DIFFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR … · DIFFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR INCREASES RESISTANCE TO EXTINCTION: CLINICAL DEMONSTRATION, ANIMAL MODELING,

DIFFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR INCREASES RESISTANCE TOEXTINCTION: CLINICAL DEMONSTRATION, ANIMAL MODELING, AND CLINICAL TEST OF

ONE SOLUTION

F. CHARLES MACE1, JENNIFER J. MCCOMAS

2, BENJAMIN C. MAURO3, PATRICK R. PROGAR

4, BRIDGET TAYLOR5,

RUTH ERVIN6, AND AMANDA N. ZANGRILLO

1

1UNIVERSITY OF SOUTHERN MAINE2UNIVERSITY OF MINNESOTA

3POSITIVE BEHAVIORAL DYNAMICS4CALDWELL COLLEGE

5ALPINE LEARNING GROUP6UNIVERSITY OF BRITISH COLUMBIA

Basic research with pigeons on behavioral momentum suggests that differential reinforcement ofalternative behavior (DRA) can increase the resistance of target behavior to change. This findingsuggests that clinical applications of DRA may inadvertently increase the persistence of target behavioreven as it decreases its frequency. We conducted three coordinated experiments to test whether DRAhas persistence-strengthening effects on clinically significant target behavior and then tested theeffectiveness of a possible solution to this problem in both a nonhuman and clinical study. Experiment1 compared resistance to extinction following baseline rates of reinforcement versus higher DRA ratesof reinforcement in a clinical study. Resistance to extinction was substantially greater following DRA.Experiment 2 tested a rat model of a possible solution to this problem. Training an alternative responsein a context without reinforcement of the target response circumvented the persistence-strengtheningeffects of DRA. Experiment 3 translated the rat model into a novel clinical application of DRA. Trainingan alternative response with DRA in a separate context resulted in lower resistance to extinction thanemploying DRA in the context correlated with reinforcement of target behavior. The value ofcoordinated bidirectional translational research is discussed.

Key words: translational research, behavioral momentum, persistence-strengthening effects of DRA,resistance to change, alternative reinforcement

_______________________________________________________________________________

Applied Behavior Analysis has proven re-markably effective at reducing a wide range ofunwanted target behaviors and replacing thesebehaviors with prosocial alternatives (Iwata etal., 1997). Differential reinforcement of alter-native behavior (DRA) has been one of theprincipal tools used to effect these improve-ments (Petscher, Rey & Bailey, 2009). Despitethe development and refinement of behavior-change technologies over the decades, Ap-plied Behavior Analysis has made limitedprogress in achieving long-term maintenanceand generalization of treatment gains. Osnes

and Lieblein (2003) reported that mostclinical studies fail to demonstrate the dura-bility and generality of treatment effects. Oneimpediment to resolving the problem oftreatment relapse, or maintenance and gener-alization failure, is the lack of understandingwhy target behaviors that have been eliminatedin one context later resume in that or othercontexts. However, developments in behavior-al momentum theory during the past twodecades may provide a model for understand-ing treatment relapse and the conditionsunder which it is more and less likely to occur.

Nevin’s formulation of behavioral momen-tum theory has shown there are two indepen-dent features of the discriminated operant: (a)its ongoing response rate, and (b) the resis-tance of this response rate to change whendisrupted by operations such as extinction,satiation, alternative reinforcement, and dis-traction (Nevin, 1992; Nevin & Grace, 2000;Nevin, Tota, Torquato & Shull, 1990). Ongo-ing response rate is known to be a function ofthe contingency between responding and itsconsequences (i.e., response–reinforcer rela-

This research was presented at an invited symposium atthe 30th annual Convention of the Association of BehaviorAnalysis International in Phoenix, Arizona (May, 2009). Anontechnical summary of this research appears in Mace,F.C., McComas, J.J., Mauro, B.C., Progar, P.R., Taylor, B.A.,Ervin, R. & Zangrillo, A.N. (2009). The persistence-strengthening effects of DRA: An illustration of bi-directional translational research. The Behavior Analyst,32, 293–300.

Address correspondence to F. Charles Mace, 407 BaileyHall, University of Southern Maine, Gorham, ME 04038 (e-mail: [email protected]).

doi: 10.1901/jeab.2010.93-349

JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR 2010, 93, 349–367 NUMBER 3 (MAY)

349

tions). By contrast, the resistance of baselineresponse rate to change is a function of thetotal rate of reinforcement present in thecontext in which reinforced behavior occurs(i.e., context–reinforcer relations). Theseresponse–reinforcer and context–reinforcerfunctional relations have been shown to berobust and general across species and popula-tions including pigeons (Nevin et al., 1990),rats (Mauro & Mace, 1996), goldfish (Igaki &Sakagami, 2004), adults with developmentaldisabilities (Mace et al., 1990) and collegestudents (Cohen, 1996).

Demonstrations that resistance to change isa function of context–reinforcer relations haveinvolved the use of alternative reinforcementin the form of time-contingent reinforcerdeliveries and differential reinforcement ofalternative behavior (DRA). Nevin et al. (1990,Experiment 2) illustrated the persistence-strengthening effects of DRA. Figure 1 dia-grams their baseline multiple concurrentschedule of reinforcement. Pigeons werepresented with two color-illuminated responsekeys mounted on an operant panel. Pecks onthe right key were designated the targetbehavior and pecks on the left key weredefined as alternative behavior. Condition Arepresents a typical DRA arrangement inwhich comparatively high rate reinforcementis provided for an alternative response (45reinforcers/hr in this case) while the targetbehavior produces a lower rate of reinforce-ment (15 reinforcers/hr). With the alternativeand target response keys illuminated green,the total reinforcement available in the greencontext, or the context–reinforcer relation,was 60 reinforcers/hr. Conditions B and C ofthe baseline schedule represented differentrates of reinforcement for the target responsewithout alternative reinforcement. ConditionsA and C contained the same total number ofcontext–reinforcer contingencies (60 reinforc-ers/hr) and Conditions A and B contained thesame number of response–reinforcer contin-gencies for the target right key response (15reinforcers/hr). This arrangement permitteda direct test of whether resistance to changewas a function of response–reinforcer orcontext–reinforcer contingencies.

Nevin et al. (1990) measured resistance tochange during the response disruptors ofextinction and short and long periods ofsatiation. Although DRA in Condition A

resulted in lower target response rates, resis-tance to change was comparable in ConditionsA and C and greater than Condition B in alltests, thus supporting the hypothesis thatresistance to change is a function of context–reinforcer contingencies. These findings dem-onstrated that although DRA can reduceoccurrences of a target response it can alsomake the response more persistent when theDRA contingency is degraded in some fashion.

This finding from basic research has con-siderable implications for clinical applicationsof DRA and possible reduction or avoidance oftreatment relapse. One factor contributing totreatment relapse may be that DRA strength-ens the persistence of unwanted target behav-ior such that when there are compromises toDRA treatment integrity there is resurgence intarget behavior. Basic research supports thisspeculation. For example, Podlesnik and Sha-

Fig. 1. Diagram of the 3-component multiple concur-rent schedule of reinforcement from Experiment 2 ofNevin et al. (1990). Condition A (green key condition)models DRA where reinforcement on the alternative leftkey is three times greater than reinforcement of the targetright key. Adapted from Nevin et al. (1990) and reprintedby permission.

