leisure items had a lower mean
select any toy, game, or activity available in the break point (M = 3, SD = 2.3) than edible
classroom. However, the edibles available in the (M = 5.1, SD = 2.9) and token (M = 5.2,
edible condition remained unchanged. Subse- SD = 3.1) reinforcers. The highest mean
quently, we conducted a third presession access response rate occurred in the edible condition
condition with Carmen only, in which we (M = 5.7 rpm, SD = 1.2), followed by the lei-
didn’t set specific time limits for her to con- sure items (M = 5.2 rpm, SD = 1.1), and
sume the edibles prior to the session tokens (M = 4.5 rpm, SD = 1.5).
8 DANIELLE RUSSELL et al.

8 15
Damien

6
10
4
5
2 Token
Edible
Leisure
0 0

8 15
Carmen
Responses per minute

Break point
6
10
4
5
2

0 0

15 30
Zane

10 20

5 10

0 0
10 20 30
Sessions

Figure 1. Rate of responding and break points during the progressive ratio assessment (phase 1).

The highest mean break point for Zane was difference in break points across the edible rein-
observed for the token condition (M = 15.7, forcers condition (M = 13.4, SD = 5.7) and the
SD = 5.2), followed by the edibles (M = 11.7, token reinforcers condition (M = 15.4, SD =
SD = 3.3), and then leisure items (M = 7.4, 6.3). These undifferentiated results between
SD = 2.8). The highest response rates were the edibles (M = 9.3, SD = 1.2) and tokens
observed in the edible condition (M = 10.7 (M = 9, SD = 1) maintained after the 5-min
rpm, SD = 1.3). The token condition was next presession access to the edible items was intro-
(M = 10 rpm, SD = 1.4), and the lowest duced. Following the introduction of the 10-
response rates occurred in the leisure item con- min presession access condition, Damien
dition (M = 6.5 rpm, SD = .7). contacted the ratio requirements in only one
session during the edible reinforcement condi-
tion (M = 1.75, SD = 3.5). However, break
Phase 2: Presession Access Comparison points remained stable and high during the
During the presession access baseline for token reinforcement condition (M = 10.5,
Damien (Figure 2, top panel), there was little SD = 2.6). Break points during the edible
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 9

Figure 3. Cumulative selections of backup reinforcers


Figure 2. Break points for Damien (top panel) and
that were exchanged during each session in which tokens
Carmen (bottom panel) across varying durations of preses-
were earned in phase 2.
sion access to edible reinforcers in phase 2.

reinforcement condition (M = 5.5, SD = 3.8) Carmen during the 5-min (M = 3.5, SD = .6)
returned to previous levels during the second and 10-min (M = 3.7, SD = .6) presession
baseline. access condition. However, level of break
For Carmen, higher break points were points reduced dramatically and responding
achieved in the token reinforcement condition was eliminated altogether in many sessions dur-
(M = 7.1, SD = 2.1) in comparison to the edi- ing the unlimited access condition (M = 1.4,
ble reinforcement condition (M = 6.4, SD = SD = 1.8). Higher break points were again
2.9) during baseline (Figure 2, bottom panel). obtained in the edible (M = 3.7, SD = .6) and
Although there was a downward trend in token token (M = 6.7, SD = 2.9) conditions during
condition break points during the 5-min pre- the return to baseline.
session access period (M = 6, SD = 2.2), this Figure 3 shows whether the participants
condition was largely unaffected by the changes selected edible or leisure backup reinforcers
in presession access. High break points contin- during token reinforcement sessions. Both
ued to occur during the 10 min (M = 6.7, Damien and Carmen exclusively exchanged
SD = .6) and unlimited conditions (M = 6.8, their tokens for edible items during baseline.
SD = 1.5). A slight decrease in the level of edi- During the 5-min presession access, Damien
ble condition break points was observed for refused to exchange his tokens for either edibles
10 DANIELLE RUSSELL et al.

