JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2000, 33, 411–418 NUMBER 4 (WINTER 2000)

CURRENT RESEARCH ON THE INFLUENCE OF ESTABLISHING OPERATIONS ON BEHAVIOR IN

APPLIED SETTINGS

BRIAN A. IWATA THE UNIVERSITY OF FLORIDA

RICHARD G. SMITH UNIVERSITY OF NORTH TEXAS

AND

JACK MICHAEL WESTERN MICHIGAN UNIVERSITY

This article provides commentary on research published in the special section of the Journal of Applied Behavior Analysis devoted to establishing operations (EOs). Three major themes are highlighted: (a) identification of the influence of EOs on behavior in applied settings, (b) the use of EO manipulation as an assessment tool, and (c) the development of interventions based on the alteration of EO influences. Methodological issues pertain- ing to research on EOs are addressed, and suggestions for future investigation are pro- vided.

DESCRIPTORS: establishing operations, antecedent events

Applied researchers have long recognized the fact that antecedent events other than discriminative stimuli can influence the oc- currence of operant behavior; however, the exact nature of that influence and what to call it have been ambiguous throughout much of the field’s history. As a result, in most of the early research published in the Journal of Applied Behavior Analysis (JABA), authors referred to antecedent conditions in general terms such as setting events or con- textual variables, emphasized the procedural aspects of antecedent manipulations (e.g., as in deprivation, satiation, or reinforcer sam- pling), and often attributed observed chang- es in behavior to the process of stimulus control. A conceptual foundation for orga- nizing research on antecedent variables was

We thank Eileen Roscoe and April Worsdell for compiling the citation data presented here.

Correspondence should be addressed to Brian A. Iwata, Psychology Department, The University of Florida, Gainesville, Florida 32611.

provided by Michael (1982), who presented a much-needed distinction between the dis- criminative and motivational properties of antecedent events and proposed the term es- tablishing operation (EO) as a functional de- scription for events of the latter type. Ap- plied researchers gradually came to recognize practical implications of the EO and, since the late 1980s, have found the concept of the EO increasingly helpful in describing, studying, and arranging antecedent influenc- es over behavior, as evidenced by the grow- ing number of citations to Michael’s (1982, 1993) articles and the use of the term in JABA (see Figure 1).

Current applied research on EOs tends to fall into three broad categories: (a) general demonstrations of the influence of an EO on behavior, (b) the use of EO manipula- tions to clarify results of behavioral assess- ments, and (c) attempts to improve (increase or decrease) behavior by incorporating EO manipulations as treatment components. Ar-

412 BRIAN A. IWATA et al.

Figure 1. Cumulative number of articles published in JABA in which the term EO has been used or in which citation to Michael’s (1982, 1993) articles has appeared.

ticles in this special issue illustrate all three types of research, although the boundaries between these categories are sometimes blurred because discovery of an EO effect often has immediate implications for assess- ment or treatment.

Identification of EO Influences Two studies provide basic demonstrations

of EO influences. Klatt, Sherman, and Shel- don present data showing that, when access to reinforcing activities is restricted, persons with developmental disabilities subsequently engage in those activities for longer periods of time. These results are helpful in suggest- ing ways to schedule leisure and educational activities so as to increase participation. A somewhat darker implication is that, when the occurrence of problem behavior is re- stricted, it too may increase subsequently. For example, if protective equipment is used

to reduce stereotypic self-injury, individuals may engage in the behavior more often when the protective equipment is removed. Thus, extension of the Klatt et al. method to problem behavior may reveal a limitation of interventions that reduce behavior tem- porarily but do not alter EOs permanently or disrupt maintaining contingencies.

Another interesting aspect of the Klatt et al. study is that the dependent variable was consumption of a reinforcer rather than the rate of a response that produced that rein- forcer. An analogous basic experiment might measure the effects of food deprivation on food consumption rather than on food- maintained bar pressing. This raises a con- ceptual question about the characteristic of the EO reflected in the Klatt et al. data. The evocative effect of the EO was not shown, which would require a change in the fre- quency of behavior that historically had been

413RESEARCH ON ESTABLISHING OPERATIONS

followed by the reinforcer, because, in this case, the activities were the reinforcers. It could be argued that changes in reinforcer consumption reflected a change in reinforcer effectiveness. However, a test of reinforcer effectiveness—the extent to which behavior that has been followed by a reinforcer in- creases—was not conducted. It may be that increased consumption is specific to partic- ular types of reinforcers rather than a general property of EOs. For example, although the sight of a slotted screw can serve as an EO for asking for a screwdriver, it is unlikely that receiving two screwdrivers would be more reinforcing than receiving one screw- driver.