350 F. CHARLES MACE et al.

han (in press) arranged VI 120-s food rein-forcement for pigeons in a two-componentmultiple schedule that also included additionaltime-contingent reinforcer deliveries in onecomponent. Extinction eliminated respondingin both components, but when baseline rein-forcement was reinstated, responding recov-ered more quickly in the component with theadded reinforcement. Clinical studies havereported similar recurrence of target behaviorfollowing its reduction via time-contingentreinforcer deliveries (Ahearn, Clark, Garden-ier, Chung, & Dube, 2003; DeLeon, Williams,Gregory & Hagopian, 2005).

The purpose of the present investigation wasthreefold. First, we sought to establish whetherDRA increases the resistance to extinction oftarget behavior in individuals with developmen-tal disabilities (Experiment 1). Second, givendemonstration of the persistence-strengthen-ing effects of DRA, we developed an animalmodel of one possible solution to this clinicalproblem (Experiment 2). We chose an animalmodel to test the effectiveness of this solutionin order to avoid placing both unwanted targetbehavior and prosocial alternative behavior in aclinical population on extinction without evi-dence that the proposed solution to persistencestrengthening would be effective. Finally, inExperiment 3 the procedure used in the animalmodel was translated into a novel clinicalintervention and its effectiveness at avoidingor reducing the persistence-strengthening ef-fects of DRA was evaluated.

EXPERIMENT 1METHOD

Participants and Setting

Three children with developmental disabil-ities who were referred for treatment of severebehavior disorders participated in the study.Andy, a 7-year-old boy, was diagnosed withautism and severe mental retardation. He wasable to follow simple, one-step instructions byusing gestures and pointing to pictures. Andyattended a private day school for children withautism and pervasive developmental disorders.Tom and Jackie were inpatients in a children’shospital unit. Tom, a 7-year-old boy, wasdiagnosed with Down’s Syndrome and Atten-tion Deficit Hyperactivity Disorder (ADHD).He was admitted to the unit for the treatmentof food stealing, elopement and tantrums.

Jackie, a 4-year-old girl, was diagnosed withsevere mental retardation, microcephaly, andmoderate hearing impairment who was hospi-talized for the treatment of aggression andself-injurious behavior. She was responsive to afew instructions delivered by manual signs.

Sessions for Andy were conducted in hisclassroom in a partitioned instructional cubi-cal measuring 6 m by 3 m containing a tableand two chairs. Other students and teacherspresent in the classroom participated in theirregularly scheduled activities, but were notwithin view of the cubical during experimentalsessions. Sessions for Tom and Jackie wereconducted in a hospital room (3 m by 5 m)with two beds and a table and chairs or in atreatment room (3 m square) with a one-wayobservation window.

Target Behaviors, Measurement, andInterobserver Agreement

Andy’s target behavior was hair pulling,defined as an extension of the hand towardthe hair of another person, including contactwith the hair or attempts at contact. Tom’starget behavior was food stealing, defined as anextension of the hand toward another per-son’s plate of food during mealtimes. Foodcontact and attempts at food contact wererecorded. The target behavior for Jackie wasaggression, defined as hair pulling, scratching,hitting, and head butting. Alternative respons-es to be reinforced during the DRA phase ofthe study were appropriate toy play (touching orinteracting with a toy for a minimum of 5 s) forAndy and Jackie, and appropriate requests for food(asking an adult for a snack item) for Tom.

Two independent observers collected dataon the target and alternative behaviors for all 3participants. Data on these behaviors wererecorded with a count within a 10-s intervalrecording procedure using either computersoftware or paper and pencil. The secondobserver collected data on an average of 40%of the sessions across conditions and partici-pants (range 5 31% to 45%). Occurrenceagreement for the target behaviors, calculatedon a point-by-point basis, averaged 93% for all3 participants (range 5 89% to 100%).

Procedures

Pre-study functional analysis. Functional be-havioral assessment interviews (adapted from

DRA INCREASES RESISTANCE TO EXTINCTION 351

Iwata, Wong, Riordan, Dorsey & Lau, 1982)and a functional analysis (Iwata, Dorsey, Slifer,Bauman and Richman (1982/1994) indicatedthat Andy’s hair pulling was maintained byadult attention and Jackie’s aggression wasmaintained by access to food. An interview andhome visit suggested that Tom’s food stealingwas maintained by access to food.

Baseline (BL). For Andy, two teachers in thepreschool served as experimenters. Andy sat athis desk in the cubical with several toysavailable. One teacher sat on the floor 1 mfrom Andy and read a magazine. Anotherteacher was positioned behind Andy on theother side of the partition and was able toobserve his behavior. Contingent on hairpulling or attempts to pull the hair of theteacher present in the cubicle, the teacherpositioned behind the partition entered thecubicle and provided a verbal reprimand on avariable ratio 3 (VR 3) schedule. Immediatelyfollowing the reprimand the teacher returnedto her position behind the partition. Nodemands or other forms of attention wereprovided during this condition.

For Tom and Jackie, a therapist served asexperimenter. Tom and the therapist sat at atable in a therapy room. Tom was given ahospital meal to eat. Approximately once every30 s during Tom’s meal, the therapist atesnack foods that Tom preferred from a platepositioned approximately 0.67 m from Tom.Food stealing resulted in access to bite-sizepieces of food on a fixed ratio 1 (FR 1)schedule. Attempts at food stealing were notblocked and the therapist did not provideattention contingent on food stealing.

Jackie and the therapist were in a hospitalroom and the therapist held a box of preferredsnack food in her hand. Aggressive responsesresulted in the therapist providing Jackie withone piece of the snack food on a variableinterval 60-s schedule (VI 60 s); a limited holdwas not employed. Interval values for all VIschedules in the experiment were calculatedusing the Fleshler-Hoffman (1962) VI genera-tor formulas and were signaled to the instruc-tor via audiotape with earphones.

Differential reinforcement of alternative behav-ior (DRA). An alternative to target behavior wasreinforced under DRA schedules that weresubstantially richer than those supporting thetarget behavior. These programmed valuesexpressed as the percentage of the baseline

rate of reinforcement for target behaviors were195% for Andy, 165% for Tom, and 185% forJackie.

Andy sat at his desk with several toysavailable as in baseline. Immediately beforeeach session, the teacher who had beenpositioned behind the partition used a grad-uated prompt hierarchy (gesture, verbal, phys-ical prompts) to assist Andy to interactappropriately with the toys for approximately15 s. Following this assistance, the teacherwent to her position behind the partition.During the session, the teacher present in thecubicle provided descriptive praise statementscontingent on appropriate play on a VI 30-sschedule. The baseline procedures and con-tingencies for reinforcing hair pulling re-mained in effect to more closely replicateExperiment 2 of Nevin et al. (1990).

For Tom, appropriate requests for food werereinforced on an FR 1 schedule. Contingent onan appropriate request for food, the therapistpraised the request and handed one piece ofpreferred snack food to Tom. Food theftscontinued to be reinforced as in baseline.

