Noncontingent reinforcement as a treatment for severe behavior disorders

MISSING IMAGE

Material Information

Title:
Noncontingent reinforcement as a treatment for severe behavior disorders
Physical Description:
v, 56 leaves : ; 29 cm.
Language:
English
Creator:
Kahng, SungWoo
Publication Date:

Subjects

Subjects / Keywords:
Psychology thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 51-55).
Statement of Responsibility:
by SungWoo Kahng.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030475451
oclc - 43164371
System ID:
AA00013615:00001


This item is only available as the following downloads:


Full Text













NONCONTINGENT REINFORCEMENT AS A TREATMENT
FOR SEVERE BEHAVIOR DISORDERS












By

SUNGWOO KAHNG


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA














ACKNOWLEDGMENTS
I would like to express my gratitude to everyone who has helped make this

dissertation possible. Thanks to all the current and former graduates students at the

Florida Center on Self-Injury who helped to plan and conduct these studies: Juliet

Conners, Iser DeLeon, Han-Leong Goh, Gregory Hanley, Jana Lindberg, Eileen Roscoe,

Rachel Thompson, Michele Wallace, and April Worsdell. I would also like to

acknowledge the support and assistance of those who served on my committee: Marc

Branch, Shari Ellis, Timothy Hackenberg, and Cecil Mercer. I am particularly indebted

to my committee chair and advisor, Brian Iwata, for his patience and invaluable guidance

throughout my graduate career.

I would further like to express gratitude to my family for providing so much

support and encouragement. Finally, I wish to extend my gratitude to Lisa Toole for

everything she has done in helping me through these past several years. Without her

unconditional love and support, none of this would have been possible.















TABLE OF CONTENTS



A CKN O W LED GM EN TS................................................................................................. ii

A B STR A C T .................................... ............................................................................. iv

IN TROD U CTION ...................... .......................................................................... 1

NCR as a Control Procedure .............................................................................. 2
Therapeutic Effects of NCR................................................................................7...
NCR as a Function-based Intervention.................................. 9
NCR: Parametric Manipulation and Procedural Variations............................... 12
Lim stations of N CR .............................................................. ........................... 14
Mechanisms Underlying Effectiveness........................................................... 16

EXPERIMENT 1: FUNCTIONAL ANALYSIS....................................................... 18

M ethod ......................................................... ................................................... 18
Results and Discussion.............................................................................. 22

EXPERIMENT 2: A COMPARISON OF PROCEDURES
FOR THINNING NONCONTINGENT
REINFORCEMENT SCHEDULES.............................................. 26

M ethod.............................. ....................................................... ........................ 27
Results and Discussion ...................... ............................................................. 30

EXPERIMENT 3: SATIATION AND EXTINCTION DURING
NONCONTINGENT REINFORCEMENT.................................. 34

M ethod ......................................................... .................................................. 36
Results and Discussion............................................................................... 38

GENERAL DISCUSSION.......................................................................................... 48

R E FER EN C E S ......................................................... ................................................... 51

BIOGRAPHICAL SKETCH...................................................................................... 56


iii















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

NONCONTINGENT REINFORCEMENT AS A TREATMENT FOR SEVERE
BEHAVIOR DISORDERS

By

SungWoo Kahng

August 1999


Chairman: Brian A. Iwata
Major Department: Psychology

Noncontingent reinforcement (NCR) has become an increasingly popular

treatment for severe behavior disorders. NCR involves the delivery of reinforcers

independent of responding, typically on a time- or trial-based schedule. The present

research examined several aspects of NCR when used as a treatment for severe behavior

disorders.

During Experiment 1, functional analyses of self-injurious behavior (SIB) or

aggression were conducted for five individuals diagnosed with mental retardation. The

assessment was used as a screening procedure to identify those individuals who engaged

in problem behavior maintained by social-positive reinforcement.

Two methods of NCR schedule thinning were compared in Experiment 2. The

fixed procedure consisted of increasing the NCR schedule in fixed increments. Under the

adjusting schedule, changes were based on the mean interresponse time (IRT) of

problem behaviors: As problem behaviors decreased, the mean IRT increased, resulting

in a leaner NCR schedule. Results showed that both procedures were effective in








suppressing problem behaviors. Furthermore, the adjusting IRT procedure led to

thinning of the NCR schedule that was slightly more efficient than the fixed procedure.

Experiment 3 examined if the mechanism of effectiveness of NCR changed from

satiation to extinction as the schedule was thinned. Determination of the operative

mechanism was based on examination of within-session response patterns observed

during the NCR session and the extinction session that immediately followed it.

Satiation during NCR would lead to a temporary increase in responding during the Post-

NCR (extinction) session. Alternatively, if extinction was responsible for behavioral

suppression during NCR, responding should remain low during the Post-NCR

(extinction) session because the contingency for the problem behavior would remain the

same across the NCR and Post-NCR (extinction) sessions. Results showed that, for

two individuals, the mechanism changed from satiation to extinction as the NCR

schedule was thinned. The results for the other participants showed that satiation for

one individual and extinction for the other were responsible for behavioral suppression

throughout NCR schedule thinning.















INTRODUCTION
Research on the treatment of severe behavior disorders (e.g., self-injurious

behavior or SIB, aggression, property destruction, etc.) has a long-standing history in

the field of behavior analysis. Not surprisingly, much of this research has focused on

interventions based on the manipulation of reinforcement contingencies that may

maintain problem behavior. Generally, these interventions consist of the weakening of

response-reinforcer relationships and/or establishment of alternative responses that

replace the function of the problem behavior. One example of this type of intervention

is differential reinforcement of other behavior (DRO), in which reinforcement is

delivered for the absence of responding (Catania, 1998). DRO, along with differential

reinforcement of alternative behavior (DRA) and extinction, are the most common

interventions for severe behavior disorders that involve contingency manipulation

(Kahng & Iwata, 1999).

In contrast to the above treatments for problem behaviors, noncontingent

reinforcement (NCR) involves the delivery of reinforcers according to a schedule that is

independent of responding. In recent years, there has been a renewed interest in the use

of NCR as a therapeutic intervention (Vollmer, Iwata, Zarcone, Smith, & Mazaleski,

1993). The present research examines several aspects of the delivery of reinforcers in

the absence of a known contingency (i.e., NCR) when used as a treatment for severe

behavior disorders.
It should be noted that the term noncontingentt reinforcement" is inaccurate

because it contradicts the definition of reinforcement. That is, reinforcement by
definition, is contingent upon a response. Additionally, NCR does not lead to a








reinforcement effect of increased responding; in fact, NCR often leads to a reduction in

responding. Several alternatives to the term NCR have been proposed; these include

fixed-time (FT) reinforcement, response-independent reinforcement, free reinforcement,

and random reinforcement However, these alternative terms, all of which contain the

term "reinforcement," are as problematic as NCR. Additionally, there are several

advantages to the term NCR over the suggested alternatives. First, NCR unambiguously

conveys the absence of a contingency between a response and the delivery of a reinforcer.

Second, NCR is more flexible than FT reinforcement because NCR can also be delivered

on variable-time (VT) or trial-based schedules. Third, NCR is synonymous with

response-independent reinforcement. Finally, NCR is more specific than free and

random reinforcement. In fact, random reinforcement suggests the absence of a schedule,

which is not true of NCR.

Although the term NCR may not be an ideal term, it is sufficiently descriptive, as

good or better than the alternative terms, and has precedent in use in both basic (Lane,

1960) and applied research (Bushell, Wrobel, & Michaelis, 1968; Hart, Reynolds, Baer,

Brawley, & Harris, 1968). Moreover, there is precedent for the use of the label
"reinforcement" in reference to procedures that do not produce a reinforcement effect.

For example, "DRO" is a widely accepted term. Yet, reinforcement is not delivered

contingent upon any particular response. Rather, reinforcement is contingent on the

absence of a particular response and elapsed time. Furthermore, as in the case of NCR,

DRO does not lead to an increase in responding.



NCR as a Control Procedure
There has been a long history of the use of NCR in the field of behavior analysis.

One such application of NCR, particularly in basic operant research, has been as a

contingency-control procedure (e.g., Edwards, Peek, & Wolfe, 1970; Lattal, 1972). One

of the first studies that used NCR as a contingency-control procedure in operant








research was conducted by Rescorla and Skucy (1969). They examined the effects of
NCR on bar-pressing exhibited by male rats in a series of experiments. During
Experiment 1, bar pressing was maintained on a variable interval (VI) schedule. During

Experiment 2, the rats were divided into four groups. The first group (Group VI)
continued to receive reinforcement (food pellets) on the same VI schedule as in
Experiment 1. The second group (Group E) no longer received any reinforcement (i.e.,

extinction). Group three (Group R) continued to receive food pellets. However, the

schedule of reinforcer delivery was yoked to the VI schedule of the first group (Group
VI); therefore, this group received food pellets independent of bar pressing (i.e., NCR).

The procedures for the last group (Group D) was identical those of Group R, except

that reinforcers were never delivered within 5 seconds of a bar press. This condition

was designed to rule out the possibility of adventitious reinforcement. Results showed

that Group VI continued to respond at rates similar to those exhibited during

Experiment 1, whereas responding decreased for the other three groups (Groups E, R,
and D). Furthermore, the overall decrease in responding observed in these three groups

was similar, although the decrease in responding for Group R (i.e., NCR) was somewhat

slower than that of Group E (i.e., extinction).

Hammond (1980) examined the parametric effects of a contingency on free-

operant responding by varying the probabilities of reinforcement for responding and not
responding. A contingency was defined as the difference between the probability of

reinforcement for responding and the probability of reinforcement for not responding.
Thus, a difference greater than zero was a positive contingency, a difference equaling
zero was the contingency-control condition (i.e., NCR), and a difference less than zero

was a negative contingency.

In Experiment 1, Hammond examined the effects of a shift from a positive
contingency to a zero contingency on response rates using 10 experimentally-naive male
rats. The probabilities of reinforcement for responding and not responding under the








positive contingency condition were 0.05 and 0.0, respectively. Under the contingency-

control condition, the probabilities of reinforcement given a response and no response

were both 0.05. Thus, the difference between the two equaled zero. Results showed a

substantial decrease in responding following the introduction of the zero contingency.
Experiment 2 examined response rates following a shift from a high positive contingency

to a lower positive contingency, to a zero contingency, or to a high negative

contingency. Results of this experiment showed a decrease in responding that was much

greater following a shift from a positive contingency to a negative contingency than from

a positive contingency to a zero contingency.

The use of NCR as a contingency-control procedure has not been limited to basic
research. Several studies have demonstrated the contingency-control feature of NCR in

applied research. For example, Bushell et al. (1968) examined the use of a token

economy to increase the studying behavior of preschool age children. Delivery of

tokens was contingent upon actively engaging in individual activities (e.g., Spanish) and

participating in study teams (i.e., study behavior), which resulted in high levels of

studying. Bushell et al. then delivered tokens noncontingently, which resulted in a

subsequent decrease in studying.