Friman presents a demonstration that ‘‘transitional objects’’ (i.e., inanimate objects for which a strong preference is shown by young children) may occasion behaviors in a way that suggests the influence of an EO. Friman observed a child’s thumb sucking when social stimulation was available and unavailable and when the child’s favored ob- ject (a cloth) was available and unavailable (consequences for thumb sucking were held constant across conditions). Results show that thumb sucking occurred only in the presence of the cloth, suggesting that thumb sucking was neither maintained by social re- inforcement nor occasioned by discrimina- tive properties of the cloth. Instead, the mere presence of the cloth seemed to function as an EO for thumb sucking. Friman correctly notes that the process by which the presence of the cloth altered the reinforcing effects of thumb sucking was unknown. Because the relationship does not appear to be an un- learned one, it seems plausible to attribute its influence to the effects of a conditioned EO. In addition to expanding the scope of analysis and interpretation to areas more common in traditional child development, Friman hints at an interesting possibility for research by extending the concept of the EO to account for complementary reinforcer re-

lationships (i.e., situations in which con- sumption of one reinforcer is correlated with increased consumption of a different rein- forcer). For example, it is likely that access to certain reinforcers (e.g., television) creates a condition (inactivity) that increases the ef- fectiveness of some reinforcers (consumption of snacks) but not others (exercise, which also could be performed while watching the television).

EO Manipulations During Assessment

As Michael (2000) noted, the EO has be- come an important focus in research on be- havioral assessment, especially in functional analysis methodologies, which require care- ful consideration of and control over ante- cedent and consequent events. Four of the articles in this issue involve identification of EO influences within the context of func- tional analyses. Berg et al. present data in- dicating that the sequence of assessment conditions may create a situation in which exposure to an EO in one condition influ- ences behavior in a subsequent condition. The authors illustrate this effect with atten- tion as the reinforcer. Three participants were exposed to different test conditions in which attention was (a) delivered contingent on problem behavior, (b) withheld, or (c) available as one alternative in a concurrent- choice arrangement. Behaviors of interest were observed to occur at high levels during these conditions when they were immediate- ly preceded by a condition in which atten- tion was generally unavailable but not when they were preceded by a condition in which attention was delivered frequently. These re- sults are important because they indicate that the reinforcing effects of consequences in a given condition may be strengthened (established) or weakened (abolished), re- spectively, by presession deprivation or ex- posure. As noted by the authors, their results are similar to those reported by Vollmer and Iwata (1991), but the implication of the

414 BRIAN A. IWATA et al.

Berg et al. data is timely because the context in which their demonstration was conducted may exemplify certain types of functional analysis arrangements. It is important to note, however, that assessment conditions were very brief (5 min) and that the pre- session and test conditions were run back to back with no time separating the two. Thus, it is unclear if similar results would be ob- tained under other arrangements (i.e., longer sessions, breaks between sessions).

Worsdell, Iwata, Conners, Kahng, and Thompson also examine general EO influ- ences during functional analyses but focus on the effects of within-session, rather than between-sessions, manipulations. They ex- posed 6 individuals whose self-injurious be- havior (SIB) was maintained by social-posi- tive reinforcement to functional analysis conditions in which the EO and reinforce- ment contingency were present and absent in various combinations. High rates of SIB were observed consistently (across all 6 par- ticipants) only when the EO and its relevant contingency were both present. None of the participants engaged in high rates of SIB when the contingency was absent, and in- consistent results were obtained when the EO was absent. These results underscore the importance of including contingency manip- ulations in tests of behavioral function; they also indicate that consistent performance is most likely to be observed when antecedent and consequent events are manipulated in concert. The authors are generally critical of assessment models in which EO influences are examined under extinction, although as Michael (2000) noted and as Berg et al. ob- served with 1 participant, EO exposures may evoke the behavior of interest regardless of the availability of reinforcement, at least temporarily.

