JOURNAL OF APPLIED BEHAVIOR ANALYSIS

NEGATIVE REINFORCEMENT IN APPLIED BEHAVIOR ANALYSIS:AN EMERGING TECHNOLOGY

BRIAN A. IWATAUNIVERSITY OF FLORIDA

Although the effects of negative reinforcement on human behavior have been studied for a numberof years, a comprehensive body of applied research does not exist at this time. This artide describesthree aspects of negative reinforcement as it relates to applied behavior analysis: behavior acquiredor maintained through negative reinforcement, the treatment of negatively reinforced behavior, andnegative reinforcement as therapy. A consideration of research currently being done in these areassuggests the emergence of an applied technology on negative reinforcement.DESCRIPTORS: aversive stimulation, avoidance, escape, negative reinforcement

Research published in the Journal of AppliedBehavior Analysis (JABA) has, for 20 years now,demonstrated how knowledge about environment-behavior interactions, particularly those involvingresponse-contingent events and correlated stimuli,may be used for the benefit of individuals and thelarger society. In doing so, applied research has alsomade significant contributions to the general scienceof behavior by providing extension and externalvalidation of experimental findings from the basicresearch laboratory (Baer, 1978).

Along with the development of the applied fieldand its expansion into a number of areas in whichthe outcome of an experiment often has immediatesocial implications (e.g., business and industry, de-velopmental disabilities, education, medicine, men-tal health, public affairs), there has been growingconcern of a widening gap between basic and ap-plied behavior analysis. Critics (Deitz, 1978; Pierce& Epling, 1980) have indicated that the emphasisofcontemporary applied behavior analysis has shift-ed away from the study of conditions that producechange to the production of change per se, and that"relevance to basic principle," a supposed char-

Appreciation is expressed to Ed Malagodi, Jack Michael,Gary Pace, Terry Page, and Terri Rodgers, whose work orhelpful suggestions assisted in the preparation of this article.Support was provided by Grant HD 16052 from the NationalInstitute of Child Health and Human Development.

Reprints may be obtained from Brian A. Iwata, Depart-ment of Psychology, University of Florida, Gainesville, Flor-ida 32611.

acteristic of applied behavior analysis (cf. Baer,Wolf, & Risley, 1968), is reflected less and less inthe research that journals such as JABA publish.The general accuracy of these criticisms, as well astheir basis and implications, will continue to be thesubject of periodic debate (Baer, 1981; Cullen,1981; Michael, 1980). Nevertheless, it is possibleat this point in the development of our field toidentify specific and well-established areas of basicresearch for which little parallel exists in the appliedliterature, and vice versa.

The thesis of this article is that research on neg-ative reinforcement provides one of the dearest andmost immediately relevant examples of a case inwhich consideration, replication, and extension ofbasic research would benefit the applied area. Alongwith positive reinforcement and punishment, neg-ative reinforcement has long been considered oneof the elementary principles of behavior. A volu-minous amount of research on negative reinforce-ment exists in the basic literature (see reviews byBolles, 1970; Herrnstein, 1969; Hineline, 1977,1981, 1984; Hoffman, 1966; Schoenfeld, 1969;Sidman, 1966), and its inclusion as a distinct topicin texts on experimental analysis (e.g., Honig &Staddon, 1977) justifies its status as a major or-ganizing principle. For example, acquisition, main-tenance, extinction, and stimulus control all havebeen studied using negative reinforcement as theoperant mechanism of interest.

Sandler and Davidson (1973) reviewed some ofthis basic research and discussed its relevance to the

361

1987,209,361-378 NUMBER4 (wiNnR 1987)

BRIAN A. IWATA

development and treatment of pathological humanbehavior. They concluded that ". . . the escape andavoidance paradigms are still plagued by a numberof unresolved issues .. . " (p. 254) that they hopedwould be darified by additional basic research andextension to the world of humans. Since that time,a number of investigations on negative reinforce-ment with humans have been conducted, yet asystematic and comprehensive body of applied re-search still does not exist. Consider the two mostrecent texts on aversive control with humans (Ax-elrod & Apsche, 1983; Matson & DiLorenzo,1984). Both provide thorough discussion of topicssuch as positive reinforcement, extinction, time-out,response cost, and contingent aversive stimulation.Thus, one might expect that these texts would bethe most likely sources of information on negativereinforcement as well, but this is not the case. Onetext (Axelrod & Apsche, 1983) devotes less thana half dozen of over 300 pages to the topic ofavoidance, and the discussion always is limited toavoidance as a side effect of punishment. No men-tion is made of escape or avoidance as directlyproduced performances. The second (Matson &DiLorenzo, 1984) describes the hypothetical fea-tures of escape and avoidance training on two pagesbut does not cite any applied references.The relative absence of integrated material on

negative reinforcement with humans raises severalquestions concerning generality and utility. Is hu-man behavior relatively insensitive to contingenciesof negative reinforcement? Are naturalistic humansituations typically characterized by the absence ofstimuli that can function as negative reinforcers, oropportunities to escape from or avoid these stimuli?What types of performances are likely to be ac-quired through negative reinforcement? Finally, doprocedures based on the application of negativereinforcement, unlike those based on positive re-inforcement and punishment, have little therapeuticor pragmatic value?

For the past few years my students, colleagues,and I have been conducting a series of investigationsin two areas-self-injurious behavior and pediatricfeeding disorders. Curiously, these very differentproblems have brought us into direct contact with

situations involving the use of negative reinforce-ment and have forced us to consider more generallythe relevance of negative reinforcement in appliedbehavior analysis. Our experience and our exami-nation of the basic and applied research suggestthat the answer to each of the above questions is"No." In fact, it appears that negative reinforce-ment plays a central role in the development ofmany behaviors, appropriate as well as inappro-priate, and that its application in a number ofstudies has not been formally acknowledged. Inwhat follows, I will describe three aspects of neg-ative reinforcement as it relates to the applied sit-uation: first, undesirable behavior acquired or main-tained through negative reinforcement; second, thetreatment of negatively reinforced behavior; andthird, negative reinforcement as therapy. This or-ganization departs somewhat from that used inreviews of the basic research literature and has beenadopted here to highlight the relevance ofparticularissues to the applied researcher. Much of the re-search included here has been done with the de-velopmentally disabled population because there isa high prevalence of significant behavioral disordersin this group and because it provides a narrow butadequate focus for discussion.

CURRENT CONCEPTUALIZATION OFNEGATIVE REINFORCEMENT

Before proceeding, it may be helpful to clarifyterminology and to delineate the defining featuresof negative reinforcement. The purpose of thisdigression is to show that the task of determiningwhether a given contingency is an example of neg-ative reinforcement may not always be a simpleone. Although there has been little confusion re-garding the effect of negative reinforcement, de-scribing its operations has posed a challenge tomany beyond the level of the beginning student.

The process of negative reinforcement typicallyinvolves the removal, reduction, postponement, orprevention of stimulation; these operationsstrengthen the response on which they are contin-gent (Hineline, 1977). Removal and reduction ofongoing stimulation typically produce behavior that

362

NEGATIVE REINFORCEMENT

is called "escape," whereas postponement and pre-vention of stimulus presentation produce behaviorthat is called "avoidance." "Typically" is used asa qualifier throughout because the terms negativereinforcement, escape, and avoidance are subject toconfusion under certain conditions, as the followingwill illustrate.

In commenting on the distinction between pos-itive and negative reinforcement, Michael (1975)reviewed a number of historical points related toterminological usage. More important, he notedthat some stimulus changes associated with an in-crease in behavior are difficult to dassify as "pre-sentation" (positive reinforcement) versus "remov-al" (negative reinforcement), and that the use ofeither description may be nothing more than anarbitrary and incomplete abbreviation for the static"prechange" and "postchange" stimulus condi-tions as well as for what transpires in between. Forexample, is a change in temperature more accuratelycharacterized as the presentation of cold (heat) orthe removal of heat (cold)? Problems such as thisled Michael to suggest that "The distinction be-tween two types of reinforcement (positive vs. neg-ative], based in turn upon the distinction betweenpresentation and removal simply can be dropped"(p. 44). An additional basis for distinguishing be-tween positive and negative reinforcement was sug-gested first by Catania (1973) and later by Hineline(1984), who noted that ". . . if a stimulus or sit-uation is to be reducible or removable by someresponse, that response must occur in its presence.In contrast, positively reinforced responses neces-sarily occur in the absence ofthe stimuli upon whichreinforcement is based" (pp. 496-497). Such adistinction is not without its own problems, as canbe seen in the previous example. Is responding priorto a temperature change more accurately describedas responding in the presence versus the absence ofheat (cold)? Another problem with this distinctionis encountered when one considers the differencebetween escape (responding in the presence of stim-ulation), and avoidance (responding in the absenceof stimulation), both of which are examples ofnegative reinforcement.