Hart et al. (1968) evaluated the effects of contingent and noncontingent social

reinforcement on cooperative play (e.g., handing objects to other children, sharing toys,

etc.) exhibited by a five-year old girl. During the contingent reinforcement phase, a

teacher smiled at, laughed and conversed with, and admired the participant when she

engaged in cooperative play. The same reinforcers were delivered at random intervals

throughout the school day in the noncontingent reinforcement condition. Results

showed an increased rate of cooperative play when reinforcers were delivered
contingently and lower rates when reinforcers were delivered noncontingently.

Finally, Goetz, Holmberg, and LeBlanc (1975) compared DRO and NCR as
control procedures while evaluating a treatment to increase compliance with adult








requests. Their participant was a 3-year-old child attending preschool. The authors

initially increased compliance by presenting verbal praise contingent upon each

occurrence of compliance. They compared these results with those obtained under NCR

(reinforcement delivered following compliance and noncompliance) and DRO

(reinforcement delivered contingent on every instance of noncompliance). Responding

was more variable and suppression was slower under NCR than under DRO. Goetz et

al. concluded that DRO was the preferred control procedure. However, one limitation

of this study was the delivery of NCR in discrete trials in which reinforcers were

delivered following compliance on some trials and following noncompliance on others.

Thus, results of the NCR/DRO comparison were somewhat confounded because NCR

may have functioned as intermittent reinforcement for compliance.

These studies highlight the potential for the use of NCR as a contingency-control

procedure. Despite numerous demonstrations of the utility of NCR in both basic and

applied research, extinction continues to be the most common control procedure. This

may be a function of the amount of effort required to implement extinction versus NCR.

It is likely that less effort is required to discontinue the delivery of reinforcers than to

deliver reinforcers independent of responding. The latter requires that the experimenter

develop an NCR schedule and deliver reinforcers based on that schedule, whereas the

former typically requires little or no response by the experimenters. However, NCR

may ultimately be better than extinction as a contingency-control procedure because of

the reduced likelihood of side effects (e.g., extinction bursts) often associated with

extinction.

One area in which NCR is commonly used as a contingency-control procedure is

in the functional analysis of severe behavior disorders. This analysis typically involves
the systematic manipulation of reinforcement contingencies suspected of maintaining the

problem behavior. Test conditions contain contingencies for the emission of the target

behavior (i.e., problem behavior), whereas the control condition does not. Therefore,








increased rates of responding in a test condition relative to the control condition would

be indicative of sensitivity to a particular contingency.

Iwata, Dorsey, Slifer, Bauman, and Richman (1982/1994) presented a

methodology for conducting a functional analysis of SIB. The participants were

exposed to four conditions (three test conditions and one control). The social

disapproval condition, which involved the delivery of brief attention by the

experimenter to the participant contingent upon SIB, was a test for SIB maintained by

social-positive reinforcement. During the demand condition, instructional tasks were

presented by the experimenter to the participant. SIB resulted in a brief escape from

tasks. This condition tested for SIB maintained by social-negative reinforcement. The

final test condition was the alone condition. The individual was alone in a room, and no

social reinforcement was delivered. Persistence of SIB in the absence of social
reinforcement suggested that SIB was maintained by automatic reinforcement (Vaughn &

Michael, 1982). The control condition was referred to as the unstructured play

condition. Leisure items were freely available, demands were not presented, and

attention was delivered on an FT schedule (i.e., NCR). Thus, both positive and negative

social reinforcement were available independent of responding, resulting in a lower

probability of the establishment of a functional relationship between the response and

the reinforcer. NCR, rather than extinction, was used as the control because of the

potential for a temporary increase in responding sometimes associated with extinction

(i.e., extinction burst). Therefore, the unstructured play condition was a direct extension

of research on the contingency-control features of NCR, and its use during assessment

suggested that access to reinforcers on a response-independent basis may serve as an

alternative means of decreasing behavior. That is, the use of NCR as a control in

functional analysis suggests that it may have beneficial therapeutic effects.








Therapeutic Effects of NCR
In one of the earliest demonstrations of the use of the principles of behavior

analysis on an applied problem, Ayllon and Michael (1959) used NCR for the treatment

of problem behaviors. In one experiment, the authors focused on the treatment of

hoarding behavior (e.g., collecting paper, trash, and magazines) exhibited by four
patients living in a psychiatric hospital. Treatment consisted of ignoring the behavior

and giving unlimited access to materials by continually replenishing the supply (i.e.,

satiation), which resulted in a decrease of the hoarding behavior. Ayllon and Michael

concluded that reinforcer satiation can be an effective intervention for problem

behaviors. One limitation of the study was that the effects of NCR were confounded

because the function of the hoarding behavior was unknown; thus, ignoring the behavior

(i.e., extinction) may have also led to the decrease.

Satiation procedures have also been used as a common treatment for vomiting

and rumination exhibited by individuals diagnosed with developmental disabilities (e.g.,

Johnston, Greene, Rawal, Vazin, & Winston, 1991; Rast, Johnston, & Drum, 1984;

Rast, Johnston, Drum, & Conrin, 1981; Wilder, Draper, Williams, & Higbee, 1997).

One of the first studies examining NCR as a treatment for this topography of SIB was

conducted by Jackson, Johnson, Ackron, and Crowley (1975). They examined the

effects of food satiation (i.e., NCR) on rumination exhibited by two individuals

diagnosed with developmental disabilities. Baseline consisted of the delivery of the

participants' regular meal portions. During the treatment condition, the individuals were

given unlimited access to food until they emitted food refusal behavior (e.g., pushing the

food away) twice during a 1-min interval. Results showed that the satiation procedure

was highly effective in decreasing rumination for one individual and moderately effective
for the other.

A common treatment for other types of problem behaviors has been to provide
continuous access to sources of reinforcement that may not function to maintain the








problem behavior. For example, Bailey and Meyerson (1970) evaluated the effects of an

alternative source of reinforcement (i.e., vibration) for the treatment of SIB. Their

participant was a child diagnosed with developmental disabilities who engaged in biting.

In condition one, the participant received 6 seconds of access to the vibratory stimulus

contingent upon pressing a lever. In the second condition, the individual received 10

minutes of continuous access to the vibrator independent of responding (i.e., NCR).

Both conditions resulted in a decrease in SIB, although continuous access resulted in a

greater reduction.

Favell, McGimsey, and Schell (1982) also examined the efficacy of alternative

sources of reinforcement as treatment for SIB in a series of experiments. Six individuals

diagnosed with developmental disabilities participated in their study. All participants

engaged in SIB consisting of hand mouthing, eye poking, and/or pica. During

Experiment 1, the participants were provided continuous access to toys (NCR), which

resulted in a decrease in SIB relative to baseline. Subsequently, the experimenters made

access to the toys contingent upon the appropriate manipulation of those toys (DRA)

rather than the less-desirable topography of toy mouthing. This resulted in a slightly

greater decrease in SIB than did free access to reinforcers. The authors replicated these

findings with a different topography (eye poking) in Experiment 2. During Experiment

3, the participants were given free access to toys, popcorn, or both for the treatment of

pica (NCR), and results indicated that continuous access to either stimulus was effective

in reducing pica. Finally, the experimenters made access to popcorn contingent upon

appropriate toy manipulation (DRA). This was only effective in reducing pica for one
individual. The authors concluded that continuous access to alternative sources of

reinforcement may be an effective treatment for SIB.








NCR as a Function-based Intervention
Operant assessment techniques based on functional analysis methodology have

become common over the last 15 years in research on the treatment of SIB and other
behavior disorders (i.e., aggression and property destruction) exhibited by individuals

diagnosed with developmental disabilities (Pelios, Morren, Tesch, & Axelrod, in press).
This methodology has aided in the development of function-based interventions that

have several advantages over interventions that are selected arbitrarily (Iwata, Vollmer,
& Zarcone, 1990). First, the identification and alteration of antecedent conditions (i.e.,

establishing operations and discriminative stimuli) is made possible with the use of a

pretreatment functional analysis. Second, once the reinforcers maintaining the behaviors

are identified, the source of reinforcement can be minimized or eliminated (i.e.,

extinction). Third, the same reinforcer that maintains the problem behavior may be used
to establish an alternative response. Finally, the information provided by a functional

analysis may permit one to eliminate treatment components (e.g., reinforcers) that may

be irrelevant to the overall reduction of the target behavior.

The renewed interest in NCR as a function-based treatment for severe problem

behaviors may be attributed to a study conducted by Vollmer et al. (1993), who

compared the effectiveness of DRO and NCR in reducing SIB exhibited by three

individuals diagnosed with developmental disabilities. A functional analysis indicated

that all individuals engaged in SIB maintained by attention, which was delivered

contingent upon each occurrence of SIB during baseline. Following baseline, the authors

compared the efficacy of DRO and NCR using a multielement experimental design.
During the DRO condition, reinforcement was delivered contingent upon the absence of
SIB during brief intervals of time (10 s). In the NCR condition, reinforcers were
delivered independent of the individual's responding on an FT 10 s schedule. The DRO
and NCR schedules were eventually thinned to 5 minutes as the individuals continued to
emit low rates of SIB. The results of the comparison indicated that DRO and NCR








were equally effective in suppressing SIB. The authors concluded that NCR may be a
viable alternative to DRO because of its ease of implementation.



Social-positive Reinforcement
The Vollmer et al. (1993) study was one of the first to use NCR as a treatment
for behavior maintained by social-positive reinforcement. Since its publication, there
have been numerous replications and extensions of the applicability of NCR (e.g.,

Derby, Fisher, & Piazza, 1996; Hagopian, Fisher, & Legacy, 1994; Hanley, Piazza, &

Fisher, 1997; Hanley, Piazza, Fisher, Contrucci, & Maglieri, 1997; Lalli, Mace, Livezey,

& Kates, 1998; Marcus & Vollmer, 1996). For example, Fischer, Iwata, and Mazaleski
(1997) implemented NCR as a treatment for SIB exhibited by two individuals diagnosed

with mental retardation. A functional analysis indicated that both participants' engaged

in SIB maintained by social-positive reinforcement (attention and access to a tangible
item). Several food items (arbitrary reinforcers), which were demonstrated to be

irrelevant to behavioral function, were identified through a preference assessment. The

authors subsequently evaluated the effectiveness of two NCR interventions. During the

first NCR condition (NCR without extinction), arbitrary reinforcers were delivered on
an FT 10 s schedule while each occurrence of SIB resulted in access to the maintaining

reinforcer. Following this condition, arbitrary reinforcers were again delivered

independent of responding. However, the NCR schedule was progressively thinned and

occurrences of SIB no longer produced access to the maintaining reinforcer (NCR with
extinction). Results showed that the delivery of arbitrary reinforcers during NCR was
effective in decreasing SIB, even when the target behavior was concurrently reinforced.
These findings indicated that the noncontingent delivery of preferred items, which were
irrelevant to behavioral function, could be effective in reducing problem behavior.








