Upload
independent
View
0
Download
0
Embed Size (px)
Citation preview
EFFECTS OF AND PREFERENCE FOR CONDITIONS OF TOKEN EARNVERSUS TOKEN LOSS
JEANNE M. DONALDSON
TEXAS TECH UNIVERSITY
ISER G. DELEON
KENNEDY KRIEGER INSTITUTE AND THE JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE
ALYSSA B. FISHER
KENNEDY KRIEGER INSTITUTE
AND
SUNGWOO KAHNG
KENNEDY KRIEGER INSTITUTE AND THE JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE
The effects of earning and losing tokens on the disruptive behavior of 12 first-grade students wereevaluated under symmetrical contingencies of earn and loss. Both contingencies produced decreasesin disruptive behavior. For some participants, more consistent decreases were observed duringthe loss contingency. In addition, participants generally earned or kept more tokens during theloss contingency. When offered a choice of contingencies, most participants preferred the losscontingency. The results showed some consistency with behavioral economic principles of lossaversion and the endowment effect.Key words: positive reinforcement, response cost, tokens, loss aversion
Token systems are commonly used in class-rooms, clinics, psychiatric hospitals, and prisons,and various token system packages have beendemonstrated to be effective at reducing problembehavior, increasing appropriate behavior, orboth (Kazdin, 1982). However, more appliedresearch is needed to examine specific features oftoken systems. One feature in need of additionalresearch is the effect of earn and loss contingen-
cies in token systems that are designed to reduceproblem behavior.
Several previous studies have compared tokenearn and loss contingencies (Broughton &Lahey, 1978; Conyers et al., 2004; Iwata &Bailey, 1974; Kaufman & O’Leary, 1972).Broughton and Lahey (1978) and Kaufmanand O’Leary (1972) compared the effects of earnand loss contingencies across groups on correctresponding to math problems and disruptivebehavior, respectively. Individuals within eachgroup experienced only one contingency. Therewere no apparent differences in the effects of earnand loss contingencies between groups in eitherstudy. Iwata and Bailey (1974) and Conyers et al.(2004) exposed all participants to both earn andloss contingencies but presented their data asgroup averages. Iwata and Bailey found bothcontingencies to be equally effective at reducingdisruptive behavior. Conyers et al. found both
We thank Lindsay Chen for her assistance in collectingdata and Ms. Oakley for allowing us to conduct research inher classroom. The contributions of Iser DeLeon andSungWoo Kahng were supported by Grants R01HD049753 and P01 HD055456 from the Eunice K.Shriver National Institute of Child Health and HumanDevelopment (NICHD). The contents are solely theresponsibility of the authors and do not necessarily representthe official views of NICHD.
Correspondence should be addressed to JeanneDonaldson,Texas Tech University, College of Education, 3008 18th St.,Lubbock, Texas 79409 (e-mail: [email protected]).
doi: 10.1002/jaba.135
JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2014, 47, 537–548 NUMBER 3 (FALL)
537
contingencies to be effective at reducing disrup-tive behavior, but the loss contingency produceda greater and more sustained change in behavioracross sessions. However, the loss contingency inthe Conyers et al. study also included feedbackwhen tokens were lost that was not providedwhen students did not earn a token in the earncondition, perhaps contributing to the differ-ences in effectiveness between the earn and losscontingencies.Comparisons between reinforcement and
punishment have also been made in the contextof the Good Behavior Game, which is a classroommanagement system, first evaluated by Barrish,Saunders, and Wolf (1969), that involves settingrules about behavior, delivering points contin-gent on breaking or following the rules, andproviding rewards contingent on meeting aspecified point criterion. Both Tanol, Johnson,McComas, and Cote (2010) and Wright andMcCurdy (2012) compared the effects of theGood Behavior Game when points were deliv-ered contingent on disruptive behavior (andrewards were contingent on earning few points)to a condition in which points were deliveredcontingent on the absence of disruptive behavior(and rewards were contingent on earning aminimum point amount). These studies reportedcomparable reductions in disruptive behaviorduring both the reinforcement and punishmentcontingencies. Wright and McCurdy assessedparticipants’ acceptability of both versions of thegame (contingencies for appropriate behavior andcontingencies for disruptive behavior) via a ratingscale in which the mean ratings suggested theparticipants, on average, found both proceduresapproximately equally acceptable.Previous applied comparisons of earn and loss
have not examined differences between earn andloss contingencies within individual participantsor assessed preference directly by allowingstudents to select and subsequently experiencetheir preferred contingency. The purpose of thecurrent study was to compare the effects of earnand loss conditions on disruptive classroom
behavior within individual students and assessindividual preference for earn and loss con-ditions. We also compared the duration ofimplementation of each condition as a measureof implementation effort.