For Jackie, social reinforcement followedalternative behavior even though her targetbehavior, aggression, was maintained by accessto food. We used this arrangement to repre-sent the use of DRA in cases where thereinforcer maintaining the target behavior isunknown or maintained by automatic rein-forcement. Five seconds of toy play wasreinforced on a VR 2 schedule and aggressioncontinued to produce access to food on a VI60-s schedule as in baseline.

Extinction (EXT). For all 3 participants,response blocking was used to effect anextinction operation. Circumstances replicat-ed those of baseline, with two exceptions. First,attempts to pull hair, steal food, or engage inaggression were blocked by the teacher ortherapist in the room. Blocking consisted ofthe adult placing his/her hand or arm in frontof the participants’ hand each time theparticipant attempted to engage in the targetbehavior. Second, the reinforcer maintainingeach participant’s target behavior was with-held. No other interactions took place be-tween the experimenters and participants.

Experimental Design

The experimental conditions describedabove were presented to participants in a


sequence of phases that arranged extinction toalternately follow baseline and DRA reinforce-ment in a counterbalanced order acrossparticipants. The sequence order for Andywas BL-DRA-EXT-BL-EXT. For Tom andJackie, the order was BL-EXT-BL-DRA-EXT.The duration of each phase was determined byevidence of visual stability in the data duringbaseline and DRA and an apparent extinctionprocess during extinction.

RESULTS AND DISCUSSION

Figure 2 presents target response rates forthe 3 participants across all phases of thestudy. In the initial baseline, Andy’s hair pullsoccurred at high rates (M 5 186.7/hr; range 5172 to 209; top panel). The addition of theDRA intervention resulted in a sharp decreasein hair pulling after 2 sessions that generallycontinued over 19 sessions (M 5 65.0/hr;range 5 29 to 171). Implementation ofextinction via response blocking in the thirdphase resulted in a large increase in hairpulling during the first extinction session(266/hr) that decreased steadily to near-zerolevels after 36 sessions of extinction (M 551.7/hr; range 5 7.8 to 266). The return tobaseline conditions in the fourth phase result-ed in an immediate resumption of high andsteady rates of hair pulling, although to a lowerlevel than the initial baseline (M 5 103.8/hr;range 5 76 to 175). The second implementa-tion of extinction in the subsequent phaseresulted in an immediate reduction in hairpulling without an extinction burst. Hairpulling decreased to near-zero levels afterthree sessions and remained at those levelsfor the remaining eight sessions of extinction(M 5 14.8/hr; range 5 7.8 to 46.9). Tom’sfirst baseline phase resulted in variable butstable levels of food thefts (M 5 74.2/hr;range 5 20 to 119). The first session of theinitial extinction phase produced a burst offood thefts to 371 per hour followed by animmediate decrease in food thefts at the thirdsession to near-zero levels that remained lowduring the next 10 sessions (M 5 52.6/hr;range5 1 to 371). The return to baselinerecaptured high rates of food thefts (M 5106.3/hr; range 5 66.7 to 166.6) and thesubsequent DRA intervention produced sub-stantial reductions in food stealing over thecourse of 19 sessions, although food theftsresurged temporarily on the 17th session of

DRA (M 5 24.7/hr; range 5 1 to 150). Tom’ssecond extinction phase did not result in aninitial burst but food thefts continued inter-mittently in the baseline range over the first 22sessions and took 30 sessions to consistentlyoccur at near-zero rates. Finally, Jackie’s initialbaseline rates of aggression were highly vari-able but stable (M 5 285/hr; range 5 45.5 to636.4). Blocking aggression produced animmediate reduction in the target behaviorduring the first extinction phase that re-mained consistently at zero or near-zero levelsafter 8 sessions. The second baseline phase sawa resumption of high and variable rates ofaggression (M 5 427.02/hr; range, 5 72.7 to854.6). Although the DRA intervention thatfollowed the second baseline reduced rates ofaggression, treatment effects were not as largeas those observed for Andy and Tom (M 5257.0/hr; range 5 62 to 632). The extinctionphase that followed DRA required 34 sessionsbefore aggression was reduced consistently tozero or near-zero levels (M 5 135.3/hr; range5 0 to 895.9). Rates of the alternative behaviorduring the DRA intervention varied widelyacross participants. The occurrence of appro-priate toy play was 383/hr and 17/hr for Andyand Jackie, respectively. Mean appropriaterequests for food were 177/hr for Tom.

The comparison of primary interest inthis experiment is the relative resistance toextinction following baseline versus DRArates of reinforcement. Figure 3 superimposestarget behavior response rates during thetwo extinction phases expressed as the pro-portion of the preceding baseline responserate for all 3 participants. Reductions in thetarget behaviors during extinction took threetimes or more longer for all 3 participants.The number of extinction sessions followingDRA versus baseline was 36 versus 10 for Andy(360% more following DRA); 36 versus 12 forTom (300%), and 37 versus 12 for Jackie(308%).

The results of Experiment 1 demonstratedthe persistence-strengthening effects of DRAon target behavior. Although DRA reducedoccurrences of target behaviors for all 3participants, DRA was associated with greaterresistance to change during the adjacentextinction phase compared to extinctionfollowing baseline rates of reinforcement.These results are consistent with and predictedby Experiment 2 of Nevin et al. (1990).


EXPERIMENT 2

Experiment 1 illustrated the capacity of DRAto increase resistance to extinction of problembehavior. Experiment 2 evaluated a rat modelof a possible solution to this untoward sideeffect of DRA. According to behavioral mo-mentum theory, DRA increases persistence byadding reinforcement to the context in whichthe target behavior has a history of reinforce-ment. As Nevin et al. (1990, Experiment 2)demonstrated, reinforcing alternative behavior(the left key in condition A shown in Figure 1)

increased the persistence of the target behav-ior during extinction tests (the right key incondition A, Figure 1). In our rat model, analternative response was first trained in acontext separate from the one in which thetarget behavior occurred. In the extinctiontest, the discriminative stimuli correlated withthe alternative response were introduced tothe context correlated with the target behav-ior. This practice is contrary to the widespreaduse of DRA in the context in which the targetbehavior occurs (e.g., Carr & Durand, 1985;Fisher et al., 1993). We hypothesized that the

Fig. 2. Experiment 1: Rates of target behaviors during baseline, DRA and extinction conditions across participants.


separate context training procedure wouldresult in less resistance to extinction thanDRA implemented in the target behaviorcontext.

We chose to test our hypothesis with a ratmodel for two reasons. First, it permittedprecise control of the subjects’ historicalexperience with specific discriminative stimu-li–reinforcer delivery pairings, which remainslargely unknown in human participants. Sec-ond, by first testing this hypothesis with rats,

we did not have to place alternative prosocialbehavior on extinction with human clinicalparticipants without having a basis for expect-ing the hypothesis would be supported.

METHOD

Subjects

Four experimentally-naı̈ve male CharlesRiver CD rats (MV 54 to MV 57) were usedin this experiment. The animals were betweenthe ages of 2 and 4 months at the beginning of

Fig. 3. Experiment 1: Comparison of rates of target behaviors during extinction following baseline reinforcementand DRA reinforcement. Data are expressed as the proportion of the preceding baseline response rate.


the experiment and were housed individuallythroughout the experiment. The rats weremaintained at 80% of their free-feeding weightthrough supplemental feeding. Water wascontinuously available in each animal’s homecage.