In many applied situations, it is possible to iden-

tify unambiguously a stimulus change as one in-volving presentation (e.g., of physical contact) orremoval (e.g., of a token) and to determine whetheror not the response of interest occurs in the presenceor the absence of stimulation. However, becauseresearch outside of the laboratory is subject to great-er variation of and less control over a multitude ofpotentially relevant stimuli, the motivational fea-tures of some stimulus changes are difficult to spec-ify. Consider, as a case in point, Osborne's (1969)"Free-time as a reinforcer in the management ofclassroom behavior," which examined the out-of-seat behavior of six students. During the baselinecondition, the students worked for approximately4 hours daily without recess, and data showed thatthe target behavior occurred frequently. Duringtreatment, students could earn 5 min of free timeat the end of every 1 5-min work period by re-maining in their seats, and the data showed a de-crease in out-of-seat. It is interesting to note thetarget behavior. Defined and recorded as "out-of-seat," free time was made available for its absence;this type of contingency usually is described asdifferential reinforcement of other behavior (DRO).However, the instructions given to students spec-ified that they were to remain in their seats, sug-gesting "in-seat" as the functional target. If so, freetime was made available for the occurrence of in-seat behavior; this type of contingency is not con-sidered an example of DRO. Depending on howone characterizes "free time," (i.e., the availabilityof preferred activities vs. the termination of non-preferred activities), we would label the contingencyas one involving positive or negative reinforcementfor in-seat behavior. Osborne suggested both ofthese possibilities in his discussion and perhaps thisis why he did not place an adjective in front of theterm "reinforcer" in the title of the article.

As a field, we have not attended carefully to theimportant distinction that Osborne drew. His studyis regarded as a seminal piece of research in theapplied literature for expanding our notion of whatconstitutes a reinforcing event and for demonstrat-ing very nicely the effects of group contingencies,although the exact nature of the contingency is stillunclear. A number of interesting replications and

363

BRIAN A. IWATA

extensions have appeared inJABA (e.g., Aaron &Bostow, 1978; Baer, Rowbury, & Baer, 1973;Barrish, Saunders, & Wolf, 1969; Harris & Sher-man, 1973, 1974; Long & Williams, 1973; Ma-loney & Hopkins, 1973; Medland & Stachnik,1973), but none have included further discussionor analysis of free time contingencies as positiveversus negative reinforcement. Although such anal-yses may have little or no impact on outcome (i.e.,in either case, behavior will have been increased),our general tendency to overlook a negative rein-forcement interpretation may lead to undue em-phasis on the numerous forms that free time maytake at the expense of considering important fea-tures of the environment that free time replaces.That is, if free time serves as negative reinforcement,its only essential component may be alteration ortermination of the preceding aversive situation.

In a more general sense, the complete analysisand specification of conditions in effect prior andsubsequent to responding was the primary basisunderlying Michael's (1975) suggestion to elimi-nate the distinction between positive and negativereinforcement. It appears unlikely that the terms"positive" and "negative" will be deleted from ourtechnical vocabulary in the near future; neverthe-less, researchers should be cognizant of the fact thatthe two are potentially interchangeable and thatfailure to consider both possibilities may have alimiting effect on experimental procedure, inter-pretation, and subsequent application.

UNDESIRABLE BEHAVIOR ACQUIRED ANDMAINTAINED BY NEGATIVE

REINFORCEMENT

Hineline (1977) noted that a typical negativereinforcement paradigm includes three features: thepresence of aversive stimulation, the availability ofa response, and a suitable contingency between theresponse and the stimulation. Any behavior thusdeveloped or maintained, including a variety ofdisruptive, destructive, aggressive, self-injurious, andotherwise problematic acts, could be considered"normal" or "adaptive" in that it is the orderlyoutcome of specific conditioning operations. The

term "undesirable" is used here only as a meansof classifying behaviors that are considered inap-propriate given the usual social context.An initial question of particular interest to those

working in applied areas relates to factors that de-termine the form of the response. Acquisition ofnegatively reinforced behavior has been a subjectof interest to basic researchers as well because it hasbeen found that some topographies are more readilyproduced than others. For example, the treadle-press and shuttle responses of pigeons are moreeasily controlled by negative reinforcement than isthe key peck, which is highly responsive to positivereinforcement (Ferrari, Todorov, & Graeff, 1973;Foree & LoLordo, 1970; MacPhail, 1968; Rachlin& Hineline, 1967; Smith & Keller, 1970). Similardata based on the study of different species haveprovided some support for the hypothesis that neg-ative reinforcement involves selective control overpreexisting "species-specific defense reactions" toaversive stimulation (Bolles, 1970, 1971). Thisaccount, however, does not provide adequate ex-planation for the wide range of human behaviorsthat apparently is succeptable to negative reinforce-ment. A more likely explanation is that aversivestimulation initially produces one or more of avariety of responses characteristic of both humanand nonhuman subjects, including flinching, freez-ing, jumping, visual scanning, and related and dif-fuse motor activity (see reviews by Davis, 1979;Hutchinson, 1977; Myer, 1971), and that theeventual and more elaborate form of the behavioris determined by the individual's previous historyand the prevailing contingency.

Thus, many of the serious behavioral disordersthat are seen in, for example, mentally retardedindividuals may be a function of negative rein-forcement applied to a particular behavioral rep-ertoire and shaped over time. It is possible thatcertain instructional sequences (e.g., requests or eventhe appearance of specific training materials or theinstructor) become discriminative for aversive stim-ulation in the form of physical contact, which is acommon element in many teaching routines. Atfirst, the stimulation and its associated cues mayproduce behaviors similar to those noted above. If,

364

NEGATIVE REINFORCEMENT

however, other behaviors have been successful ineliminating similar types of stimulation in the past,their eventual occurrence should not be surprising.Tantrums, attempts to flee, or destruction of ma-terials are examples of such behavior, particularlyif the individual is unskilled at more subtle orsocially acceptable forms of escape. Although dis-ruptive, these behaviors usually are not consideredinsurmountable barriers to instruction. A numberof informal and formal interventions (e.g., pro-ceeding in spite of the tantrum, "scooting" theindividual's chair under a table and backing bothagainst a wall, bolting the materials to the table,etc.) are successful in managing disruptive behaviorin some cases. In other cases, the interventions mayprovide a means for shaping more serious forms ofescape. The immediate result of aggression for theindividual toward whom it is directed suggests thatphysically harmful acts could serve as very effectiveescape behaviors, and their ability to terminate aver-sive instruction is most likely a function of therelative size and strength of client and trainer. Fi-nally, self-injurious behavior, if severe enough, willquickly terminate any situation.

Data relevant to a negative reinforcement hy-pothesis for the development of behavior disordersin the mentally retarded exist in retrospective formonly because it would be unethical to producepathological behavior in humans when it does notalready exist. Nevertheless, support for such a hy-pothesis can be found in several studies. Carr andDurand (1985) and Weeks and Gaylord-Ross(1981) showed that several different topographiesof inappropriate behavior occurred more frequentlyduring a "difficult task" condition when comparedto an "easy task" condition, suggesting that theformer condition contained aversive properties andthat the resulting behavior was escape- or avoid-ance-motivated. Carr, Newsom, and Binkoff(1976)examined variables that apparently exerted stim-ulus control over the self-injurious behavior of apsychotic boy. In one of their experiments, theypresented the boy with three alternating situations:a free-play period, a condition in which the exper-imenter spoke descriptive sentences to the child(e.g., "The sky is blue")-this was called the "tact"

condition, and a third condition in which the ex-perimenter spoke instructions to the child-this wascalled the "mand" condition. Higher levels of self-injurious behavior were associated with the mandcondition. Carr, Newsom, and Binkoff(1980) con-ducted a similar analysis of aggressive behavior intwo boys, showing that aggression was more likelyto occur when demands were present than whenthey were absent. Finally, Iwata, Dorsey, Slifer,Bauman, and Richman (1982) described a generalmethodology that allowed one to differentiate self-injury associated with positive versus negative re-inforcement. In one of the conditions, self-injurywas followed by adult attention; in another, self-injury produced brief escape from adult demands.Some subjects consistently exhibited self-injury dur-ing the latter condition, suggesting that their be-havior was more sensitive to and maintained bynegative reinforcement.