METHOD
Participants and SettingParticipants were 12 students in a general
education first gradeclassroom.Fivegirls andsevenboys, all either 6 (n¼ 4) or 7 (n¼ 8) years old,participated. Eight participants were AfricanAmerican, three participants were Hispanic, andone participant was biracial. The classroomconsisted of 23 students: 61% female, 49%male; 48% African American, 35% Hispanic,9% Caucasian, 4% South Asian, and 4% biracial.Studentswere selectedtoparticipate in thestudybymeeting the following criteria during baseline: (a)The student engaged in disruptive behaviorduringbaseline observations, and (b) the trend of thestudent’s baseline data was not decreasing. Anystudents in the classwhodidnotmeet those criteriawere excluded from the evaluation. Although notall 23 students in the class participated in the study,all students received the study contingencies.All sessions occurred in the classroom during
either the seat-work center (in small-grouprotations) or independent reading (whole class).The student groups for small-group rotationswere determined by reading level, so the groupsdid not necessarily stay the same across sessions. Ifstudents were determined to read at a higher orlower level by teacher assessments, they weremoved to a different small group. Sessions wereconducted with all participants during both typesof activities during all phases of the study. Duringboth session times, students were expected to sitin their assigned seats and complete work quietlyor read silently. They were allowed to work onseat work with other students at their table aslong as they whispered. During centers, theteacher worked with a small group at a separatetable. During independent reading, the teacher
538 JEANNE M. DONALDSON et al.
conducted reading evaluations with individualstudents.
Response Measurement and InterobserverAgreementThe dependent variables were responses per
minute of disruptive behavior across all conditions,the number of tokens earned or kept in eachcondition in which token earning or keeping waspossible, the percentage selection of earn andloss conditions during the choice phase, and theduration of intervention implementation forearn and loss sessions.Disruptive behavior includedspeaking above a whisper without permissionfrom the teacher, standing up and moving awayfrom the student’s assigned seat, rocking back inthe chair such that at least one leg of the chair wasno longer touching the ground, loudly tappingobjects (e.g., pencils) on the table, banging onthe table, stomping feet, and manipulatingobjects that were not relevant to the assignedwork (e.g., playing with a toy from the student’sbackpack during seat work or drawing in thestudent’s journal during independent reading).Responses that could occur continuously (e.g.,rocking back in the chair, playing with a toy)were scored once when the response was initiatedand only scored a second time if the participantdiscontinued the response for at least 3 s andbegan again.In the tokens: choice phase, the selection of earn
or loss was recorded for each participant before thestart of the session. The number of tokens earned(or kept) for each participant was recorded at theend of the session from the checkmarkswritten oneach participant’s token board. The duration ofintervention implementation (i.e., monitoringbehavior according to the DRO and deliveringor removing tokens) was recorded from the timethe clicker sounded until the experimentersignaled to the data collector that she had finisheddelivering or removing tokens. Data for theduration of intervention implementation werecollected during a single session of each of thefollowing types: small-group earn, small-group
loss, whole-class earn, and whole-class loss. Theestimates of intervention implementation dura-tionwere based on implementing the interventionfor the entire class, not just the participants.A second independent observer recorded
disruptive behavior during 73% of baselinesessions and 31% of token sessions across allparticipants. Average interobserver agreement fordisruptive behavior was calculated using theproportional agreement method, in which eachsession was divided into 10-s intervals, the smallernumber of responses recorded by an observer wasdivided by the larger number of responsesrecorded by an observer within each interval (ifboth observers recorded no responses in aninterval, that interval was counted as 1), addingthe proportions from each interval, and dividingby the total number of intervals. During baseline,interobserver agreement averaged 93% (range,82% to 100%). During token sessions, agree-ment averaged 99% (range, 97% to 100%).