McComas, Hoch, Paone, and El-Roy show that the conditions that establish es- cape as negative reinforcement may be high- ly idiosyncratic. Their results extend a grow-

ing body of research on a wide range of po- tential EOs for escape and highlight the util- ity of systematic EO (demand) assessments within functional analyses of problem be- havior. The study also extends previous work in the area of treatment by showing positive effects with EO interventions that did not include extinction. The results of Charlie’s analysis are particularly interesting. Whereas interventions for the other 2 participants in- volved the alteration of response require- ments, Charlie was permitted to decide the order in which tasks would be performed. This was sufficient to decrease his escape be- havior despite the fact that he always per- formed the same tasks in the same way. Thus, although task demands appeared to function as reflexive conditioned EOs for Eli and Ben because the demands reliably pre- ceded aversive task requirements, a different and unknown process resulted in Charlie’s sensitivity to predetermined task sequences as EOs for escape.

O’Reilly, Lacey, and Lancioni contribute another report in an interesting series of demonstrations of EO influences on prob- lem behavior by showing that background noise exacerbated the escape-maintained problem behavior of a child with Williams syndrome. One of the characteristics of Wil- liams syndrome, hyperacusis, or hypersensi- tivity to sound, was evidenced by the occur- rence apparent pain-attenuating behaviors (e.g., placing hands over the ears) across as- sessment conditions. Problem behaviors, on the other hand, occurred primarily during demand conditions and increased dramati- cally in the presence of background noise. These results suggest that noise had multiple influences: It evoked pain-related behavior (probably as an EO) and also altered the ef- fects of demands as an EO. This second ef- fect suggests a relationship in which noise apparently strengthened further an existing EO, in that the effects of demands as an EO for escape behavior were enhanced in the

415RESEARCH ON ESTABLISHING OPERATIONS

presence of noise. Subsequent work in this area might clarify the nature of relationships between independent EOs. In the present study, one EO strengthened another and evoked higher rates of problem behavior. It would be interesting to determine the extent to which the opposite arrangement might produce therapeutic benefit. That is, would exposure to an EO for appropriate behavior cancel out the effects of an existing EO for problem behavior?

Treatment of Problem Behavior with EO Manipulations

The final two studies in this special sec- tion illustrate attempts to determine, within the context of treatment, whether observed changes in behavior can be attributed to an EO manipulation. The studies by Kahng, Iwata, Thompson, and Hanley and by Ha- gopian, Crockett, van Stone, DeLeon, and Bowman are related in another way, in that the intervention of interest was noncontin- gent reinforcement (NCR), whose therapeu- tic effects have been attributed to both sa- tiation (elimination of the EO for problem behavior via frequent reinforcer delivery) and extinction (termination of the contin- gency between the occurrence of problem behavior and the delivery of a reinforcer).

After determining (via functional analysis) that 3 participants’ problem behaviors were maintained by social-positive reinforcement, Kahng et al. delivered the maintaining re- inforcers under rich fixed-time schedules (NCR) but not as consequences for problem behavior (extinction), and observed rapid re- ductions in problem behavior. The authors continued to collect data during extinction periods (no reinforcement available) follow- ing each NCR session in an attempt to iden- tify the functional properties of NCR (see the article for details on experimental logic). One participant showed no increase in prob- lem behavior at the end of NCR sessions, suggesting that the behavior had been extin-

guished during those sessions. The other 2 participants, however, showed increases in problem behavior at the end of NCR ses- sions, suggesting that the transition from the availability (NCR) to the unavailability (ex- tinction) of reinforcement reinstated an in- fluential EO and, as a result, evoked prob- lem behavior. Other interesting effects (an apparent transition from satiation to extinc- tion and failure to obtain extinction even under thin NCR schedules) are reflected in the data for these 2 participants. Kahng et al. pose questions about data interpretation in their discussion that illustrate difficulties in attempting to identify processes by which NCR affects behavior and suggest that mul- tiple influences related to both satiation and extinction might also account for behavior change under differential-reinforcement-of- other-behavior schedules.

Hagopian et al. present a more direct ap- proach than that taken by Kahng et al. They implemented NCR without extinction (i.e., maintaining reinforcers were delivered under a fixed-time schedule and also were contin- gent on the occurrence of problem behavior) and observed immediate reductions in the problem behavior of 3 participants. These results suggest that behavior change was the result of satiation; however, no change was observed in reinforcer consumption, raising the question of whether satiation had, in fact, occurred. Hagopian et al. later suggest the possibility that NCR may decrease be- havior by altering its EO yet may not pro- duce some of the other behavioral effects as- sociated with satiation. This seems quite plausible and consistent with Michael’s (2000) comment that, whereas deprivation can be easily operationalized through refer- ence to duration of restricted access, satia- tion cannot, due to its association with other behavioral effects. The Hagopian et al. data are thus helpful in illustrating the EO-alter- ing effects of NCR in the apparent absence of satiation. The authors provide additional

416 BRIAN A. IWATA et al.

data related to the consideration of extinc- tion effects.

Some Suggestions for Future Research

We have attempted to interject questions for consideration in future research within our discussion of each of the articles that appear in this special issue. We fear that a more exhaustive attempt would be incom- plete, given the extensive treatment of meth- odological and conceptual issues found in the reviews by McGill (1999) and by Smith and Iwata (1997). Instead, we close this in- troduction by suggesting several general themes for research that are raised through consideration of the articles as a group.

A methodological question arises in con- ducting research on EOs: Should their evoc- ative effects be examined in the presence or absence of reinforcement? Although, as not- ed previously, the presence of an EO should evoke behavior regardless of the availability of reinforcement, EO evaluations under con- ditions of extinction may prevent detection of an EO influence if the unavailability of reinforcement is readily discriminated or if extinction occurs quickly (as has been ob- served in a number of studies). The absence of reinforcement also raises the question of whether the behavior of interest is, in fact, maintained by the putative reinforcer. For example, being left alone for a period of time may evoke behavior that has been main- tained by nonsocial (automatic) as well as social reinforcement. Although the evocative effect on behavior of being left alone is un- ambiguous in this case, identification of the behavior’s maintaining reinforcer requires a further demonstration: differential control over the behavior in the presence of the EO. For these reasons, we believe that EO inves- tigations are best undertaken in the presence of relevant reinforcement contingencies.

A second general issue related to the iden- tification of EO influences is specification of the historical basis of the EO, to the extent

that it is possible. Given certain methodo- logical controls (e.g., manipulation of ante- cedent and consequent events, ruling out discrimination as a source of influence), it is possible to attribute changes in behavior to the presence and absence of an EO with a relatively high degree of confidence. A more difficult yet helpful step would involve ad- ditional analysis to determine how the EO acquired and maintains its reinforcer-estab- lishing and evocative properties. Was the EO unconditioned or conditioned? If condi- tioned, is its influence due to pairing with another EO or some other consequence? Does the influence appear to be reflexive or transitive? Answers to questions such as these would give applied researchers additional options for intervention as well as some helpful insights for prevention.

Antecedent manipulations that influence behavior by way of an EO are fairly straight- forward when they involve the mere presen- tation or removal of an event whose struc- tural features are invariant. However, be- cause most EOs currently being studied in applied research have multidimensional characteristics (quality, magnitude, duration, rate, etc.), attempts to identify the influence of EOs might benefit from both qualitative and quantitative analyses. Social interaction provides a good example, and parallels can be drawn with other types of events. A con- clusion that deprivation from social inter- action does or does not function as an EO requires qualification: What kind of social interaction? With whom? For what dura- tion? Researchers would do well to remem- ber that identification of an apparent EO influence (or failure to find such an influ- ence) may be a function of particular fea- tures of procedural implementation with re- spect to the class of variables being manip- ulated. Similarly, procedures that diminish the effectiveness of a positive reinforcer may involve reinforcer removal rather than pre- sentation. Consider the reinforcing effects of

417RESEARCH ON ESTABLISHING OPERATIONS

parental attention for a child who has just been put to bed. Is the child less likely to cry for parental attention after receiving a lot of it just prior to bedtime? Or would a pe- riod of diminished attention be more likely to attenuate its reinforcing effects, perhaps by strengthening the reinforcing effects of an alternative behavior (e.g., lying quietly, shut- ting one’s eyes, etc.)?

Significant refinements in the functional analysis of behavior disorders have already been realized through incorporation of EO manipulations, which improve the accuracy or efficiency of assessment by increasing the likelihood that a target behavior of interest will occur in the presence of its maintaining reinforcer. However, technical details are too scattered, perhaps among dozens of reports, to be of much immediate use to those in- terested in clinical implementation. In ad- dition, some EOs that influence behavior on either a within- or between-session basis are just now being investigated, as evidenced by articles in this issue. Thus, research in the area of functional analysis might benefit fur- ther from a series of simple demonstrations of how to maximize or minimize the influ- ence of EOs during the course of assessment, followed by integration of this information with what is already known about maximiz- ing contingency strength.

Research on the treatment of behavior disorders also has benefited from systematic EO manipulations. For example, the rapidly growing number of studies incorporating NCR as a treatment is directly related to conceptual and empirical work on the EO, and the same might be said more generally for the larger class of antecedent interven- tions. As an aside, the Hagopian et al. and Kahng et al. studies on NCR in this issue raise interesting questions with respect to terminology. As already noted, it is unclear whether NCR, as a result of diminishing the effectiveness of reinforcement for problem behavior, produces a state of satiation; if not,

it may be better to merely refer to such an effect as an EO or ‘‘abolishing’’ operation. The other effect of NCR—discontinuation of the response–reinforcer relation—is also subject to terminological ambiguity. For purposes of improving procedural specificity, Michael (2000) prefers to distinguish be- tween the operations of (a) discontinuing re- inforcement altogether (the first type of un- pairing that results in extinction of R2 in his example) and (b) delivering reinforcement irrespective of the occurrence of behavior (the second type of unpairing). Because NCR involves the latter procedure, adopting Michael’s distinction means that NCR may disrupt a response–reinforcer relation by un- pairing the two elements but that this does not constitute extinction. (It should be not- ed that Catania, 1992, prefers to call both operations extinction.) Regardless of whether Michael’s distinctions lead to changes in ter- minological usage, they provide a useful framework for classifying operational differ- ences.

Of greater concern than terminology is the fact that the procedural elements of cer- tain treatments remain somewhat elusive. For example, independent variables contin- ue to be described in rather global terms, such as curriculum revision, choice making, and simply antecedent intervention. Al- though data resulting from the use of these interventions often show evidence of behav- ior change, it is unclear how this change was effected or if it resulted from an ante- cedent manipulation rather than a change in behavioral consequences. For over 30 years, the JABA editorial process has been instrumental in promoting continued re- finement in the specification of consequenc- es. Now that research on EOs seems to have been firmly established in this journal, we look forward to further improvement in the specification, measurement, and evaluation of EO effects.

418 BRIAN A. IWATA et al.

REFERENCES

Catania, A. C. (1992). Learning (3rd ed.). Englewood Cliffs, NJ: Simon & Schuster.

McGill, P. (1999). Establishing operations: Implica- tions for the assessment, treatment, and preven- tion of problem behavior. Journal of Applied Be- havior Analysis, 32, 393–418.

Michael, J. (1982). Distinguishing between discrim- inative and motivational functions of stimuli. Journal of the Experimental Analysis of Behavior, 37, 149–155.

Michael, J. (1993). Establishing operations. The Be- havior Analyst, 16, 191–206.

Michael, J. (2000). Implications and refinements of the establishing operation concept. Journal of Ap- plied Behavior Analysis, 33, 401–410.

Smith, R. G., & Iwata, B. A. (1997). Antecedent influences on behavior disorders. Journal of Ap- plied Behavior Analysis, 30, 343–375.

Vollmer, T. R., & Iwata, B. A. (1991). Establishing operations and reinforcement effects. Journal of Applied Behavior Analysis, 24, 279–291.

Received August 29, 2000 Final acceptance August 29, 2000 Action Editor, F. Charles Mace

Applied Sciences
Architecture and Design
Biology
Business & Finance
Chemistry
Computer Science
Geography
Geology
Education
Engineering
English
Environmental science
Spanish
Government
History
Human Resource Management
Information Systems
Law
Literature
Mathematics
Nursing
Physics
Political Science
Psychology
Reading
Science
Social Science
Home
Homework Answers
Blog
Archive
Tags
Reviews
Contact
twitterfacebook
Copyright © 2022 SweetStudy.com