Social-neeative Reinforcement

Relative to studies investigating behaviors maintained by social-positive

reinforcement, fewer studies have examined the use of NCR for behavior maintained by

social negative reinforcement (e.g., Kahng, Iwata, DeLeon, & Worsdell, 1997; Piazza,

Contrucci, Hanley, & Fisher, 1997). Vollmer, Marcus, and Ringdahl (1995) conducted

one of the first demonstrations of NCR as a treatment for escape-maintained problem

behaviors. They provided noncontingent breaks from task engagement for two

individuals for whom escape was a reinforcer (as determined from a functional analysis).

During baseline, a therapist delivered instructions to the participants. Contingent upon

SIB, the individual briefly escaped the demand. Treatment consisted of the delivery of

demands, similar to baseline. However, both participants received a break immediately

following each instruction, independent of SIB (i.e., NCR). Finally, the effects of NCR

were compared to DRO for one participant. During DRO, escape was provided

contingent upon the absence of SIB for a specified period of time. Results indicated that

NCR was effective in reducing escape-maintained SIB for both participants.

Additionally, NCR was as effective as DRO for one individual.



Automatic Reinforcement

Shore, Iwata, DeLeon, Kahng, and Smith (1997) examined the use of NCR for

the treatment of SIB maintained by automatic reinforcement by three individuals. A

functional analysis indicated that SIB occurred most often in the absence of programmed

social consequences (i.e., the alone condition), suggesting that the response was
maintained by automatic reinforcement. During Experiment 1, the participants were

given continuous access to preferred toys (based on a stimulus preference assessment),

which resulted in the near elimination of SIB. Additionally, the authors noted an

increase in the manipulation of toys to nearly 100% of the intervals. Experiment 2

consisted of the use of the same reinforcers (i.e., toys) in a DRO procedure. The toys








were delivered contingent upon the absence of SIB for a specified DRO interval.
Results of this experiment showed little decrease in SIB relative to baseline under the

DRO condition. Shore et al. concluded that for their participants, continuous access to

toys effectively competed with SIB, whereas the use of the same reinforcers in a DRO

procedure was ineffective.

These studies demonstrated the efficacy of NCR as a treatment for problem
behaviors maintained by social reinforcement (positive and negative) as well as by

automatic reinforcement. Additionally, several studies have examined the use of NCR

with other problem behaviors (e.g., Day, Homer, & O'Neill, 1994; Hanley et al., 1997)

and different populations (e.g., Bloxham, Long, Alderman, & Hollin, 1993; Knight &
McKenzie, 1974).



NCR: Parametric Manipulation and Procedural Variation



Schedules

Research suggests that NCR is more effective in suppressing responding under
dense rather than lean schedules. Dense schedules have typically consisted of

continuous or near continuous delivery of reinforcers (e.g., FT 10 s), whereas lean

schedules have been set at an arbitrarily determined FT 5-min. Hagopian et al. (1994)

examined if a lean NCR schedule would result in decreases in problem behaviors. Their

participants were quadruplets diagnosed with developmental disabilities whose problem
behaviors (i.e., SIB, aggression, and property destruction) were maintained by social-
positive reinforcement (i.e., attention). During baseline, attention was delivered

contingent upon problem behaviors. Following baseline, Hagopian et al. compared the
dense (FT 10 s) and lean (FT 5-min) NCR schedules in a multielement design. Their








results indicated that, for all participants, the dense NCR schedule resulted in greater
reductions in the problem behaviors.



Reinforcer Magnitude

Carr, Bailey, Ecott, Lucker, and Weil (1998) examined the effects of magnitude
of reinforcement, while holding rate constant, on the suppressive effects of NCR. Their
participants were five individuals diagnosed with developmental disabilities. Rather

than using problem behaviors as the dependent variable, Carr et al. decreased arbitrary

responses (dropping poker chips into a plastic cylinder) with NCR. During baseline,

the arbitrary response was reinforced on an FR 1 schedule, which was eventually
thinned to variable ratio (VR) 3 or VR 5. Following baseline, the authors evaluated three

NCR conditions based on magnitude (high, medium, and low) of reinforcement. In the

high magnitude condition, a predetermined quantity of food was delivered independent
of responding. Under the medium magnitude condition, half of the amount of food

(relative to the high magnitude condition) was delivered. Finally, during the low

magnitude condition, a third of the amount of food (relative to the high magnitude

condition) was delivered independent of responding. Results showed that the arbitrary

response was suppressed during the high magnitude condition for all five participants.
Under the medium magnitude condition, responding decreased with only two

participants. Finally, there was little decrease in responding for any participants under

the low magnitude condition.



Schedule Thinning
To date, most studies examining NCR have thinned the schedule irrespective of
the individual's rates of responding (e.g., Hagopian et al., 1994; Vollmer et al., 1993).
That is, the initial NCR schedule is typically based on an arbitrarily determined FT








value (e.g., FT 10 s) and eventually thinned by increasing the FT value according to fixed

increments. The second experiment of this dissertation compared two methods of NCR
schedule thinning one based on procedures typically used (i.e., fixed increments of

time) and the other based on the individual's rate of responding (i.e., interresponse time

or IRT) -- to determine which procedure would result in quicker thinning of the NCR
schedule.



Limitations of NCR



Adventitious Reinforcement
Although no explicit contingencies are established with NCR, the possibility

exists that some proportion of responses may be followed by reinforcers. This

potential correlation between response and reinforcer may lead to response maintenance

through adventitious reinforcement (Skinner, 1948). Vollmer, Ringdahl, Roane, and

Marcus (1997) documented adventitious reinforcement as a potential limitation of NCR.

Their initial NCR schedule resulted in near continuous access to reinforcers, which led to

a decrease in the problem behavior exhibited by an individual diagnosed with

developmental disabilities. As the NCR schedule was thinned, Vollmer et al. noted a
gradual increase in responding as the schedule reached one reinforcer per minute. A

within-session analysis revealed that problem behavior often occurred just before the

delivery of scheduled reinforcers, and Vollmer et al. concluded that adventitious

reinforcement during NCR may have been responsible for the increase in responding

because of the correlation between response and reinforcers.








Alternative Response

Another potential limitation to NCR may be that it does not explicitly result in
an increase of alternative, more appropriate behaviors. In fact, it is possible that NCR

may interfere with the acquisition of alternative responses. That is, NCR may reduce

(through satiation) the effectiveness of a reinforcer (i.e., alter the establishing operations)
for the problem behavior as well as for the alternative response. Marcus and Vollmer

(1996) examined if satiation effects may contraindicate the use of NCR when also
attempting to increase appropriate behaviors with the same reinforcer maintaining the

problem behavior (e.g., differential reinforcement of alternative or DRA). The

participants were 3 individuals diagnosed with developmental disabilities who engaged

in SIB or aggression. A functional analysis suggested that their problem behaviors were
maintained by social-positive reinforcement. Two of the three participants were

exposed to NCR plus DRA, in which the same reinforcers were delivered

noncontingently on an FT schedule (i.e., NCR) and contingent on each occurrence of an

alternative response (i.e., DRA).

During baseline, the preferred stimuli were delivered contingent upon each
occurrence of SIB or aggression. Following baseline, the participants were taught to

emit an alternative response to receive reinforcement. Following this training phase, the
authors evaluated the effects of NCR in combination with DRA. The initial NCR

schedule was based on the approximate mean of the baseline rate of reinforcement (3

reinforcers per minute). SIB and aggression were extinguished, whereas the alternative

response resulted in reinforcement on an FR 1 schedule. Thus, if differential
reinforcement failed to produce increases in the alternative behavior, this would suggest
that NCR may have interfered with the emission of the alternative response by leading
to satiation to the reinforcer. The results of their study showed that the combination of
NCR and DRA led to continued suppression of problem behaviors. Furthermore, the
alternative response increased as the NCR schedule was systematically thinned. The








authors concluded that NCR may not interfere with the emission of alternative

responses. Additionally, these data suggest that satiation was not a factor in the

effectiveness of NCR.



Mechanisms Underlying Effectiveness
The use of NCR as a treatment for severe behavior disorders typically consists

of dense schedules of reinforcer delivery. Therefore, the noncontingent delivery of

reinforcers may lead to the momentary alteration of establishing operations (Michael,

1982). That is, the continual delivery of reinforcers may lead to satiation to the

reinforcer, which may account for the suppression in responding (Vollmer et al., 1993).

Lalli et al. (1997) conducted a series of experiments related to this issue. The final

experiment evaluated the effectiveness of NCR without extinction as a treatment for

SIB. The participant was a 7-year-old boy diagnosed with severe mental retardation and

severe behavior disorders. During the NCR condition, the participant was given 30

second access to a preferred object on an FT schedule. Additionally, SIB continued to

be concurrently reinforced on an FR 1 schedule. Results showed an initial increase in

the rate of SIB (relative to baseline), which eventually decreased to near zero levels.

Thus, given that extinction was not necessary for response suppression, the authors

concluded that satiation was the mechanism responsible for behavior reduction.

An alternative operative mechanism may be extinction. That is, the disruption

of the response-reinforcer contingency may lead to what amounts to extinction

(Rescorla & Skucy, 1969). This hypothesis may be supported by the findings of

Marcus and Vollmer (1996). Given that the participants' alternative responses

increased when both NCR and DRA were in effect, it would suggest that the individuals

were not satiated (via NCR) to the reinforcers, but rather, the reduction was a result of

extinction.





17

One explanation for the discrepant findings of Lalli et al. (1997) and Marcus and
Vollmer (1996) may be that both satiation and extinction are responsible for the

effectiveness of NCR. That is, initially dense NCR schedules may result in satiation to

the reinforcer. However, satiation becomes unlikely as the NCR schedule is thinned

because fewer reinforcers are delivered. Therefore, the operative mechanism may change

from satiation to extinction as the NCR schedule is thinned. The third experiment of

this dissertation evaluated whether the mechanism of NCR changes from satiation to

extinction and attempted to identify the point at which this change may occur.















EXPERIMENT 1: FUNCTIONAL ANALYSIS
Over the past two decades, functional analysis has become the most common
method of assessing problem behaviors (Didden, Duker, & Korzilius, 1997). Functional

analysis involves observation of behavior under a series of controlled environmental

conditions to identify maintaining contingencies for problem behavior (Iwata et al.,

1990). Because the use of NCR in Experiments 2 and 3 required the delivery of

reinforcers that maintained problem behavior, functional analyses were conducted for all

participants in Experiment 1. The functional analysis was used as a screening procedure

to identify those individuals who engaged in problem behavior maintained by social-

positive reinforcement.



Method



Participants and Settings

The participants (Table 1) were five individuals diagnosed with developmental

disabilities. All participants lived in a state residential facility and had limited

communication skills. None emitted vocal speech, although they did emit simple

gestures (e.g., pointing) to indicate some of their needs. All participants were diagnosed

with severe or profound mental retardation. Julia was a 43-year-old female diagnosed
with a seizure disorder. She received dilantin and phenobarbitol throughout this study

(at a constant dose) to control seizures. Her SIB consisted of skin picking. Lisa was a

27-year-old woman who engaged in SIB consisting of head/body hitting and hand biting.