ProcedureThe effects of earn and loss conditions were
evaluated using a combination of a multielementand ABAC reversal design in which A phases werebaseline, the B phase was a multielementcomparison of earn and loss conditions, andthe C phase was a choice condition in which eachparticipant was given the choice to work in eitherthe earn or the loss condition. Sessions lasted10min and were conducted once or twice perday, 3 to 5 days per week. The first authorimplemented all study contingencies. The firstauthor did not serve as a data collector during anytoken sessions or any sessions during the secondbaseline but did serve as a data collector duringsome initial baseline sessions.Baseline. No programmed consequences were
provided to the participants for appropriate ordisruptive behavior. The classroom teacher wasasked to continue doing what she would normallydo during sessions. On several occasions, theteacher reminded the class to be quiet and remainin their appropriate centers.
TOKEN EARN VERSUS TOKEN LOSS 539
Tokens. At the start of each session, eachparticipant was given a laminated piece of whitepaper with 10 open circles that represented tokenslots. Tokens were check marks in the token slotsmade with dry erase markers. An experimenterdescribed the contingencies to the participantsbefore the start of the session and reviewedbehavior considered to be on task that could earnor keep tokens and behavior that was consideredto be disruptive and could not earn or could losetokens. Token earn and loss schedules werearranged as a momentary differential-reinforce-ment-of-other-behavior schedule (mDRO) with10 intervals per session signaled by a clickersound. This schedule was used to simulate acommon classroom schedule in which the teacheris reading a book, writing on the board, orworking with a small group of students and looksup to see who is on task to deliver or removepoints in the classroom point system. Theintervals were based on a random-time 1-minschedule, with the caveat that clicks could notoccur within 20 s of each other (to allow theexperimenter enough time to deliver all checks inthe earn condition).During all token sessions, the experimenter
sounded the clicker at the predeterminedintervals, scanned the room to determine whichstudents should earn or lose tokens, and deliveredthe consequences table by table. The table thatreceived consequences first rotated after eachclick. At the end of each session, participantsexchanged their tokens for back-up reinforcers. Avariety of food items were selected as back-upreinforcers. The prices for these reinforcers wereas follows: 1 Goldfish¼ 1 token, 1 pretzel¼ 2tokens, 1 M&M¼ 5 tokens, and 1 Skittles¼ 5tokens. The prices were set so that students couldeasily calculate howmany of each item they couldpurchase with their tokens (counting by 1, 2, and5 has typically been mastered by first grade).Candy items were valued higher because youngchildren generally prefer them to pretzels andGoldfish. Pretzels were priced higher thanGoldfish because they are larger.