It is important for us to identify how environ-ments that we create may provide negative rein-forcement for undesirable behaviors. When facedwith situations in which our students and clientsare disruptive, we should immediately examine theantecedent as well as the consequent conditions todetermine ifthe difference between the two providesreduction of aversive stimulation, keeping in mindthat negative reinforcers may be just as idiosyncraticas positive ones. Ifwe condude that our dients andstudents exhibit bizarre and potentially dangerousbehaviors to terminate instruction, we might ques-tion whether or not our well-intentioned efforts toteach are in our dients' best interest; at the veryleast, we must question one or more aspects of ourteaching technique. Perhaps most important fromthe standpoint of contingencies, our ability to iden-tify negative reinforcement as a maintenance vari-able for undesirable behavior may directly influencetreatment selection and outcome. This is particu-larly true with respect to extinction and time-out.Their use typically calls for one or more therapistresponses (e.g., turning away from the client, re-moving stimul from immediate access, removingthe dient from the setting, etc.) that terminate theongoing situation. Studies conducted with non-human (Appel, 1963; Azrin, 1961; Thompson,

BRIAN A. IWATA

1964) and human (Plummer, Baer, & LeBlanc,1977; Solnick, Rincover, & Peterson, 1977) sub-jects, however, indicate that the effects of time-outare highly dependent on features of the "time-in"environment. Thus, although time-out might bean effective means of extinguishing most positivelyreinforced behavior, it might directly strengthennegatively reinforced behavior.

TREATMENT OF NEGATIVELYREINFORCED BEHAVIOR

A number ofprocedures based on the applicationof extinction, differential reinforcement, and pun-

ishment have been evaluated as treatments forproblematic behavior of unspecified origin. Theiruse with behavior maintained by negative rein-forcement will be discussed in this section, alongwith an additional procedure involving stimulusfading.

Extinction

Traditional time-out will not provide for theextinction of behavior that has been maintained bynegative reinforcement, but other procedures might.One rather obvious possibility is elimination of thesupposed aversive stimulation and its related cues,

which should produce a reliable decrease in escape

or avoidance behavior (Boren & Sidman, 1957;Shnidman, 1968). However, as Hineline (1977)has noted, this procedure may not be a true ex-

tinction operation. The complete removal of aver-

sive stimulation during extinction of negativelyreinforced behavior can be considered analogous to

the continuous presence of, for example, food dur-ing extinction of positively reinforced behavior. Bothprocedures amount to noncontingent reinforce-ment, which removes the basis for responding andindirectly reduces the frequency of behavior. Thatis, if food is always present during extinction offood-maintained behavior, there is no basis for re-

sponding; a similar situation exists ifshock is alwaysabsent during extinction of shock-avoidance be-havior. Following these procedures, food removalor, alternatively, reappearance of the shock shouldimmediately produce the target response (see Mi-

chael, 1982, for an extended discussion of thistopic).A more appropriate extinction procedure would

entail continued presentation of the aversive stim-ulus or its cue and elimination of the consequencethat was provided formerly (i.e., avoidance or es-cape). In this manner, the basis for responding(aversive stimulation) remains, but reinforcementdoes not (Bankart & Elliott, 1974; Coulson, Coul-son, & Gardner, 1970; Davenport, Coger, & Spec-tor, 1970; Schiff, Smith, & Prochaska, 1972).Techniques derived from this type of extinctionactually have been used for a number of years inthe treatment of clinical phobias and provide themajor theoretical basis for interventions collectivelyknown as "implosion therapies" (Levis, 1979).An example of extinction for negatively rein-

forced behavior was reported recently by Heidornand Jensen (1984). After noting that demand-related situations were associated with an increasein their subject's self-injurious behavior, a treatmentwas developed that included the following: (a) con-tinued presentation of demands, (b) physical guid-ance to complete the requested performance con-tingent on the occurrence of self-injury, (c)termination of the session contingent on compli-ance, and (d) gradual increase in performance cri-teria across sessions. Positive reinforcement in theform of praise, food, and physical contact also wasprovided, but its role as an active component oftreatment may have been minimal. A similar pro-cedure was used in one of the experiments reportedby Carr et al. (1980) on the treatment of aggression.Extinction consisted of belting the subject in a chairto prevent escape while a therapist wearing pro-tective gear sat across a table from him. The in-tervention differed from that used by Heidorn andJensen in that no attempt was made to deliverinstructions during extinction sessions; instead, de-mands were introduced after aggressive behaviorwas eliminated almost completely.

As with extinction of positively reinforced be-havior, it is possible to foresee situations in whichextinction of negatively reinforced behavior mightnot be in the immediate best interest of either theclient (as in the case of severe self-injury) or the

366

NEGATIVE REINFORCEMENT

therapist (as in the case of aggression). Extinctionprocedures may be compromised further by thepotential effects of what procedurally may resemblenoncontingent aversive stimulation (see earlier dis-cussion on acquisition of avoidance responding).To the extent that these "elicited" responses occurduring the extinction of negatively reinforced be-havior in applied situations, attempts to increasealternative behaviors, as well as to reduce the targetbehavior, may be disrupted. Finally, research show-ing that time-based delivery of aversive stimulationcan maintain (Powell & Peck, 1969) and evenincrease (Kelleher, Riddle, & Cook, 1963; Sidman,Herrnstein, & Conrad, 1957) the rates of avoidancebehavior suggests that schedule-related variablesand the subject's previous history may be importantconsiderations in the use of extinction.

Differential ReinforcementApplications ofreinforcement to decrease a target

behavior (differential reinforcement of other be-havior [DROI, differential reinforcement of incom-patible behavior [DRI], etc.) are well documentedin the applied literature, although the maintainingvariable for the target behavior rarely is noted. Thereinforcement contingency itself typically involvesthe use of positive reinforcement, and discussionhere will be similarly confined. Applications of neg-ative reinforcement will be addressed separately.An experiment designed to examine the sup-

pressive effects of differential reinforcement on neg-atively reinforced behavior may take several forms.First, access to an appetitive reinforcer (e.g., food)could be made contingent on the absence of thetarget behavior (DRO) while the escape/avoidancecontingency is still operative. Although this ap-proach might be considered unusual, it may resem-ble very closely situations in the natural environ-ment in which DRO is implemented withoutattempting to identify the behavior-maintainingcontingency. To my knowledge, this study has notbeen reported in the basic literature, probably dueto difficulties associated with equating reinforce-ment. It is possible that this type of study has beenreported in the applied literature but that it wasnot explicitly identified.

A second approach might involve appetitive re-inforcement for a competing behavior (DRI) withthe escape/avoidance contingency again operative.Ruddle, Bradshaw, Szabadi, and Foster (1982)studied human operant performance (button press-ing) using exactly this procedure. They presentedsubjects with concurrent avoidance/positive rein-forcement schedules, and obtained matched re-sponding when the schedules were equated (thiswas made possible by using points exchangeablefor money). Performance shifts were correlated withschedule shifts roughly in a manner predicted byHerrnstein's (1961) matching law. Our assessmentresearch on self-injury (Iwata et al., 1982) providesan approximation to the Ruddle et al. methodol-ogy. During one condition, we presented to subjectsa series of instructional demands. Compliance wasfollowed by praise and physical contact from theexperimenter, whereas the occurrence of self-injuryproduced a 30-s time-out. Data gathered duringthat study, as well as those collected since, indicatethat both responses are likely to occur; in otherwords, positive reinforcement for compliance alonedoes not suppress avoidance-motivated self-injury.Another example of differential reinforcement wasreported by Kelley, Jarvie, Middlebrook, McNeer,and Drabman (1984). They provided token rein-forcement (stars) for reductions in the pain behavior(screaming, interfering, etc.) of two childrenundergoing open burn treatment. The procedurewas moderately effective in that reductions in painbehavior averaged less than 50%. The findings ofIwata et al. and Kelley et al. are consistent withthose of Ruddle et al., indicating that positivelyreinforced behavior competes with but does notsuppress avoidance or escape responding that isreinforced concurrently. In contrast, Carr et al.(1980) were able to obtain almost complete elim-ination of aggression in one of their subjects byintroducing positive reinforcement to an existingdemand situation. They did note, however, thattheir second subject was not responsive to the pos-itive reinforcement and that a different treatment(see previous discussion of extinction) was used.A third experiment might examine reinforce-

ment, as described in either of the above examples,

367

BRIAN A. IWATA

combined with extinction (continued presentationof aversive stimuli and prevention of escape). TheHeidorn and Jensen (1984) study on self-injury,described previously, is an example of this ap-proach. From an applied perspective, their proce-dures represent optimal treatment because contin-gencies were provided for the inappropriate as wellas the appropriate behavior. However, here theeffects of reinforcement are inseparable from thoseof extinction, and a dearer interpretation wouldrequire comparative analysis (reinforcement plusextinction vs. reinforcement alone vs. extinctionalone).

The studies described here remain prototypicalfor the most part because very little research hasbeen reported on the use of differential (positive)reinforcement with escape and avoidance behavior.On purely ethical grounds, and for the purposes ofestablishing and strengthening alternative behav-iors, the use of positive reinforcement seems criti-cally important. On the other hand, its therapeuticeffects as primary treatment for negatively rein-forced behavior have yet to be demonstrated. Basedon the small amount of data available, one mightexpect that positive reinforcement is more likely toproduce beneficial results if the negatively rein-forced behavior is extinguished concurrently or ifthe density of positive reinforcement is noticeablygreater than that of the negative reinforcement.