or tokens, choosing instead to return to his conditioned reinforcers with two of the partici-
classroom. Carmen continued exchanging her pants. The results showed that presession access
tokens for edibles (with the exception of the to food detrimentally affected progressive ratio
last session in this phase). Their selections break points only when edibles were used as
shifted to leisure items when the 10-min pre- reinforcers, but not when tokens (exchangeable
session access to edibles was introduced (which for leisure items and edibles) were used as rein-
coincided with a broader range of leisure items forcers. Further, when presession access to food
being made available), and Damien continued was implemented, the participants chose to
to select leisure items even during the second exchange their tokens for leisure items rather
baseline. Carmen chose leisure items in the first than edibles. These results suggest that tokens
token session in this phase, edibles in the sec- functioned as effective generalized conditioned
ond session, but declined both in the third and reinforcers in the current study.
final session, opting to go to her classroom The results of Phase 1 suggested that tokens
instead. After the range of available leisure maintained responding for a longer period and
items was increased, both participants selected at leaner schedules of reinforcement (for two
leisure items that had not been available before. participants) but at a somewhat slower pace rel-
Damien chose iPod Touch® six times. How- ative to edible reinforcement. However,
ever, it should be noted Damien played the because the token, edible, and leisure reinforce-
same game on his personal iPod Touch as he ment conditions were not replicated within
had on the iPad in the leisure condition; the participants, conclusions regarding the differen-
difference was that a higher game level was tial effectiveness of tokens cannot be made.
available on his personal iPod. Interestingly, Further research should incorporate a sound
there were three occasions on which he chose experimental design to examine if tokens are
not to engage in leisure activities, but requested indeed differentially effective.
to go back to working on academic math tasks. We found that presession access to edibles
Carmen chose LEGO® bricks (an item that abolished the value of the edible reinforcers.
was available during the initial preference The tokens, on the other hand, were unaffected
assessment, but had not been highly preferred by the presession access and continued to
at that time) seven times, but also picked one maintain high break points. It is likely that the
of the original items (iPad®) three times. Car- combination of multiple categories of backup
men declined leisure items and chose to rejoin reinforcers, both leisure and edible items,
scheduled activities in the classroom on one resulted in the tokens’ resilience to singular
occasion. abolishing operations. Additionally, the shift in
the exchange for backup reinforcers from edi-
bles to leisure items during the token condition
DISCUSSION supports this interpretation.
In the first phase of the current study, we The current results differ somewhat from
evaluated leisure items, tokens, and edibles as those found by Vargo and Ringdahl (2015;
reinforcers with three children with autism using Experiment 4). In their study with four typi-
progressive ratio schedules. Results indicated cally developing preschoolers, these authors
that all influenced high rates of behavior, with found that presession access to edible items had
slightly higher break points occurring during the a more deleterious effect on responding main-
token reinforcement condition for two partici- tained by token reinforcement than responding
pants. In the second phase, we evaluated the maintained by edible reinforcement. However,
extent to which tokens functioned as generalized the tokens in the Vargo and Ringdahl study
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 11

were exchangeable only for the same edible reinforcers by including more categories of
items that were delivered in the presession reinforcers and measuring the effects on tokens
period. Therefore, it seems unlikely that the when abolishing operations are manipulated for
tokens functioned as generalized conditioned multiple categories. In addition, we have seen
reinforcers. Instead, the delivery of the tokens only deleterious effects of satiation, but the
may have signaled the availability of delayed inverse may be possible as well. In other words,
edibles. Further, the participants received only presession consumption of one category of rein-
a set amount of edibles presession, which may forcers may increase the break points of tokens
not have been sufficient to produce satiation. exchangeable for different, independent backup
Thus, the presession period in the Vargo and reinforcers. However, this remains an empirical
Ringdahl study may have served to reduce the question.
value of delayed edible reinforcement more Although the results suggest that tokens func-
than the value of immediate reinforcement, tioned as generalized conditioned reinforcers,
resulting in greater reduction in response rates there is a slight possibility that the tokens them-
in the token condition. The results of the cur- selves were reinforcers. Due to practical con-
rent study, however, suggest that token rein- straints, there was no baseline token condition
forcement is likely to be resistant to disruption included. In such a condition, the tokens would
by presession access to one type of reinforce- be earned, but no backup reinforcers would be
ment, if the tokens are exchangeable for a vari- available for exchange. If high break points were
ety of backup reinforcers. maintained in such contexts, we could say that
Because access to a greater variety of leisure
the value of the tokens might be somewhat inde-
items was instituted at the same time as preses-
pendent of their relation to the backup rein-
sion access to edibles, we cannot be certain
forcers. During the 5-min presession access
whether the generalized function of the tokens
period, Damien continued to achieve high break
was established by the enrichment of the leisure
points without exchanging the tokens for edibles
items, substitution of multiple dissimilar rein-
or leisure items. This might indicate that the
forcers, or both. Nevertheless, the fact remains
that breakpoints occurring in the edible condi- tokens functioned as reinforcers, independent of
tion were disrupted as a result of presession the backup reinforcers. However, this is difficult
access to edibles, whereas breakpoints occurring to determine because both participants had an
in the token condition were unaffected, and this extensive history with token economies and
effect was demonstrated in a reversal design. backup reinforcers. Along similar lines, a no-
The current findings add to those of Moher reinforcement baseline condition was not con-
et al. (2008), in that the properties of token ducted, leaving the possibility of the behavior
reinforcers may extend across individual stimuli being maintained in the absence of programmed
creating categories that are affected by a more reinforcement. Although inclusion of a no-
encompassing MO. Moher et al. included only reinforcement baseline would reduce speculation
edible items and manipulated the value of each about the tasks being automatically reinforcing,
item, whereas the current study included a we can affirm the consequent. During presession
three-item array of the most highly preferred access periods, responding was effectively elimi-
edibles and leisure items. However, the current nated even though the contingency was still in
study is limited in that the value of only one place. In other words, if completing the tasks
category of backup reinforcers was manipu- was maintained by automatic reinforcement,
lated. Future research could focus on determin- task engagement should have continued irre-
ing the boundaries of generalized conditioned spective of the MO manipulations.
12 DANIELLE RUSSELL et al.

Additional inferences can be drawn from consuming and may lead to lean ratios that are
previous studies that examined the use of PR aversive, they should not be used unless there is
schedules in reinforcer assessments. Roane a clear benefit of doing so and participants are
et al. (2001) found that stimuli with higher free to opt out of the arrangement at any time
break points were overall more effective in (cf. Poling, 2010). Future research could help
decreasing problem behavior and increasing guide the appropriate application of different
appropriate behavior. These findings, combined reinforcer assessments.
with the outcomes of the current study, suggest
that tokens might be the most effective rein-
forcers to increase appropriate behavior under REFERENCES
certain conditions. In addition, Penrod, Wal- Ayllon, T., & Azrin, N. H. (1968). The token economy: A
lace, and Dyer (2008) found that more pre- motivational system for therapy and rehabilitation.
New York, NY: Appleton-Century-Crofts.
ferred stimuli resulted in higher break points.
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007).
All the items used in the current study were Applied behavior analysis (2nd ed.). Upper Saddle
ranked as the top three preferred items accord- River, NJ: Pearson.
ing to the initial preference assessments. Thus, Cowles, J. T. (1937). Food-tokens as incentives for learn-
ing by chimpanzees. Comparative Psychological
break points on PR schedules may prove to be Monographs, 12, 1-96. https://doi.org/10.1037/
a more sensitive measure of reinforcer efficacy 14268-000
than assessments including a singular dense DeFulio, A., Yankelevitz, R., Bullock, C., &
schedule of reinforcement. This is not to say Hackenberg, T. D. (2014). Generalized conditioned
reinforcement with pigeons in a token economy.
that PR schedules should replace all other rein- Journal of the Experimental Analysis of Behavior, 102,
forcer assessments. Which assessment to con- 26-46. https://doi.org/10.1002/jeab.94
duct depends on time restrictions, the DeLeon, I. G., Iwata, B. A., Goh, H., & Worsdell, A. S.
participant’s response when exposed to thin (1997). Emergence of reinforcer preference as a func-
tion of schedule requirements and stimulus similarity.
reinforcement schedules, and relevance to the Journal of Applied Behavior Analysis, 30, 439-449.
task. For example, if brevity is required or if https://doi.org/10.1901/jaba.1997.30-439
the participant is unlikely to contact thin rein- Fisher, W. W., & Mazur, J. E. (1997). Basic and applied
forcement schedules in their everyday environ- research on choice responding. Journal of Applied
Behavior Analysis, 30, 387-410. https://doi.org/10.
ment, a simple comparison using dense 1901/jaba.1997.30-387
reinforcement schedules may be sufficient. Fisher, W., Piazza, C. C., Bowman, L. G.,
However, if the participant is likely to contact Hagopian, L. P., Owens, J. C., & Slevin, I. (1992).
A comparison of two approaches for identifying
a much broader array of reinforcement sched- reinforcers for persons with severe and profound dis-
ules, an assessment using a PR schedule may be abilities. Journal of Applied Behavior Analysis, 25,
more appropriate. It is also important to note 491-498. https://doi.org/10.1901/jaba.1992.25-491
that the rapidly thinning reinforcement sched- Hackenberg, T. D. (2009). Token reinforcement: A
review and analysis. Journal of the Experimental Analy-
ules characteristic of PR arrangements can sis of Behavior, 91, 257-286. https://doi.org/10.1901/
result in potentially aversive conditions. In the jeab.2009.91-257
current study, we ensured that the participants Hodos, W. (1961). Progressive ratio as a measure of
could emit a low-effort escape response at any reward strength. Science, 134, 943-944. https://doi.
org/10.1126/science.134.3483.943
time. The fact that Damien and Carmen some- Kazdin, A. E., & Bootzin, R. R. (1972). The token econ-
times opted to end the sessions very quickly omy: An evaluative review. Journal of Applied Behav-
(sometimes without emitting any target ior Analysis, 5, 343-372. https://doi.org/10.1901/
jaba.1972.5-343
responses) suggests that high schedule values
Kelleher, R. T. (1956). Intermittent conditioned rein-
became aversive or less preferred. Due to the forcement in chimpanzees. Science, 124, 679-680.
fact that PR arrangements are relatively time- https://doi.org/10.1126/science.124.3224.679
TOKENS AS GENERALIZED CONDITIONED REINFORCERS 13

Kelleher, R. T. (1958). Fixed-ratio schedules of condi- of Applied Behavior Analysis, 29, 1-9. https://doi.org/
tioned reinforcement with chimpanzees. Journal of 10.1901/jaba.1996.29-1
the Experimental Analysis of Behavior, 1, 281-289. Poling, A. (2010). Progressive-ratio schedules and applied
https://doi.org/10.1901/jeab.1958.1-281 behavior analysis. Journal of Applied Behavior Analysis,
Laraway, S., Snycerski, S., Michael, J., & Poling, A. 43, 347-349. https://doi.org/10.1901/jaba.2010.
(2003). Motivating operations and terms to describe 43-347
them: Some further refinements. Journal of Applied Roane, H. S., Lerman, D. C., & Vorndran, C. M.
Behavior Analysis, 36, 407-414. https://doi.org/10. (2001). Assessing reinforcers under progressive sched-
1901/jaba.2003.36-407 ule requirements. Journal of Applied Behavior Analysis,
McAdam, D. B., Klatt, K. P., Koffarnus, M., 34, 145-167. https://doi.org/10.1901/jaba.2001.
Dicesare, A., Solberg, K., Welch, C., & Murphy, S. 34-145
(2005). The effects of establishing operations on pref- Roane, H. S., Vollmer, T. R., Ringdahl, J. E., &
erences for tangible items. Journal of Applied Behavior Marcus, B. A. (1998). Evaluation of a brief stimulus
Analysis, 38, 107-110. https://doi.org/10.1901/jaba. preference assessment. Journal of Applied Behavior
2005.112-03 Analysis, 31, 605-620. https://doi.org/10.1901/jaba.
Moher, C. A., Gould, D. D., Hegg, E., & 1998.31-605
Mahoney, A. M. (2008). Non-generalized and gener- Tustin, R. D. (1994). Preference for reinforcers under
alized conditioned reinforcers: Establishment and val- varying schedule arrangements: A behavioral eco-
idation. Behavioral Interventions, 23, 13-38. https:// nomic analysis. Journal of Applied Behavior Analysis,
doi.org/10.1002/bin.253 27, 597-606. https://doi.org/10.1901/jaba.1994.
Murphy, E. S., McSweeney, F. K., Smith, R. G., & 27-597
McComas, J. J. (2003). Dynamic changes in rein- Vargo, K. K., and Ringdahl, J. E. (2015), An evaluation
forcer effectiveness: Theoretical, methodological, and of resistance to change with unconditioned and con-
practical implications for applied research. Journal of ditioned reinforcers. Journal of Applied Behavior
Applied Behavior Analysis, 36, 421-438. https://doi. Analysis, 48, 643-662. https://doi.org/10.1002/
org/10.1901/jaba.2003.36-421 jaba.226
Pace, G. M., Ivancic, M. T., Edwards, G. L.,
Windsor, J., Piche, L. M., & Locke, P. A. (1994). Prefer-
Iwata, B. A., & Page, T. J. (1985). Assessment of ence testing: A comparison of two presentation
stimulus preference and reinforcer value with pro-
methods. Research in Developmental Disabilities, 15,
foundly retarded individuals. Journal of Applied 439-455. https://doi.org/10.1016/0891-4222(94)
Behavior Analysis, 18, 249-255. https://doi.org/10. 90028-0
1901/jaba.1985.18-249
Wolfe, J. B. (1936). Effectiveness of token rewards for
Penrod, B., Wallace, M. D., & Dyer, E. J. (2008). Asses-
chimpanzees. Comparative Psychological Monographs,
sing reinforcer potency of high preference and low-
12, 1-72. https://doi.org/10.1037/h0093425
preference reinforcers with respect to response rate
and response patterns. Journal of Applied Behavior Woodcock, R. W., McGrew, K. S., & Mather, N.
Analysis, 41, 177-188. https://doi.org/10.1901/jaba. (2001). Woodcock-Johnson III. Itasca, IL: Riverside.
2008.41-177
Piazza, C. C., Fisher, W. W., Hagopian, L. P., Received April 9, 2016
Bowman, L. G., & Toole, L. (1996). Using a choice Final acceptance October 9, 2017
assessment to predict reinforcer effectiveness. Journal Action Editor, Anthony DeFulio

Você também pode gostar