Nancy was a 50-year-old woman who engaged in SIB consisting of head/body hitting.

Susan was a 31-year-old female diagnosed with seizure disorder, cerebral palsy, and

Angelman Syndrome. She received a constant dose of depakote throughout the study to

control seizures. She engaged in SIB consisting of arm hitting and aggression consisting

of hitting, scratching, pinching, and kicking others. Matt was a 25-year-old male who

engaged in SIB consisting of hand or arm biting.



Table 1
Subject characteristics and NCR experiment in which they participated


Age Diagnosis

43 severe MR

seizure disorder

27 profound MR


50 profound MR

31 profound MR

cerebral palsy

seizure disorder

Angelman Syndrome


25 profound MR


Topography

skin picking


head/body hit

hand bite

head/body hit

arm hitting

hitting others

scratching others

pinching others

kicking others

hand/arm bite


All individuals were referred to a day clinic, located on the grounds of the

residential facility, for the assessment and treatment of their SIB. Sessions were

conducted in therapy rooms located in the clinic. Each room contained several chairs, a


Name

Julia


Lisa


Nancy

Susan


Matt


Experiment

2, 3


2, 3


2

3








3








table, and other materials necessary to run therapy sessions (e.g., work material, toys,

etc.).



Data Collection and Reliability

An observer used a hand-held computer (Assistant, Model AST102) to record
frequencies of SIB during continuous 10-s intervals throughout the sessions. Data were

summarized as the number of self-injurious or aggressive responses per minute.

Additionally, data were collected on experimenters' implementation of assessment and

treatment procedures (e.g., delivery of instructions, prompting, and consequences).

A second observer simultaneously but independently collected data. Agreement

percentages were then calculated based on an interval-by-interval comparison of

observers' records (Table 2). The smaller number of responses in each interval was

divided by the larger number of responses in each interval. These proportions were

summed across all intervals and divided by the total number of intervals in the session to

obtain the percentage agreement between the observers.



Table 2
Percentage of sessions with interobserver agreement and interobserver
agreement scores for all participants during the functional analysis


Percent of Sessions with
Participant Interobserver Agreement

Julia 20.0%

Lisa 46.7%

Nancy 37.5%

Susan 36.0%

Matt 39.3%


Interobserver Agreement Scores

99.7% (range, 98.9% to 100.0%)

95.4% (range, 69.0% to 100.0%)

96.9% (range, 94.9% to 97.8%)

99.0% (range, 95.6% to 100.0%)

94.2% (range, 74.6% to 100.0%)








Functional Analysis

Each individual was exposed to four assessment conditions (Alone, Attention,
Demand, and Play). Two individuals were exposed to a fifth condition (Tangible). The

procedures were similar to those described by Iwata et al. (1982/1994). All sessions

were 15 minutes in length.

Alone. The individual was in a therapy room. An experimenter was not present
nor were leisure materials available. The purpose of this condition was to determine if
problem behavior persisted in the absence of social reinforcement, which would suggest

that problem behavior was maintained by automatic reinforcement.

Attention. An experimenter was present, and the room contained various leisure
materials. At the beginning of the session, the experimenter prompted the individual
toward the materials. No social interaction was provided to the participant unless the
individual engaged in problem behaviors, at which time the experimenter approached the

participant and provided brief social attention (e.g., "Don't do that, you'll hurt

yourself') and light physical contact. The purpose of this condition was to determine if

problem behaviors were maintained by social-positive reinforcement in the form of
attention.

Demand. The participant was seated at a table with the experimenter.
Instructional trials were presented on an FT 30 s schedule using a three-step graduated

prompting procedure (verbal instruction, demonstration, physical guidance). Prompts

were delivered at 5 s intervals if compliance did not occur. Brief praise was delivered

contingent upon compliance with the request. If the individual emitted problem
behaviors during the instructional trial, the experimenter terminated all interaction and
turned away from the individual until the beginning of the next trial. The purpose of
this condition was to determine if SIB was maintained by social-negative reinforcement
in the form of escape from tasks.








Tangibl. This condition was conducted only if previous information suggested
that a tangible item may function to maintain problem behavior. An experimenter was

present throughout the session. A tangible item (i.e., food) was given to the individual

contingent upon problem behaviors. No other form of attention was delivered. This

condition tested for access to food as a reinforcer for problem behaviors.

Play. Leisure items were available throughout the session. An experimenter
delivered noncontingent attention on an FT 30 s schedule. All problem behaviors were

ignored (i.e., extinction). This condition served as the control condition.



Results and Discussion
Julia's SIB (Figure 1, top panel) occurred most often during the attention
condition of the functional analysis and rarely occurred during the other conditions,

suggesting that her SIB was maintained by attention. Lisa (Figure 1, middle panel)

exhibited the most SIB during the tangible condition, suggesting that her SIB was

maintained by access to food (a small amount of pickles). Nancy (Figure 1, bottom

panel) engaged in SIB in all four conditions of the functional analysis. However,

responding was substantially higher during the attention condition, indicating that her

SIB was maintained primarily by attention. Finally, both Susan and Matt (Figure 2)

exhibited initially low rates of destructive behaviors during their functional analyses.

Susan's rates of SIB and aggression (top panel) during the attention condition increased

(relative to the other conditions) by the fourth attention session and continued to remain

high during the remainder of her functional analysis. Therefore, Susan's functional
analysis suggested that her SIB and aggression were maintained by attention. Similarly,

Matt's SIB (bottom panel) increased by the fourth tangible session of his functional

analysis and maintained for the remainder of the assessment. Therefore, the results of

his functional analysis suggested that his SIB was maintained by access to food

(M&Ms).





23

The results of all five individuals' functional analyses indicated that their
problem behaviors were maintained by social-positive reinforcement in the form of
attention or access to tangible items (i.e., food). The results of the functional analysis in

Experiment 1 served as the basis for the NCR evaluations in Experiment 2 and 3.






24








5-

4-

3-

2-





2 4 6 8 10 12 14 16



O6 iSa Alone

----0--- Attention
4 Play

C/ Demand
M-. 2-
2- ---- Tangible
0 o- -- ----

V) 2 4 6 8 10 12 14 16
8-


6- Nancy


4


2-



2 4 6 8 10 12 14 16

SESSIONS


Figure 1. Responses per minute of SIB during functional analysis for Julia (top panel),
Lisa (middle panel), and Nancy (bottom panel).




















30-

0
> 20-


o10



0


z
o-



S10-

z 8.




(j M


SESSIONS


Figure 2. Responses per minute of problem behaviors during functional analysis for
Susan (top panel) and Matt (bottom panel).














EXPERIMENT 2: A COMPARISON OF PROCEDURES FOR THINNING
NONCONTINGENT REINFORCEMENT SCHEDULES

One of the primary advantages of NCR over other reinforcement-based
interventions, such as DRO, is ease of use (Vollmer et al., 1993). In DRO, a caregiver

must continuously observe the individual to ensure that the criterion for reinforcement
(i.e., no responding) has been met. However, this is unnecessary under NCR because

the schedule is not affected by the individual's behavior. Although NCR may be easier

to implement than DRO, NCR may still be cumbersome to implement given that it

typically requires dense initial schedules (Hagopian et al., 1994). Therefore, thinning

NCR schedules may be an important determinant of procedural integrity.

Typically, the initial NCR schedule is often set at a dense value (e.g., FT 10 s).
The schedule is then thinned according to specified criteria (e.g., Vollmer et al., 1993)

using fixed increments of time. However, this is a somewhat arbitrary process that does

not take into account the individual's rate of responding. This arbitrary process may

result in an unnecessarily slow thinning process, or to schedules that are hastily

increased, which may lead to an increase in problem behavior.

Lalli et al. (1997) examined several procedural variations of NCR, one of which

was the determination of the initial NCR schedule. The participants were two children

admitted to an inpatient unit specializing in the treatment of severe problem behaviors.
Both were diagnosed with developmental disabilities and engaged in aggression and SIB.
Their initial NCR schedule was based on the mean latency to the first problem behavior
during the functional analysis. This resulted in initial schedules of FT 90 s for one
individual and FT 120 s for the other. There was a substantial decrease in the problem








behavior for both participants following the introduction of NCR, even though the initial

NCR schedule was much leaner than those previously used (e.g., Hagopian et al., 1994;
Vollmer et al., 1993). These data suggest that leaner NCR schedules, based on rates of

responding, may be effective in reducing problem behaviors.

Although Lalli et al. used relatively lean NCR schedules to decrease problem

behaviors, latency to the first response may not be an accurate reflection of the

individual's rates of responding throughout an entire session. Furthermore, Lalli et al.

subsequently thinned their NCR schedule based on fixed increments of time. An

alternative method for both determining the initial NCR schedule and subsequently

thinning reinforcement may be to derive the schedules from the mean interresponse time

(IRT) for the target behavior. That is, the NCR schedule could be determined prior to

each session by calculating the mean IRT of previous sessions. Thus, the mean IRT

increases as responding decreases, resulting in a thinning of the NCR schedule. The
purpose of Experiment 2 was to examine the utility of thinning the NCR schedule based

on the mean IRT of problem behavior. This alternative method of NCR schedule

thinning was compared to a more common thinning procedure that was based on fixed

increments (Vollmer et al., 1993).



Method



Participants and Settings

The participants were Julia, Lisa, and Nancy from Experiment 1. Settings and

materials were identical to those of Experiment 1.








Data Collection and Reliability

Data collection and calculation of interobserver agreement were identical to those
of Experiment 1. Interobserver agreement was assessed for Julia, Lisa, and Nancy

during 35.3%, 51.8%, and 33.3% of the baseline and treatment sessions, respectively.

Mean interobserver agreement scores for Julia, Lisa, and Nancy were 98.7% (range,

95.6% to 100.0%), 95.9% (range, 72.1% to 100.0%), and 95.9% (range, 89.2% to

100.0%), respectively.



Experimental Sequence and Design

Following Experiment 1, two simultaneous baselines were implemented for all

three participants. Each baseline was conducted in a different therapy room with a

different experimenter to enhance discrimination between the two NCR schedule

thinning conditions. Following baseline, the NCR intervention was introduced according

to a multiple baseline across subjects design. The two NCR schedule thinning

procedures were evaluated using a multielement design.


Treatment Conditions

Baseline. An experimenter was present throughout the session, and leisure

materials were available. The baseline contingencies were identical to those of the

functional analysis condition in which responding was highest. Julia and Nancy received

approximately 5 to 10 s of attention (e.g., "Don't do that, you will hurt yourself")

contingent upon SIB, and Lisa received a small amount of pickles contingent upon SIB.

Reinforcement was delivered on an FR 1 schedule for all participants during baseline.

Noncontingent reinforcement. During the two NCR conditions, the therapist
delivered the same reinforcers as in baseline. However, reinforcers were delivered on an

FT schedule, independent of the individual's responding. Additionally, there were no

programmed consequences for SIB.