PunishmentContingent aversive stimulation for negatively

reinforced behavior is the functional complementof DRO for positively reinforced behavior, in thatprevention of aversive stimulation (negative rein-forcement) is contingent on the absence of respond-ing. Procedural curiosities aside, we know very littleabout the effects of punishment on human escapeand avoidance, in spite of the many applied studieson punishment published to date. For example, theliterature on self-injury, in which most ofthe currentapplied research on punishment can be found, con-tains only two studies reporting the use of punish-ment for behavior described as avoidance motivat-ed. One of the elements in the Heidorn and Jensen(1984) multiple-treatment approach consisted of

physical guidance to complete a requested perfor-mance, contingent on the occurrence of self-injury.It is interesting to note that the particular form ofstimulation used as punishment may have beenexactly the same aversive stimulation whose priorremoval served as negative reinforcement; if so, thetreatment amounted to a perfect reversal of themaintaining contingency. Borreson (1980) also re-ported a case study of multiple treatment for avoid-ance-motivated self-injury; however, the punishingstimulus-"forced running" up and down a stair-way-appeared to be unrelated to the prior func-tion of the behavior.

Punishment of negatively reinforced behaviorpresents significant complexities not found withpositively reinforced behavior because it involvesaversive stimulation following responses for whichsuch stimulation already plays an important role asan eliciting, discriminative, and motivating event(for extensive reviews of this topic, see Davis, 1979;Fowler, 1971; Hineline, 1981; and Morse & Kel-leher, 1970, 1977). The major issues are sum-marized here. First, the eliciting properties of aver-sive stimulation, described previously with respectto acquisition and extinction, are relevant in thecase of punishment. Although elicited behavior maynot necessarily compromise the use of punishment,it may have a deleterious effect on the overall treat-ment program. Second, punishment with the samestimulus used during escape or avoidance trainingmay acquire discriminative properties for respond-ing as a result of reinstating the conditions underwhich escape originated, thereby occasioning thevery behavior being punished. For example, severalstudies have shown response maintenance and evenfacilitation when shock-preventing behavior wasfollowed by the presentation of shock (e.g., Appel,1960; Sandier, Davidson, & Malagodi, 1966).Third, schedule-related variables can determinewhether contingent stimulation serves as either pun-ishment or reinforcement. Kelleher and Morse(1968) and McKearney (1972) found that re-sponding developed as avoidance behavior was sup-pressed under dense schedules of punishment butfacilitated under thinner schedules. Finally, it hasbeen noted that punishment intensity and the pres-

368

NEGATIVE REINFORCEMENT

ence or absence of avoidance contingencies mayhave an interactive effect on behavior. Sandier, Da-vidson, Greene, and Holzschuh (1966) imposedhigh-, intermediate-, and low-intensity shock aspunishment for ongoing avoidance behavior andfound greater response persistence under the high-intensity condition. However, when the avoidancecontingency was later removed (i.e., respondingproduced shock but no longer prevented it), thehigh-intensity condition produced the most rapidresponse suppression.

The use of punishment should always be con-sidered very carefully, and even greater cautionshould be taken when there is reason to believethat the target behavior has been maintained bynegative reinforcement. Findings from the basicresearch literature suggest, although in a very ten-tative manner, that a stimulus different from thatassociated with prior avoidance should be used,that the schedule of punishment should be a con-tinuous one, and that "mildly aversive" stimulimay produce greater response suppression than moreintense stimulation when the prevailing avoidancecontingency remains operative. On the other hand,data from the Heidorn and Jensen (1984) studyindicated that, within the context of their multiple-treatment approach, the relationship between stim-ulation used as negative reinforcement and punish-ment may not be an important one.

Stimulus FadingIn contrast to approaches in which a contingency

is directly manipulated, fading consists of alteringone or more features of stimuli that occasion thetarget behavior. Various types of stimulus fadinghave been used for many years in the treatment ofclinical fears and phobias, dating back to the workofJones (1924). Contemporary formulations differgreatly along procedural dimensions (actual vs. rep-resentational stimulus presentation, the presence orabsence of reinforcement and punishment) as wellas on underlying theory (respondent vs. operantconditioning). The operant model of stimulus fad-ing to reduce escape or avoidance behavior involves(a) initial identification of response-producing stim-uli, (b) stimulus alteration to the point where re-

sponding does not occur, (c) presentation of thealtered stimuli with a gradual return to their orig-inal state, and (d) extinction of escape behavior.

Approximations to the stimulus fading approachcan be found in two studies previously discussed.Heidorn and Jensen (1984) decreased and thengradually increased the frequency with which re-sponse-producing stimuli (demands) were pre-sented, although they did not withdraw them en-tirely at the beginning of treatment. In contrast,Carr et al. (1976) were able to reduce self-injuryby embedding demands within entertaining stories,although the stories were never faded out nor wereadditional demands faded in. The results of bothstudies suggest that more complete evaluations oftreatment based on fading are warranted. One po-tential advantage of fading over extinction and pun-ishment might be the complete elimination of es-cape behavior from the outset of treatment.

NEGATIVE REINFORCEMENT ASTHERAPY

I have noted previously that free-time contin-gencies might function as negative reinforcement,although that possibility has been seldom acknowl-edged. Free time also may be one of the few con-tingencies in the applied literature that representspure escape in that the stimulation (work) is rel-atively continuous and can be reduced or terminatedbut not avoided. A great majority of applicationsmake use of time- or trial-based presentation ofstimuli preceded by cues, which produces avoidancebehavior. Examples of negative reinforcement usedto strengthen desirable behavior will be discussedin this section, grouped according to similarities ineither procedure or problem.

Behavioral EngineeringThe earliest examples of negative reinforcement

to develop or maintain appropriate behavior pub-lished in JABA made use of apparatus-deliveredstimulation. Azrin and his colleagues conductedtwo such studies. The first (Azrin, Rubin, O'Brien,Ayllon, & Roll, 1968) established automated mea-surement and control over postural slouching. An

369

BRIAN A. IWATA

apparatus built into a shoulder harness operatedclosure of a circuit when slouching occurred. Thisaction produced an audible click, followed 3 s laterby a 5 5-db tone. The contingency was an ingeniousone in that it included aspects of both free-operantand discriminated avoidance plus escape. Mainte-nance of correct posture (free-operant) avoided thedick, postural correction (discriminated) during the3 s following the dick avoided the tone, and cor-rection during the tone provided escape. Subjectsconsisted of 25 adults, all of whom showed re-ductions in slouching while the device was worn.When the contingency was reversed for two of thesubjects, both showed increases in slouching. In alater study, Azrin and Powell (1969) evaluated anapparatus to increase pill taking in six subjects. Thepill dispenser produced a 50-db tone every 30 min,an arbitrary between-pill interval. The tone couldbe turned offby pushing a knob on the case, whichalso delivered two pills. One might expect that thisarrangement would produce escape behavior ini-tially, followed by free-operant avoidance, althoughdata to that effect were not presented. A third studyinvolving apparatus-delivered negative reinforce-ment was conducted by Greene and Hoats (1969).Their subject was an adult male assigned to a cor-rectional unit whose task was to sort computercards, for which he earned cigarettes. He also wasallowed to watch TV while performing the task.During the treatment condition, if the subject didnot complete a task cycle within a specified intervalof time (avoidance), visual and auditory outputfrom the TV were distorted and remained that wayuntil the work cycle was completed (escape).

Toilet Training and IncontinenceThe presence or absence of elimination is more

than the occurrence or nonoccurrence of a response.Sphincter contraction as well as relaxation is in-volved, and negative reinforcement has been usedin the management of both behaviors. Hansen(1979) incorporated escape and avoidance contin-gencies in the treatment of nocturnal enuresis intwo children. When a device placed in the beddetected urination, an apparatus located 4 ft awayproduced a 70-db tone, which was followed in 7

s by a 95-db tone. By remaining dry, the childcould avoid the first tone; if the child urinated, heor she could escape the first tone and avoid thesecond by immediately getting out of bed and turn-ing the unit off. O'Brien, Ross, and Christophersen(1986) recently used negative reinforcement to pro-duce elimination. As part of an overall treatmentprogram for four encopretic children, the authorswanted to establish morning control over bowelmovements. To do so, parents had the children siton the toilet each morning for 5 min; failure todefecate a minimum equivalent of one-fourth of acup during that time was followed immediately byinsertion of a suppository. A second administrationwas given if the first did not produce the desiredoutcome.

OvercorrectionOriginally designed as a means for eliminating

accidents during toilet training (Azrin & Foxx,1971), overcorrection consists of a group of tech-niques whose common feature is repetitive perfor-mance of motor activity. Overcorrection is one ofthe most thoroughly studied and frequently usedmethods for reducing the frequency of a wide rangeof undesirable behaviors (see Foxx & Bechtel, 1983,for a review). The procedures are considered to bederivatives of punishment and are applied contin-gent upon the occurrence of a target behavior. Atthe same time, overcorrection can serve as negativereinforcement in at least two ways. First, becausethe procedure calls for performance of activities thatapparently are aversive, a therapist always is at handto ensure compliance through continued instructionand, if necessary, physical guidance. Thus, the clientcan avoid repeated instructions and potentially in-trusive physical contact through continued perfor-mance of the required activity. Avoidance behavioris also produced when overcorrection is appliedcontingent upon the absence of a desirable response.For example, Foxx (1977) showed that 5 min offunctional movement training, involving the prac-tice of varying head positions, was superior to foodand praise in developing and maintaining eye con-tact in three retarded children. Examination of theovercorrection literature yields a number of in-

370

NEGATIVE REINFORCEMENT

stances in which the procedure was used to increaserather than decrease behaviors, including class at-tendance (Foxx, 1976), repetitive tasks (Carey &Bucher, 1983), sharing (Barton & Osborne, 1978),speech (Matson, Esveldt-Dawson, & O'Connell,1979), and spelling accuracy (Foxx & Jones, 1978;Ollendick, Matson, Esveldt-Dawson, & Shapiro,1980).

Error Correction During InstructionRodgers and Iwata (1987) recently conducted a

survey whose initial focus was on response prompt-ing as an adjunct during behavioral acquisition.We quickly determined that most prompting oc-curs following an error and expanded the analysisto include all events that can be made contingenton incorrect responses. Negative reinforcement wasnot a subject of interest at the outset, but some ofthe techniques that were found suggest that it playsa much more prominent role in the instructionalprocess than is currently acknowledged.

The most dramatic example is a study by Kirch-er, Pear, and Martin (1971) entitled "Shock aspunishment [emphasis added) in a picture-namingtask with retarded children." In one experiment,two children were exposed to the following twotreatments: (a) token reinforcement for correct pic-ture-naming responses, and (b) token reinforcementfor correct responses plus shock for either errors ora response latency greater than 5 s. The shockcondition produced superior results. Because thetoken reinforcement remained constant across thetwo conditions and because, regardless of how onedefines the target behavior (i.e., errors vs. correctresponses, inattention vs. attention), the desirableperformance was a correct picture name, the pro-cedure dearly represents an avoidance contingencyin that correct responses made within 5 s of thecue prevented the delivery of the shock.

The Kircher et al. (1971) study represents arather extreme use of negative reinforcement toincrease desirable behavior, one that cannot be de-fended on ethical grounds today. However, lessdramatic but analogous situations are quite com-mon in the literature on instructional technology.It has become standard practice to follow errors

with statements ofdisapproval (Dunlap &Johnson,1985; Rincover & Newsom, 1985; Schreibman,1975), physical guidance (Haring, 1985; Luyben,Funk, Morgan, Clark, & Delulio, 1986; Sprague& Homer, 1984), session-lengthening proceduresconsisting of either time-out (Barrera & Sulzer-Azaroff, 1983; O'Brien & Azrin, 1972) or remediallearning trials (Nutter & Reid, 1978; Page, Iwata,& Neef, 1976; Richman, Reiss, Bauman, & Bailey,1984), and so on. Thus, in addition to producingpositive reinforcement in the form of experimenterpraise, correct responses also may function to avoidaversive social and physical stimulation and to ef-fectively reduce the duration of training sessions(this latter point is potentially significant, for it hasbeen shown that complex setting events or stimulussituations, and not just discrete stimuli, can functionas negative reinforcers [Krasnegor, Brady, & Find-ley, 1971], and that reduction of avoidance-sessiondurations can itself serve as negative reinforcement[Mellitz, Hineline, Whitehouse, & Laurence,1983]).

Behavioral Replacement StrategiesGiven that aversive stimuli are ubiquitous and

that escape is highly adaptive in their presence, itis usually the form, rather than the function, ofescape and avoidance behavior that presents a prob-lem. This raises the possibility of eliminating in-appropriate forms of escape by negatively reinforc-ing appropriate alternatives; in essence, replacingone behavior with another but not eliminating thefunction of the original. The concept is ratherstraightforward in principle and is analogous tolever switching by nonhuman subjects following achange in reinforcement schedule (e.g., De Villiers,1974).

Carr and Durand (1985) recently provided anexample of this strategy by teaching three children,who had tantrums when faced with difficult tasks,how to request help from the teacher. When a childexhibited the appropriate response ("I don't un-derstand"), brief escape was provided in the formof teacher assistance.

Another example is drawn from our work onfeeding disorders. While attempting to increase the

371

BRIAN A. IWATA

oral acceptance of food in four children, Riordan,Iwata, Finney, Wohl and Stanley (1984) foundthat one child, who was fed through a gastrostomytube, did not respond well to positive reinforcementbecause her baseline rate of acceptance was virtuallynonexistent. She also had resisted a number offorced feeding regimens; these practices were aver-sive and it appeared that her success in defeatingthem constituted negative reinforcement. Her treat-ment consisted of the following components: (a)the presentation of a redundant cue-"Take abite"- immediately followed by (b) the presen-tation of food on a spoon. If acceptance of fooddid not occur within 3 s, (c) her mouth was heldopen and the food was deposited. This procedurethus resembled very dosely a discriminated avoid-ance contingency in which one avoidance behavior(active food refusal) was replaced with another(opening the mouth and accepting food) by allow-ing it to prevent forced feeding.

Other ExamplesIn addition to the Kircher et al. (1971) study

on academic performance described earlier, one canfind instances in which negative reinforcement hasbeen used-in a highly intrusive manner-to in-crease appropriate social behaviors. Lovaas, Schaef-fer, and Simmons (1965) used escape and avoid-ance in training two autistic children to approachadults. Prior to treatment, the children frequendyengaged in stereotypic behavior and showed nosocial responsivity or appropriate play. Treatmentconsisted ofpresenting the instruction, "Come here,"followed by shock delivered through a floor grid.The shock was terminated when the child movedtoward the therapist (escape). Both children quicklylearned to approach the adult in response to theverbal instruction (avoidance). In defense of theiruse of electric-shock avoidance, the authors pre-sented data indicating that increases in approachbehavior were accompanied by more frequent dis-plays of affection and decreases in stereotypy andaggression.A less intrusive but similar procedure was used

by Fichter, Wallace, Liberman, and Davis (1976)in an attempt to improve the social skills of a

chronic and withdrawn schizophrenic male. Theresearchers targeted three aspects of his conversa-tional behavior: voice loudness, duration of verbalresponding, and keeping his hands on the armrestsof his chair while speaking. A therapist approachedthe client, called his name, and asked him to con-verse about one of several predetermined topics.During treatment, the staff member continuouslymonitored the target behaviors; if any failed to meetcriterion, the staff member would nag loudly oneor more of the following: "Longer!," "Louder!,"or "Put your hands on the armrests of the chair!,"and would continue to nag at 3-s intervals untilthe target behavior occurred. Although effective, itshould be noted that the contingencies used byFichter et al. would not be considered typical con-sequences for social interaction; in fact, it is entirelypossible that the appearance of the therapist wouldbecome discriminative for withdrawal. Data similarto those provided by Lovaas et al. (1965) showingthat social behavior (e.g., conversations initiated bythe subject or his response to approach by a ther-apist) improved outside of treatment sessions wouldhave been informative with respect to this question.One can only assume that there was no generalizedimprovement, based on a comment made by theauthors:

i.... his [the subject's) last interaction before... [being discharged from the)... unit wasto tell one staffmember how much he dislikedthe unit and the staff" (Fichter et al., 1976,pp. 384-385).

SUMMARY

In this artide I have attempted to point out anumber of ways in which negative reinforcementis relevant to behavioral development and its sub-sequent modification in the applied situation. Myreview has been a selective one in that I made noattempt to summarize the large and varied literatureon aversion, implosion, and desensitization thera-pies often used in the treatment of alcoholism,smoking, phobic reactions, sexual disorders, andrelated clinical problems. Still, the applied examples

372

NEGATIVE REINFORCEMENT

represent a thorough cross section of research pub-lished in JABA over the past 20 years, and anumber of general condusions and implications canbe drawn from the work described here.

Historically, applied analyses of behavior havefailed to acknowledge escape and avoidance as po-tentially common and powerful sources of rein-forcement. Evidence of this can be found in workon severe behavioral disorders such as self-injury,in which discussions of etiology have focused pri-marily on attentional factors rather than on thoserelated to escape; in research on contingencies suchas free time, in which free time as negative rein-forcement has not been a subject of analysis; instudies in which avoidance contingencies have beeninaccurately described as punishment; and in re-search on instructional processes, in which a varietyof avoidance contingencies have been used but notevaluated or even described as such.A second and more optimistic condusion sup-

portable by work described here is that an appliedtechnology of negative reinforcement is emerging.The work is somewhat scattered at present andlittle is known in some areas. Nevertheless, undereach of the topics induded in the present discus-sion-behavioral development, treatment of neg-atively reinforced behavior, and therapeutic uses ofnegative reinforcement-research activity has in-creased in recent years, and we are beginning to seeinvestigations of common procedures to the pointwhere categorization is both possible and useful.This evidence ofgrowing interest suggests that neg-ative reinforcement may be one of the most sig-nificant areas of applied research during the comingyears. Having made that prediction, what remainsis to offer some prompts to help ensure its accuracy.

The area of behavioral development is particu-larly problematic because applied researchers oftenare faced with situations in which the behavior ofinterest has a long, complex, and unknown history.In fact, the most important difference between lab-oratory- and field-based research, at least from abehavioral standpoint, is the lack of control overhistory that is characteristic of applied research. Insome cases (perhaps even most), behavioral historymay be irrelevant if a sufficiently powerful contin-

gency can be found. In other cases, however, im-plementation of treatment without consideration ofdevelopmental factors, or treatment selection basedon a consideration of topography alone, may pro-duce a number of unnecessary failures and subse-quently may limit our ability to determine the basisfor differential outcome.

Laboratory researchers solve problems related tohistory by controlling its course in a naive animal.Although applied researchers rarely can exercise thisoption, they can make a unique contribution bydeveloping methods for "unravelling" behavioralhistory, to the extent that it is possible. Our real-ization that behavioral development through neg-ative reinforcement can produce the same topog-raphy as that resulting from positive reinforcement,time and time again, may provide the impetus forcontinued refinement in the analysis of behavioralfunction. For example, a number of researchershave conduded that unitary accounts of severe be-havioral disorders in the developmentally disabledare unsatisfactory and have begun to establishmethodologies for identifying the functional prop-erties (one being negative reinforcement) of disor-ders such as pica (Mace & Knight, 1986), self-injury (Carr & Durand, 1985; Iwata et al., 1982),and stereotypy (Durand & Carr, 1987). Most re-cently, Bailey (1987) has proposed the term "be-havioral diagnostics" to describe a general strategyfor isolating the bases of problematic behavior.Continued work in this area and extension of therelevant methodologies to other human problemsare essential if we are to develop a mature tech-nology of behavior. In the meantime, researchersshould be encouraged (perhaps by editors) to seekout and indude more detail on subjects' behavioralhistories. In addition to the usual demographicsoffered (e.g., age, sex, grade or functioning level,etc.), which provide little information relevant toa behavioral analysis, it would be helpful to providesome account of factors related to behavioral de-velopment and maintenance. As evidence support-ing a negative reinforcement interpretation accu-mulates, we will be increasingly compelled toformalize our anecdotal observations and to confirmthese observations through manipulation.

373

374 BRIAN A. IWATA

Because applied research often is concerned witha problem as it actually exists, the treatment ofnegatively reinforced behavior will provide perhapsthe greatest opportunity for creative work. For ex-ample, research on the extinction of positively rein-forced behavior has induded variation and exten-sion (e.g., time-out, exclusion, sedusion, contingentobservation, time-out ribbon, movement suppres-sion), parametric analysis (e.g., duration, delay,schedule, changeover requirement), and compari-son (e.g., with differential reinforcement, responsecost, and punishment). None of these questionshave been addressed adequately in applied researchon the extinction of negatively reinforced behavior,and a similar situation exists with respect to dif-ferential reinforcement and punishment.

Research on the treatment of negatively rein-forced behavior will require consideration of issuesthat are different than those relevant to the treat-ment of positively reinforced behavior. These issueshave been noted previously, and some have beenthe focus of laboratory research for several years.Stimulus selection, schedules, and intensity, for ex-ample, may differentially affect the outcome of con-tingent aversive stimulation for ongoing avoidancebehavior. Therefore, it will be important for appliedresearchers to become acquainted with basic find-ings on negative reinforcement. As a result, we mayfind that methodologies and procedures developedin the laboratory can be extended to the appliedsituation so as to facilitate analysis and treatment.

Research on the use of negative reinforcementmay take several interesting directions. First, neg-ative reinforcement may provide an alternativemeans for establishing behavior when attempts touse positive reinforcement fail (e.g., as in the caseof Riordan et al., 1984, in which a child's operantlevel of eating was nonexistent). If so, we will wantto know the behaviors for which specific contin-gencies are useful and the conditions under whichthey should be applied. Second, it appears that theacquisition of adaptive behavior in our trainingprograms is at least partially a function of negativereinforcement. Future research must evaluate theroles of escape and avoidance within the trainingcontext so that (a) we will have a proper estimate

of the effectiveness ofcommonly used positive rein-forcers (the results of this estimate may indicatethat more potent reinforcers are needed), (b) wecan determine whether procedures such as remedialtrials, physical assistance, and so on, serve any usefilfunction and if that function is one of negativereinforcement, and (c) we can base future trainingsuccesses on the planned rather than the accidentaluse of negative reinforcement. A third promisingapplication involves further elaboration of behav-ioral replacement strategies. If we are willing toentertain the assumption that it is impossible toeliminate all sources of aversive stimulation, the useofsuch stimulation to alter the topography ofescapeand avoidance behavior, from an undesirable oneto a tolerable one, makes eminent sense from aclinical standpoint.A final cautionary note. Some of the applied

research induded in this review was selected spe-cifically to show that negative reinforcement canform the basis of highly intrusive intervention. Inat least one sense, negative reinforcement might beconsidered more intrusive than punishment be-cause, with negative reinforcement, presentation ofthe aversive stimulus is contingent on the absence,rather than the occurrence, of behavior. Therefore,as with punishment, we should conduct researchon negative reinforcement with great care and underthe appropriate conditions to determine how it mightbe used effectively and humanely, its limitations,and its proper role within the larger realm of cur-rently available treatment.

REFERENCES

Aaron, B. A., & Bostow, D. E. (1978). Indirect facilitationof on-task behavior produced by contingent free-time foracademic productivity. Journal of Applied BehaviorAnalysis, 11, 197.

Appel, J. B. (1960). Some schedules involving aversivecontrol. Journal of the Experimental Analysis of Be-havior, 3, 349-359.

Appel, J. B. (1963). Aversive aspects of a schedule ofpositive reinforcement. Journal of the ExperimentalAnalysis of Behavior, 6, 423-430.

Axeirod, S., & Apsche,J. (1983). The efects ofpunishmenton human behavior. New York: Academic Press.

Azrin, N. H. (1961). Time-out from positive reinforce-ment. Science, 133, 382-383.

NEGATIVE REINFORCEMENT 375

Azrin, N. H., & Foxx, R. M. (1971). A rapid method oftoilet training the institutionalized retarded. Journal ofApplied Behavior Analysis, 4, 89-99.

Azrin, N. H., & Powell,J. (1969). Behavioral engineering:The use of response priming to improve prescribed self-medication. Journal of Applied Behavior Analysis, 1,99-108.

Azrin, N., Rubin, H., O'Brien, F., Ayllon, T., & Roll, D.(1968). Behavioral engineering: Postural control by aportable operant apparatus.Journal ofApplied BehaviorAnalysis, 2, 39-42.

Baer, A. M., Rowbury, T., & Baer, D. M. (1973). Thedevelopment of instructional control over classroom ac-tivities of deviant preschool children.Journal ofAppliedBehavior Analysis, 6, 289-298.

Baer, D. M. (1978). On the relation between basic andapplied research. In A. C. Catania & T. A. Brigham(Eds.), Handbook of applied behavior analysis: Socialand instructional processes (pp. 11-17). New York:Irvington.

Baer, D. M. (1981). A flight of behavior analysis. TheBehavior Analyst, 4, 85-9 1.

Baer, D. M., Wolf, M. M., & Risley, T. R. (1968). Somecurrent dimensions of applied behavior analysis. Journalof Applied Behavior Analysis, 1, 91-97.

Bailey, J. S. (1987, May). Behavioral diagnostics: Newtools for applied behavior analysis. Presented as an in-vited address at the Association of Behavior AnalysisConvention, Nashville, TN.

Bankart, B., & Elliott, R. (1974). Extinction of avoidancein rats: Response availability and stimulus presentationeffects. Behaviour Research and Therapy, 12, 53-56.

Barrera, R. D., & Sulzer-Azaroff, B. (1983). An alternatingtreatment comparison of oral and total communicationtraining programs with echolalic autistic children. Jour-nal of Applied Behavior Analysis, 16, 379-394.

Barrish, H. H., Saunders, M., & Wolf, M. M. (1969).Good behavior game: Effects of individual contingenciesfor group consequences on disruptive behavior in a class-room. Journal of Applied Behavior Analysis, 2, 119-124.

Barton, E. S., & Osbome,J. G. (1978). The developmentof classroom sharing by a teacher using positive practice.Behavior Modification, 2, 231-250.

Bolles, R. G. (1970). Species-specific defense reactions andavoidance learning. Psychological Review, 77, 32-48.

Bolles, R. G. (1971). Species-specific defense reactions. InF. R. Brush (Ed.), Aversive conditioning and learning(pp. 183-233). New York: Academic Press.

Boren,J. J., & Sidman, M. (1957). A discrimination basedon repeated conditioning and extinction of avoidancebehavior. Journal of Comparative and PhysiologicalPsychology, 50, 18-22.

Borreson, P. M. (1980). The elimination of a self-injuriousavoidance response through a forced running conse-quence. Mental Retardation, 18, 73-77.

Carey, R. G., & Bucher, B. (1983). Positive practice over-correction: The effects of duration of positive practice onacquisition and response reduction. Journal of AppliedBehavior Analysis, 16, 101-109.

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

Carr, E. G., Newsom, C. D., & Binkoff, J. A. (1976).Stimulus control of self-destructive behavior in a psy-chotic child. Journal ofAbnormal Child Psychology, 4,139-153.

Carr, E. G., Newsom, C. D., & Binkoff, J. A. (1980).Escape as a factor in the aggression of two retardedchildren. Journal of Applied Behavior Analysis, 13,101-117.

Catania, A. C. (1973). The nature of learning. In J. A.Nevin & G. S. Reynolds (Eds.), The study of behavior:Learning, motivation, emotion, and instinct (pp. 31-68). Glenview, IL: Scott, Foresman.

Coulson, G., Coulson, V., & Gardner, L. (1970). The effectof two extinction procedures after acquisition on a Sid-man avoidance contingency. Psychonomic Science, 18,309-310.

Cullen, C. (1981). The flight to the laboratory. The Be-havior Analyst, 4, 81-83.

Davenport, D. G., Coger, R. W., & Spector, 0. J. (1970).The redefinition of extinction applied to Sidman free-operant avoidance responding. Psychonomic Science, 19,181-182.

Davis, H. (1979). Behavioral anomolies in aversive situ-ations. In J. D. Keehn (Ed.), Psychopathology in ani-mals: Research and clinical implications (pp. 197-222). New York: Academic Press.

Deitz, S. M. (1978). Current status of applied behavioranlaysis: Science versus technology. American Psychol-ogist, 33, 805-814.

De Villiers, P. A. (1974). The law of effect and avoidance:A quantitative relationship between response rate andshock-frequency reduction. Journal ofthe ExperimentalAnalysis of Behavior, 21, 223-235.

Dunlap, G., & Johnson, J. (1985). Increasing the inde-pendent responding of autistic children with unpredict-able supervision.Journal ofApplied Behavior Analysis,18, 227-236.

Durand, V. M., & Carr, E. G. (1987). Social influenceson "self-stimulatory" behavior. Journal of Applied Be-havior Analysis, 20, 119-132.

Ferrari, E. A., Todorov, J. C., & Graeff, F. G. (1973).Nondiscriminated avoidance ofshock by pigeons peckinga key. Journal of the Experimental Analysis of Behav-ior, 19, 211-218.

Fichter, M. M., Wallace, C. J., Liberman, R. P., & Davis,J. R. (1976). Improving social interaction in the chron-ic psychotic using discriminated avoidance ("nagging"):Experimental analysis and generalization.Journal ofAp-plied Behavior Analysis, 9, 367-386.

Foree, D., & LoLordo, V. (1970). Signalled and unsignalledfree-operant avoidance in the pigeon. Journal of theExperimental Analysis of Behavior, 13, 283-290.

Fowler, H. (1971). Suppression and facilitation by responsecontingent shock. In R. F. Brush (Ed.), Aversive con-ditioning and learning (pp. 537-604). New York: Ac-ademic Press.

Foxx, R. M. (1976). Increasing a mildly retarded woman's

376 BRIAN A. IWATA

attendance at self-help dasses by overcorrection and in-struction. Behavior Therapy, 6, 390-396.

Foxx, R. M. (1977). Attention training: The use of over-correction avoidance to increase the eye contact of autisticand retarded children.Journal ofApplied Behavior An-laysis, 10, 488-499.

Foxx, R. M., & Bechtel, D. R. (1983). Overcorrection: Areview and analysis. In S. Axelrod & J. Apsche (Eds.),The efects ofpunishment on human behavior (pp. 133-220). New York: Academic Press.

Foxx, R. M., &Jones,J. R. (1978). A remediation programfor increasing spelling achievement of elementary andjunior high school students. Behavior Modification, 2,211-230.

Greene, R. R., & Hoats, D. L. (1969). Reinforcing ca-pabilities of television distortion. Journal of AppliedBehavior Analysis, 2, 139-141.

Hansen, G. D. (1979). Enuresis control through fading,escape, and avoidance training. Journal of Applied Be-havior Analysis, 12, 303-307.

Haring, T. G. (1985). Teaching between-dass generaliza-tion of toy play behavior to handicapped children. Jour-nal of Applied Behavior Analysis, 18, 127-139.

Harris, V. W., & Sherman,J. A. (1973). Use and analysisof the "good behavior game" to reduce disruptive dass-room behavior. Journal of Applied Behavior Analysis,6, 405-417.

Harris, V. W., & Sherman, J. A. (1974). Homeworkassignments, consequences, and classroom performancein social studies and mathematics. Journal of AppliedBehavior Analysis, 7, 505-519.

Heidorn, S. D., & Jensen, C. C. (1984). Generalizationand maintenance of the reduction of self-injurious be-havior maintained by two types of reinforcement. Be-havior Research and Therapy, 22, 581-586.

Herrnstein, R. J. (1961). Relative and absolute strengthof response as a function of frequency of reinforcement.Journal of the Experimental Analysis of Behavior, 4,267-272.

Herrnstein, R. J. (1969). Method and theory in the studyof avoidance. Psychological Review, 76, 49-69.

Hineline, P. N. (1977). Negative reinforcement and avoid-ance. In W. K. Honig & J. E. R. Staddon (Eds.), Hand-book of operant behavior (pp. 364-414). EnglewoodCliffs, NJ: Prentice-Hall.

Hineline, P. N. (1981). Several roles of stimuli in negativereinforcement. In P. Harzem & M. D. Zeiler (Eds.),Advances in analysis of behavior: Vol. 2. Predict-ability, correlation, and contiguity (pp. 203-246).Chichester, England: Wiley.

Hineline, P. N. (1984). Aversive control: A separate do-main?Journal of the Experimental Analysis of Behav-ior, 42, 495-509.

Hoffman, H. S. (1966). The analysis of discriminatedavoidance. In W. K. Honig (Ed.), Operant behavior:Areas ofresearch and application (pp. 499-530). NewYork: Appleton.

Honig, W. K., & Staddon,J. E. R. (Eds.). (1977). Hand-book ofoperant behavior. Englewood Cliffs, NJ: Prentice-Hall.

Hutchinson, R. R. (1977). By-products of aversive control.In W. K. Honig & J. E. R. Staddon (Eds.), Handbookof operant behavior (pp. 415-431). Englewood Cliffs,NJ: Prentice-Hall.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E.,& Richman, G. S. (1982). Toward a functional analysisof self-injury. Analysis and Intervention in Develop-mental Disabilities, 3, 1-20.

Jones, M. C. (1924). Elimination of children's fears. Jour-nal of Experimental Psychology, 7, 382-390.

Kelleher, R. T., Riddle, W. C., & Cook, L. (1963). Per-sistent behavior maintained by unavoidable shocks.Jour-nal ofthe Experimental Analysis of Behavior, 6, 507-517.

Kelleher, R. T., & Morse, W. H. (1968). Schedules usingnoxious stimuli, III: Responding maintained with re-sponse-produced electric shocks. Journal of the Exper-imental Analysis of Behavior, 11, 819-838.

Kelley, M. L., Jarvie, G. J., Middlebrook, J. L., McNeer,M. F., & Drabman, R. S. (1984). Decreasing burnedchildren's pain behavior: Impacting the trauma of hy-drotherapy. Journal ofApplied Behavior Analysis, 17,147-158.

Kircher, A. S., Pear,J.J., & Martin, G. L. (1971). Shockas punishment in a picture-naming task with retardedchildren.Journal ofApplied Behavior Analysis, 4, 227-233.

Krasnegor, N. A., Brady, J. V., & Findley, J. D. (1971).Second-order optional avoidance as a function of fixed-ratio requirements. Journal of the Experimental Anal-ysis of Behavior, 15, 181-187.

Levis, D. J. (1979). The infrahuman avoidance model ofsymptom maintenance and implosive therapy. In J. D.Keehn (Ed.), Psychopathology in animals: Research andclinical implications (pp. 257-277). New York: Aca-demic Press.

Long, J. D., & Williams, R. W. (1973). The comparativeeffectiveness ofgroup and individual contingent free timewith inner-city junior high school students. Journal ofApplied Behavior Analysis, 6, 465-474.

Lovaas, 0. I., Schaeffer, B., & Simmons, J. Q. (1965).Building social behavior in autistic children by use ofelectric shock.Journal ofExperimental Research in Per-sonality, 1, 99-109.

Luyben, P. D., Funk, D. M., Morgan, J. K., Clark, K. A.,& Delulio, D. W. (1986). Team sports for the retarded:Training a side-of-the-foot soccer pass using a maximum-to-minimum prompt reduction strategy. Journal ofAp-plied Behavior Analysis, 19, 431-436.

Mace, F. C., & Knight, D. (1986). Functional analysis andtreatment of severe pica. Journal of Applied BehaviorAnalysis, 19, 411-416.

MacPhail, E. M. (1968). Avoidance responding in pigeons.Journal of the Experimental Analysis of Behavior, 11,629-632.

Maloney, K. B., & Hopkins, B. L. (1973). The modifi-cation of sentence structure and its relationship to sub-jective judgments of creativity in writing. Journal ofApplied Behavior Analysis, 6, 425-433.

Matson, J. L., & DiLorenzo, T. M. (1984). Punishment

NEGATIVE REINFORCEMENT 377

and its alternatives: A new perspective for behaviormodification. New York: Springer.

Matson,J. L., Esveldt-Dawson, K., & O'Connell, D. (1979).Overcorrection, modeling, and reinforcement proceduresfor reinstating speech in a mute boy. Child BehaviorTherapy, 1, 363-371.

McKearney, J. W. (1972). Maintenance and suppressionof responding under schedules of electric shock presen-tation. Journal of the Experimental Analysis of Be-havior, 17, 425-432.

Medland, M. B., & Stachnik, T. J. (1973). Good behaviorgame: A replication and systematic analysis. Journal ofApplied Behavior Analysis, 6, 45-51.

Mellitz, M., Hineline, P. N., Whitehouse, W. G., & Lau-rence, M. T. (1983). Duration-reduction of avoidancesessions as negative reinforcement. Journal of the Ex-perimental Analysis of Behavior, 40, 57-67.

Michael, J. (1975). Positive and negative reinforcement:A distinction that is no longer necessary; Or a better wayto talk about bad things. Behaviorism, 3, 33-44.

Michael, J. (1980). Flight from behavior analysis. TheBehavior Analyst, 3, 1-24.

Michael, J. (1982). Distinguishing between discriminativeand motivational functions of stimuli. Journal of theExperimental Analysis of Behavior, 37, 149-155.

Morse, W. H., & Kelleher, R. T. (1970). Schedules asfundamental determinants ofbehavior. In W. N. Schoen-feld (Ed.), The theory of reinforcement schedules (pp.139-185). New York: Appleton.

Morse, W. H., & Kelleher, R. T. (1977). Determinantsof reinforcement and punishment. In W. K. Honig &J. E. R. Staddon (Eds.), Handbook of operant behavior(pp. 174-200). Englewood Cliffs, NJ: Prentice-Hall.

MyerJ. S. (1971). Someeffectsofnoncontingentaversivestimulation. In R. F. Brush (Ed.), Aversive conditioningand learning (pp. 469-536). New York: Academic Press.

Nutter, D., & Reid, D. H. (1978). Teaching retardedwomen a dothing selection skill using community norms.Journal of Applied Behavior Analysis, 11, 475-487.

O'Brien, F., & Azrin, N. H. (1972). Developing propermealtime behaviors of the institutionalized retarded.Journal of Applied Behavior Analysis, 5, 389-399.

O'Brien, S., Ross, L. V., & Christophersen, E. R. (1986).Primary encopresis: Evaluation and treatment. Journalof Applied Behavior Analysis, 19, 137-145.

Ollendick, T. H., Matson, J. L., Esveldt-Dawson, K., &Shapiro, E. S. (1980). Increasing spelling achievement:An analysis of treatment procedures utilizing an alter-nating treatments design. Journal of Applied BehaviorAnalysis, 13, 645-654.

Osborne, J. G. (1969). Free-time as a reinforcer in themanagement of classroom behavior. Journal of AppliedBehavior Analysis, 2, 113-118.

Page, T. J., Iwata, B. A., & Neef, N. A. (1976). Teachingpedestrian skills to retarded persons: Generalization fromthe dassroom to the natural environment. Journal ofApplied Behavior Analysis, 9, 433-444.

Pierce, W. D., & Epling, W. F. (1980). What happenedto analysis in applied behavior analysis? The BehaviorAnalyst, 3, 1-10.

Plummer, S., Baer, D. M., & LeBlanc,J. M. (1977). Func-tional considerations in the use oftimeout and an effectivealternative. Journal of Applied Behavior Analysis, 10,689-705.

Powell, R. W., & Peck, S. (1969). Persistent shock-elicitedresponding engendered by a negative reinforcement pro-cedure. Journal of the Experimental Analysis of Be-havior, 12, 1049-1062.

Rachlin, H., & Hineline, P. N. (1967). Training andmaintenance of keypecking in the pigeon using negativereinforcement. Science, 157, 954-955.

Richman, G. S., Reiss, M. L., Bauman, K. E., & Bailey, J.S. (1984). Teaching menstrual care to mentally re-tarded women: Acquisition, generalization, and main-tenance. Journal of Applied Behavior Analysis, 17,441-451.

Rincover, A., & Newsom, C. D. (1985). The relativemotivational properties of sensory and edible reinforcersin teaching autistic children. Journal of Applied Be-havior Analysis, 18, 237-248.

Riordan, M. M., Iwata, B. A., Finney, J. W., Wohl, M. K.,& Stanley, A. E. (1984). Behavioral assessment andtreatment of chronic food refusal in handicapped chil-dren. Journal ofApplied Behavior Analysis, 17, 327-341.

Rodgers, T. A., & Iwata, B. A. (1987, September). Analysisof error correction procedures during behavioral acqui-sition. In J. S. Bailey (Chair), Training research inmental retardation. Symposium presented at the FloridaAssociation for Behavior Analysis Convention, Sarasota.

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.

Sandier, J., & Davidson, R. S. (1973). Psychopathology:Learning theory, research and applications. New York:Harper & Row.

Sandier, J., Davidson, R. S., Greene, W. E., & Holzschuh,R. D. (1966). Effects of punishment intensity on in-strumental avoidance behavior. Journal of Comparativeand Physiological Psychology, 61, 212-216.

Sandier, J., Davidson, R. S., & Malagodi, E. F. (1966).Durable maintenance of behavior during concurrentavoidance and punished-extinction conditions. Psycho-nomic Science, 6, 105-106.

Schiff, R., Smith, N., & Prochaska, J. (1972). Extinctionof avoidance in rats as a function of duration and numberof blocked trials. Journal of Comparative and Physio-logical Psychology, 81, 356-359.

Schoenfeld, W. N. (1969). "Avoidance" in behavioraltheory. Journal of the Experimental Analysis of Be-havior, 12, 669-674.

Schreibman, L. (1975). Effects of within-stimulus and ex-tra-stimulus prompting on discrimination learning in au-tistic children. Journal of Applied Behavior Analysis,8, 91-112.

Shnidman, S. R. (1968). Extinction of Sidman avoidancebehavior. Journal of the Experimental Analysis of Be-havior, 11, 153-156.

Sidman, M. (1966). Avoidance behavior. In W. K. Honig

378 BRIAN A. IWATA

(Ed.), Operant behavior: Areas of research and appli-cation (pp. 448-498). New York: Appleton.

Sidman, M., Herrnstein, R. J., & Conrad, D. G. (1957).Maintenance ofavoidance behavior by unavoidable shocks.Journal of Comparative and Physiological Psychology,50, 553-557.

Smith, R., & Keller, F. (1970). Free-operant avoidance inthe pigeon using a treadle response. Journal of the Ex-perimental Analysis of Behavior, 13, 211-214.

Solnick, J. V., Rincover, A., & Peterson, C. R. (1977).Some determinants of the reinforcing and punishingproperties of timeout. Journal of Applied BehaviorAnalysis, 10, 410-424.

Sprague, J. R., & Horner, R. H. (1984). The effects ofsingle instance, multiple instance, and general case train-

ing on generalized vending machine use by moderatelyand severely handicapped students. Journal of AppliedBehavior Analysis, 17, 273-278.

Thompson, D. M. (1964). Escape from SD associated withfixed-ratio reinforcement. Journal of the ExperimentalAnalysis of Behavior, 7, 1-8.

Weeks, M., & Gaylord-Ross, R. (1981). Task difficultyand aberrant behavior in severely handicapped students.Journal of Applied Behavior Analysis, 14, 449-463.

Received August 25, 1987Initial editorial decision August 30, 1987Revision received September 14, 1987Final acceptance September 15, 1987Action Editor, Jon S. Bailey