Apparatus

Experimental sessions were conducted infour BRS/LVE two-lever operant chambers forrats (Model RTC-002). The left and right leversprotruded 2.2 cm into the chamber from theright wall (operant panel) and each leverrequired a force of 0.15 N to operate. A bankof three jeweled lights was located 4.3 cmabove each lever. Only the center light of eachbank of lights was used as the visual stimulusand provided either flickering or steadyillumination. An automatic food dispenserdelivered 45-g Noyes pellets to a scoop-typechute centered on the operant panel 1.2 cmabove the grid floor.

The experimental chambers were placed inMED Associates sound and light attenuatingenclosures (Model ENV-015M). Additionalsound attenuation was provided by a ventila-tion fan and a Grason-Stadler noise generator(Model 901B) that produced masking noisewithin 1 m of the enclosures. A computerusing MED Associates notation was used toprogram stimuli and record responses througha MED Associates interface.

Procedure

Preliminary training. Responding on the leftand right levers was separately establishedusing differential reinforcement of successiveapproximations. This was followed by contin-uous reinforcement (CRF) and fixed-ratio(FR) training in which the reinforcementschedule was gradually thinned to FR15 across3 to 9 days. The order in which trainingoccurred on the left and right levers wascounterbalanced across rats. Subsequently,a series of concurrent VI VI scheduleswas introduced and the schedules were grad-ually thinned from VI 10 s to VI 30 s across10 days.

Baseline. Baseline sessions consisted of threecycles of a three-component multiple concur-rent schedule of reinforcement. Each compo-nent was 6 min in duration and was presentedquasirandomly without replacement within a

session. Components were separated by a 30-sintercomponent period without any experi-mentally-programmed events occurring. Thus,each cycle lasted 19.5 min and the total sessionduration was 58.5 min. The stimulus arrange-ments for the three components duringbaseline are illustrated in the top panel ofFigure 4. Component 1 was accompanied by adarkened left light and a right flicker light at arate of 1 flick per second (1 f/s). Both left andright lights flickered at 5 f/s during Compo-nent 2. During Component 3, the right lightwas darkened and the left light was oncontinuously. Baseline sessions were conduct-ed daily. The experimental design involvedseveral baseline sessions followed by a singleextinction session. The number of baselinesessions for MV 54 to MV 57 was 41, 39, 20, and39, respectively.

The arrangement of concurrent reinforce-ment for each of the three baseline compo-nents is also illustrated in the top panel ofFigure 4. A single VI 30-s schedule operatedcontinuously throughout the session andarranged reinforcers for either the left orright lever press through a probability gate (p).In Component 1, extinction was arranged onthe left lever (p 5 0) and the programmedreinforcement rate per hr on the right leverwas 24 food pellets (p 5 .2). During Compo-nent 2, the programmed rate of reinforcementwas 96 (p 5 .8) food pellets per hr on the leftlever and 24 (p 5 .2) on the right lever. InComponent 3, 96 (p 5 .8) food pellets per hrwere programmed on the left lever andextinction was arranged on the right lever (p5 0). Thus, the combined reinforcementarranged in Components 1 and 3 equaledthe rate of reinforcement provided in Com-ponent 2. Reinforcers programmed but notearned when each component ended werecancelled. For each individual animal, theresponse rate per hr during baseline wasjudged as stable by visual inspection beforethe extinction test was conducted. Additional-ly, performance was deemed stable when itmet both of the following criteria: (a) responserates did not exceed or fall below 6 15responses/hr on the right and 6 10 re-sponse/hr on the left lever in all threecomponents for five consecutive sessions, and(b) stable ordinal relationships across thethree components on the left and right leversfor five consecutive sessions.


Extinction test. A single extinction test sessionwas conducted after baseline performance wasjudged to be stable. All food reinforcementwas discontinued during the extinction test byrendering the pellet dispenser inoperative. Amultiple concurrent schedule similar to thatused in baseline was employed during the

extinction test. To allow responding in allthree components to contact extinction dur-ing the beginning, middle and end of the test,the components were presented quasiran-domly within blocks of three componentsand data were recorded at the end of each19.5 min block. Components 1, 2, and 3 werepresented randomly once per block. Theextinction session was terminated at thecompletion of the block in which there hadbeen no response on the right lever for each ofthe three components. Visual stimuli present-ed during the extinction test are displayed inthe lower panel of Figure 4. Components 1and 2 were signaled by the same visual stimuliused in baseline. A stimulus compound wasintroduced in a third component that com-bined the stimuli correlated with Components1 and 3 during baseline. This stimuluscompound modeled the clinical situation inwhich DRA is used to train an alternativebehavior in a separate context (baselineComponent 3) and the discriminative stimulicorresponding to DRA training are introducedinto the context in which the target problembehavior has been reinforced (baseline Com-ponent 1). Thus, the signaled reinforcementhistory in Component 2 and Component 3(the Stimulus Compound of Baseline Compo-nents 1 and 3) were equal during extinction(120/hr), and were greater than the signaledreinforcement history correlated with Compo-nent 1 (24/hr).


The mean terminal absolute response andreinforcement rates for the left and rightlevers and the relative response and reinforce-ment rates for the right lever (target response)are displayed in Table 1. Obtained rates ofreinforcement for both response alternatives(left and right levers) across all three compo-nents approximated the programmed rates ofreinforcement for each alternative, suggestingexperimental control over the absolute andrelative rates of reinforcement. For all subjects,within-component response rates on the leftlever were comparable for Components 2 and3 that programmed 96 pellets/hr and werelowest in Component 1 (0 pellets/hr). Re-sponse rates on the right lever were highest inComponent 1 (24 pellets/hr with no alterna-tive reinforcement) and lowest in Component3 (0 pellets/hr) for all subjects. An examina-

Fig. 4. Experiment 2: Diagram of the discriminativestimuli and reinforcement rates in the three-componentmultiple concurrent schedule of the rat model. The toppanel depicts the baseline arrangement and the bottompanel represents arrangements during the extinction tests.


tion of the absolute and relative response ratesin Components 1 and 2 (both arranging 24pellets/hr) shows that for all subjects targetresponding was lower in the component thatarranged response-contingent alternative rein-forcement (i.e., DRA in Component 2).

The log proportion of baseline respondingon the right (target) lever across successiveblocks of extinction for all 4 rats is presentedin Figure 5. For all 4 rats, responding took theshortest amount of time to extinguish inComponent 1 where the visual stimuli (1 f/s)were correlated with the least amount ofprogrammed reinforcement (24/hr) and noalternative reinforcement. Conversely, theproportion of baseline target response ratesfor all subjects took longest to extinguish inComponent 2 in which the visual stimulicorrelated with the overall programmed ratesof reinforcement (120/hr) were the same forthe target and alternative responses (5 f/s).Target responding in Component 3 thatcompounded the stimuli signaling baselinereinforcement for Component 1 (1 f/s) andComponent 3 (constant light) was markedlylower than that in Component 2, even thoughthe total programmed reinforcement correlat-ed with the discriminative stimuli duringextinction was equal (120/hr) (for MV-54,MV-55, MV-56, and MV-57, respectively, meanproportion of baseline log response rate onthe target lever 5 .64, .54, .76, and .52 forComponent 2 and .37, .33, .33, and .31 forCompound C1 and C3). Of relevance to the

main experimental hypothesis, respondingduring extinction was comparable duringComponent 1 and the Compound Stimulusfor MV 54 and MV 56, and lower in theCompound Stimulus than Component 1 forMV 55.

The purpose of Experiment 2 was to modela possible solution to the problem of DRAincreasing the resistance to extinction ofunwanted target behavior. Baseline Compo-nent 1 represented reinforcement of thetarget response only, Component 2 simulatedDRA reinforcement of alternative behavior inthe same context in which the target responseis also reinforced, and Component 3 repre-sented establishing an alternative responsewith high rate reinforcement in a contextseparate from the one correlated with rein-forcement of the target behavior. When thediscriminative stimuli in baseline Component3 were placed in compound with the stimuli inComponent 1, increased resistance to extinc-tion was avoided in 3 of 4 subjects. For the 4thsubject, responding during extinction was stillless than during Component 2.

EXPERIMENT 3

Experiment 2 modeled a method of imple-menting DRA that attenuated the persistence-strengthening effects of DRA. The modelsuggests that a homologous clinical interven-tion may effectively avoid the persistence-strengthening effects of DRA and, hence,

Table 1

Absolute and relative response and reinforcement rates for the target (right) and alternativeresponses (left) during Baseline in Experiment 2. Each value is the mean of the final five sessions.

Rat

Response rate (min) Reinforcement rate (hr) Relative rates

Left Right Left Right Total Resp. Reinf.

Component 1MV 54 4.89 19.30 0.00 24.67 24.67 0.81 1.00MV 55 6.09 27.79 0.00 18.67 18.67 0.84 1.00MV 56 6.50 13.29 0.00 18.00 18.00 0.68 1.00MV 57 4.82 21.52 0.00 26.00 26.00 0.80 1.00




justify a clinical trial in which both the targetand alternative responses are placed onextinction in a clinical population. As apreliminary test of this proposition, in Exper-iment 3 we taught a communication responseusing functional communication training (aform of DRA) to children with severe disrup-tive behavior in a context in which disruptivebehavior had no history of reinforcement. Asin Experiment 2, we then compared targetresponding during extinction in the separatecontext DRA to the context in which DRA wasimplemented in the same context in which

target behaviors had a history of reinforce-ment.

METHOD

Participants and Setting

The participants were 2 males with develop-mental disabilities who were admitted to aninpatient hospital program for the treatmentof severe behavior disorders. Mickey was11 years old and diagnosed with moderatemental retardation. He spoke in simple sen-tences and engaged in severe disruptivebehavior during performance of daily living

Fig. 5. Experiment 2: Log proportion of baseline response rates during six blocks of a single extinction session forComponents 1, 2 and Compound of Baseline Component 1 and Component 3 stimuli across 4 rats.


skills. Paul was 21 years old, and diagnosedwith severe mental retardation. He was non-verbal and engaged in severe disruptivebehavior during vocational tasks. Sessions wereconducted in three different rooms on aninpatient hospital unit, each containing a tableand task materials.

Target Behaviors, Measurement, andInterobserver Agreement

Disruptive behavior was the target for bothparticipants. Mickey’s disruptive behavior wasdefined as throwing task items, pounding onthe table, shouting and slapping instructors.Paul’s disruptive behavior was defined asthrowing or destroying task materials, pound-ing on the table, kicking the underside of thetable with his knees, and slapping instructors.Requests were defined as speaking the phrase‘Break please’ (Mickey) or touching a 7.6 cmby 20.3 cm card with the printed word ‘Stop’on it (Paul).

Data on target behaviors, requests forbreaks, and reinforcer deliveries were collect-ed using a continuous count within a 10-sinterval recording procedure during 10-minsessions. Interobserver agreement was calcu-lated on a minimum of 37% of the sessions.Mean occurrence agreement was 87% (range5 61% to 100%) across participants.

Pre-Study Functional Analysis

Prior to the study, functional analysis pro-cedures described for Andy and Jackie ofExperiment 1 indicated that both participants’disruptive behavior was maintained primarilyby negative reinforcement in the form ofescape from demands.

Baseline Multiple Concurrent Schedule

Baseline conditions employed the multipleconcurrent schedule of reinforcement depict-ed in the top portion of Table 2. Componentswere presented to participants in randomorder without replacement with 5 to 10 minbetween components, during which the ex-perimenters did not interact with the partici-pants. One or two sets of three componentswere presented per day, 5 days per week.Therapists wore different colored hospitalgowns in the three schedule components.Gown colors varied by component as notedin Table 2. Each component was also conduct-

ed in a different room to promote componentdiscrimination. In order to control for thepresence of two instructors in Component 2,two instructors were present during all threecomponents for Mickey and both wore thesame colored hospital gowns.

Component 1. This component representedreinforcement of target behavior withoutreinforcement of alternative prosocial behav-ior. Different task demands were presented toparticipants. Mickey was asked to sort and foldshirts and pants and Paul was instructed towash tables and windows using paper towelsand a spray bottle. Two instructors stood oneither side of Mickey. The instructor on theright side presented task demands while theother instructor did not interact with theparticipant. One instructor was present forPaul. Sessions began with the instructorplacing the task materials on the table andsaying, ‘‘Do your work, please.’’ Following thisinitial instruction, instructions to performrelevant task steps were presented using agraduated three-step prompt hierarchy (verbalprompt, modeling, physical prompt) withprompts separated by 5 s. Contingent ondisruptive behavior, the instructor said, ‘‘Okay,take a break,’’ and moved at least 2 m awayfrom the participant. No programmed inter-action took place during the break. Escapefrom demands was arranged on a VI 75-sschedule of negative reinforcement for whichinterval values were calculated using theprocedure of Fleshler and Hoffman (1962).The completion of each interval was signaledto the instructor via earphones.

Component 2. This component representednegative reinforcement of alternative behaviorin the same context in which disruptivebehavior was negatively reinforced. Proceduresin Component 2 were identical to those ofComponent 1 except that an instructor pro-vided prompts to request a break frominstruction on a fixed interval 20-s (FI 20-s)schedule. Prompts were ‘‘Ask if you want abreak’’ for Mickey and ‘‘Point to the card ifyou want a break’’ for Paul. Contingent onrequests for breaks, the instructor provided abreak as described in Component 1. If theparticipant requested a break prior to theelapse of the FI 20-s interval, the instructorsaid, ‘‘Okay, just a little while longer,’’ andcontinued instruction until the reinforcementinterval elapsed. Disruptive behavior contin-


ued to be reinforced on the VI 75-s scheduledescribed in Component 1. For Mickey, theinstructor to his left provided instruction andreinforcement of requests for a break. Theinstructor to Mickey’s right reinforced disrup-tive behavior. For Paul, the single instructorreinforced both requests and disruptive behav-ior on their respective schedules.

Component 3. This component representedreinforcement of alternative behavior in acontext that provided no reinforcement ofdisruptive behavior. The task instruction de-scribed in Component 1 and reinforcement ofrequests for breaks described in Component 2were replicated in the third component.However, the instructor did not provide breaksfrom instruction contingent on disruptivebehavior.

Extinction Test

Instruction of each participant’s daily livingtask continued as in baseline during all threecomponents. However, all reinforcement oftarget disruptive behavior and requests wasdiscontinued in order to test the hypothesisthat reinforcement of prosocial alternativebehavior in a separate context from one thatalso reinforced disruptive behavior could

reduce or avoid the persistence-strengtheningeffects of DRA. The discriminative stimuli andarranged baseline rates of reinforcement arepresented in the lower portion of Table 2. Thediscriminative stimuli for Components 1 and 2were identical to those in baseline. Thediscriminative stimuli in a third componentduring extinction comprised a compound ofthe baseline stimuli for Components 1 and 3.This represented the clinical situation inwhich a prosocial alternative behavior, suchas communication, is first established in acontext with no history of reinforcement ofunwanted target behavior prior to introducingthe discriminative stimuli correlated withalternative behavior into the context that hasa history of reinforcement for target behavior(see lower portion of Table 2).


Figure 6 presents baseline data from thethree components of the multiple concurrentschedule. For both participants, the highestrates of disruptive behavior occurred inComponent 1, in which only disruptive behav-ior was reinforced (Mickey, M 5 50.4 disrup-tions /hr, range 5 30.0 to 93.5; Paul, M 5135.7, range 5 35.0 to 292.0). Rates of

Table 2

Discriminative stimuli (gown color and room) and programmed rates of reinforcement per hourduring the baseline multiple concurrent schedule of reinforcement and extinction tests inExperiment 3.

Component Discriminative Stimuli Rft/hr Requests Rft/hr Disruption Total BL Rft

Baseline

1 Yellow (Mickey) none 48 48Blue (Paul)Room 1

2 Blue (Mickey) 180 48 228White (Paul)Room 2

3 White (Mickey) 180 none 180Yellow (Paul)Room 3

Extinction Test

1 Yellow (Mickey) EXT EXT 48Blue (Paul)Room 1

2 Blue (Mickey) EXT EXT 228White (Paul)Room 2

Compound BL C1 + C3 Yellow/White (Mickey) EXT EXT 228Blue/Yellow (Paul)Room 1


disruptive behavior were substantially lower inComponent 2, in which requests were rein-forced 3.75 times more often than disruptivebehavior (Mickey, M 5 22.9 disruptions/hr,range 5 5.5 to 48.3; Paul, M 5 34.06, range 55.0 to 173.0). The lowest rates of disruptivebehavior occurred in Component 3, in whichrequests only were reinforced (Mickey, no

disruptive behavior; Paul, M 5 15.2 disrup-tions/hr, range 5 0 to 75.0).

Resistance to extinction expressed as propor-tion of baseline response rate was greatest inComponent 2 for both participants (see Fig-ure 7) where the discriminative stimuli wereidentical to those in baseline (see lower portionof Table 2). Both participants showed an

Fig. 6. Experiment 3: Disruptions per hour during baseline sessions of the multiple concurrent schedule ofreinforcement. Instructors wore different colored hospital gowns and conducted sessions in different rooms todifferentiate the schedule components.


extinction burst in Component 2. Mickey’sdisruptive behavior reduced to low rates afterfour extinction sessions; however, Paul contin-ued to engage in high rates of disruptivebehavior in Component 2 during nine sessionsof extinction. Component 2 sessions werediscontinued for Paul prior to effecting anextinction process for ethical reasons (hisdisruptive behavior had reduced to belowbaseline rates in Components 1 and the

Compound Stimulus Component). Resistanceto extinction was considerably lower in Com-ponent 1 where the discriminative stimuli werethe same as baseline and baseline rates ofreinforcement were approximately one-quarterof those in Component 2. Of primary interest tothe experimental hypothesis of the study arethe data on resistance to extinction during theCompound Stimulus of baseline Components 1and 3. This compound stimulus was correlated

Fig. 7. Experiment 3: Rates of disruptive behavior during extinction sessions expressed as proportion of baselineresponse rates for Components 1 and 2 and a Compound of Baseline Component 1 and Component 3 stimuli.


with the same reinforcement rates as Compo-nent 2. However, rates of disruptive behaviorduring extinction were much lower than inComponent 2 and were comparable to those inComponent 1 for both participants where theadded reinforcement of DRA was absent. Theseresults replicate the rat model of Experiment 2by showing that the persistence-strengtheningeffects of DRA can be reduced or avoided byusing DRA in a context separate from thatcorrelated with a history of reinforcement ofunwanted target behavior.

GENERAL DISCUSSION

We conducted two clinical studies and onebasic research experiment that were coordi-nated to demonstrate the persistence-strength-ening effects of DRA and develop a possiblesolution to the problem. Experiment 1 showedthat clinical applications of DRA can increasethe resistance of unwanted target behavior toextinction. The duration and response levelsof the extinction processes were substantiallygreater following DRA reinforcement thanafter baseline reinforcement of target behav-ior. In Experiment 2, we devised a DRAtreatment for target behavior in a rat modelthat was aimed at avoiding or reducing thepersistence-strengthening effects of DRA. Thesource of the problem is that typical applica-tions of DRA involve adding reinforcement tothe context in which target behavior is, or hasbeen, reinforced. When a new alternativeresponse such as communication is establishedwith DRA, it is generally reinforced at a highrate both to teach the response and tocompete with reinforcement for target behav-ior. For example, Fisher et al. (1993) usedfunctional communication training (FCT) toreduce unwanted target behavior with andwithout extinction of problem behavior. Com-munication responses were reinforced on a FR1 schedule and this resulted in rates ofcommunication ranging from 4.5 per min to7.0 per min for one study participant. To avoidadding reinforcement to the context in whichthe modeled target behavior occurs, we rein-forced an alternative response in a differentcontext in our rat model of DRA treatment(Component 3). During the extinction tests,we arranged a stimulus compound composedof the stimuli correlated with reinforcement ofthe target behavior (Component 1) and those

correlated with reinforcement of alternativebehavior in a context without reinforcementof the target behavior (Component 3). Al-though the total reinforcement in the com-pound stimulus equaled that of the compo-nent that included DRA and reinforcement oftarget behavior (Component 2), resistance toextinction was much lower in the componentwith the compound stimulus. We translatedthese findings into a novel approach to DRAtreatment in a clinical study in Experiment 3.Consistent with our rat model, using FCT toestablish communication in a context withoutreinforcement of target behavior circumvent-ed the persistence-strengthening effects ofDRA during extinction tests.

These findings have implications for resolv-ing some longstanding limitations to AppliedBehavior Analysis interventions. Long-termmaintenance and generalization of treatmentgains are often not reported or achieved(Osnes & Leiblien, 2003; Stokes & Osnes,1989). One possible reason for this is thatlapses in treatment integrity can result inrecurrence of problem behavior. Lapses intreatment integrity can occur when rates ofDRA reinforcement drop precipitously ordiminish over time resulting in a temporaryextinction operation or an increase in relativereinforcement for target behavior. Burst re-sponding with the initiation of extinction iswidely reported. For example, Lerman, Iwataand Wallace (1999) reported extinction burstsin 62% of clinical cases in which extinction ofself-injurious behavior followed reinforce-ment-based treatments. Partial compromisesto treatment integrity have also been related tothe resumption of problem behavior. DiGen-naro, Martens and Kleinmann (2007) identi-fied 13 procedural steps for teachers to followto implement a DRA intervention. They foundsignificant inverse correlations between thepercentage of accurately implemented treat-ment procedures and occurrences of targetbehavior for three-quarters of their partici-pants, ranging from -.45 to -.78. Vollmer,Roane, Ringdahl and Marcus (1999) systemat-ically varied DRA treatment integrity for 3children with developmental disabilities. For 2participants, varied levels of treatment integri-ty were 0/100 (baseline), 25/75, 50/50, 75/25,and 100/0 (full treatment). In their notation,75/25 denotes 75% of occurrences of proso-cial alternative behaviors were reinforced and


25% of target behaviors were reinforced.Resumption of target behavior was observedwhen treatment integrity dropped to or below50/50.

Basic research on resurgence predicts thatsuch lapses in treatment integrity would resultin recurrence of target behavior after success-ful treatment (Podlesnik, Jimenez-Gomez, &Shahan, 2006; Podlesnik & Shahan, in press).In the resurgence paradigm, after a targetresponse is established, it is then extinguishedwhile also reinforcing an alternative response(analogous to DRA plus extinction). If rein-forcement of the alternative response issubsequently diminished, recovery of thetarget response is observed even thoughreinforcement of the target response doesnot resume.

A second factor known to contribute to themaintenance and generalization failure is thereintroduction of stimuli that were previouslycorrelated with pretreatment reinforcement oftarget behavior (Osnes & Lieblein, 2003;Stokes & Baer, 1977; Stokes & Osnes, 1989).Transferring treatment effects from behavioranalysts to parents or educational personnel isimpeded by the phenomena of reinstatementand renewal. In the reinstatement paradigm, aresponse reinforced in baseline is then elim-inated by extinction or alternative reinforce-ment. Subsequent time-contingent delivery ofthe reinforcer that maintained baseline re-sponding then results in recovery of the targetresponse despite ongoing extinction (Shaham,Shalev, Lu, de Wit, & Stewart, 2003). DeLeonet al. (2005) demonstrated reinstatementeffects in a clinical population. After extinc-tion had reduced target behavior to low levels,introduction of reinforcers on a time-contin-gent schedule increased rates of target behav-ior despite ongoing extinction. The clinicalproblem this poses is that parents and educa-tional personnel will almost certainly deliverreinforcers noncontingently that previouslymaintained target behavior; thus, reinstate-ment effects should be expected.

Renewal occurs when a response is elimi-nated in a context different from the onecorrelated with baseline reinforcement asmight occur when an individual receivestreatment in a clinic for unwanted behaviorthat occurs at home. After successful reductionof target responding, recurrence of the targetresponse is observed when the baseline con-

text is reintroduced. Poldesnik & Shahan (inpress) demonstrated renewal effects withhoming pigeons. A baseline multiple scheduleof reinforcement (mult VI 120 s VI 120 s VT20 s) was presented with a steady houselightilluminated, followed by extinction during aflashing houselight. After responding waseliminated in the presence of the flashinglight, extinction was then applied with thebaseline steady houselight in an A-B-A arrange-ment. Recovery was observed in both compo-nents of the second A phase but was greater inthe schedule component with the added time-contingent reinforcers.

Results from Experiment 1 demonstrate thatDRA can strengthen the persistence of un-wanted target behavior. Basic research onreinstatement, resurgence and renewal furthersuggest the conditions under which thesepersistence-strengthening effects are likely tooccur. Resurgence effects may result fromlapses in treatment integrity. Reinstatementand renewal effects may interfere with effortsto transfer treatment gains first obtained in atreatment setting to a natural setting with ahistory of reinforcement of target behavior.The growing number of clinical demonstra-tions that alternative reinforcement canstrengthen persistence (Ahearn et al., 2003;DeLeon et al., 2005) points to a significantclinical problem in need of solution.

In addition to the separate context solutiondemonstrated in Experiments 2 and 3, we cansuggest two additional remedies that differfrom conventional Applied Behavior Analysispractice. First, when a pretreatment functionalanalysis identifies a socially mediated reinforc-er that maintains problem behavior, it canoften be withheld in an extinction operation.Matching theory predicts that reductions inproblem behavior may be achieved with lowrates of DRA reinforcement when juxtaposedwith extinction for target behavior (i.e., thelow-rate DRA solution). This remedy contrastswith conventional practice that typically ar-ranges high-rate reinforcement for the proso-cial alternative behavior (e.g., Fisher et al.,1993; Vollmer et al., 1999).

A second possible solution is a variant of theseparate context solution. Establishing a newalternative response often requires high-ratereinforcement to compete with the reinforce-ment history for unwanted target behavior.However, if high-rate DRA is introduced in a


separate context and then faded to low-rateDRA prior to introducing the DRA discrimi-native stimuli to the target context, this mayavoid adding to the context–reinforcer con-tingencies that would contribute to persistencestrengthening. This separate context schedulethinning solution is suggested because persis-tence-strengthening effects in humans appearto be sensitive to recent rates of alternativereinforcement. In Experiment 1, resistance toextinction was greater when extinction imme-diately followed DRA reinforcement thanwhen it immediately followed baseline rein-forcement. Ahearn et al. (2003) reportedsimilar findings in a clinical population usingresponse-independent reinforcer deliveries.This remedy departs from the usual practiceof DRA schedule thinning in the context inwhich target behavior has a history of rein-forcement.

The three experiments reported here rep-resent an explicitly bidirectional approach totranslational research. While the past threedecades have seen growing recognition of theimportance of linking basic and appliedresearch to promote advancements in behav-ioral technologies and establish the broadgenerality of behavioral laws (e.g., Dietz,1978; Hake, 1982; Lerman, 2003; Mace, 1994;Mace & Wacker, 1994), most translationalresearch has focused on establishing thegenerality of basic research to human affairs.However, coordinated basic and applied re-search can stimulate both sectors to pursueresearch questions and paradigms less likely tobe followed when basic and applied research-ers work separately. Experiment 1 was stimu-lated by our collaboration with Tony Nevin onclinical applications of behavioral momentum.Through this collaboration it became appar-ent that Condition A in Experiment 2 of Nevinet al. (1990) represented a clinical applicationof DRA (see Figure 1). That high-rate rein-forcement of prosocial behavior could inad-vertently strengthen the persistence of thesame unwanted target behavior it reduces wascounterintuitive. Our clinical demonstrationthat DRA can increase resistance to extinction,in turn, stimulated further collaborationamong basic and applied researchers toattempt a solution to this problem.

The present series of experiments is notwithout limitations. First, Experiment 1 didnot replicate the extinction following baseline

and extinction following DRA conditionswithin participants. We chose not to replicatethe extinction conditions so as not to subjectthe participants to additional lengthy extinc-tion processes. However, the principal findingin Experiment 1 that DRA can increase thepersistence of target behavior was replicated inboth participants in Experiment 3 in themultiple schedule design employed. Second,the extinction conditions in Experiments 2and 3 presented the discriminative stimulicorrelated with the target response in Compo-nent 1 and the compound of Components 1and 3 twice as many times as Component 2.Although greater resistance to extinction wasapparent early in the extinction process forboth rats and humans, replications of thisresearch should conduct extinction only withthe discriminative stimuli in Component 2 andcompound of Components 1 and 3 to elimi-nate this possible confound.

The present investigation suggests severalareas for future study that extend the bidirec-tional nature of this research and address somelimitations of our studies. First, we did notmeasure the resistance of alternative behaviorto extinction. Because strengthening the per-sistence of prosocial behavior is an importantclinical goal, this should be integral to futureresearch protocols. Second, the generality ofthe present findings on DRA and those usingtime-contingent schedules (Ahearn et al., 2003;DeLeon et al., 2005) should be examined forother forms of alternative reinforcement usedin clinical practice (such as differential rein-forcement of other behavior, differential rein-forcement of high rates, and differentialreinforcement of low rates). Finally, the persis-tence-strengthening effects of alternative rein-forcement may be best understood within thecontexts of the reinstatement, resurgence andrenewal paradigms discussed above. Explicitlinkage of this research to the clinical problemof failures of treatment maintenance andgeneralization may lead to novel solutions toan important limitation of Applied BehaviorAnalysis interventions.

REFERENCES

Ahearn, W. H., Clark, K. M., Gardenier, N. C., Chung, B. I.,& Dube, W. V. (2003). Persistence of stereotypicbehavior: Examining the effects of external reinforc-ers. Journal of Applied Behavior Analysis, 36, 439–448.


Carr, E. G., & Durand, V. M. (1985). Reducing behaviorproblems through functional communication train-ing. Journal of Applied Behavior Analysis, 18, 111–126.

Cohen, S. L. (1996). Behavioral momentum of typingbehavior in college students. Journal of BehaviorAnalysis and Therapy, 1, 36–51.

DeLeon, I. G., Williams, D. C., Gregory, M. K., &Hagopian, L. P. (2005). Unexamined potential effectsof the noncontingent delivery of reinforcers. EuropeanJournal of Behavior Analysis, 5, 57–69.

Dietz, S. M. (1978). Current status of applied behavioranalysis: Science versus technology. American Psycholo-gist, 33, 805–814.

DiGennaro, F. D., Marten, B. K., & Kleinmann, A. E.(2007). A comparison of performance feedbackprocedures on teachers’ treatment implementationintegrity on students’ inappropriate behavior inspecial education classrooms. Journal of Applied Behav-ior Analysis, 40, 447–461.

Fisher, W. W., Piazza, C. C., Cataldo, M., Harrell, R.,Jefferson, G., & Conner, R. (1993). Functionalcommunication training with and without extinctionand punishment. Journal of Applied Behavior Analysis,26, 23–36.

Fleshler, M., & Hoffman, H. S. (1962). A progression forgenerating variable-interval schedules. Journal of theExperimental Analysis of Behavior, 5, 529–530.

Hake, D. F. (1982). The basic–applied continuum and thepossible evolution of human operant social and verbalresearch. The Behavior Analyst, 5, 21–28.

Igaki, T., & Sakagami, T. (2004). Resistance to change ingoldfish. Behavioural Processes, 66, 139–152.

Iwata, B. A., Bailey, J. S., Neef, N. A., Wacker, D. P., Repp,A. C., & Shook, G. L. (1997). Behavior analysis indevelopmental disabilities, 1968–95 from the Journal ofApplied Behavior Analysis. Bloomington, IN: Society forthe Experimental Analysis of Behavior.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., &Richman, G. S. (1994). Toward a functional analysis ofself-injury. Journal of Applied Behavior Analysis, 27,197–209.

Iwata, B. A., Wong, S. E., Riordan, M. M., Dorsey, M. F., &Lau, M. M. (1982). Assessment and training of clinicalinterviewing skills: Analogue analysis and field repli-cation. Journal of Applied Behavior Analysis, 15, 191–203.

Lerman, D. C. (2003). From the laboratory to communityapplication: Translational research in behavior anal-ysis. Journal of Applied Behavior Analysis, 36, 415–419.

Lerman, D. C., Iwata, B. A., & Wallace, M. D. (1999). Sideeffects of extinction: Prevalence of bursting andaggression during the treatment of self-injuriousbehavior. Journal of Applied Behavior Analysis, 32, 1–8.

Mace, F. C. (1994). Basic research needed for stimulatingthe development of behavioral technologies. Journal ofthe Experimental Analysis of Behavior, 61, 529–550.

Mace, F. C., Lalli, J. S., Shea, M. C., Lalli, E. P., West, B. J.,Roberts, M., et al. (1990). The momentum of humanbehavior in a natural setting. Journal of the ExperimentalAnalysis of Behavior, 54, 163–172.

Mace, F. C., & Wacker, D. P. (1994). Toward greaterintegration of basic and applied behavioral research:An introduction. Journal of Applied Behavior Analysis,27, 569–574.

Mauro, B. C., & Mace, F. C. (1996). Differences in theeffect of Pavlovian contingencies upon behavioralmomentum using auditory versus visual stimuli.Journal of the Experimental Analysis of Behavior, 65,389–399.

Nevin, J. A. (1992). An integrative model for the study ofbehavioral momentum. Journal of the ExperimentalAnalysis of Behavior, 57, 301–316.

Nevin, J. A., & Grace, R. C. (2000). Behavioral momentumand the Law of Effect. Behavioral and Brain Sciences, 23,73–130.

Nevin, J. A., Tota, M. E., Torquato, R. D., & Shull, R. L.(1990). Alternative reinforcement increases resis-tance to change: Pavlovian or operant contingen-cies? Journal of the Experimental Analysis of Behavior, 53,359–379.

Osnes, P. G., & Lieblein, T. (2003). An explicit technologyof generalization. The Behavior Analyst Today, 3,364–374.

Petscher, E. S., Rey, C., & Bailey, J. A. (2009). A review ofempirical support for differential reinforcement ofalternative behavior. Research in Developmental Disabil-ities, 30, 409–425.

Podlesnik, C. A., Jimenez-Gomez, C., & Shahan, T. A.(2006). Resurgence of alcohol seeking produced bydiscontinuing non-drug reinforcement as an animalmodel of drug relapse. Behavioural Pharmacology, 17,369–374.

Podlesnik, C. A., & Shahan, T. A. (in press). Behavioralmomentum and relapse of extinguished operantresponding. Learning and Behavior.

Shaham, Y., Shalev, U., Lu, L., de Wit, H., & Stewart, J.(2003). The reinstatement model of drug relapse:History, methodology and major findings. Psychophar-macology, 168, 3–20.

Stokes, T. F., & Baer, D. M. (1977). An implicit technologyof generalization. Journal of Applied Behavior Analysis,10, 349–367.

Stokes, T. F., & Osnes, P. G. (1989). An operant pursuit ofgeneralization. Behavior Therapy, 20, 337–355.

Vollmer, T. R., Roane, H. S., Ringdahl, J. E., & Marcus, B.A. (1999). Evaluating treatment challenges withdifferential reinforcement of alternative behavior.Journal of Applied Behavior Analysis, 32, 9–23.

Received: July 30, 2009Final Acceptance: February 5, 2010


Documents

DIFFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR … · DIFFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR INCREASES RESISTANCE TO EXTINCTION: CLINICAL DEMONSTRATION, ANIMAL MODELING,