The Fixed NCR schedule thinning procedure was based on Vollmer et al. (1993).

The initial NCR schedule was always set at FT 10 s. If SIB was at or below 0.5

responses per minute for a given session, the NCR schedule was increased in the

subsequent session according to fixed increments (Table 3). The schedule of reinforcer

delivery was thinned from an initial rate of 6 per minute (FT 10 s) to a rate of 0.2

reinforcers per minute (FT 5-min).

During the Adjusting IRT NCR schedule thinning condition, the initial schedule

was based on the mean IRT of the last 3 baseline sessions. Before each subsequent NCR

session, the schedule was determined by calculating the mean IRT of the previous 3



Table 3
Schedule Thinning Increments (Fixed Condition)


Reinforcers NCR schedule Reinforcers NCR schedule

6/min FT 10 s 1/min FT 1-min

5/min FT 12 s .5/min FT 2-min

4/min FT 15 s .33/min FT 3-min

3/min FT 20 s .25/min FT 4-min

2/min FT 30 s .2/min FT 5-min



sessions. In order to prevent large, sudden decreases in responding from dramatically

lengthening the NCR schedule, the schedule of reinforcer delivery for a given session

was limited to a maximum of a 100% increase from the previous session's NCR

schedule.

The terminal NCR schedule was FT 5-min for both NCR conditions. Therefore,

schedule thinning was stopped once the NCR schedule reached FT 5-min. The criterion

for completion of the NCR condition was SIB at or below 0.5 responses per minute for 5








consecutive sessions at the FT 5-min schedule. All sessions (baseline and treatment)

were 15 minutes in duration.


Results and Discussion

Figure 3 shows data from the comparison of the two NCR schedule thinning

procedures. Julia's responding during both baselines was variable (Figure 3, top panel).

During baseline for the Fixed and Adjusting IRT conditions, Julia averaged 3.1 and 3.5

SIBs per minute, respectively. Following the introduction of NCR, there was an

immediate decrease in SIB under both the Fixed and Adjusting IRT conditions M = 0.1

per minute and 0.2 per minute, respectively). The initial NCR schedule under the

Adjusting IRT condition was FT 15 s based on the mean IRT of the last three sessions

of baseline. As previously stated, the initial NCR schedule under the Fixed condition

was always FT 10 s. The terminal NCR schedule (FT 5-min) was reached in 6 sessions

under the Adjusting IRT condition, compared to 10 sessions under the Fixed condition.

The reduction in Lisa's SIB (Figure 3, middle panel) following the introduction

of NCR was more gradual than that observed for Julia. During baseline, Lisa averaged

9.4 SIBs per minute in the Fixed condition and 8.0 SIBs per minute in the Adjusting IRT

condition. NCR resulted in a decrease in the Fixed condition to 1.8 SIBs per minute and

in the Adjusting IRT condition to 2.0 SIBs per minute. The terminal NCR schedule of

FT 5-min was reached quicker under the Adjusting IRT condition (11 sessions)

compared to the Fixed condition (15 sessions).

Finally, NCR led to an immediate reduction in Nancy's SIB (Figure 3, bottom

panel). During baseline, Nancy averaged 4.1 and 5.8 SIBs per minute in the Fixed

Adjusting IRT conditions, respectively. SIB decreased to 0.6 per minute during the

Fixed NCR condition and to 1.2 per minute during the Adjusting IRT condition. The

NCR schedule was thinned to the terminal schedule in 15 sessions under the Adjusting
IRT condition compared to 25 sessions under the Fixed condition.











Noncontingent Reinforcement


FT5 I Julia

"-"0
20-

I 15- 6"
1Adjusting
10 IRT Fixed


5 / i5 F T 5
S5- T T' FT5'

0o y






FT 11"


0-
10 20 30 40 50 60 70 80
SESSIONS

Figure 3. Responses per minute of SIB during baseline and treatment conditions for all
participants. The initial NCR schedule under the Fixed Condition was always
FT 10 s. Italicized text represents the NCR schedule under the Adjusting IRT
condition.








The results of this study indicated that both NCR schedule thinning procedures
were effective in reducing problem behavior. Additionally, the NCR schedule was
thinned to a terminal value somewhat quicker using a procedure based on the
individual's rate of responding (i.e., the Adjusting IRT procedure). That is, for all
participants, it took fewer sessions to reach FT 5-min using the Adjusting IRT
procedure when compared to the Fixed method of schedule thinning. Therefore, the
Adjusting IRT procedure may be a viable alternative to the Fixed procedure, especially
in light of the fact that the Adjusting IRT procedure was somewhat more efficient.
For two of the three participants (Julia and Nancy), NCR produced immediate reductions
in SIB, resulting in increases of the NCR schedule under the Adjusting IRT condition that
were much greater than the those resulting from the Fixed method of schedule thinning.
Because of such low rates of SIB, it is possible that the terminal NCR schedule of FT 5-

min using the Adjusting IRT schedule thinning procedure could have been reached in
even fewer sessions. However, a limit of a 100% increase (i.e., double) from the
previous NCR schedule was set to prevent dramatic increases in the NCR schedule from
adversely affecting rates of responding. Hypothetically, it is conceivable that the
terminal NCR schedule of FT 5-min could be reached by the fourth NCR session using
the Adjusting IRT procedure if there was no limit to the increase in NCR schedule from

previous session. That is, if no SIB was emitted during the first three NCR sessions,
the mean IRT would equal 5 minutes resulting in an NCR schedule of FT 5-min. Future
studies should focus on what effect this limit may have on NCR schedule thinning.

One limitation to this study may have been the minimum number of sessions
required to reach the terminal schedule with both procedures. The Adjusting IRT
procedure required a minimum of 4 sessions to reach the terminal schedule whereas the
Fixed procedure required at least 10 sessions. It is conceivable that had the Fixed
procedure been limited to 4 steps (i.e., schedule values) the outcome, in terms of the
difference in the number of sessions it took to reach the terminal schedule, may have








been different. However, we compared the Adjusting IRT procedure to the most

commonly used NCR schedule thinning procedure (e.g., Vollmer et al., 1993; Hagopian

et al., 1994).

Finally, one limitation to the use of NCR as a treatment for severe behavior
disorders is the risk of adventitious reinforcement (Vollmer et al., 1997). In fact, there
may be a gra~gr risk of adventitious reinforcement associated with the Adjusting IRT

procedure. There is a possibility that the correlation between the occurrence of SIB and
the delivery of reinforcers may be higher given that the NCR schedule is based on the

individual's rate of responding. This may be one possible explanation for the gradual
decrease observed in Lisa's SIB following the introduction of the Adjusting IRT NCR
condition. Her responding during the first 6 Adjusting IRT NCR sessions was higher
than that obtained during the first 6 Fixed NCR sessions. However, the rapid thinning

of the NCR schedule may have led to subsequent extinction of this adventitiously

reinforced behavior.

One area of future research may be to examine the degree to which the NCR
schedule may be thinned. Studies have not examined the possibility of thinning the
NCR schedule beyond those typically used (i.e., FT 5-min). If it is possible to extend

the NCR schedule beyond FT 5-min, the efficiency of the Adjusting IRT procedure

(relative to the Fixed procedure) would also increase. Thus, it is possible that the

difference in efficiency between the two schedule thinning procedures may also continue

to increase so that the Adjusting IRT procedure would result in NCR schedules that are

significantly leaner than those produced by the Fixed procedure.














EXPERIMENT 3: SATIATION AND EXTINCTION DURING
NONCONTINGENT REINFORCEMENT
Because NCR typically involves dense schedules of reinforcer delivery,
responding may be reduced through elimination of the behavior's establishing operation

(Michael, 1982). A closer examination of the rate of reinforcer delivery for two of the

participants in the Hagopian et al. (1994) study lends support to this hypothesis. That
is, the mean rate of reinforcer delivery during the dense NCR condition was 6.0 per

minute for two of the individuals, which was greater than the rate of reinforcer delivery

during baseline. Thus, it is possible that the overabundance of reinforcers during NCR

resulted in satiation to the reinforcer. Furthermore, results of two studies (Fischer et al.,
1997; Lalli et al., 1997) showed that dense NCR schedules suppressed problem

behaviors, even when each occurrence of problem behavior also produced reinforcement

indicating that mechanisms other than extinction (presumably satiation) were
responsible for behavioral suppression.

An alternative explanation for the effectiveness of NCR is that problem behavior

decreases because it is extinguished (Vollmer et al., 1993). NCR contains an extinction

component because NCR results in the removal of the response-reinforcer contingency

(Rescorla & Skucy, 1969). One participant in the Vollmer et al. study emitted high rates

of SIB during baseline relative to treatment. The mean rate of reinforcement during
baseline was 22 per minute, whereas the densest rate of reinforcement during NCR was
6 per minute. Vollmer et al. noted that the rate of reinforcement during NCR may not
have been sufficient to produce satiation, leading them to suggest that the effectiveness

of NCR may have been due to extinction.








The findings of Marcus and Vollmer (1996) lend further support for extinction

as the mechanism of effectiveness for NCR. They evaluated the effects of NCR when
used in conjunction with DRA. Attention, which was the reinforcer maintaining

problem behavior, was delivered noncontingently (NCR) and contingent upon emission
of an alternative response (DRA). NCR + DRA led to a decrease in the problem

behaviors. Additionally, the participants continued to engage in the alternative

response, suggesting that they were not satiated to the reinforcer. Therefore, it is likely

that some other mechanism (presumably extinction) was responsible for behavioral

suppression during NCR.

An explanation for these seemingly discrepant results may be provided by
further examination of the findings of the Hagopian et al. (1994) study. Although the

rate of reinforcer delivery was much higher during the dense NCR condition (6
reinforcers per minute or FT 10 s) relative to baseline (suggesting satiation), response

suppression was also observed in three of the four participants during the lean NCR

condition, in which the rate of reinforcer delivery was 0.2 per minute (FT 5-min). This

comparison of the dense and lean NCR schedules was relatively brief; had Hagopian et

al. continued with the lean NCR condition, the problem behaviors may have decreased
further from continued exposure to extinction. As Hagopian et al. thinned the dense

NCR schedule, they noted that responding continued to be suppressed throughout the

schedule thinning process. Thus, it is possible that dense and lean NCR schedules affect

behavior differently but produce the same outcome. That is, removal of the establishing
operation (i.e., satiation) may have been responsible for the reduction in problem

behavior under dense NCR schedules; but, as the NCR schedule was thinned, the
operative mechanism may have changed to extinction. Therefore, the purpose of
Experiment 3 was to examine if the mechanism of effectiveness for NCR changes from

satiation to extinction and to identify the point at which this change may occur.








Method



Participants and Settings

The participants were Julia, Lisa, Susan, and Matt from Experiment 1. Settings
and materials were identical to those of Experiment 1.