Earn. Before the start of the session, the classwas told that they would be starting with zerotokens, their goal was to try to earn tokens, andthey would earn a token if they were on task eachtime they heard a click. Each participant’s tokenboard displayed 10 empty token slots at the startof the session. Participants earned one token forthe absence of disruptive behavior each time theclicker sounded. The experimenter delivered atoken by drawing a check mark in the designatedtoken slot for that interval. Participants who wereengaging in disruptive behavior when the clickersounded did not earn a token, and theexperimenter reminded the participant what heor she needed to be doing to earn tokens.Loss. Before the start of the session, the class
was told that they would be starting with 10tokens, their goal was to try to keep their tokens,and they would lose a token if they were not ontask each time they heard a click. Eachparticipant’s token board displayed 10 tokens atthe start of the session. The timing of clickersounds and token loss was yoked to the previousearn session. Participants who were engaging indisruptive behavior when the clicker sounded losta token, and the participant was reminded whathe or she needed to be doing to keep tokens. Theexperimenter removed a token by erasing thecheck mark from the designated token slots forthat interval. Participants who were not engagingin disruptive behavior when the clicker soundeddid not lose a token.Tokens: Choice. Before the start of each
session, an experimenter showed each participanta board with 0 tokens and a board with 10 tokensand asked each participant individually if he orshe wanted to either “start with zero tokens andtry to earn tokens” or “start with 10 and try tokeep tokens.” Participants chose by saying, “startwith 0” or “start with 10.” Although eachparticipant was asked to choose individually,other students at their table could hear theirchoices. Therefore, it was possible for selection tobe influenced by peer selection. The contingen-cies during the sessions corresponded with each
540 JEANNE M. DONALDSON et al.
participant’s selection. The day after the lastsession was conducted, all students in attendancewere asked two questions: “Which did you likebetter: starting with 0 or starting with 10?” and“why?” Eight participants (Talia, Zane, April,Erik, Evan, Shania, Damon, and Tabitha)responded to both questions.
RESULTS
Figures 1 through 3 display rates (responses perminute) of disruptive behavior for all participantsduring all phases. Data points are missing forsome participants for some sessions because theparticipant was absent, was working individuallywith the teacher, or had left the classroom duringthe session (e.g., some students were removedfrom class to receive special instruction in Englishas a second language or to work 1:1 with avolunteer). All participants displayed higher ratesof disruptive behavior during baseline phasesthan in tokens phases. There were often nodifferences between earn and loss conditionswithin the tokens phases, although greatervariability in responding during the earn phasewas observed for Talia, Erik, Tucker, andLamar.Table 1 displays the average number of tokens
earned or kept for each participant in eachcondition in which tokens could be earned orkept. In the majority of both earn and losssessions, students were able to earn or keep all 10tokens. During the tokens phase (in which thecondition was selected by the experimenter), sixof the 12 participants (Talia, Tanya, Shania,Tucker, Damon, and Shaun) earned or kept moretokens, on average, in the loss condition than inthe earn condition. Only one participant (Erik)earned or kept more tokens, on average, in theearn condition, and five participants (Zane,April, Evan, Lamar, and Tabitha) earned or keptall 10 tokens during every session. Takentogether, the average number of tokens earnedor kept was slightly greater in the loss condition.There was no apparent relation between the
number of tokens earned or kept and partic-ipants’ selection of earn or loss in the tokens:choice condition.Figure 4 depicts the percentage of selections for
earn and loss conditions in the tokens: choicephase for each participant. Participants arearranged in order from strongest preference forthe earn condition (Talia) to lowest preference forthe earn condition (Tabitha). One participant(Talia) displayed exclusive preference for the earncondition. Two participants (Zane and April)displayed a preference for the earn condition butdid not select it exclusively. Two participants(Erik and Evan) showed indifference (i.e.,selected both conditions approximately equally).Three participants (Tanya, Shania, and Tucker)showed a preference for the loss condition but didnot select it exclusively. Four participants(Damon, Shaun, Lamar, and Tabitha) displayedexclusive preference for the loss condition.When participants were asked which condi-
tion they liked better (“starting with 0 or startingwith 10”), their responses corresponded withtheir choices. The two participants whose choiceswere indifferent (Erik and Evan) reportedpreference for the earn condition. When askedwhy he preferred the earn condition, Evanresponded, “Because so I know if I’m going toget a negative or not a negative, and it’s just funlooking at it.” Talia, the only participant withexclusive preference for the earn condition,responded “I don’t know” when asked “why?”The other participants’ answers to “why?” couldbe categorized as a description of the condition(e.g., “because you can earn them” or “because Idon’t hardly lose none”) or a statement thatreflected loss aversion (e.g., “I don’t want to gettokens taken away” or “because I don’t like losingtokens.”). Interestingly, participants with no clearpreference for earn or loss conditions reportedloss aversion when asked why they liked onecondition better.Figure 5 shows the percentage of participants
who selected the loss condition in each session ofthe tokens: choice phase. On average, 67.3% of
TOKEN EARN VERSUS TOKEN LOSS 541
participants selected the loss condition (range,50% to 85.7%). In general, participants weremore likely to select the loss condition whengiven the choice.Table 2 displays the duration of intervention
implementation (for the entire class) for earn andloss sessions during both small-group and whole-class sessions. The earn conditions required
considerably more time for the experimenter toimplement. When the earn contingency was ineffect for the whole class at once, the experimenterspent nearly the entire session deliveringtokens. The loss contingency, however, requiredrelatively little time to implement, even when thecontingency was in effect for the whole class atonce.
0
0.5
1
Earn
Loss
Baseline
0
2
4
Talia
Baseline Tokens Tokens: Choice
Zane
0
0.5
1
1.5
2
0 5 10 15 20Sessions
0
0.5
1
1.5
April
Erik
Res
pons
espe
r Min
ute
of D
isru
ptiv
e B
ehav
ior
Figure 1. Responses per minute of disruptive behavior across sessions for Talia, Zane, April, and Erik. Earn sessions aredenoted by the triangles, and loss sessions are denoted by the squares.
542 JEANNE M. DONALDSON et al.
DISCUSSION
Both earn and loss conditions effectivelyreduced the disruptive behavior of all partic-ipants. Both contingencies resulted in equivalenteffects for eight participants. These results areconsistent with the notion that reinforcement
and punishment produce symmetrical effects onbehavior (Balsam & Bondy, 1983). However, thedisruptive behavior of seven of those participants(all but Tanya) was reduced to near-zero levels,which hinders the detection of differences.Previous research that has examined symmetricaleffects of reinforcement and punishment has
0
2
4
Earn
Loss
Baseline
0
1
2
3
Evan
Baseline Tokens Tokens: Choice
Tanya
0
1
2
3
0 5 10 15 20Sessions
0
0.5
1
1.5
Shania
Tucker
Res
pons
espe
r Min
ute
of D
isru
ptiv
e B
ehav
ior
Figure 2. Responses per minute of disruptive behavior across sessions for Evan, Tanya, Shania, and Tucker. Earn sessionsare denoted by triangles, and loss sessions are denoted by squares.
TOKEN EARN VERSUS TOKEN LOSS 543
found differing results with respect to symmetry.Both Ruddle, Bradshaw, Szabadi, and Foster(1982) and Magoon and Critchfield (2008)found symmetrical effects of reinforcement andan avoidance contingency on the rate of anarbitrary response (button pressing and mouseclicking, respectively) in adult participants.However, Rasmussen andNewland (2008) found
asymmetrical effects of reinforcement and pun-ishment in adult participants when concurrentvariable-interval schedules of reinforcement werearranged and a punishment schedule was super-imposed on one alternative. One importantdifference between the human operant studiesthat compared reinforcement and punishmentand the current study was that the punished
0
1
2
3Earn
Loss
Baseline
0
1
2
3
0
0.5
1
1.5
2
Damon
Tabitha
Baseline Tokens Tokens: Choice
Shaun
0
0.5
1
1.5
2
0 5 10 15 20Sessions
0
2
4
Lamar
Tabitha
Res
pons
espe
r Min
ute
of D
isru
ptiv
e B
ehav
ior
Figure 3. Responses per minute of disruptive behavior across sessions for Damon, Shaun, Lamar, and Tabitha. Earnsessions are denoted by triangles, and loss sessions are denoted by squares.
544 JEANNE M. DONALDSON et al.
response had an uncontrolled history of rein-forcement in a different form (i.e., disruptivebehavior was shaped and maintained by some-thing other than tokens).Four participants (Talia, Erik, Tucker, and
Lamar) engaged in high rates (in the range ofbaseline levels) of disruptive behavior during atleast one earn session, suggesting that the earncontingency did not exert as much control overtheir behavior as the loss contingency. The datafrom these participants could be interpreted asconsistent with the notion of loss aversion(Kahneman & Tversky, 1979), which suggeststhat losses are weighted greater than gains when the amount is equivalent. Anecdotally, partic-
ipants nearly always reported to the experimenterthat she “forgot to give me a token” when theyfailed to earn a token in the earn condition(despite the experimenter’s explanation of whythey did not earn the token during tokendelivery) but never reported suspected treatmentintegrity failures when a token was removed in
Table 1Average Number of Tokens Earned or Kept
Tokens Tokens: Choice
Earn Loss Earn Loss
Talia 9.3 10 10Zane 10 10 9.5 9.5April 10 10 9.8 —
a
Erik 10 9.7 8.3 10Evan 10 10 10 9.7Tanya 9 10 9 10Shania 9.7 10 9.5 10Tucker 8.7 9.7 9 9.5Damon 9.3 9.8 9.9Shaun 9.5 10 9.5Lamar 10 10 9.8Tabitha 10 10 10
aApril was pulled from the classroom for special instruction before the end of the sessions both times she selected loss in thetokens: choice phase.
0
25
50
75
100
snoitcel eS fo eg atn ecr ePfo
r Ear
n an
d Lo
ss
EarnLoss
Figure 4. Percentage of selections for earn and lossconditions for each participant during the tokens: choicephase.
0
25
50
75
100
14 16 18 20 22
ohw stnapic itraP fo egatnecreP
Sele
cted
Los
s
Sessions
Figure 5. The percentage of participants who selectedthe loss condition across tokens: choice sessions.
Table 2Duration (in Minutes) of Intervention Implementation
Small group Whole class
Earn 4.52 7.68Loss 1.18 0.42
TOKEN EARN VERSUS TOKEN LOSS 545
the loss condition. Although perhaps not a directeffect of the contingency per se, differences indisruptive behavior in the earn condition couldhave occurred because failure to earn is a lesssalient consequence than loss.One possible contributing factor to the
effectiveness of both contingencies might havebeen extinction of disruptive behavior main-tained by attention from peers, because deliveryof attention to a peer who is disruptive could alsoresult in token loss or failure to earn a token.Because the contingencies were symmetrical inthe earn and loss conditions, this is less of alimitation and perhaps more of an interesting sideeffect of arranging classwide (albeit individual)contingencies. Future research should comparethe effects of classwide contingencies to contin-gencies arranged specifically for one student (e.g.,behavior intervention plans) on the behavior ofthat targeted student and the class as a whole, aswell as assess the effort required to implementboth contingencies.Participants usually earned or kept more
tokens in the loss condition. In addition, bothparticipants with a preference for earn andparticipants with a preference for loss verballyreported finding token loss aversive when askedwhy they preferred one condition over the other.Taken together, these findings are consistent withthe notion that losses are valued more highly thangains of equal magnitude (i.e., loss aversion).Despite some consistencies with loss aversion,most participants preferred the loss condition.This finding was consistent with Hanley, Piazza,Fisher, and Maglieri (2005), who demonstratedan overall preference for the treatment thatincluded a punishment contingency. One poten-tial way to conceptualize selection of the losscontingency when losses are weighted moreheavily than gains is as a self-control response.Participants selected the loss contingency beforethe start of the session, therefore increasing thelikelihood that they would engage in appropriatebehavior during the session to avoid losingtokens. Another possibility is that the value of
tokens already in the participants’ possession,delivered at the start of the loss sessions, held agreater value than tokens that had not yet beendelivered in the earn session. The latter possibilityis consistent with behavioral economic researchon the endowment effect, which demonstratesthat items hold greater value when already inone’s possession (see Kahneman, Knetsch, &Thaler, 1991).The duration of implementation was mea-
sured for both conditions, and the loss contin-gencies required considerably less time toimplement than the earn contingencies. Becausenearly all of the students earned all of their tokensin the earn sessions, the experimenter spent mostof the earn sessions delivering tokens. Alterna-tively, because nearly all of the students kept all oftheir tokens in the loss sessions, the experimenterspent very little time removing tokens. Therefore,the loss contingency was much more practical.Considering that both contingencies were ap-proximately equally effective, participants had apreference for the loss contingency, and the losscontingency required less time to implement,teachers who use a token system as a behaviormanagement strategy might find arranging losscontingencies to be a better option than arrangingearn contingencies.Although the current data, taken together,
favor arranging loss contingencies in tokensystems, additional research that compares tokenearn and loss should be conducted to address thelimitations of the current study. The tokencontingencies in the current study were arrangedfor 10-min sessions; future research shouldevaluate the efficacy of and preference for earnand loss contingencies when implemented for theentire school day. The multielement design of theinitial tokens phase could have affected prefer-ence during the tokens: choice phase. Iwata andBailey (1974) examined choice after repeatedexposure to both earn and loss contingencies(order of exposure was counterbalanced acrossparticipants) and reported that some participantsselected earn during all choice sessions, some
546 JEANNE M. DONALDSON et al.
selected loss during all choice sessions, and someselected both conditions across choice sessions.However, the authors did not report whether thecondition immediately before the choice phaseaffected preference. That is, it is not clear if all ofthe participants who were exposed to the losscontingency immediately before the choice phaseshowed a preference for earn. Future researchshould examine whether repeated exposure toone contingency affects preference.Future research should also examine the side
effects of earn and loss contingencies on teacherbehavior. If a classroom token system werearranged in which only loss contingencies werein effect, the desired effect would be that teacherscould allocate the majority of their time used forbehavior management to provide attention forappropriate behavior and would no longer needto spend time reprimanding students or provid-ing attention for minor inappropriate behavior.However, arranging a token system in which onlyloss contingencies are in effect could produce theopposite effect: an exclusive focus on inappropri-ate behavior. Future research should determinewhat training, if any, would be required, inaddition to learning to use the token system, toproduce the former effect on teacher behavior.There were several additional limitations that
warrant consideration. Because all students werein the same class, phase changes had to occur atthe same time for all participants. Decisions onwhen to change phases were made based on themajority of participants’ data, resulting in weakerdemonstrations of experimental control for someparticipants (e.g., Zane and Evan). Also, manyparticipants did not contact the loss contingencyor the contingency of not earning tokens. Thiswas due to two procedural components: (a) Ruleswere stated to the participants before each session,and changes in behavior may have been a result ofrule governance; and (b) the mDRO was notsensitive enough to capture disruptive behaviorthat occurred throughout the intervals. Therewere also two components related to the choiceprocedure that may have affected participant
choices: (a) The presence of other students nearbyand the choices of other students could haveaffected participants’ choices, and (b) seeing 10tokens already on the token board could haveinfluenced some students to select the losscontingency. Also, data were not collected ontreatment integrity.In the current evaluation, an experimenter
implemented the contingencies during all tokensessions. Although the researchers were familiarto all of the participants (i.e., they had beeninvolved in their classroom 3 to 5 days per weekfor at least 5 months before the start of thisstudy), future research should have the classroomteacher implement all of the contingencies todetermine teacher preference for implementingearn versus loss contingencies and shouldevaluate whether teachers are more or less likelyto make treatment integrity errors with earn orloss contingencies. The current study mayprovide a bridge between the basic and appliedresearch that has compared earn and losscontingencies, but there is still more work todo to determine the most appropriate way toarrange token systems in classrooms.
REFERENCES
Balsam, P. D., & Bondy, A. S. (1983). The negative sideeffects of reward. Journal of Applied Behavior Analysis,16, 283–296. doi: 10.1901/jaba.1983.16-283
Barrish, H. H., Saunders, M., &Wolf,M.M. (1969). GoodBehavior Game: Effects of individual contingencies forgroup consequences on disruptive behavior in aclassroom. Journal of Applied Behavior Analysis, 2,119–124. doi: 10.1901/jaba.1969.2-119
Broughton, S. F., & Lahey, B. B. (1978). Direct andcollateral effects of positive reinforcement, responsecost, and mixed contingencies for academic perfor-mance. Journal of School Psychology, 16, 126–136. doi:10.1016/0022-4405(78)90051-1
Conyers, C., Miltenberger, R., Maki, A., Barenz, R.,Jurgens, M., Sailer, A., … Kopp, B. (2004). Acomparison of response cost and differential reinforce-ment of other behavior to reduce disruptive behavior ina preschool classroom. Journal of Applied BehaviorAnalysis, 37, 411–415. doi: 10.1901/jaba.2004.37-411
Hanley, G. P., Piazza, C. C., Fisher,W.W., &Maglieri, K. A.(2005). On the effectiveness of and preference forpunishment and extinction components of function-based
TOKEN EARN VERSUS TOKEN LOSS 547
interventions. Journal of Applied Behavior Analysis, 38, 51–65. doi: 10.1901/jaba.2005.6-04
Iwata, B. A., & Bailey, J. S. (1974). Reward versus cost tokensystems: An analysis of the effects on students andteacher. Journal of Applied Behavior Analysis, 7, 567–576. doi: 10.1901/jaba.1974.7-567
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1991).Anomalies: The endowment effect, loss aversion, andstatus quo bias. The Journal of Economic Perspectives, 5,193–206. doi: 10.1257/jep.5.1.193
Kahneman, D., & Tversky, A. (1979). Prospect theory: Ananalysis of decision under risk. Econometrica, 47, 263–292. doi: 10.2307/1914185
Kaufman, K. F., &O’Leary, K. D. (1972). Reward, cost, andself-evaluation procedures for disruptive adolescents ina psychiatric hospital school. Journal of Applied BehaviorAnalysis, 5, 293–309. doi: 10.1901/jaba.1972.5-293
Kazdin, A. E. (1982). The token economy: A decade later.Journal of Applied Behavior Analysis, 15, 431–445. doi:10.1901/jaba.1982.15-431
Magoon, M. A., & Critchfield, T. S. (2008). Concurrentschedules of positive and negative reinforcement:Differential-impact and differential-outcomes hypoth-eses. Journal of the Experimental Analysis of Behavior, 90,1–22. doi: 10.1901/jeab.2008.90-1
Rasmussen, E. B., &Newland,M. C. (2008). Asymmetry ofreinforcement and punishment in human choice.Journal of the Experimental Analysis of Behavior, 89,157–167. doi: 10.1901/jeab.2008.89-157
Ruddle, H. V., Bradshaw, C. M., Szabadi, E., & Foster,T. M. (1982). Performance of humans in concurrentavoidance/positive-reinforcement schedules. Journal ofthe Experimental Analysis of Behavior, 38, 51–61. doi:10.1901/jeab.1982.38-51
Tanol, G., Johnson, L., McComas, J., & Cote, E. (2010).Responding to rule violations or rule following: Acomparison of two versions of the Good Behavior Gamewith kindergarten students. Journal of School Psychology,48, 337–355. doi: 10.1016/j.jsp.2010.06.001
Wright, R. A., & McCurdy, B. L. (2012). Class-widepositive behavior support and group contingencies:Examining a positive variation of the Good BehaviorGame. Journal of Positive Behavior Interventions, 14,173–180. doi: 10.1177/1098300711421008
Received August 21, 2013Final acceptance March 25, 2014Action Editor, David Wilder
548 JEANNE M. DONALDSON et al.