Data Collection and Reliability

Data collection and interobserver agreement procedures were similar to those
used in Experiment 1. Interobserver agreement was assessed during 20.0%, 42.9%,

42.9%, and 34.6% of the sessions during baseline and treatment for Julia, Lisa, Susan,

and Matt, respectively. Mean interobserver agreement scores during baseline and

treatment were 99.3% (range, 96.1% to 100.0%), 95.9% (range, 91.8% to 100.0%),

97.0% (range, 91.0% to 100.0 %), and 97.5% (range, 79.3% to 100.0%) for Julia, Lisa,

Susan, and Matt, respectively.



Experimental Sequence and Design

Following baseline, the effects of NCR were evaluated using a multiple-baseline

across subjects design.


Treatment Conditions

Baselin. All sessions were 10 minutes in length. Baseline contingencies were
identical to the functional analysis conditions in which problem behavior was highest.

Brief attention was delivered contingent upon SIB or aggression for Julia and Susan.

Lisa and Matt were given a small amount of edibles (pickles and M&Ms, respectively)

contingent upon SIB. All reinforcers were delivered on an FR 1 schedule.








Noncontingent reinforcement. Reinforcers identical to those used in baseline were
delivered independent of problem behaviors on an FT schedule. There were no

programmed social consequences for problem behaviors. The initial NCR schedules for

Julia, Lisa, and Susan were based on the mean IRT of SIB or aggression during the last 3

baseline sessions. The NCR schedule was thinned in each subsequent NCR session

based on the mean IRT of SIB or aggression during the previous 3 sessions. Therefore,

as problem behaviors decreased the mean IRT increased, resulting in a leaner NCR

schedule. Schedule thinning was discontinued once the NCR schedule reached FT 300 s.

NCR schedule thinning for Matt was more gradual because of a history of

difficulty with thinning schedules of reinforcement. That is, previous thinning of

reinforcement schedules (e.g., DRA) was unsuccessful in that problem behavior

increased as reinforcement schedules were thinned. Therefore, Matt's initial NCR

schedule was FT 5 s (12 reinforcers/min), which resulted,in continuous access to

reinforcers. Thinning of the NCR schedule was started after the third NCR session. If

SIB was at of below 0.5 per minute, the delivery of 1 reinforcer per minute was deleted in

the subsequent session. This continued until the NCR schedule reached FT 60 s or the
delivery of 1 reinforcer per minute. At this point, the delivery of 1 reinforcer per session

was deleted in each subsequent session until the terminal schedule of FT 300 s (i.e., 2

reinforcers delivery per session) was reached. All NCR sessions were 10 minutes in

length.

Post-NCR (extinction). Immediately following each NCR session, a 20-minute

extinction session was conducted. This condition was similar to baseline (i.e., same

room, therapist present, etc.), except that no reinforcers were delivered.

Evaluation of the mechanism underlying the effectiveness of NCR was based on
cumulative records of minute-by-minute patterns of responding during the NCR session

and extinction session that immediately followed it. If the individual was satiated to the

reinforcer during NCR, the Post-NCR (extinction) period would result in deprivation of









the reinforcer, which may lead to a temporary increase in responding during the Post-

NCR (extinction) session (Figure 4, left panel). Alternatively, if extinction was

responsible for behavioral suppression during NCR, responding should continue to

remain low during the Post-NCR (extinction) session because the contingency (i.e.,

extinction) for the problem behaviors would remain the same across the NCR and Post-

NCR (extinction) sessions (Figure 4, right panel).




16 SATIATION EXTINCTION
14- 14-
Z 12- 12-
--- NCR
10- 10- Extinction
8- 8
S 6- 6-
S 4- 4-
S 2- 2
U 0 0 I II I I
5 10 15 20 25 30 5 10 15 20 25 30
MINUTES

Figure 4. Hypothetical graphs representing cumulative records of within-session patterns
of responding expected from satiation (top panel) and extinction (bottom panel)
effects during NCR.



Results and Discussion

NCR data across sessions for all participants are presented in Figure 5. Lisa

averaged 10.6 SIBs per minute during baseline (top panel). Initially, NCR (FT 5 s) did

not result in a reduction of SIB. It was possible that the dense NCR schedule resulted in

adventitious reinforcement of SIB. Therefore, an omission contingency, in which

reinforcer delivery was deleted if SIB immediately preceded the scheduled delivery, was








implemented in sessions 10 through 12. Lisa averaged 0.7 SIBs per minute during the
last 9 NCR sessions (without the omission contingency).
NCR was effective in immediately decreasing SIB for the other three
participants. Julia averaged 3.9 SIBs per minute (second panel). Following the
introduction of NCR, SIB decreased to near zero levels (-a = .03). Susan and Matt
averaged 15.0 and 6.4 destructive responses per minute, respectively (third and fourth
panels). NCR decreased responding to 4.4 and 0.4 responses per minute, respectively.
For Matt, a brief extinction condition was conducted following the NCR phase (see
Results for further details), which resulted in the continued suppression in responding
(M= 0.04).

Given that conclusions about the effectiveness of NCR for Lisa could not be
made when SIB continued to occur at high rates (i.e., NCR was not effective) or when
the omission contingency was in place (i.e., no longer NCR), Lisa's within-session
evaluation (Figure 6, top panel) was started at session 13 (FT 17 s). During session 13,

responding decreased throughout the NCR session until no SIB was emitted during the
last 4 minutes of the session. Termination of the delivery of reinforcers in the
subsequent Post-NCR (extinction) session was followed by an increase of SIB,
suggesting that satiation was responsible for behavioral suppression during the NCR
session (session 13). Within-session patterns of responding indicative of satiation
continued (e.g., see session 16, FT 128 s) until the NCR schedule was thinned to FT

235 s. At this schedule, the discontinuation of reinforcer delivery during the Post-NCR

(extinction) session did not lead to an increase in responding, suggesting that the
mechanism of effectiveness during NCR had changed to extinction. A similar response
pattern was observed during the replication of the NCR FT 235 s schedule (e.g., session
21). Thus, it appears that the operative mechanism of NCR changed from satiation to
extinction between sessions 16 (FT 128 s) and 17 (FT 235 s).













NONCONTINGENT REINFORCEMENT


20 25 30
SESSIONS


Figure 5. Responses per minute of problem behaviors (SIB for Lisa, Julia, Susan, and
Matt; aggression for Susan) during baseline and the NCR condition.








Figure 6 (middle panel) shows representative samples of the within-session
response patterns for Julia. As previously mentioned, NCR led to the near elimination

of her SIB, which was also evident in her within-session patterns of responding (session

11, FT 15 s). The first and second Post-NCR (extinction) sessions were extended to 30

and 45 minutes during sessions 11 and 12, respectively, to ensure a sufficient period of

deprivation to the reinforcer. However, responding did not increase during these Post-

NCR (extinction) sessions, suggesting that extinction had been responsible for behavioral

suppression during the preceding NCR sessions. Given that responding did not increase

during the extended Post-NCR (extinction) sessions, the length of subsequent Post-NCR

(extinction) sessions was decreased to 20 minutes for the remainder of Julia's sessions.

Furthermore, the NCR schedule remained at FT 15 s to replicate the previous effects of

extinction. SIB remained low throughout the NCR and Post-NCR (extinction) sessions

(e.g., sessions 16 and 22).

Susan's within-session response patterns are shown in Figure 6 (bottom panel).
The NCR schedule during the first NCR session (session 15) was FT 4 s. Responding

in the subsequent Post-NCR (extinction) session showed a sharp increase in destructive

behaviors within the first few minutes following the termination of NCR, suggesting that

satiation was responsible for the effectiveness of NCR. This pattern of responding

representative of satiation is also shown in session 17, although the increase in

responding was more gradual than that observed during the first Post-NCR (extinction)

session. Because of an increase in destructive behaviors during session 21 (not shown),

the NCR schedule was changed to FT 9 s (session 23). Susan's responding during the

subsequent Post-NCR (extinction) session showed that problem behavior did not
increase, suggesting that extinction had been responsible for behavioral suppression

during the preceding NCR session. NCR schedule thinning was discontinued at this

point (FT 9 s), and the effects observed during FT 9 s were subsequently replicated

(e.g., session 26).










Session 13 i Session 16 Session 17 Session 21
V, FT 17s FT 128s FT 235s FT 235s
5 30-





1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25
Session 11 Session 12 Session 16 Session 22
O FF 15s FT 15s FT 15s FT 15s
30-
S20 --*-- NCR
10o- --o-- Extinction


1 5 10 15 20 25 30 35 1 5 10 15 20 25 30 35 40 45 50 1 5 10 15 20 25 1 5 10 15 20 25
Session 15 Session 17 Session 23 Session 26
W 80 F FT 15s FT 9s FT 9s


60 **,
20- | Susan I
^3 '* .. I 000000000000


1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25
MINUTES

Figure 6. Representative samples of within-session response patterns (Lisa, Julia, and Susan) during NCR and Post-NCR
(extinction) sessions.


3-








Figure 7 shows representative samples of Matt's responding during the NCR
and Post-NCR (extinction) sessions. The initial NCR schedule was FT 5 s (session 15).

SIB increased during the subsequent Post-NCR (extinction) session, indicating that

satiation may have been responsible for the efficacy of the preceding NCR session. As

the NCR schedule was thinned, responding continued to increase during the Post-NCR

(extinction) sessions (sessions 24, 27, 29, 31, 35, 39, 41, 42, and 43), suggesting that

satiation was responsible for behavioral suppression during NCR. This pattern

continued through FT 300 s (session 43). At this point, NCR was discontinued and five

extinction sessions were conducted, which resulted in low rates of SIB throughout the

sessions.

The results of this study suggest that the mechanism of effectiveness for NCR

changed from an alteration of establishing operations (i.e., satiation) to extinction as the

NCR schedule was thinned for two participants. That is, Lisa's and Susan's within-

session response patterns during NCR and the Post-NCR (extinction) sessions

suggested that satiation was initially responsible for behavioral suppression during

NCR. However, as the NCR schedule was thinned, the mechanism of effectiveness

during NCR changed to extinction.

It appears that extinction was responsible for the efficacy of NCR from the outset
for Julia. That is, responding continued to remain low throughout the NCR and Post-

NCR (extinction) sessions, even when the period of deprivation (i.e., Post-NCR

[extinction]) was extended to 45 minutes. It is conceivable that the initial NCR schedule

of FT 15 s was not sufficient to produce satiation, and had we used a denser schedule

(e.g., FT 5 s), within-session patterns would have resembled satiation. However, the

delivery of 5 to 10 seconds of attention during the NCR session (FT 15 s) resulted in the

near continuous delivery of attention. Finally, results of Matt's within-session analysis

suggested that satiation was responsible for the effectiveness of NCR throughout the

schedule thinning. We included the final extinction phase to compare with the within-






















04
5
4











3

2
1


50- Session 15 Session 24 Session 27 Session 29 Session 31
FF 5s FT 10s FT20s FT30s FT 60s

20 I





1 5 1 15 20 25 1 5 10 1 20 25 1 5 10 15 2 1 5 10 152025 1 5 10 15 20 25
Session 35 Session 39 Session 41 Session 42 Session 43
S F 75s | FT 100s FT 150s FT 200s Ff 300s
0o-
0 -





1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25 15 10 15 20 251 5 10 15 20 25
0 Session 45 Session 46 Session 47 Session 48 Session 49
TI X EXT EXT TEXT


0-
0-
0-
0-


1 5 10 15 20 25 1


SNCR

I IExtinction
I I I I1 I*1*I t*1 II

5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25 1 5 10 15 20 25
MINUTES


Figure 7. Representative samples of within-session response patterns (Matt) during NCR and Post-NCR (extinction) sessions.








session patterns during the NCR phase. Response patterns during the extinction phase

(i.e., responding low throughout the session) were markedly different from those

obtained during the NCR phase, further adding support to our conclusion that satiation

was responsible for the efficacy of NCR throughout the thinning process. It is

conceivable that, had we continued thinning Matt's NCR schedule, the mechanism of

effectiveness may have changed to extinction. However, given the lengthy phase and the

fact that numerous studies have stopped thinning at FT 5-min, we discontinued the

schedule thinning. Alternatively, the outcome of Matt's within-session evaluation may

have been affected by the different schedule thinning procedure. That is, Matt's NCR

schedule thinning was more gradual than that used with other participants; had we thinned

Matt's NCR schedule based on mean IRTs, the outcome may have been different.

Although the results for Lisa, Nancy, and Matt suggested that satiation was

partly responsible for the effectiveness of NCR, one cannot rule out the possibility that

extinction was a factor in the outcome of NCR. Extinction may have been responsible

for behavioral suppression during the dense NCR session, and the increase in responding

during the subsequent Post-NCR (extinction) session may have been due to a

generalization decrement (Mazur, 1998). That is, there was a change in responding

when the conditions of the test stimuli (i.e., the Post-NCR [extinction] session) became

less and less similar to the conditions of the training stimuli (i.e., the NCR session). A

change from a dense NCR schedule (e.g., FT 10 s) to the Post-NCR (extinction) session

resulted in a greater change in stimulus conditions than a change from a lean NCR

schedule (e.g., FT 5-min) to the Post-NCR (extinction) session. Therefore, there have

been a higher probability that responding would change (i.e., increase) in the Post-NCR

(extinction) session following this large change in stimulus conditions.

Several studies have demonstrated that NCR can be effective without extinction

(e.g., Fischer et al., 1997; Lalli et al., 1997), suggesting that satiation may be responsible
for behavioral suppression under dense NCR schedules. Given that Fischer et al. (1997)








only implemented NCR at very dense FT schedules, little can be said about the

mechanisms underlying the effectiveness of NCR under leaner schedules. It is

conceivable that had Fischer et al. thinned their NCR schedule, there would have been a

deterioration of problem behaviors because the individual would no longer be satiated to

the reinforcer, and problem behaviors would continued to be reinforced (i.e., no

extinction).

Lalli et al. (1997) also demonstrated with one participant that NCR can be
effective without extinction (i.e., problem behavior continues to be reinforced).
However, it took several sessions before NCR produced a suppression in responding.

This gradual reduction in SIB may have been a function of their lean, initial NCR

schedule (FT 120 s). Unlike Fischer et al. (1997), Lalli et al. thinned the NCR schedule

to FT 300 s without the use of extinction, which resulted in continued suppression of

problem behavior. This outcome may be similar to that observed with our fourth
participant, Matt. Matt's response pattern continued to resemble those of satiation as

the NCR schedule was thinned to FT 300 s, suggesting that satiation was responsible

for the effectiveness of NCR throughout the schedule thinning process. Similarly, Lalli

et al. demonstrated that extinction was not a necessary component of NCR for their

participant, even as the schedule was thinned, leading the authors to conclude that

satiation was responsible for behavioral suppression.

This study adds to the existing body of research by demonstrating that NCR
may affect behavior through different mechanisms (satiation and extinction), which may

be related to the density of the NCR schedule. Not only is the identification of

mechanisms underlying the effectiveness of NCR important from a conceptual

viewpoint, but it may also have practical implications. These data suggest that, in some
cases, it may be unnecessary to extinguish the problem behaviors at dense NCR
schedules. This may prove to be beneficial to caregivers and individuals, given that it is

sometimes difficult to implement extinction. For example, the nature of the problem





47

behavior (e.g., severe SIB or aggression) may make ignoring the problem behavior

impossible. Alternatively, there may be limitations to the caregiver's behaviors (e.g.,

physical limitations) that would make extinction impractical. These findings also

suggest that, in some cases, it may be imperative that the problem behaviors are not

reinforced when implementing lean NCR schedules to ensure that NCR remains effective

in suppressing problem behaviors throughout the thinning process. That is, intermittent

reinforcement of the problem behavior may lead to maintenance of the problem behavior

during the NCR schedule thinning process.















GENERAL DISCUSSION
The development of NCR as a function-based intervention for severe behavior

disorders has taken a unique path. Its use in behavior analysis originated as a

contingency-control procedure in basic research, which eventually led to its use as a

control procedure in applied research. The ability of NCR to suppress responding

ultimately led researchers to use it as a treatment for severe behavior disorders.

Current NCR research has focused on procedural variations such as its use with

a variety of behavioral functions (Shore et al., 1997; Vollmer et al., 1995), different

topographies (Lalli, Livezey, & Kates, 1996; Piazza et al., 1998), and its use in

conjunction with other interventions (Piazza, Moes, & Fisher, 1996). Additionally,

recent research has focused on the identification of the mechanisms underlying the

effectiveness of NCR by concurrently reinforcing problem behaviors (Fischer et al.,

1997; Lalli et al., 1997), altering the magnitude of reinforcement (Carr et al., 1998), and

examining its effects on alternative behaviors (Marcus & Vollmer, 1996).

Research on the use of NCR as a treatment for severe behavior disorders

suggests the NCR schedule should be initially dense (Hagopian et al., 1994). This dense

NCR schedule may lead to the momentary alteration of the behavior's establishing

alteration (i.e., satiation). Alternatively, the effectiveness of NCR may be based on

extinction, especially at leaner schedules. In these cases, NCR may lead to reductions of

side effects (e.g., extinction burst) sometimes associated with extinction (Vollmer et al.,

1998).

Research (Fischer et al., 1997; Lalli et al., 1997) also suggests that NCR may be
effective at dense schedules without extinction (i.e., problem behavior continues to be








reinforced). However, extinction should be incorporated into NCR as a method of

eliminating ongoing reinforcement for problem behavior. If extinction cannot be

implemented (e.g., severe aggression or SIB, behavior maintained by automatic

reinforcement, etc.), it may be necessary to use very dense NCR schedules to eliminate

the problem behavior.

Finally, Marcus and Vollmer (1996) suggested that NCR may not affect the

occurrence of alternative responses maintained by the same reinforcers as the problem

behavior. However, closer examination of their data showed that occurrence of the

alternative behavior may have been correlated with gradual thinning of the NCR

schedule. That is, at dense NCR schedules, the alternative response occurred at low

rates; the alternative response did not increase until the NCR schedule was thinned.

Therefore, an alternative explanation may be that dense NCR schedules may interfere

with alternative responding; but, as the schedule is thinned, the likelihood that NCR will

interfere with the alternative response decreases. This hypothesis was supported by

the findings of Goh, Iwata, DeLeon, and Sorenson (1997). They used procedures

similar to those described by Marcus and Vollmer (1996). However, their initial

examination of NCR + DRA was conducted at a dense NCR schedule, which remained

in effect throughout the entire phase. Their results showed that NCR suppressed both

the problem behaviors and the alternative response. During the second NCR + DRA

phase, Goh et al. thinned the NCR schedule and observed increases of the alternative

behavior at NCR values. These findings suggest that one limitation to the use of NCR

as a treatment for severe behavior disorders is that it may interfere with the emission of
more appropriate behaviors that have the same function as the problem behavior.

However, the findings of Goh et al. also suggest that the effects of NCR on the

alternative response may be attenuated by thinning the NCR schedule.

Another limitation of NCR may the possibility of adventitious reinforcement of
the problem behavior. Vollmer et al. (1997) observed an increase in SIB with one








participant as the NCR schedule was thinned. A within-session examination showed

that the participant's responding was closely followed by reinforcer delivery, leading

Vollmer et al. to conclude that behavioral maintenance may have been due to

adventitious reinforcement. The findings of Vollmer et al. are somewhat surprising

because the probability that a reinforcer will follow a response decreases as the NCR

schedule is thinned. One would expect that the risk of adventitious reinforcement would

be greater earlier in the thinning process because the probability that a reinforcer will

follow a response is greater at dense NCR schedules. This may be evident in Lisa's

(Experiment 2 and 3) NCR data. Responding did not initially decrease during the NCR

condition until the schedule was sufficiently lean to produce extinction (Experiment 2)

or until reinforcer delivery no longer followed SIB (Experiment 3). An alternative

explanation to the findings of Vollmer et al. may be that the increase in problem

behaviors was a result of an extinction burst. This is particularly conceivable in light of

the results of Experiment 3.

Finally, future studies may wish to examine the long-term effects of NCR. That
is, most NCR studies have examined NCR during brief (10-15 minutes) sessions.

However, it is unclear what the effects of implementing NCR throughout the entire day

may have on problem behaviors. This may be particularly important when using a

reinforcer that does not maintain the problem behavior (e.g., Fischer et al., 1997; Hanley

et al., 1997). The continuous delivery of these arbitrary reinforcers may eventually lead

to an increase in the problem behavior because the individual may become satiated to the

reinforcer.














REFERENCES


Ayllon, T., & Michael, J. (1959). The psychiatric nurse as a behavioral engineer.
Journal of the Experimental Analysis of Behavior. 2. 323-334.

Bailey, J., & Meyerson, L. (1970). Effect of vibratory stimulation on a
retardate's self-injurious behavior. Psychological Aspects of Disability, 17, 133-137.

Bloxham, G., Long, C. G., Alderman, N., & Hollin, C. R. (1993). The behavioral
treatment of self-starvation and severe self-injury in a patient with borderline
personality disorder. Journal of Behavior Therapy and Experimental Psychiatry. 24.
261-267.

Bushell, D., Jr., Wrobel, P. A., & Michaelis, M. L. (1968). Applying "group"
contingencies to the classroom study behavior of preschool children. Journal of Applied
Behavior Analysis. 1. 55-61.

Carr, J. E., Bailey, J. S., Ecott, C. L., Lucker, K. D., & Weil, T. M. (1998). On
the effects of noncontingent delivery of differing magnitudes of reinforcement. Journal
of Applied Behavior Analysis. 31 313-321.

Catania, C. A. (1998). Learning. Upper Saddle River, NJ: Prentice Hall.

Day, H. M., Homer, R. H., & O'Neill, R. E. (1994). Multiple functions of
problem behaviors: Assessment and intervention. Journal of Applied Behavior
Analysis. 27, 279-289.

Derby, K. M., Fisher, W. W., & Piazza, C. C. (1996). The effects of contingent
and noncontingent attention on self-injury and self-restraint. Journal of Applied
Behavior Analysis. 29 107-110.

Didden, R., Duker, P. C., & Korzilius, H. (1997). Meta-analytic study on
treatment effectiveness for problem behaviors with individuals who have mental
retardation. American Journal on Mental Retardation. 101 387-399.

Edwards, D. D., Peek, V., & Wolfe, F. (1970). Independently delivered food
decelerates fixed-ratio rates. Journal of the Experimental Analysis of Behavior. 14.301-
307.








Favell, J. E., McGimsey, J. F., & Schell, R. M. (1982). Treatment of self-injury
by providing alternate sensory activities. Analysis and Intervention in Developmental
Disabilities. 2, 83-104.

Fischer, S. M., Iwata, B. A., & Mazaleski, J. L. (1997). Noncontingent delivery
of arbitrary reinforcers as treatment for self-injurious behavior. Journal of Applied
Behavior Analysis. 30, 239-249.

Goetz, E. M., Holmberg, M. C., & LeBlanc, J. M. (1975). Differential
reinforcement of other behavior and noncontingent reinforcement as control procedures
during the modification of a preschooler's compliance. Journal of Applied Behavior
Analysis. 8. 77-82.

Goh, H., Iwata, B. A., DeLeon, I. G., & Sorenson, T. J. (1997). Is satiation a
potential limiting factor of NCR and, if so, can it be attenuated?. Unpublished
manuscript.

Hagopian, L. P., Fisher, W. W., & Legacy, S. M. (1994). Schedule effects of
noncontingent reinforcement on attention-maintained destructive behavior. Journal of
Applied Behavior Analysis. 27, 317-325.

Hammond, L. J. (1980). The effect of contingency upon the appetitive
conditioning of free-operant behavior. Journal of the Experimental Analysis of
Behavior. 34. 297-304.

Hanley, G. P., Piazza, C. C., & Fisher, W. W. (1997). Noncontingent
presentation of attention and alternative stimuli in the treatment of attention-maintained
destructive behavior. Journal of Applied Behavior Analysis. 30, 229-237.

Hanley, G. P., Piazza, C. C., Fisher, W. W., Contrucci, S. A., & Maglieri, K. A.
(1997). Evaluation of client preference for function-based treatment packages. Journal
of Applied Behavior Analysis. 30, 459-473.

Hart, B. M., Reynolds, N. J., Baer, D. M., Brawley, E. R., & Harris, F. R.
(1968). Effect of contingent and non-contingent social reinforcement on the cooperative
play of a preschool child. Journal of Applied Behavior Analysis. 1, 73-76.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S.
(1994). Toward a functional analysis of self-injury. Journal of Applied Behavior
Analysis. 27, 197-209. (Reprinted from Analysis and Intervention in Developmental
Disabilities. 2. 3-20, 1982)

Iwata, B. A., Vollmer, T. R., & Zarcone, J. R. (1990). The experimental
(functional) analysis of behavior disorders: Methodology, applications, and limitations.
In A. C. Repp & N. N. Singh (Eds.), Perspectives on the use of nonaversive and








aversive interventions for persons with developmental disabilities, (pp. 301-330).
Sycamore, IL: Sycamore Publishing Co.

Jackson, G. M., Johnson, C. R., Ackron, G. S., & Crowley, R. (1975). Food
satiation as a procedure to decelerate vomiting. American Journal of Mental Deficiency.
80 223-227.

Johnston, J. M., Greene, K. S., Rawal, A., Vazin, T., & Winston, M. (1991).
Effects of caloric level on ruminating. Journal of Applied Behavior Analysis. 24, 597-
603.

Kahng, S., & Iwata, B. A. (1999). A Quantitative Review of Research on the
Treatment of Self-Injurious Behavior. Unpublished manuscript.

Kahng, S., Iwata, B. A., DeLeon, I. G., & Worsdell, A. (1997). Evaluation of the
"control over reinforcement" component in functional communication training. Journal
of Applied Behavior Analysis, 30, 267-277.

Knight, M. F., & McKenzie, H. S. (1974). Elimination of bedtime thumbsucking
in home settings through contingent reading. Journal of Applied Behavior Analysis, 7,
33-38.

Lalli, J. S., Casey, S. D., & Kates, K. (1997). Noncontingent reinforcement as
treatment for severe problem behavior: Some procedural variations. Journal of Applied
Behavior Analysis. 30 127-137.

Lalli, J. S., Livezey, K., & Kates, K. (1996). Functional analysis and treatment
of eye poking with response blocking. Journal of Applied Behavior Analysis. 29, 129-
132.

Lalli, J. S., Mace, F. C., Livezey, K., & Kates, K. (1998). Assessment of
stimulus generalization gradients in the treatment of self-injurious behavior. Journal of
Applied Behavior Analysis. 31 479-483.

Lane, H. (1960). Control of vocal responding in chickens. Science. 132. 37-39.

Lattal, K. A. (1972). Response-reinforcer independence and conventional
extinction after fixed-interval and variable-interval schedules. Journal of the
Experimental Analysis of Behavior. 18. 133-140.

Marcus, B. A., & Vollmer, T. R. (1996). Combining noncontingent
reinforcement and differential reinforcement schedules as treatment for aberrant
behavior. Journal of Applied Behavior Analysis. 29, 43-51.

Mazur, J. E. (1998). Learning and Behavior. Upper Saddle River, NJ: Prentice








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

Pelios, L., Morren, J., Tesch, D., & Axelrod, S. (in press). The impact of
functional analysis methodology on treatment choice for self-injurious and aggressive
behavior. Journal of Applied Behavior Analysis.

Piazza, C. C., Contrucci, S. A., Hanley, G. P., & Fisher, W. W. (1997).
Nondirective prompting and noncontingent reinforcement in the treatment of destructive
behavior during hygiene routines. Journal of Applied Behavior Analysis. 30, 705-708.

Piazza, C. C., Fisher, W. W., Hanley, G. P., LeBlanc, L. A., Worsdell, A. S.,
Lindauer, S. E., & Keeney, K. M. (1998). Treatment of pica through multiple analyses
of its reinforcing functions. Journal of Applied Behavior Analysis. 31, 165-189.

Piazza, C. C., Moes, D. R., & Fisher, W. W. (1996). Differential reinforcement
of alternative behavior and demand fading in the treatment of escape-maintained
destructive behavior. Journal of Applied Behavior Analysis, 29, 569-572.

Rast, J., Johnston, J. M., & Drum, C. (1984). A parametric analysis of the
relationship between food quantity and rumination. Journal of the Experimental
Analysis of Behavior. 41, 125-134.

Rast, J., Johnston, J. M., Drum, C., & Conrin, J. (1981). The relation of food
quantity to rumination behavior. Journal of Applied Behavior Analysis. 14, 121-130.

Rescorla, R. A., & Skucy, J. C. (1969). Effect of response-independent
reinforcers during extinction. Journal of Comparative and Physiological Psychology. 67,
381-389.

Shore, B. A., Iwata, B. A., DeLeon, I. G., Kahng, S., & Smith, R. G. (1997). An
analysis of reinforcer substitutability using object manipulation and self-injury as
competing responses. Journal of Applied Behavior Analysis. 30 21-41.

Skinner, B. F. (1948). 'Superstition' in the pigeon. Journal of Experimental
Psychology. 38. 168-172.

Vaughn, M. E., & Michael, J. L. (1982). Automatic reinforcement: An
important but ignored concept. Behaviorism. 10. 217-227.

Vollmer, T. R., Iwata, B. A., Zarcone, J. R., Smith, R. G., & Mazaleski, J. L.
(1993). The role of attention in the treatment of attention-maintained self-injurious
behavior: Noncontingent reinforcement and differential reinforcement of other behavior.
Journal of Applied Behavior Analysis. 26, 9-21.





55


Vollmer, T. R., Marcus, B. A., & Ringdahl, J. E. (1995). Noncontingent escape
as treatment for self-injurious behavior maintained by negative reinforcement. Journal
of Applied Behavior Analysis. 28 15-26.

Vollmer, T. R., Progar, P. R., Lalli, J. S., Van Camp, C. M., Sierp, B. J., Wright,
C. S., Nastasi, J., & Eisenschink, K. J. (1998). Fixed-time schedules attenuate
extinction-induced phenomena in the treatment of severe aberrant behavior. Journal of
Applied Behavior Analysis. 31. 529-542.

Vollmer, T. R., Ringdahl, J. E., Roane, H. S., & Marcus, B. A. (1997). Negative
side effects of noncontingent reinforcement. Journal of Applied Behavior Analysis. 30,
161-164.

Wilder, D. A., Draper, R., Williams, W. L., & Higbee, T. S. (1997). A
comparison of noncontingent reinforcement, other competing stimulation, and liquid
rescheduling for the treatment of rumination. Behavioral Interventions. 12, 55-64.















BIOGRAPHICAL SKETCH

SungWoo Kahng was born in Seoul, South Korea, and emigrated to the United States in
1973. He attended Kalamazoo College in Kalamazoo, MI, and graduated in 1990 with a
degree in psychology. After graduation, he found employment at the Kennedy Krieger

Institute in Baltimore, MD. There he gained his first exposure to behavior analysis and

individuals diagnosed with developmental disabilities. He served as a clinical specialist

on the Neurobehavioral Unit conducting assessments and treatments of severe behavior
disorders. His experience at the Kennedy Krieger Institute led him to seek graduate
training in the field of behavior analysis. In the Fall of 1993, he enrolled in the

experimental analysis of behavior program at the University of Florida to pursue a

doctoral degree in psychology. He worked as a graduate research assistant at the Florida

Center on Self-Injury from 1993 to 1998 under the supervision of Dr. Brian Iwata.

Since Fall 1998, Mr. Kahng has been employed as a program specialist at the New
England Center for Children located in Southborough, MA. Expecting graduation in
summer 1999, he will return to the Kennedy Krieger Institute, where he will continue

clinical work and research in the application of behavior analytic principles to the

treatment of behavior disorders.








I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.


Brian A. Iwata,/thairperson
Professor of Psychology

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.


Marc N. Bran"
Professor of Psychology

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy


Shat A Ellis
Assistant Professor of Psychology

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor f P osoph .


rim thy D. H nberg
tsiate Professor of Psychology

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.


Cecil D. MIercer -
Distinguished Professor of
Special Education








This dissertation was submitted to the Graduate Faculty of the Department of
Psychology in the College of Liberal Arts and Sciences and to the Graduate School and
was accepted as partial fulfillment of the requirements for the degree of Doctor of
Philosophy.


August 1999
Dean, Graduate School















LD

1991










12 20lllllll llil il ill II'lllilliii8 III 1
3 1262 08556 6635




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ELVUD50JX_L5GNZB INGEST_TIME 2013-03-25T15:51:45Z PACKAGE AA00013615_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES