<%BANNER%>

Parametric Evaluation of the Differential Reinforcement of Alternative Behavior Procedure

Permanent Link: http://ufdc.ufl.edu/UFE0022097/00001

Material Information

Title: Parametric Evaluation of the Differential Reinforcement of Alternative Behavior Procedure
Physical Description: 1 online resource (85 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: dra, extinction, reinforcement
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The purpose of this study was to determine if children who exhibit problem behavior would allocate responding according to the differences in reinforcement following appropriate and problem behavior. Relative reinforcer values were manipulated in relation to the parameters of reinforcement using a differential reinforcement of alternative behavior (DRA) procedure without an extinction (EXT) component. A total of 6 individuals diagnosed with developmental disabilities who engaged in severe problem behavior participated. In Experiment I, functional analyses were conducted for all 6 participants to identify the reinforcers for problem behavior. Results showed that problem behavior was sensitive to social positive reinforcement in the form of access to tangible items for two participants and access to attention for one participant, social negative reinforcement in the form of escape from instructional demands for two other participants, and social positive reinforcement in the form of attention and social negative reinforcement in the form of escape from instructional demands for the sixth participant. Two individuals participated in each subsequent experiment. In Experiments II-V, concurrent variable-interval (VI) schedules of reinforcement were in place for problem and appropriate behavior. In Experiment II, problem behavior resulted in access to a relatively lower quality reinforcer, and appropriate behavior in access to a relatively higher quality reinforcer. In Experiment III, problem behavior resulted in a relatively shorter duration of access to a reinforcer and appropriate behavior in a relatively longer duration of access to a reinforcer. In Experiment IV, problem behavior was followed by reinforcement after a delay, and appropriate behavior resulted in immediate access to a reinforcer. In Experiment V, problem behavior resulted in delayed access to a relative lower quality reinforcer for a shorter duration, and appropriate behavior resulted in immediate access to a relatively higher quality reinforcer for a longer duration. Results showed that problem behavior was reduced even though extinction was not in effect.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Vollmer, Timothy R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022097:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022097/00001

Material Information

Title: Parametric Evaluation of the Differential Reinforcement of Alternative Behavior Procedure
Physical Description: 1 online resource (85 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: dra, extinction, reinforcement
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The purpose of this study was to determine if children who exhibit problem behavior would allocate responding according to the differences in reinforcement following appropriate and problem behavior. Relative reinforcer values were manipulated in relation to the parameters of reinforcement using a differential reinforcement of alternative behavior (DRA) procedure without an extinction (EXT) component. A total of 6 individuals diagnosed with developmental disabilities who engaged in severe problem behavior participated. In Experiment I, functional analyses were conducted for all 6 participants to identify the reinforcers for problem behavior. Results showed that problem behavior was sensitive to social positive reinforcement in the form of access to tangible items for two participants and access to attention for one participant, social negative reinforcement in the form of escape from instructional demands for two other participants, and social positive reinforcement in the form of attention and social negative reinforcement in the form of escape from instructional demands for the sixth participant. Two individuals participated in each subsequent experiment. In Experiments II-V, concurrent variable-interval (VI) schedules of reinforcement were in place for problem and appropriate behavior. In Experiment II, problem behavior resulted in access to a relatively lower quality reinforcer, and appropriate behavior in access to a relatively higher quality reinforcer. In Experiment III, problem behavior resulted in a relatively shorter duration of access to a reinforcer and appropriate behavior in a relatively longer duration of access to a reinforcer. In Experiment IV, problem behavior was followed by reinforcement after a delay, and appropriate behavior resulted in immediate access to a reinforcer. In Experiment V, problem behavior resulted in delayed access to a relative lower quality reinforcer for a shorter duration, and appropriate behavior resulted in immediate access to a relatively higher quality reinforcer for a longer duration. Results showed that problem behavior was reduced even though extinction was not in effect.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Vollmer, Timothy R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022097:00001


This item has the following downloads:


Full Text

PAGE 1

PARAMETRIC EVALUATION OF THE DI FFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR PROCEDURE By ELIZABETH S. ATHENS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 ELIZABETH S. ATHENS 2

PAGE 3

ACKNOWLEDGMENTS I would like to thank those i ndividuals who made my dissert ation possible. I would like to express my sincere appreci ation to my mentor throughout graduate school, Dr. Timothy Vollmer, for all of his help and support throughout this process. I would also like to thank the members of my committee, Drs. Lise Abrams Brian Iwata, and Stephen Smith, for their comments. Additional thanks are given to Drs. Michael Kelley and Nathan Call, and Lillie Wilson, who assisted in various aspects of the im plementation, and data analysis for this project. Finally, I would like to thank my parents, Loui se and James Athens, sister, Mary Smiley, and grandmother, Bessie Keding, for their unwav ering love, support, and encouragement. 3

PAGE 4

TABLE OF CONTENTS page LIST OF FIGURES.........................................................................................................................6 ABSTRACT.....................................................................................................................................7 CHAPTER 1 INTRODUCTION................................................................................................................. ...9 Overview....................................................................................................................... ............9 Historical Overview............................................................................................................ ....11 Purpose...................................................................................................................................22 2 EXPERIMENT I: EXPERIMENTAL ANALYSES OF PROBLEM BEHAVIOR...............25 Method....................................................................................................................................25 Participants......................................................................................................................25 Setting........................................................................................................................ ......26 Informed Consent............................................................................................................26 Procedure.........................................................................................................................26 Results and Discussion......................................................................................................... ..29 3 EXPERIMENT II: ANALYSIS OF QUALITY OF REINFORCEMENT............................35 Method....................................................................................................................................35 Participants......................................................................................................................35 Setting........................................................................................................................ ......35 Procedure.........................................................................................................................35 Results and Discussion......................................................................................................... ..39 4 EXPERIMENT III: ANALYSIS OF REINFORCEMENT DURATION..............................47 Method....................................................................................................................................47 Participants......................................................................................................................47 Setting........................................................................................................................ ......47 Procedure.........................................................................................................................47 Results and Discussion......................................................................................................... ..49 5 EXPERIMENT IV: ANALYSIS OF REINFORCMENT DELAY........................................55 Method....................................................................................................................................55 Participants......................................................................................................................55 Setting........................................................................................................................ ......55 Procedure.........................................................................................................................55 Results and Discussion......................................................................................................... ..58 4

PAGE 5

6 EXPERIMENT V: ANALYSIS OF REINFORCMENT QUALITY, DURATION, AND DELAY...................................................................................................................................64 Method....................................................................................................................................64 Participants......................................................................................................................64 Setting........................................................................................................................ ......64 Procedure.........................................................................................................................64 Results and Discussion......................................................................................................... ..66 7 GENERAL DISCUSSION.....................................................................................................70 LIST OF REFERENCES...............................................................................................................79 BIOGRAPHICAL SKETCH.........................................................................................................85 5

PAGE 6

6 LIST OF FIGURES Figure page 2-1. Overall response rates duri ng the functional analysis for Justin, Corey, and Kenneth. A) Responses per min of aggression, disrupti on, and inappropriate sexual behavior for Justin. B) Responses per min of aggressi on and disruption for Corey. C) Responses per min of disruption and aggression for Kenneth.............................................................33 2-2. Overall response rates for Henry, Lana, a nd George during the functional analysis. A) Responses per min of aggre ssion and disruption for Henry. B) Responses per min of aggression for Lana. B) Responses per min of ag gression for George..............................34 3-1. Overall response rates for problem a nd appropriate behavior for Justin................................45 3-2. Kenneths overall response rates for probl em and appropriate behavior. A) Response per minute of problem and appropriate be havior during all assessment phases for Kenneth. B) Responses per min of problem and appropriate behavi or during the final 5 sessions of each condition for Kenneth..........................................................................46 4-1. Justins overall response rates fo r problem and appropriate behavior...................................53 4-2. Lanas overall response rates for problem and appropriate behavior.....................................54 5-1. Coreys overall response rates for problem and appropriate behavior. A) Responses per minute of problem behavior and appropria te behavior during all assessment phases for Corey. B) Responses per min of probl em and appropriate behavior during the final 5 sessions for Corey...................................................................................................62 5-2. Henrys overall responses rates fo r problem and appropriate behavior.................................63 6-1. Georges overall response rates for problem and appropriate behavior.................................69

PAGE 7

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PARAMETRIC EVALUATION OF THE DI FFERENTIAL REINFORCEMENT OF ALTERNATIVE BEHAVIOR PROCEDURE By Elizabeth S. Athens May 2008 Chair: Timothy R. Vollmer Major: Psychology The purpose of this study was to determine if children who exhibit problem behavior would allocate responding according to the differ ences in reinforcement following appropriate and problem behavior. Relative reinforcer values were manipulated in relation to the parameters of reinforcement using a differential reinforcem ent of alternative behavior (DRA) procedure without an extinction (EXT) component. A total of 6 individuals diagnosed with developmental disabilities who engaged in severe problem behavior participated. In Experiment I, functional analyses were c onducted for all 6 particip ants to identify the reinforcers for problem behavior. Results showed that problem behavior wa s sensitive to social positive reinforcement in the form of access to tang ible items for two participants and access to attention for one participant, social negative reinforcement in the form of escape from instructional demands for two other participants, a nd social positive reinforcement in the form of attention and social negative reinforcement in the form of escape from instructional demands for the sixth participant. Two individuals participated in each subsequent experiment. In Experiments II-V, concurrent variable-interval (VI) schedules of reinforcement were in place for problem and appropriate behavior In Experiment II, problem be havior resulted in access to a relatively lower quality reinforcer, and appropr iate behavior in access to a relatively higher 7

PAGE 8

quality reinforcer. In Experiment III, problem beha vior resulted in a relatively shorter duration of access to a reinforcer and appropriate behavior in a relatively longer duration of access to a reinforcer. In Experiment IV, problem behavior was followed by reinforcement after a delay, and appropriate behavior resulted in immediate access to a reinforc er. In Experiment V, problem behavior resulted in delayed access to a relative lower quality reinforcer for a shorter duration, and appropriate behavior resulted in immediate access to a relativel y higher quality reinforcer for a longer duration. Results showed that problem behavior was reduced even though extinction was not in effect. 8

PAGE 9

9 CHAPTER 1 INTRODUCTION Overview Differential reinforcement is a fundamental prin ciple of behavior analysis that has led to the development of a set of procedures used as treatment for problem behavior (Cooper, Heron, Heward, 1987). One of the most frequently used of these procedures is the differential reinforcement of alternative behavior procedure (D RA). Differential reinfo rcement of alternative behavior typically involves withholding reinforcers following problem behavior (extinction, EXT), and providing reinforcers following a ppropriate behavior (D eitz & Repp, 1983). An example of a DRA treatment might involve ignoring the behavior of a child who is throwing his toys, and providing attention following an appropriate behavior, such as when the child says, Lets play. The success of DRA treatments appears to be at least partially dependent on pretreatment identification of the reinforcers maintaining problem behavior. Identifying reinforcers allows EXT procedures to be effectively utilized in that the therapist wi ll know which reinforcer or reinforcers to withhold following the occurre nce of problem behavior (Iwata, Pace, Cowdery, & Miltenberger, 1994). In addition, the same rein forcer can be delivered contingent on the occurrence of an alternative, mo re appropriate response. Under these conditions DRA procedures have been largely successful at reducing probl em behavior (Vollmer & Iwata, 1992). When the response-reinforcer contingency is not severed following problem behavior, however, DRA has been found to be less effective at decreasing problem behavior (Fishe r et al., 1993; Hagopian, Fisher, Sullivan, Acquisto, & LeBlanc, 1998). While EXT is an important and powerful co mponent in the DRA procedure, it is unfortunately not always possible to implement EXT. Of prime concer n are cases in which treatment integrity failures with EXT may be in evitable. For example, in the event of escape-

PAGE 10

maintained behavior, a caregiver may be physica lly unable to prevent escape with a large or combative individual, leading to compromises in treatment integrity. Similarly, it would be difficult to completely withhold reinforcement for behavior maintained by attention in the form of physical contact if physical bl ocking is required to protect the individual or others. For example, if an individuals a ttention-maintained eye gouging is a threat to his eyesight, caregivers may have to intervene in order to protect his vision. It is also likely that in the natural environment, caregivers may not always implemen t EXT procedures accurately (Shores et al., 1993). Given the potential integrity failures a ssociated with implementing EXT, several researchers have examined variations of the DRA procedure that exclude the EXT component. Differential reinforcement procedures have been conceptualized in te rms of a concurrentoperants arrangement (e.g. Fisher et al., 1993; Mace & Roberts, 1993). Concurrent schedules are two or more schedules simultaneously in e ffect. Each schedule independently arranges reinforcement for different responses (Ferster & Skinner, 1957). In this arrangement an individual can engage in either an appropriate response or prob lem behavior to produce the same reinforcer. When the rate and immediacy of rein forcement of the two re sponses are equivalent under this arrangement, the respons e that is less effortful should occur. Research has indicated that the appropriate resp onse is not necessarily the less effortful (Hagopian et al., 1998; Horner & Day, 1991). Thus, several researchers have manipul ated the parameters of reinforcement such that the consequences following appropriate behavior are of a greater value along some dimension than the consequences following prob lem behavior (Piazza, 1999). Variables that have been shown to affect relative response ra tes in concurrent-operants arrangements include response effort or difficulty, a nd reinforcer delay, amount, rate magnitude, and quality (Mazur, 10

PAGE 11

1994). Research evaluating the effectiveness of parametric manipulations of reinforcement during DRA procedures has been mixed, such that additional resear ch is required (Hoch, McComas, Thompson, & Paone, 2002; Piazza et al., 1999). The present study was designed to evaluate DRA when providing some combination of higher quality, long er duration, or more immediate reinforcers for appropriate behavior re lative to some combin ation of lower quality, shorter duration, or less immediate reinforcers for problem behavior. In the event that initial parametric manipulations failed to result in a th erapeutic decrease in problem behavior, further experimental manipulations were conducted in order to increase the difference between reinforcement following appropriate behavior and reinforcement following problem behavior. Historical Overview In one early example of a differential re inforcement procedure without a programmed EXT component, Russo, Cataldo, and Cushing (1981) examined the effects of providing reinforcement following compliance to requests and following problem behavior. In baseline conditions, no consequences were provided following compliance with requests, while problem behavior resulted in the therapist moving away from the participant. In treatment conditions, problem behavior continued to re sult in the therapist moving aw ay; compliance resulted in the delivery of a small piece of food, physical contact, and verbal praise. Results for 3 participants showed that, relative to baseline conditions compliance increased and problem behavior decreased during treatment condi tions. One limitation of the experiment, however, was that Russo et al. did not confirm that escape was in fact a reinforcer for problem behavior. Research on functional analysis methods (see Hanley, Iwata, & McCord, 2003, for a comprehensive review) demonstrates the utility of identifyi ng events as reinforcers before implementing differential reinforcement. In fact, previous research has indicated that a failure to accurately identify the reinforcer or reinforcers maintain ing problem behavior could lead to inaccurate 11

PAGE 12

execution of EXT (Iwata, Pace, Cowdery, & Miltenberger, 1994). Given the lack of identification of the reinforcers maintaining prob lem behavior in the Russo et al. experiment, it cannot be said conclusively that EXT was not in place for problem behavior. In addition, in some cases problem behavior decreased prior to the start of treatment, further limiting the conclusions drawn from the treatment results. In another example of DRA without EXT, Pa rrish, Cataldo, Kolko, Neef, and Egel (1986) used a differential reinforcement procedure to ex amine response covariation. In initial treatment conditions, problem behavior result ed in social reinforcement in the form of statements of concern and disapproval; compliance to verb al requests resulted in no programmed consequences. Under these conditions, problem be havior occurred at a high rate and compliance at a low rate. In a brief test condition for 1 of the 4 participants, however, Parrish and colleagues found that when problem behavior resulted in soci al disapproval and compliance in social praise, occasionally edible reinforcers, and physical attention, problem be havior decreased and compliance increased. Interestingly, these results were similar in comparison to a condition where problem behavior was ignored (extinction), and compliance was reinforced with praise and edible reinforcers. The only noted difference was variability in the rate of compliance when EXT was not programmed following problem beha vior. Unfortunately, th ere was no functional analysis of problem behavior for this participant, and no attempt was made to assess the generality of these findings via replications within or across subjects. More recently, Fisher and colleagues (1993) eval uated a specific type of DRA procedure, functional communication training (FCT), alone and combined w ith EXT, punishment, or both. Functional communication training typically involves placing problem behavior on EXT, and providing the functional reinforc er contingent on an appropr iate communicative response. 12

PAGE 13

Results showed that when FCT was introduced without an EXT or punishment component for problem behavior, the pre-determined goal of a 70% reduction in problem behavior was met with only 1 of 3 participants. It is important to note, however, that each of the participants had prior histories with EXT or punishment. It is unclear how this history may have affected responding in the absence of EXT or punishment. Shirley and colleagues (1997) conducted a similar component analysis of the FCT procedure, but with the additi on of an analysis of training procedures to establish manual signing. Results of this study indicated that when both problem and appropriate behavior resulted in reinforcement, there was initially no decr ease in problem behavior for any of the 3 participants, and no acquisition of the appropriate response. As an additional analysis, following acquisition of the appropriate response and exposure to FCT with EXT, the effect of FCT without EXT was assessed again. Results indicated that, for all 3 participants, the appropriate response was maintained; however, the effects of the intervention on problem behavior were less consistent. For only 1 participant was there excl usive preference for the appropriate response; 1 participant showed an initial increase in problem behavior followed by a decrease; 1 participant showed continuous shifts in responding across alte rnatives. These results suggest that it may be difficult to establish alternative behavior if inappr opriate behavior continues to be reinforced, but that, when established with a history of EXT fo r problem behavior, alte rnative behavior might compete successfully with ongoing contingencies of reinforcement for inappropriate behavior. Hagopian et al. (1998) conducted an extensive replication of the previous examinations of FCT treatment packages to further exam ine the contribution of EXT and punishment components. These researchers found that a pre-determined goal of a 90% reduction in problem 13

PAGE 14

behavior was not achieved with any of 11 partic ipants exposed to FCT without EXT. None of these individuals had a previous experiment al history of either punishment or EXT. Collectively, the majority of research on DRA without EXT has shown a bias in responding toward problem behavior when the rate and immediacy of reinforcement of problem and appropriate behavior are equivalent under a concurrent-operants arra ngement. This research characteristically examines DRA without EX T following baseline conditions where problem behavior is reinforced continuously (i.e. on fixed-ratio [FR] 1 sche dules) and appropriate behavior is placed on EXT. It is unclear what effect, if any, this has on the results. Previous research has shown that human beha vior will co-vary based on rate and quality of reinforcement, with responding favoring the alte rnative associated with a highe r reinforcement rate, greater quality of reinforcement, or both (Conger & Killeen, 1974; Mace, McCurdy, & Quigley, 1990; Martens & Houk, 1989; Neef, Mace, Shea, & Shade, 1992). Based on this finding, several researchers have recently manipulated the parameters of reinforcement such that the rate, quality, immediacy, or magnitude of reinforcement is not equivalent across response alternatives during treatment for problem behavior. For example, in a case study on the treatment of multiply controlled problem behavior, Lalli and Casey (1996) examined the affects of reinforcement schedule. Appropriate behavior was reinforced on a FR 1 schedule, and problem behavior on a concurrent variable-ratio (VR) 5 schedule. U nder this concurrent reinforcement schedule, appropriate behavior increased and problem behavior decreased. Unfortunately, the participants problem behavior appeared sensitive to both es cape and attention as reinforcement. During treatment, appropriate behavior produced both reinforcers, and problem behavior produced escape only. Given this, it is not clear which com ponent of the treatment (s chedule or quality of reinforcement) produc ed the results. 14

PAGE 15

Piazza et al. (1997) extended the Lalli and Casey (1996) study by examining the effects of increasing the quality of reinforcement for compliance relative to reinforcement associated with problem behavior. Three individuals whos e problem behavior was sensitive to negative reinforcement (break from tasks) and positive re inforcement (access to tangible items, attention, or both) participated. Piazza and colleagues system atically evaluated the effects of reinforcing appropriate behavior with one, tw o, or three of the reinforcing consequences (a break, tangible items, attention), both when problem behavior produced a break and when it did not (escape extinction). For 2 of the 3 part icipants, appropriate behavior increased and problem behavior decreased when appropriate behavior produced a 30-s break with access to tangible items, and problem behavior only produced a 30-s brea k. The authors suggested that one potential explanation for these findings is that the rela tive rates of appropriate behavior and problem behavior were a function of the relative value of the reinforcement produced by escape. It is unclear, however, whether the intervention would be effective with individuals whose problem behavior was sensitive to only one type of reinforcement. Piazza et al. (1999) assessed FCT without EXT when a more effective reinforcer followed appropriate behavior relative to th e reinforcer following pr oblem behavior. This analysis took place with 1 participant engaging in problem behavior maintained by attention in the form of verbal reprimands. As part of the initial analysis, the reinforcing properties of two types of attention (e.g. physical and verbal ) were examined in a concurrent operants arrangement. This analysis showed physical atten tion was a more effective reinforcer than verbal attention. Piazza and colleagues then examined an FCT without EXT procedure that consisted of delivery of a verbal reprimand contingent on prob lem behavior, and physical attention or praise contingent on appropriate behavior. Results indicated that when praise and verbal reprimands 15

PAGE 16

were available concurrently for appropriate a nd problem behavior respectively, responding was allocated almost exclusively toward problem behavior. When physical attention followed appropriate behavior and verb al reprimands followed problem behavior, however, responding was allocated almost exclusively toward approp riate behavior. These results supported previous research showing that responding will favor the alternative associated with a higher quality reinforcer. Based on the results, the authors suggested EXT may be a critical treatment component only when the alternative forms of rein forcement used in treatment do not effectively compete with the reinforcement contingent on problem behavior. Unfortunately, these procedures were assessed with only 1 partic ipant, limiting the generality of the results. Vollmer, Roane, Ringdahl and Marcus (1999) exposed participants to DRA that was implemented at less than optimal parameters. Fo r example, during DRA some problem behavior was reinforced, and some appropriate behavior was not reinforced. After a baseline condition, individuals were exposed to a traditional DRA procedure in wh ich problem behavior was never reinforced, and appropriate behavi or was always reinforced. Partic ipants showed a bias toward appropriate behavior at perfect implementati on. Lower levels of treatment implementation eventually reduced treatment efficacy if the sc hedule favored inappropriate behavior, however participants showed a general bi as toward appropriate behavior While this study showed that EXT was not a required component for DRA, it is unclear if the bias toward appropriate responding seen during treatment ch allenges would have been obtained if the participants had not been exposed to perfect treatment implemen tation prior to the introduction of treatment challenges. Lalli and colleagues (1999) showed that EX T was not necessary in the treatment of escape maintained behavior when positive reinforcement was made contingent on appropriate 16

PAGE 17

behavior. They compared the effects of reinforcing appropriate behavior with either positive reinforcement in the form of edible items or ne gative reinforcement in the form of a break from demands with 5 participants whose problem behavior was maintained by escape. These procedures were assessed both w ith and without EXT of problem behavior. Results showed that, for all participants, appropriate behavior increased and probl em behavior decreased when appropriate behavior resulted in positive reinforcem ent rather than negative reinforcement. These results were achieved w ithout the use of EXT. Worsdell, Iwata, Hanley, Thompson and Kahng (2000) examined the effect of reinforcement rate on response allocation. Five individuals whose problem behavior was reinforced by social positive reinforcement were first exposed to a baseline FCT condition in which both problem and appropriate behavior were reinforced on fixed-ratio (FR) 1 schedules. During subsequent FCT conditions reinforcem ent for problem behavior was made more intermittent (e.g. FR 2, FR 3, FR 5, etc.), while appropriate behavior continued to be reinforced on a FR 1 schedule. Results indicated that a sh ift in response allocati on from problem behavior to appropriate behavior occurred with 1 of the participants during the initial baseline FCT condition. The remaining participants showed sh ifts in response allocation to appropriate behavior as the schedule of reinforcement for pr oblem behavior became more intermittent. It was noted, however, that for at least 2 of the participants, small errors in reinforcement appeared to compromise treatment effects. There were seve ral limitations to this research. For example, reinforcement rate was faded in the same order for each participant. Given this, reductions in problem behavior may be due in part to sequence effects. In addition, the reinforcement schedule was thinned to FR 20 for 2 individuals. For th ese 2 participants, reinforcement for problem behavior was rarely contacted. The schedule in these cases may have been functionally more 17

PAGE 18

equivalent to EXT than intermittent reinforcement. Overall, these results suggest that individuals might acquire alternative responses during FCT in spite of inconsistencies in the application of EXT, although even small errors in reinfor cement may compromise treatment effects. Manipulations of magnitude of reinforcement have been also shown to produce shifts in response allocation to the res ponse alternative that provides the greater magnitude of reinforcement, relative to that concurrently available for other altern atives (Catania, 1963). Magnitude of reinforcement can take the form of intensity, number, or duration. For example, Lerman, Kelley, Vorndran, Kuhn, and LaRue (2002) examined the effects of magnitude of reinforcement on positively reinforced screami ng of an adult woman with severe mental retardation. Lerman and colleague s arranged two separate concurrent schedules, both of which programmed EXT for screaming. In one arrange ment, a mand produced 10-s access to the functional reinforcer (toys) and in the other, the mand produced 60-s access to toys. The results demonstrated equivocal mands across the two sc hedule arrangements, but fewer occurrences of screaming in the schedule in which mands prod uced the greater magnitude of reinforcement. Similar findings have been found when magnitude and quality of reinforcement are manipulated during play activities; that is the dimensions of magnitude and quality of reinforcement influenced choice responding in favo r of playing near a peer or si bling rather than playing alone when the reinforcement favored playing near peers or siblings (Hoch, McComas, Johnson, Faranda, & Guenther, 2002). Hoch, and colleagues (2002) conducted a parame tric analysis of DR A without EXT with 3 participants who engaged in problem behavior maintained by negative reinforcement in the form of escape from tasks. These researcher s found that when both problem and appropriate behavior produced a break from tasks, problem behavior for each participant occurred at high 18

PAGE 19

rates and few tasks were completed. In contrast when problem behavior produced a break and task completion produced both a break and access to preferred activities, problem behavior was eliminated and task completion increased. The effects were maintained when the response requirement was increased and the reinforcemen t schedule was thinned. One limitation to the design, however, is the order of experimental conditions. Specifically, the condition where breaks were provided for both pr oblem and appropriate behavior never preceded conditions where breaks followed problem behavior and breaks plus preferred activities followed appropriate behavior. It is unknown whether the escape-alone c ontingency would have been sufficient to decrease problem behavior had par ticipants not had a history of escape with access to preferred activities for task completion. Most recently, Borrero (2006) examined respons e allocation in relati on to experimentally arranged reinforcement rates. Participants we re 3 individuals whose problem behavior was multiply controlled by both social positive reinforcement (access to tangibles or attention) and social negative reinforcement (escape from demands). Throughout the experiment, concurrent schedules of reinforcement were in place for both problem behavior and appropriate behavior. In some conditions, reinforcement schedules favored problem behavior, in others, reinforcement schedules favored appropriate be havior. Results showed that re lative rates of responding on both alternatives approximated the relative rates of reinforcement available for each response. For all participants, however, full trea tment evaluations of DRA with EXT were conducted to reduce problem behavior to clinically significant levels. While it was out side the scope of the Borrero study, future research should extend on this and previous investigations to evaluate whether concurrent schedules of reinforcement can be manipulated in such a way as to obtain a therapeutic reduction in inappropriate behavior without the use of EXT. 19

PAGE 20

One component the Borrero (2006) study broug ht to the experimental analyses of responding on concurrent reinforcement schedules during a treatment for problem behavior was an analysis of responding using the matching la w as a conceptual framework. The matching law was first introduced by Herrnstei n (1961), who provided a quantit ative description of responding on concurrent schedules of reinforcement. Gene rally, the matching law states that the relative rate of responding on one alternative will approximate the relative rate of reinforcement provided on that alternative. Baum (1974a) provided an alternative formulation of the matching law, known as the generalized matching law, which acc ounted for deviations from strict matching by incorporating a bias parameter and a sensitivity parameter. The matching law has provided a conceptual framework for addressing differenc es in responding across alternatives, including alternatives where the quality, delay, or durat ion of reinforcement available for each response differs. The matching law has been evaluated in a number of investigations using both nonhumans (Baum, 1974b; Baum, 1979; Belke & Belliveau, 2001; Crowley & Donahoe, 2004; Herrnstein & Loveland, 1975; MacDonall, 1988; McSweeney, Farmer, Dougan, & Whipple, 1986) and humans (e.g., Borrero & Vollmer, 2002; Mace, Neef, Shade, & Mauro, 1994; Martens & Houk, 1989; Neef, Mace, Shea, & Shade, 1992; Oliver, Hall, & Nixon, 1999; Symons, Hoch, Dahl, & McComas, 2003; Vollmer & Bourret, 2000). A number of applied studies have evaluated naturally occurr ing situations using the matching law. Of particular importance to the current study are experimental investigations on the matching law with humans when reinforcement available for each response form is manipulated. Neef et al. (1992) conducted an investigation to demonstrate that (a) human behavior is sensitive to concurrent schedules of reinforcement when reinforcer quality is held constant, as 20

PAGE 21

suggested by the matching law, and (b) the matc hing relation would not occur when reinforcer quality was not equal, and that a bias for th e higher quality reinforcer would occur. The participants were individuals diagnosed with emotional disturba nces and learning difficulties. Before each session, the participant was asked if she preferred to work for nickels or tokens. During each session identical stacks of arithmetic problems were placed in front of the participant; each stack of cards was associated with a variable-interval (VI) schedule of reinforcement (e.g., VI 30-s, VI 120-s), and corr ect responses resulted in reinforcement (e.g., nickels or tokens) delivered according to the schedul e in place for that alternative. Sessions were first conducted to identify the participants sensitiv ity to the VI schedules of reinforcement, and a timer was included to signal the amount of time remaining in the reinforcement interval. Neef and colleagues then evaluated two additional co nditions: (a) equal-quality reinforcers, during which two stacks of cards were presented on c oncurrent VI schedules, and the reinforcers delivered were the same (i.e., either nickels or tokens were delivered for both alternatives), and (b) unequal-quality reinforcers, during which two stacks of card s were presented on concurrent VI schedules, and high-quality re inforcers (nickels) were delive red on the leaner schedule of reinforcement (i.e., VI 120-s) and low-quality rein forcers (tokens) were delivered on the richer schedule of reinforcement (i.e., VI 30-s). For a ll participants, time-allocation matching occurred following the introduction of a timer signaling the reinforcement inte rval. During the equalquality reinforcers condition, matching was obtained, with the time allocated to each response alternative closely approximating the obtained re inforcement from that alternative. During the unequal quality reinforcers condition, matching was not observed, and responding suggested a preference for one of the two alternatives (i.e ., nickels or tokens) fo r two participants, or responding that maximized the number of reinforcer s for that alternative, for one participant. 21

PAGE 22

This study provided support for the applicability of the matching re lation to socially significant human behavior, and highlighted some potentially important considerations, such as biased responding, which may occur if the quality of available reinforcers is not equal. Using the same general procedures described in prior work (i.e., N eef et al., 1992), Neef and colleagues (Mace et al., 1994; Neef, Mace, & Shade, 1993; Neef, Shade, & Miller, 1994) extended the work reported by Neef et al. (1992) and showed that response allocation under concurrent VI schedules was also sensitive to additional reinforcement parameters including reinforcer delay. Because this series of experi ments involved academic behavior of individuals with emotional and learning disabilities, the ge nerality of the matching law was extended to socially significant (appropriate) behavior. Overall, research on the matching law indicates that responding on concurrent schedule s of reinforcement should match the relative reinforcement available for each response. Manipulations of the ra te quality, delay, or duration of reinforcement across response alternatives have been found to result in switches in allocations of responding, and could indicate a treatment for severe proble m behavior. More research, however, must be conducted with individuals who engage in severe problem behavior to extend the generality of the findings. Purpose Extinction is a highly effective component of the DRA procedure. Unfortunately, problem behavior must sometimes be reinforced to avert danger, and, due to integrity failures, problem behavior is likely to be reinforced on some intermittent schedule in the natural environment. Collectively, previous research sugg ests that EXT may not always be a necessary component of differential reinforcement treatment packages targeting problem behavior. There were, however, certain limitations inherent in pr evious investigations. In addition, there is a relative paucity of research in the area of para metric manipulations of reinforcement during DRA 22

PAGE 23

without EXT procedures within a concurrent -operants arrangement. The purpose of this dissertation was to examine manipulations that could be considered in the event that EXT either cannot or will not be implemented. These mani pulations were assesse d with 6 individuals diagnosed with developmental disorders or disabilities who engaged in seve re problem behavior. In Experiment I, functional analyses were conducted for all participants to identify reinforcers for problem behavior. Specifically, conditions were included to evaluate whether problem behavior was sensitive to (a) social posit ive reinforcement, including adult attention or access to preferred tangible items (e.g., toys, ed ible items, etc.), (b) social negative reinforcement, including escape from instructiona l demands or aversive situations (e.g., hygiene tasks, daily living skills, etc.), or (c) automatic reinforcement (e.g., sensory reinforcement, pain alleviation, etc.). In Experiments II-V, equal concurrent VI schedules of reinforcement were introduced for appropriate behavior and probl em behavior (using reinforcer s previously identified in Experiment I). Appropriate behavior was identi fied during descriptive observations of each participant as well as during the functional analysis, and included requests for access to tangible items, attention, escape from demands (e.g., usin g picture cards, sign language, or vocal requests), and compliance with inst ructions. For each participant during baseline, appropriate and problem behavior produced reinforcers identical in relation to quality (Experiment II), duration (Experiment III), delay (Experiment IV), a nd quality, duration, and delay (Experiment V). During parametric manipulations, appropriate beha vior produced access to a reinforcer that was parametrically different from the reinforcer produced by problem behavior. During Experiments II-IV, the purpose was to manipulate a single feature of rein forcement such that reinforcement favored appropriate behavior. In Experiment II, appropriate behavior produced a higher quality 23

PAGE 24

24 reinforcer and problem behavior produced a lo wer quality reinforcer. In Experiment III, appropriate behavior produced a longer duration of access to the reinforcer and problem behavior produced a shorter duration of access to the reinfo rcer. In Experiment IV, appropriate behavior produced immediate access to the reinforcer and problem behavior produced access to the reinforcer after a delay. In Experiment V, th e purpose was to combine several features of reinforcement such that reinforcement favored appropriate behavior. During Experiment V, appropriate behavior produced immediate, longer duration of access to a higher quality reinforcer and problem behavior produced dela yed, shorter duration of access to a lower quality reinforcer. In each experiment, parametric mani pulations continued until problem behavior was reduced to clinically significant levels and appropriate behavior increased.

PAGE 25

CHAPTER 2 EXPERIMENT I: EXPERIMENTAL ANALYSES OF PROBLEM BEHAVIOR Method Participants Six individuals diagnosed with developmental disorders or disabilities who engaged in severe problem behavior participated. These we re the first six indivi duals who engaged in problem behavior sensitive to socially medi ated reinforcement and were admitted to an outpatient clinical unit (Justin, Henry, Corey, Kenneth, Lana), or refe rred for behavioral consultation services at his sc hool (George). Justin was a 7year-old boy diagnosed with attention deficit and hyperactivit y disorder (ADHD). His problem behavior included aggression (hitting, kicking, biting, pinching, scratching, pushing, and head butting other individuals), disruptive behavior (throwing objec ts or hitting the wall), and inappropriate sexual behavior (touching himself or the therapist in a sexual way). Henry was an 8-year-old male diagnosed with autism. His problem behavior included ag gression (hitting and kick ing other individuals) and disruptive behavior (throwi ng objects). Corey was a 9-year-o ld male diagnosed with autism and ADHD. His problem behavior included aggressi on (hitting, biting, sp itting, and kicking other individuals) and disruptive behavior (throwing, tearing, and otherwise destroying objects). Kenneth was a 6-year-old male diagnosed with autism. His problem behavior was disruption (throwing objects) and aggressi on (hitting, scratching, a nd pinching others). Lana was a 5-yearold female diagnosed with autism. Her problem behavior was aggressi on (hitting, ki cking, and scratching others). George was a 10-year-old male diagnosed with autism. His problem behavior was disruption (throwing objects) and aggression (hitting, kicking, and biting other individuals). 25

PAGE 26

Setting Sessions for Justin, Henry, Corey, Kenneth and Lana were conducted on an outpatient clinical unit to which participants were referr ed for the assessment and treatment of problem behavior. Session rooms (3 m by 3 m) were equipped with one-way mirrors and sound monitoring equipment. Each room contained materials necessary for a session, which could include toys, paper, and a wasteb asket. Sessions for George were conducted in a classroom at his elementary school. The classroom contained materials necessary for a session, and general classroom materials such as posters and tables. Informed Consent Prior to beginning each experiment informed consent was obtained for the participants. Participants were considered unabl e to give informed consent due to their disabilities so parents or legal guardians of the particip ant gave consent on his or her beha lf. Caregivers were asked to read and sign an informed consent which had be en approved by an institutional review board and which stated in detail that their child was being asked to participate in a sequence of experiments on treatments for problem behavior. Procedure All sessions were conducted by trained clinicians serving as experimenters. Observers were clinicians who received in-vivo training in behavioral observati on and had previously demonstrated high interobserver agreement (IO A) scores (> 90%) with trained observers. Observers on the outpatient clinical unit were seated behind a one-way mirror. Observers in school were seated out of the dir ect line of sight of the child. Ob servers collected data on desktop or laptop (George) computers that provided real-time data and scor ed events as either frequency (e.g., aggression, disruption, SIB, and screaming) or duration (e.g., delivery of attention, escape from instructions, etc.). Sessions were conducted four to sixteen times each day, five days per 26

PAGE 27

week. Sessions were 10 min in duration and ther e was a minimum of 5 to 10 min break between each session. Stimulus preference assessment. For each participant, prior to the functional analysis, paired stimulus preference assessments were con ducted using procedures de scribed by Fisher et al. (1992) to identify preferred items to be includ ed in the conditions of the functional analysis. A total of 7-10 leisure items were assessed with each participant (e.g., musical keyboard, drawing toys, ball, etc.). Before beginning the assessment the participant was shown the item and allowed brief (i.e., 2-3 min) contact with the item. The participant was then presented with 2 of the leisure items and told that he or she could briefly (i.e., 20-30 s) play with either of the items. The item the participant selected first was scored. Each leisure item was paired with each other leisure item at least twice. Preferred items were considered to be the three items selected most often. Functional analysis. Functional analyses were conducte d using procedures similar to those described by Iwata et al. (1982), and Day, Rea, Schussler, Larsen and Johnson (1988). Four test conditions were compared: (a) attention, (b) tangible, (c) escape, and (d) ignore, to a control condition (play) using a multielelement design for all participants. Consequences for problem behavior were provided contingent on aggression or di sruption (Henry, Corey, Kenneth, Lana, and George), and aggressi on, disruption, or inappropriate sexual behavior (Justin). During the attention condition the participant was provided with preferred tangible items, no demands were presented, and the therapist diverted her at tention to a work task. Contingent on problem behavior brief attention was provided for 30-s and consisted of a reprimand (e.g., Dont do that) followed by the therapist conversing with the participant. This condition was included to determine if problem behavior was reinforced by adult attention. During the tangible condition 27

PAGE 28

the participant was provided with adult atten tion and no demands were present, while the therapist restricted access to pr eferred tangible items. Access to preferred items was provided for 30-s contingent on problem behavior. This c ondition was included to determine if problem behavior was reinforced by access to tangible it ems. During the escape condition the therapist provided instructional demands (e .g., brushing teeth, washing face, combing hair, folding towels) using a three-prompt instructional sequence (Horner & Keilitz, 1975). Contingent on problem behavior, a 30-s break from instruct ions was provided and the task materials were removed. This condition was included to determine if problem behavior was negatively reinforced by escape from instructional demands. Du ring the ignore condition all preferred tangible items were removed and the participant recei ved no attention from the therap ist. There were no programmed consequences for problem beha vior. This condition was includ ed to determine if problem behavior persisted in the absence of programmed social consequences. Finally, during the control condition, the participant had continuous access to preferred tangible items, no demands were present, and adult attention was provided at least every 30 s. There were no programmed consequences for problem beha vior. This condition was included as a point of comparison (control) to the test conditions. Interobserver agreement (IOA). Two independent observers collected data on aggression, disruption, and inapprop riate sexual behavior for a propor tion of functional analyses sessions to assess interobserver agreement (IOA). Ob servations were divided into 10-s bins and the number of observed responses was scored for each bin. The smaller number of observed responses within each bin was divided by the larg er number of observed responses and converted to agreement percentages for frequency meas ures (Bostow & Bailey, 1969). Agreement on the nonoccurrence of behavior within any given bin was scored as 100% agreement. The bins were 28

PAGE 29

then averaged across the session. For duration me asures the smaller number of s was divided by the larger number of s for duration measures (and agreement on the nonoccurrence of behavior within any bin was scored as 100% agreement). The bin data were then averaged across the sessions. For Justin, IOA was scored for 43% of functional analysis sessions and averaged 98% for aggression (range, 87% to 100%), 96% for disruption (range, 85% to 100%), and 100% for inappropriate sexual behavior. For Henry, IOA was scored for 49% of functional analysis sessions and averaged 100% for aggression a nd 99.9% for disruption (range, 99.7% to 100%). For Corey, IOA was scored for 27% of functi onal analysis sessions and averaged 100% for aggression and disruption. For Kenneth, IOA wa s scored for 44% of functional analysis sessions, and averaged 98.3% for aggression (range, 94% to 100%) and 98.9% for disruption (range, 97% to 100%). For Lana, IOA was scored for 32% of functional analysis sessions and averaged 99% for aggression (range, 98.7% to 10 0%). For George, IOA was scored for 30% of functional analysis sessions and averaged 99.5% (range, 98 % to 100%) for aggression and 99 % for disruption (range, 97.9% to 100%). Observer agreement scores throughout 39% of all functional analysis sessions averaged 100% for therapist attention, 99.9% for access to tangible items (range, 99% to 100%), and 100 % for escape from instructions. Results and Discussion Figure 2-1 shows the results of the functional analyses for Justin, Corey, and Kenneth. Panel A of Figure 2-1 shows responses per min (rpm) of problem behavior for Justin. The highest rates of problem behavior occurred in the escape condition with a mean response rate of 13.72 rpm compared to the attention ( M = .38 rpm), tangible ( M = .44 rpm), ignore (M = .78 rpm), and control (M = .14 rpm) conditions. These results suggested that Justins aggression, disruption, and inappropriate se xual behavior were reinforced by escape from instructional demands. Data were collected for aggression, disruption, and inappropr iate sexual behavior 29

PAGE 30

separately and similar results were obtained, ther efore all topographies were combined in this analysis. Although the overall trend in the esca pe condition appears downward, inspection of the data shows that Justin was becoming more e fficient in escape be havior by responding only when demands were presented. Panel B of Figure 2-1 shows th e results of Coreys functiona l analysis. The highest rates of aggression and disruption were observed during the tangible (M = 1.03 rpm) condition when compared to the escape ( M = .07 rpm), attention ( M = .07 rpm), ignore ( M = .07 rpm), and control ( M = 0 rpm) conditions. These results suggested that Coreys aggr ession and disruption were reinforced by access to tangible items. Data were collected for aggression and disruption separately and similar results we re obtained, therefore both topogra phies were combined in this analysis. Panel C of Figure 2-1 shows th e results of Kenneths functi onal analysis. The highest rates of aggression and disruption we re observed during the attention ( M = 2.25 rpm) and escape conditions ( M = 1.54 rpm) when compared to low rates in tangible ( M = .16 rpm) and control ( M = .46 rpm) conditions. These results suggested th at Kenneths aggression and disruption were reinforced by attention and escape from dema nds. Data were collect ed for aggression and disruption separately and similar results were obtained, therefore both topographies were combined in this analysis. Figure 2-2 shows the results of the functi onal analyses for Henry, Lana, and George. Panel A of Figure 2-2 shows responses per min (rpm) of aggression and disruption for Henry. The highest rates of aggressi on and disruption occurred in the escape condition with a mean response rate of 1.75 rpm, as compared to the attention ( M = 0 rpm), tangible ( M = 0 rpm), ignore ( M = .05 rpm), and control (M = 0 rpm) conditions. These resu lts suggested that Henrys 30

PAGE 31

aggression and disruption were reinforced by es cape from instructional demands. Data were collected for aggression and disr uption separately and similar re sults were obtained, therefore both topographies were combin ed in this analysis. Panel B of Figure 2-2 shows the results of Lanas functional an alysis. The highest rates of aggression were observe d during the tangible ( M = 2.65 rpm) condition when compared to the escape (M = .02 rpm), attention ( M = 0 rpm), ignore ( M = 0 rpm), and control (M = 0 rpm) conditions. These results suggested that Lanas aggression was re inforced by access to tangible items. Panel C of Figure 2-2 shows the results of Georges functiona l analysis. The highest rates of aggression and disruption were observed during the attention ( M = 1.53 rpm) condition when compared to the escape ( M = .06 rpm), attention ( M = 0 rpm), and control ( M = 0 rpm) conditions. These results suggested that Georges aggression a nd disruption were reinforced by social attention. Data were collected for aggre ssion and disruption separate ly and similar results were obtained, therefore both topographies were combined in this analysis. In summary, results of Experiment 1 identifi ed the socially mediated reinforcers for the problem behavior exhibited by the five participants. For all but one partic ipant problem behavior was controlled by one form of environmental re inforcer; two individuals engaged in problem behavior reinforced by escape from instructi onal demands, two engaged in problem behavior reinforced by access to tangible items, and one engaged in problem behavior reinforced by therapist attention. One individual engaged in multiply controlled pr oblem behavior; that is, this individual engaged in problem behavior reinfo rced by adult attention and by escape from highly aversive instructional demands. 31

PAGE 32

This experiment was a necessary prerequi site to Experiment II-V. The results of Experiment I provided a basis for each subs equent experiment, during which parametric manipulations were made to the reinforcer s delivered on concurre nt VI schedules of reinforcement following problem and appropriate behavior. Such analyses would not have been possible without identifying the func tion(s) of problem behavior. 32

PAGE 33

Justin 2468101214161820222426 0 10 20 30 40 50 Corey 1234567891011121314151617Responses per minute 0.0 0.5 1.0 1.5 2.0 Kenneth Sessions 5101520253035404 5 0 2 4 6 8 10 Escape Ignore Tangible Attention ControlA C B Figure 2-1. Overall response rates during the functional analysis for Justin, Corey, and Kenneth. A) Responses per min of aggression, disr uption, and inappropriate sexual behavior for Justin. B) Responses per min of aggression and disruption for Corey. C) Responses per min of disruption and aggression for Kenneth. 33

PAGE 34

34 Henry 51 01 52 02 53 03 5 0 1 2 3 4 5 6 Lana 51 01 52 02 53 03 5Responses per minute 0 1 2 3 4 5 6 7 Control Escape Tangible Ignore A B George Sessions 246810121416182022 0 1 2 3 4 Attention C Figure 2-2. Overall response rates for Henry, Lana and George during the functional analysis. A) Responses per min of aggression and di sruption for Henry. B) Responses per min of aggression for Lana. B) Responses per min of aggression for George.

PAGE 35

CHAPTER 3 EXPERIMENT II: ANALYSIS OF QUALITY OF REINFORCEMENT Method Participants Participants were two of the individuals, Justin and Kenneth, in Experiment I. Problem behavior was the same for each participant as in Experiment I and appropriate behavior was also evaluated for each participant. Justins appropri ate behavior was compliance with instructional demands (e.g., tidiness training). Kenneths appropriate behavior was requests (e.g., mands) for attention through the exchange of a picture card. Due to clini cal exigencies, Kenneths escape maintained behavior was addressed outsi de the context of this research. Setting The setting was the same outpatient clinical unit as described for Experiment I. Session rooms (3 m by 3 m) were equipped with one-way mirrors and sound monitoring equipment. Each room contained materials necessary for a session, which could include toys, task materials, a picture card, or some combination of each. Procedure A reversal design was used during this and al l subsequent analyses. All sessions were conducted by trained clinicians serving as e xperimenters. Observers were clinicians who received in-vivo training in behavioral obs ervation and had previously demonstrated high interobserver agreement (IOA) scores (> 90%) with trained observers. Observers were seated behind a one-way mirror. Observers collected data on desktop computers th at provided real-time data and scored events as either frequenc y (e.g., aggression, disruption, inappropriate sexual behavior, compliance, and picture card exchange), or duration (e .g., delivery of attention, escape from instructions, etc.). Sessions were conducted four to sixteen times each day five days per 35

PAGE 36

week. Sessions were 10 min in duration and ther e was a minimum of 5 to 10 min break between each session. Interobserver agreement (IOA). Interobserver agreement was calculated as in Experiment 1. For Justin, IOA was scored for 27% of quality analysis sessions and averaged 98% for aggression (range, 93% to 100%), 98% for disruption (range, 97% to 100%), and 100% for inappropriate sexual behavi or. For Justin, agreement on the occurrence of compliance averaged 96% (range, 93% to 100%). For Kennet h, IOA was scored for 32% of quality analysis sessions and averaged 98.7% for aggression (r ange, 94% to 100%) and 96% for disruption (range, 92% to 100%). For Kenneth, agreement on the occurrence of picture card exchange averaged 98% (range, 96% to 100%). Stimulus preference assessment. Before the quality analysis with Justin a paired stimulus preference assessment was conducted as in Experiment I. Immediately prior to each session in the quality analysis Justin was al so exposed to a multiple stimulus without replacement (MSWO) preference assessment (DeLe on & Iwata, 1996). A total of 4 leisure items were assessed each MSWO. These items were the most preferred items as identified by the previous paired stimulus preference assessmen t. Before beginning the assessment Justin was shown the item and allowed brief (i.e., 2-3 min) c ontact with the item. Justin was then presented with all of the leisure items and told that he coul d play with any of the it ems; the item he selected first was scored. The most preferred item was cons idered to be the one se lected first. Depending on the condition in effect, the first item selected or the first 3 selected items in the MSWO were included in the upcoming experimental session. Reinforcer assessment. Before conducting the quality analysis with Kenneth, a reinforcer assessment was conducted using proce dures described by Piazza et al. (1999). The 36

PAGE 37

reinforcer assessment identified the relative effica cy of two reinforcers in a concurrent operants arrangement. Two sets of toys, identical except for color, were present (i.e. green and orange blocks). The session room was divided in half by blue painters tape; same colored toys were located on opposite sides of the room. During base line the therapist stood in the middle of the room but provided no social in teraction; toy contact and pr oblem behavior resulted in no arranged consequences. Before each contingent attention session, Kenneth was physically guided to interact with toys of each color and provided with the consequence associated with the color. In the first contingent attention phase of the a ssessment, contact with th e green toys resulted in praise (e.g., Good job, Kenneth) a nd contact with the orange toys resulted in reprimands (e.g., Dont do that). Praise was delivered in a high pitched, loud voice w ith an excited tone. Reprimands were delivered in a deeper pitche d loud voice, with a harsh tone. During the second contingent attention phase, the consequences associated with each color of toys were reversed such that green toys were associated with reprimands and orange toys with praise. Continuous reprimands or praise were delivered for the durati on of toy contact, and atte mpts to play with two different-colored toys simultaneously were blocked. Therapists were instructed to remain at a distance of .3 to .5 m from Kenneth and provide no physical contact. Functional Analysis Baseline. The functional analysis baseli ne condition was identical to the functional analysis condition associated with problem behavior during Experiment I. These conditions were the escape condition for Ju stin and the attention condition for Kenneth. During baseline, each instance of problem behavior resulted in de livery of the reinforcer (i.e., escape from instructions for Justin, therap ist attention for Kenneth). No programmed consequences were in place for appropriate behavior; that is, instances of appropriate behavior did not result in access to the reinforcer. 37

PAGE 38

Baseline. Equal concurrent VI schedules of rein forcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. A random number generator selected intervals between 1 and 39 s, with an average interval le ngth of 20 s. The intervals were timed using a computer generated printout and two timers. Th e printout indicated th e current interval for problem and appropriate behavior, and a therapist controlled the timers to time that interval for each behavior. When reinforcement was available fo r a response (i.e., the interval elapsed), and the behavior occurred, an observer signaled the therapist discreetly by tapping on the one-way mirror from the observation room. The first instan ce of behavior following availability of a reinforcer resulted in delivery of the reinforcer (i.e. escape from instructions for Justin, and attention in the form of neutral statements such as I really li ke your outfit today, for Kenneth). After 30 s of reinforcer access, the reinforcer was removed and the timer was reset for that response. Quality Analysis I. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in th e baseline analysis. For Justin, problem behavior produced escape with access to one low-preferred tangible item identified via the paired stimulus preference assessment. Appropriate behavior produced escape with access to one high-preferred tangible item, identified in a pre-session MS WO. For Kenneth, problem behavior produced reprimands (e.g., Dont do that, I really do not like it and you could end up hurting someone. If I see you keep doing this I will be very disappointed), which were indicated by the reinforcer assessment as a less effective form of reinfor cement than social praise Appropriate behavior produced praise (e.g., Good job handing me the car d, I really like it when you hand it to me so 38

PAGE 39

nicely.), a more effective form of reinforcem ent. For both participants, the reinforcer was removed after 30 s of reinforcer access a nd the timer was reset for that response. Quality Analysis II. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in th e baseline condition. In quality analysis II, the difference in the quality of reinforcement availa ble following appropriate behavior relative to problem behavior was increased. For Justin, problem behavior produ ced access to one lowpreferred tangible item, identified via the paired stimulus preference assessment, and a 30 s break from demands. Appropriate behavior produced access to three high-pref erred tangible items, identified in a pre-session MSWO, and a 30s break from demands. For Kenneth, problem behavior produced reprimands. Appropriate behavior produced praise and physical attention (e.g., Good job handing me the card; hugs and tick les). For both particip ants, after 30 s of reinforcer access, the reinforcer was removed and the timer was reset for that response. Results and Discussion Figure 3-1 shows the results for Justin. Res ponses per min of problem and appropriate behavior are displayed for all pha ses. There was an initially high rate of appropriate behavior during the functional analysis base line. The rate of responding de creased after th e first session, however, and ultimately Justin engaged in higher rates of problem behavior ( M = 1.58 rpm) than appropriate behavior ( M = 1.95 rpm). During the subsequent baseline condition, when problem and appropriate behavior were reinforced on equa l concurrent VI schedules Justin continued to engage in higher rates of problem behavior ( M = 2.88 rpm) than appropriate behavior ( M = 1.4 rpm). The quality of reinforcement was then manipul ated such that in quality analysis I, one low quality reinforcer was delivered on a VI 20-s schedule for problem behavior, and one high quality reinforcer was delivered on a VI 20-s sc hedule for appropriate behavior. During this 39

PAGE 40

condition, rates of problem behavior ( M = .94 rpm) decreased and appr opriate behavior increased ( M = 2.21 rpm). The initially therapeu tic results were not sustained however. During the last five sessions problem behavior increased and appropriate behavior decreased. Quality analysis II was subsequently conducted. In this analysis, one lo w quality reinforcer was delivered on a VI 20-s schedule for problem behavior, a nd three high quality reinforcer s were delivered on a VI 20-s schedule for appropriate behavior During this manipulation, lower rates of problem behavior ( M = .55 rpm) than appropriate behavior ( M = 1.39 rpm) were obtained. In the final five sessions of this condition, problem behavior continued to decrease, and appr opriate behavior increased to high rates. During the subsequent baseline reversal, we failed to recapture previous rates of problem and appropriate behavior. Instead, problem behavior occurr ed at a lower rate (1.24 rpm) than appropriate behavior ( M = 4.56 rpm). Despite the failure to replicate the previous baseline condition with respect to appropriate behavior, probl em behavior increased ( M = 1.24 rpm) relative to what was observed in the immedi ately preceding quality of reinforcement manipulation (M = .55 rpm). Following this, we returned to our second quality of reinforcement manipulation and observed a decr ease in problem behavior ( M = .41 rpm) and sustained high levels of appropriate behavior ( M = 3.59 rpm). In the final five sessions of this manipulation, problem behavior occurred at low rates (M = .08 rpm), and appropriate be havior occurred at high rates ( M = 3.7 rpm). Figure 3-2 shows the results for Kenneth. Resp onses per min of problem and appropriate behavior are displayed for all phases. Kenneth en gaged in higher rates of problem behavior (M = 9.05 rpm) than appropriate behavior ( M = .02 rpm) in the functional analysis baseline. In the subsequent baseline condition, Kenn eth continued to engage in hi gher rates of problem behavior ( M = 4.73 rpm) than appropriate behavior ( M = .05 rpm). The quality of reinforcement was then 40

PAGE 41

manipulated such that in quality analysis I, attention in the form of reprimands (e.g., a low quality reinforcer) was delivered on a VI 20-s sc hedule for problem behavior and higher quality reinforcement, in the form of social praise, wa s delivered on a VI 20-s schedule for appropriate behavior. During this condition, rates of problem behavior decreased (M = 2.58 rpm) and appropriate behavior increased ( M = .75 rpm) relative to baseline. A closer analysis of the condition shows that by the last five sessions of the analysis responding was shifting rapidly across response alternatives. Before conducting furt her analyses, during a replication of baseline, high rates of problem behavior ( M = 5.1 rpm), and relatively lower ra tes of appropriate behavior ( M = 1.18 rpm) were obtained. During a subsequent re plication of the first quality analysis and obtained slightly higher ra tes of problem behavior ( M = 2.4 rpm) than appropriate behavior ( M = 2.22), with responding rapidly shifti ng across response alternatives. In a replication of baseline, high rates of problem behavior ( M = 2.11 rpm) and relatively lower ra tes of appropriate behavior ( M = 1.8 rpm) were obtained. Following this replic ation, quality analysis II was conducted such that reprimands were delivered on a VI 20-s schedule for problem be havior, and praise as well as physical attention (e.g., hugs, tic kles) was delivered on a VI 20s schedule for appropriate behavior. During this manipulation, problem decreas ed to rates lower than observed in previous conditions, and appropriate behavi or increased to high rates ( M = 2.31 rpm). In a reversal to baseline, the findings of the previous baseline condition were replicat ed. During the baseline reversal, problem behavior o ccurred at a higher rate ( M = 3.52 rpm) than appropriate behavior ( M = .83 rpm). Following this, the second quality anal ysis condition was repl icated and a similar decrease in problem behavior (M = 1.72 rpm) and increase in appropriate behavior ( M = 1.58 rpm) was obtained. In the final five sessions of this manipulation, problem behavior occurred at 41

PAGE 42

low rates ( M = .1 rpm), and appropriate behavior at high rates ( M = 1.54 rpm). For closer visual inspection of the final five sessions for each condition, refer to Panel B of Figure 3-2. In summary, results of the quality analysis indicated that, for both participants, the relative rates of both problem behavior and approp riate behavior were sensitive to the quality of reinforcement available for each al ternative. This finding replicates the findings of previous investigations on the relative effects of quality of reinforcement on choice responding (Conger & Killeen, 1974; Hoch, McComas, Johnson, Fara nda, & Guenther, 2002; Mace, McCurdy, & Quigley, 1990; Martens & Houk, 1989; Neef, Mace, Shea, & Shade, 1992; Peck et al., 1996). In addition, the interventions successfully decreased levels of problem behavior to clinically acceptable levels. Overall, these findings are im portant in cases where EXT is not an option given the potential for practical constraints or pr ocedural failures that may be associated with EXT procedures. Several additional findings with each partic ipant require comment. In Experiment 2, quality of reinforcement was assessed using pref erence assessments (e.g., Justin) and reinforcer assessments (e.g., Kenneth). Previous researchers in this area have also used preference assessments (Hoch et al., 2002) and reinforcer assessments (Piazza et al., 1999) to assess quality of reinforcement. In the current experiment, however, the ultimate identification of a higher quality reinforcer was made based on a shift in response allocation from problem behavior to appropriate behavior. For both Ju stin and Kenneth, at least two manipula tions of quality of reinforcement were required before a sh ift in response allocation was obtained. For Justin, a greater number and higher qua lity of toys delivered contingent on appropriate behavior relative to the reinfor cement following problem behavior were required before a consistent shift in response alloca tion was obtained. In this case, magnitude of 42

PAGE 43

reinforcement as well as quality of reinforcem ent was manipulated. Ini tially, access to one high quality reinforcer contingent on appropriate behavior and one low quality reinforcer contingent on problem behavior was effective at shifting response allocation toward appropriate behavior. This effect did not maintain, possibly due to satia tion. Within session analysis showed that as the session progressed, Justin stopped playing with the toy and showed decreases in compliance. In addition to this observation, we did not have access to potentially higher quality toys Justin requested (e.g., video game systems, etc). Given our limited access to th ese potentially higher quality single toy items, we in creased the number of preferred toys we provided contingent on appropriate behavior as a way of addressing potential satiation with the toys. For Kenneth, the initial reinforcer assessmen t indicated praise was a more effective reinforcer than reprimands. This finding was obtained in relation to res ponse allocation across different colored toys. During the functional anal ysis in Experiment I, however, reprimands functioned as reinforcers for problem behavior. During the first quality analysis, while problem behavior decreased, there were continuous shifts in response allocation ac ross alternatives each session. Based on anecdotal observa tions between sessions, when Kenneth frequently requested physical attention in the forms of hugs and tickles by guiding the therapists hands around him or to his stomach, physical attention was added to the social praise available following appropriate behavior. We did not have time to assess this combination of attention in a reinforcer assessment, however after this manipulation a shift in response allocation wa s observed. Overall, data from both Justin and Kenneth show that preference and reinforcer assessments contributed to the identification of a higher and lower quality rein forcer. These results indicate the need to empirically assess the preference fo r or quality of reinforcers pr ior to implementing treatments 43

PAGE 44

for problem behavior, and to contin ue manipulations of the paramete rs of reinforcement if initial manipulations fail to result in shifts in response allocation. Another finding from the current experime nt was evidence of relatively weak experimental control. With Justin, for exampl e, we were unable to replicate the rates of appropriate behavior observed in previous baseline conditions in our final reversal to baseline. In the final reversal, responding favor ed appropriate behavior rather than problem behavior. What differed from the first to final baseline c onditions was the intervening history where the reinforcement contingencies favored appropriate behavior. It is possibl e that the effect of reinforcement history influenced the later resu lts. With Kenneth, there was an overall slow change in responding each condition. Without EX T in place, the gradual change should be expected. Responding under intermittent schedules of reinforcement can be more resistant to change (Ferster & Skinner, 1957). The point of comparison for these conditions is to conditions where EXT is not in place (i.e., baseline condit ions where problem behavior is reinforced). Given that, these gradual changes are expected, but represent positive results as behavior decreased despite an absen ce of the EXT component. With Kenneth, this history with variable rein forcement schedules may have resulted in an initial resistance to the effects of each condition. If weak expe rimental control was influenced by histories of reinforcement favoring appropriate behavior and intermittency of reinforcement, these effects may have important implications fo r caregivers. If a treatm ent can be implemented with sufficient procedural integrity to produce a bias in responding toward appropriate behavior relative to problem behavior, subs equent failures in procedural integrity may have less of an effect on behavior. This finding requ ires additional, future research. 44

PAGE 45

Justin Sessions 5101520253035404550Responses per minute 0 2 4 6 8 FA Escape VI VI BL Quality Analysis I VI VI BL Quality Analysis II Quality Analysis II Compliance Problem Behavior Figure 3-1. Overall response rates for problem and appropriate behavior for Justin. 45

PAGE 46

Kenneth Sessions 51015202530354045 0 2 4 6 8 10 12 FA Attn VI VI BL VI VI BL Quality Analysis I VI VI BL Quality Analysis I Quality Analysis II VI VI BL Quality Analysis IIB 102030405060708090Responses per minute 0 2 4 6 8 10 FA Attn VI VI BL VI VI BL VI VI BL Quality Analysis I VI VI BL Quality Analysis II Quality Analysis II Mand Problem BehaviorA Figure 3-2. Kenneths overall response rates for problem and appropriate behavior. A) Response per minute of problem and appropriate be havior during all assessment phases for Kenneth. B) Responses per min of problem and appropriate behavi or during the final 5 sessions of each condition for Kenneth. 46

PAGE 47

47 CHAPTER 4 EXPERIMENT III: ANALYSIS OF REINFORCEMENT DURATION Method Participants Participants were two of the individuals, Justin and Lana, in Experiment I. Problem behavior was the same for each participant as in Experiment I, and appropriate behavior was also assessed for each participant. Justins appropria te behavior was compliance with instructional demands (e.g., tidiness training). Lanas appropriate behavior was requests (i.e., mands) for a toy through the use of the American Sign Language sign for play. Setting The setting was the same outpatient clinical unit as described for Experiment I. Session rooms (3 m by 3 m) were equipped with one-w ay mirrors and sound monitoring equipment. Rooms contained materials necessary for a session, which could include task materials, a toy, or some combination of each. Procedure All sessions were conducted by trained clinicians serving as experimenters. Observers were clinicians who received in-vivo training in behavioral observati on and had previously demonstrated high interobserver agreement (IO A) scores (> 90%) with trained observers. Observers were seated behind a one-way mirror. Observers collected data on desktop computers that provided real-time data and scored events as either frequency (e.g., aggression, disruption, inappropriate sexual behavior, compliance, and mands), or duration (e.g., escape from instructions, delivery of toy, etc.). Sessions were conducted four to sixteen times each day, five days per week. Sessions were 10 min in duration, and there was a minimum of 5 to 10 min break between each session.

PAGE 48

Interobserver agreement (IOA). Interobserver agreement was calculated as in Experiment 1. For Justin, IOA was scored for 27% of duration analysis sessions, and averaged 97% for aggression (range, 94% to 100%), 99% for disruption (range, 98% to 100%), and 100% for inappropriate sexual behavi or. For Justin, agreement on the occurrence of compliance averaged 97% (range, 94% to 100%). For Lana, IOA was scored for 30% of duration analysis sessions, and averaged 99% for aggression (range 98% to 100%). For Lana, agreement on the occurrence of sign lang uage averaged 98% (r ange, 95% to 100%). Stimulus preference assessment. Before the duration analysis, paired stimulus preference assessments were conducted with Lana as was conducted in Experiment 1. Before each duration analysis session, Lana was also exposed to a MSWO preference assessment (DeLeon & Iwata, 1996) as was de scribed in Experiment 2. The ite m selected first in the MSWO was included in the upcoming experimental session. Functional analysis baseline. The functional analysis baseline condition was identical to the condition associated with problem behavior during the functional anal ysis. These conditions varied for each participant, and included the escape condition for Ju stin and the tangible condition for Lana. During baseline, each instance of problem behavior re sulted in delivery of the reinforcer (i.e., escape from instructions for Justin, tangible item for Lana). No programmed consequences were in place for appropriate behavior; that is, instances of appropriate behavior did not result in access to the reinforcer. Baseline. Equal concurrent VI schedules of rein forcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior dur ing the baseline analysis. The intervals were selected and timed as described in Experime nt 2. The first instance of behavior following availability of a reinforcer result ed in delivery of the reinforcer (i.e. escape from instruction for 48

PAGE 49

Justin, tangible item for Lana) for 30 s. After 30 s of reinforcer acce ss, the reinforcer was removed and the timer was reset for that response. Duration Analysis I. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in th e baseline analysis. For Justin, problem behavior produced a 10-s break from instructions. Appropr iate behavior produced a 30-s break from instructions. For Lana, problem behavior pr oduced 10 s of access to a top preferred toy. Appropriate behavior produced 30 s of access to th e top preferred toy. For both participants, after the interval of reinforcer access was complete, the reinforcer was removed and the timer was reset for that response. Duration Analysis II. Justin participated in duration an alysis II. Equal concurrent VI schedules of reinforcement (VI 20 s VI 20 s) we re in place for both problem and appropriate behavior. The intervals were timed and the ther apist signaled in the same manner as in the baseline analysis. Problem be havior produced a 5-s break fr om instructions. Appropriate behavior produced a 45-s break fr om instructions. After the interval of reinforcer access was complete, the reinforcer was removed and the timer was reset for that response. Results and Discussion Figure 4-1 shows the results for Justin. Res ponses per min of problem and appropriate behavior are displayed for all phases. During the functional analysis base line, problem behavior occurred at a high rate ( M = 13.72 rpm), and appropriate beha vior never occurred. In the subsequent baseline cond ition, problem behavior (M = 2.58 rpm) continued to occur at higher rates than appropriate behavior ( M = .28 rpm). In the initial dura tion analysis condition, there was a further decrease in the rate of problem behavior ( M =1.84 rpm). Problem behavior continued to occur, however, at higher rates than appropriate behavior ( M = .64 rpm). Following 49

PAGE 50

this, reinforcement duration was altered by decr easing the reinforcement for problem from 10 s to 5 s, and increasing the duration for appropriate behavior from 30 s to 45 s. Under these conditions (duration analysis II), pr oblem behavior decreased further ( M = 1.16rpm), and appropriate behavior increased ( M = 1.15 rpm). In the last five sessions of this condition, problem behavior decreased to its lowest observed rates. In a reversal to baseline, there was a shift in response allocation such that probl em behavior occurred more frequently ( M = 1.27 rpm) than appropriate behavior ( M = 1.03 rpm). In the subsequent reversal to the second duration analysis, there was a similar d ecrease in problem behavior ( M =.57 rpm) and increase in appropriate behavior ( M = .85 rpm). Responding stabilized in the last five sessions of this condition, with problem behavior remaining low and appropriate behavior remaining high. In a reversal to baseline, however, there was a failure to replicate previous baseline levels of responding. Instead, low rates of problem behavior ( M = .22 rpm) and high rates of appropriate behavior ( M = 1.4 rpm) were observed. Figure 4-2 shows the results for Lana. Res ponses per min of problem and appropriate behavior are displayed for all phases. During the functional analysis base line, problem behavior occurred at high rates (M = 2.48 rpm) and appropriate behavior at low rates ( M = .12 rpm). In the subsequent baseline cond ition problem behavior ( M = 1.68 rpm) continued to occur at higher rates than appropriate behavior ( M = .28 rpm). In the initial dura tion analysis condition, there was a further decrease in the rate of problem behavior ( M =.82 rpm), and an increase in appropriate behavior ( M = 1.46 rpm). In the final four sessions, problem behavior decreased even further ( M = .04 rpm) and appropriate behavior increased (M = 1.95 rpm). Following this, the baseline condition was replicated and there was an observed increase in problem behavior ( M = 1.8 rpm) and appropriate behavior decreased ( M = .9 rpm). Lanas participation was concluded 50

PAGE 51

with a return to the initial duration analysis. During this analysis there was a decrease in problem behavior ( M =.7 rpm) and increase in appropriate behavior ( M = 1.17 rpm). Problem behavior ceased to occur the final five se ssions of this condition, and appropriate behavior remained at relatively higher rates ( M = 1.8 rpm). In summary, results of the duration analysis indicate that, for both participants, the relative rates of both problem behavior and appropriate behavior were sensitive to the reinforcement duration available for each alternat ive. This finding replicates the findings of previous investigations on the effects of re inforcement duration on choice responding (Catania, 1963; Lerman, Kelley, Vorndran, Kuhn, & LaRue, 2002; Ten Eyck, 1970). The interventions also successfully decreased levels of problem be havior to clinically ac ceptable levels. Problem behavior occurred at zero to near zero rates for both Justin and La na by the final four sessions of the analysis. Appropriate behavior increased and ultimately remained at a relatively higher rate. Similar to Experiment II, these findings are impor tant in cases where EXT is not an option given the potential for practical constrai nts or procedural failures that may be associated with EXT procedures. Similar to Experiment II, there was evidence of lack of experimental control with Justin. With Justin, we were unable to recapture rates of problem and appropriate behavior in our final reversal to baseline. Similar to th e previous experiment, this failure to recapture previous rates of responding followed conditions of reinforcement that favored appropriate behavior. These results could indicate a history effect, which, alt hough not the aim of the current study, again has important implications for caregivers. If a tr eatment can be implemented with sufficient procedural integrity to produce a bias in respond ing toward appropriate behavior relative to 51

PAGE 52

problem behavior, subsequent failures in proced ural integrity may have less of a detrimental effect on behavior. Future research into this finding is warranted. One final result of the analysis requires discussion. For Justin, 10 s of access to reinforcement following problem behavior was su fficient to maintain higher rates of problem behavior. This was not the case for Lana, for whom 10 s of access to reinforcement following behavior resulted in a cessation of problem beha vior. This finding highlights the idiosyncratic nature of reinforcement for problem behavior, in dicating a need for systematic analysis of the parameters of reinforcement maintaining problem behavior on an individual basis. 52

PAGE 53

Sessions 102030405060708090Responses per minute 0 2 4 6 8 40.4 FA Escape VI VI BL VI VI BL VI VI BL Duration Analysis II Duration Analysis II Duration Analysis I Problem Behavior Compliance Justin Figure 4-1. Justins overall response rates for problem and appropriate behavior. 53

PAGE 54

Sessions 51015202530354045Responses per minute 0 1 2 3 4 FA Tangible VI VI BL Duration Analysis I Duration Analysis I VI VI BLLana Problem Behavior Mand Figure 4-2. Lanas overall response rates for problem and appropriate behavior. 54

PAGE 55

55 CHAPTER 5 EXPERIMENT IV: ANALYSIS OF REINFORCEMENT DELAY Method Participants Participants were two of the individuals, Corey and Henry, in Experiment I. Problem behavior was the same for each participant as in Experiment I, and appropriate behavior was also assessed for each participant. Coreys appropriate behavior was a vocal verb al request (mand) for his toy (e.g., May I have my toy please.). Henr ys appropriate behavior was the exchange of a picture card as a request (mand) for a break from instructional demands. Setting The setting was the same outpatient clinical unit as described for Experiment I. Session rooms (3 m by 3 m) were equipped with one-w ay mirrors and sound monitoring equipment. Rooms contained materials necessary for a session, which could include task materials, toys, a picture card, or some combination of each. Procedure All sessions were conducted by trained clinicians serving as experimenters. Observers were clinicians who received in-vivo training in behavioral observati on and had previously demonstrated high interobserver agreement (IO A) scores (> 90%) with trained observers. Observers were seated behind a one-way mirror. Observers collected data on desktop computers that provided real-time data and scored even ts as either freque ncy (e.g., aggression and disruption) or duration (e.g., deliv ery of toys, escape from instru ctions, etc.). Sessions were conducted four to sixteen times each day, five days per week. Sessions were 10 min in duration, and there was a minimum of 5 to 10 min break between each session.

PAGE 56

Interobserver agreement (IOA). Interobserver agreement was calculated as in Experiment 1. For Corey, IOA was scored for 30% of delay analysis sessions, and averaged 95% for aggression (range, 89% to 100%), and 96% fo r disruption (range, 95% to 100%). For Corey, agreement on the occurrence of mands averaged 99% (range, 98% to 100%). For Henry, IOA was scored for 31% of delay analysis sessions, and averaged 97% for aggression (range, 94% to 100%), and 92% for disruption (range, 87% to 100%). For Henry, agreement on the occurrence of mands averaged 97 % (range, 93% to 100%). Stimulus Preference Assessment. Before the reinforcement delay analyses, paired stimulus preference assessments were conducted w ith Corey as was conducted in Experiment 1. Immediately prior to each session in the delay analysis, Corey was also exposed to a MSWO preference assessment (DeLeon & Iw ata, 1996) as described in E xperiment 2. The item selected first in the MSWO was included in the upcoming experimental session. Functional analysis baseline. The functional analysis baseline condition was identical to the condition associated with problem behavior during the functional anal ysis. These conditions were the tangible condition for Corey and the escape condition for Henry. During baseline, each instance of problem behavior resulted in delivery of the reinforcer (i.e., tangible item for Corey, escape from instructions for Henry). No programme d consequences were in place for appropriate behavior; that is, instances of appropriate beha vior did not result in acce ss to the reinforcer. Baseline. Equal concurrent VI schedules of rein forcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior dur ing the baseline analysis. The intervals were selected and timed as described in Experime nt 2. The first instance of behavior following availability of a reinforcer resu lted in delivery of the reinforcer (i.e. tangible item for Corey, 56

PAGE 57

escape from instruction for Henry) for 30-s. Afte r 30-s of reinforcer access, the reinforcer was removed and the timer was reset for that response. Delay Analysis I. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in th e baseline analysis. For Corey, problem behavior produced 30 s of access to a preferred toy after a 30-s unsignaled delay. Appropriate behavior produced 30 s of immediate ac cess to the same preferred t oy. For Henry, problem behavior produced a 30-s break from instructions after a 30-s unsignaled delay. Once a delay interval was started, additional instances of pr oblem behavior did not reset the interval. Appropriate behavior produced an immediate 30 s break from instruct ions. If reinforcement was available for appropriate behavior and the be havior occurred during a delay to reinforcement for problem behavior, the delay interval was stopped and reinforcement was immediately delivered for appropriate behavior. For both participants, after the interval of reinforcer access was complete, the reinforcer was removed and the timer was reset for that response. Delay Analysis II. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in ba seline. For Corey, problem behavior produced 30 s of access to a top preferred toy after a 60-s unsig naled delay. Appropriate behavior produced 30 s of immediate access to the top preferred toy. For Henry, pr oblem behavior produced a 30-s break from instructions after a 60-s unsignaled delay. Delay intervals were timed in the same manner described in Delay Analysis I. For both participants, after the interval of reinforcer access was complete, the reinforcer was removed and the timer was reset for that response. 57

PAGE 58

Results and Discussion Figure 5-1 Panel A shows the results for Corey. Responses per min of problem and appropriate behavior are displa yed for all phases. During the functional analysis baseline condition there were high rates of problem behavior ( M = 3.35 rpm) and low rates of appropriate behavior ( M = .05 rpm). During the subsequent (VI VI) baseline condition, Corey continued to engage in higher rates of problem behavior ( M = 3.03 rpm) than appropriate behavior ( M = 1.63 rpm). A replication of the ini tial delay analysis was then c onducted. During this condition, problem behavior continued to occur at a higher rate ( M = 2.16 rpm) than appropriate behavior ( M = 1.98 rpm). Given this, the second delay anal ysis was implemented. During this condition, a slow decrease in problem behavior ( M = 1.99) and increase in appropriate behavior ( M = .91) was observed. By the final five sessions of the second delay analysis condition, problem behavior ceased and appropriate behavi or occurred at a high, steady rate (M = 1.22 rpm). To further establish contro l over responding, a baseline reversal was conducted. During this reversal to baseline there was an incr ease in problem behavior ( M = 3.45 rpm) and decrease in appropriate behavior ( M = 1.8 rpm). In the final reversal to the second delay analysis, problem behavior decreased and appropriate behavior increased. In the fina l five sessions of the condition there was a cessation in problem behavior and high, steady rates of appropriate behavior ( M = 2.52 rpm). To better show the cont rol over behavior in the final se ssions of each phase of the experiment, the last five sessions of each condition with Corey are presen ted in Figure 5-1 Panel B. Figure 5-2 shows the results for Henry. Res ponses per min of problem and appropriate behavior are displayed for all phases. During the functional analysis baseline, Henry engaged in higher rates of problem behavior ( M = 3.4 rpm) than appropriate behavior ( M = .05 rpm). During the subsequent (VI VI) baseline condition, Henry c ontinued to engage in higher rates of problem 58

PAGE 59

behavior ( M = 2.2 rpm) than appropriate behavior ( M = .03 rpm). In the in itial delay analysis, despite a slight decrease in problem behavior, He nry continued to engage in a higher rate of problem behavior ( M = 1.28 rpm) than appropriate behavior ( M = .91). In a reversal to baseline, there was a slight increase in problem behavior from the previous condition ( M = 1.75 rpm) and decrease in appropriate behavior ( M = .55 rpm). Thereafter, the second delay analysis was implemented. During this condition there wa s a decrease in problem behavior ( M = 1.6 rpm) to zero rates the last three sessions. There wa s an increase in appr opriate behavior ( M = 1.6 rpm), to steady rates of 2 per min the last three sessions. To replicate these findings, a reversal to baseline was implemented and increased rates of problem behavior ( M = 1.5 rpm) and decreased rates of appropriate behavior ( M = .7 rpm) were observed. In a reversal to the initial delay analysis, there was a decrease in the overall rate of problem behavior ( M = 1.3 rpm), however this rate was still higher than the rate of appropriate behavior ( M = .96 rpm). A final reversal to baseline was then conducted, and higher rates of problem behavior ( M = 2.05 rpm), than appropriate behavior ( M = .1 rpm) were obtained. Henrys part icipation in this experiment concluded with exposure to the second delay analysis condition. In this condit ion, there was a further decrease in problem behavior ( M = 1.0 rpm) and increase in appropriate behavior ( M = 1.58 rpm). In the last five sessions of this condition, problem behavior ceased and there was an average of 2 responses per minute of appropriate be havior each session. In summary, results of the delay analysis indicate that, for both participants, the relative rates of problem behavior and appropriate behavior were sensitive to the delay to reinforcement following each alternative. These results replicat e the findings of previous investigations on the relative disruptive effects of unsignaled delay to reinforcement (Sizemore & Lattal, 1978; Williams, 1976; Vollmer, Borrero, Lalli, & Daniel 1999). The interventions also successfully 59

PAGE 60

decreased levels of problem behavior to clinic ally acceptable levels. Problem behavior decreased to zero rates for both Corey and Henry by the conc lusion of the analysis. Appropriate behavior increased and remained at a stable rate. Similar to Experiment II and III, these findings are important in cases where EXT is not an option given the potential for pr actical constraints or procedural failures that may be associated with EXT procedures. It is important to note that the delays pr ogrammed were not necessa rily obtained in the experiment. The occurrence of problem behavior in each delay analysis started a timer that, when elapsed, resulted in delivery of reinforcement. It was therefore possible that problem behavior occurred within the delay interval and resulted in shorter delays to reinforcement. In an examination of the data, this rarely occurred with Henry. With Corey, however, this did occur when problem behavior occurred in bursts or at high rate s, and most often occu rred at the start of each delay analysis condition. The high rates of responding resulted in some problem behavior being reinforced after delays s horter than the programmed 30 or 60 s. This could explain the weak experimental control observed at the start of each delay analysis condition with Corey. In a close analysis of the data it was found that at least once per session, the interval between instances of problem behavior was sufficient such that the programmed delay to reinforcement was experienced. With this, experi mental control was eventually ob tained. This control is most easily observed in Figure 5-1 Panel B, which show s the final five sessions of each condition for Corey. Overall this finding highlights a potentia l limitation to the current method. One way to address this limitation would be to add a differen tial reinforcement of ot her behavior component with a resetting reinforcement in terval to the procedure. This component would involve the delivery of a reinforcer after a period of tim e in which no problem behavior occurred. The resetting feature would result in the occurrence of problem behavior within the interval resetting 60

PAGE 61

the interval and therefore delaying reinforcement. This component was not added in the current experiment in part because, with high rate probl em behavior, it could have resulted in a schedule of reinforcement for problem behavior that resembled EXT. Another potential limitation to th e current experiment was the possibility of adventitious reinforcement of chains of problem behavior and appropriate behavior. For example, when appropriate behavior occurred du ring the delay interval for problem behavior, and the VI schedule indicated reinforcement was availabl e for that response, there was immediate reinforcement of appropriate behavior. This re inforcement could have strengthened a chain of problem and appropriate behavior While this did not seem to be a concern in the current experiment, one way to control for this limitati on would be to add a changeover delay (COD). A COD allows a response to be reinforced only if a certain interval of time has passed since the last changeover from the other respon se alternative. The COD coul d result in longer periods of responding at the alternatives and thus greater control by the relative reinfor cement available for those alternatives (Cat ania, 1966). Each of the potential lim itations to the current experiment indicates an avenue of future research on reinforcement delay. The results of Experiments II-IV have been relatively positive, with each experiment showing the independent effects of isolated parametric manipulations. Although generally positive, the changes in behavior were very gr adual. From a treatment standpoint, one might consider combining all the parame tric manipulations to stack the deck in favor of appropriate behavior. This possibility wa s tested in Experiment V. 61

PAGE 62

62 Corey Sessions 51 01 52 02 53 0Responses per minute 0 2 4 6 8 10 FA Tangible VI VI BL Delay Analysis I Delay Analysis II VI VI BL Delay Analysis II VI BL PB VI BL AB Plot 8 Upper specification 0:30 delay PB 0:30 delay AB Plot 10 Upper specification VI BL PB VI BL PB VI BL AB 100110120130140150160170180190200210 0 2 4 6 8 10 12 FA Tang. VI VI BL Delay Analysis I Delay Analysis II VI VI BL Delay Analysis II ill B AProblem Behavior Mand Figure 5-1. Coreys overall respons e rates for problem and appropriate behavior. A) Responses per minute of problem behavior and appropriate behavior during all assessment phases for Corey. B) Responses per min of problem and appropriate behavior during the final 5 sessions for Corey.

PAGE 63

Sessions 10 20 30 40 50 60 70Responses per minute 0 1 2 3 4 5 FA Escape VI VI BL Delay Analysis I VI VI BL Delay Analysis II VI VI BL Delay Analysis I VI VI BL Delay Analysis IIHenry Problem Behavior Mand Figure 5-2. Henrys overall respons es rates for problem and appr opriate behavior for Henry. 63

PAGE 64

CHAPTER 6 EXPERIMENT V: ANALYSIS OF REIN FORCEMENT QUALITY, DURATION, AND DELAY Method Participant The participant was one of the individuals, Geor ge, in Experiment I. Problem behavior was the same for George as in Experiment I, and ap propriate behavior was al so assessed. Georges appropriate behavior was the exch ange of a picture card as a request (mand) for attention. Setting The setting was the same elementary school classroom as described for Experiment I. The classroom contained materi als necessary for a session, which included a picture card and various classroom materials su ch as posters and tables. Procedure All sessions were conducted by trained clinicians serving as experimenters. Observers were clinicians who received in-vivo training in behavioral observati on and had previously demonstrated high interobserver agreement (IO A) scores (> 90%) with trained observers. Observers were seated out of the clients direct line of sight. Observers collected data on laptop computers that provided real-time data and scor ed events as either frequency (e.g., aggression and disruption), or duration (e.g., de livery of toys, escape from inst ructions, etc.). Sessions were conducted four to sixteen times each day, five da ys per week. Sessions were 10 min in duration, and there was a minimum of 5 to 10 min break between each session. Interobserver agreement (IOA). Interobserver agreement was calculated as in Experiment 1. For George, IOA was scored for 25 % of experimental sessions, and averaged 97% for aggression (range, 90% to 100%), and 95% for disruption (range 89% to 100%). 64

PAGE 65

Reinforcer assessment. Before conducting the experiment al analyses with George, a reinforcer assessment was conducted in the same manner as Experiment II using procedures described by Piazza et al. (1999). The reinforci ng efficacy of praise (e.g., Good job, George) and physical contact (e.g. high fives, pats on th e back), was compared with reprimands (e.g., Dont do that) and physi cal contact (e.g. therapist using his hands to block aggression from George). Functional analysis baseline. The functional analysis baseline condition was identical to the condition associated with pr oblem behavior during the func tional analysis. This condition was the attention condition for George. During baseline, each instance of problem behavior resulted in delivery of attention. No programme d consequences were in place for appropriate behavior; that is, instances of appropriate beha vior did not result in acce ss to the reinforcer. Baseline. Equal concurrent VI schedules of rein forcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior dur ing the baseline analysis. The intervals were selected and timed as described in Experiment II. The first instance of behavior following availability of a reinforcer resu lted in delivery of the reinforcer (i.e. immediate social attention and physical contact) for 30-s. After 30-s of re inforcer access, the reinforcer was removed and the timer was reset for that response. Parametric Analysis. Equal concurrent VI schedules of reinforcement (VI 20-s VI 20-s) were in place for both problem and appropriate behavior. The intervals were timed and the therapist signaled in the same manner as in th e baseline analysis. Problem behavior produced 5 s of low-preferred attention in th e form of social disapproval and blocking of aggression after a 10-s unsignaled delay. Appropriate behavior produced 30 s of immediate access to highpreferred attention in the form of social praise and physical attent ion (e.g., high fives, pats on the 65

PAGE 66

back, etc.). Delays to reinforcement were timed in the same manner as described in Experiment IV. After the interval of reinfo rcer access was complete, reinforcement ceased and the timer was reset for that response. Results and Discussion Figure 6-1 shows the results for George. Res ponses per min of problem and appropriate behavior are displayed for all phases. The initia l functional analysis baseline shows sessions for the attention condition. During the functional analysis baseline, Ge orge engaged in higher rates of problem behavior ( M = 2.24 rpm) than appropriate behavior ( M = .02 rpm). George continued to engage in higher rate s of problem behavior ( M = 3.6 rpm) than appropriate behavior ( M = .07 rpm) during the VI VI baseline. In the initial para metric analysis there was a decrease in problem behavior ( M = .07 rpm) and increase in appropriate behavior ( M = .94 rpm). In a reversal to baseline, there was an increase in problem behavior ( M = 1.52 rpm) and decrease in appropriate behavior ( M = .22 rpm). In a reversal to the initial parametric analysis, there was a further decrease in problem behavior (M = .1 rpm) and increase in appropriate behavior ( M = 1.17 rpm). Georges participation in this experiment concl uded with a one month follow-up to evaluate the maintenance of these effects of the parametric an alysis. The last three se ssions of this condition were conducted by Georges teacher. In this c ondition, there was a furthe r decrease in problem behavior ( M = 0 rpm) and increase in appropriate behavior ( M = 1.64 rpm). In summary, results of the parametric analys is indicate that the relative rates of problem behavior and appropriate beha vior were sensitive to the quality, delay, and duration of reinforcement following each alternative. Thes e results replicate the findings of previous investigations on the relative eff ects of manipulation of several pa rameters of reinforcement in a concurrent-operants arrangement (Parrish, Cataldo, Kolko, Neef & Egel, 1986; Piazza et al., 1997; Russo, Cataldo, & Cushing, 1981). The intervention also successfully decreased levels of 66

PAGE 67

problem behavior to clinically acceptable levels. Problem behavior decreased to zero rates for both George by the conclusion of the analysis. Appr opriate behavior increased and remained at a stable rate. Similar to Experiments II-IV, these findings are important in cases where EXT is not an option given the potential for pr actical constraints and integrit y failures associated with EXT procedures. In further comparison to Experiments IIIV, this experiment resulted in clear experimental control; there were rapid change s in response allocation across conditions and consistent replications of res ponding under previous conditions. The changes in behavior seen in Experiments II-IV were very gradual, but the experiments showed the effects of isolated parametric manipulations. Experiment V show ed the effects of combined parametric manipulations. It is possible that the relative inse nsitivity to change associated with intermittent schedules of reinforcement was reduced with the manipulation of several parameters of reinforcement in combination. This fi nding requires additional research. Similar to Experiment IV, it is important to note that the delays programmed were not necessarily obtained in the experiment. The o ccurrence of problem beha vior in each delay analysis started a timer that, when elapsed, resulted in delivery of reinforcement. It was therefore possible that problem behavior occu rred within the delay interval and resulted in shorter delays to reinforcement. In an examination of the data, this rarely occu rred with George. Another potential limitation to th e current experiment was the possibility of adventitious reinforcement or of chaining problem behavior with appropriate behavior Appropriate behavior that occurred during the dela y interval for problem behavior resulted in immediate reinforcement. This reinforcement could have strengthened a chain of problem and appropriate behavior. In an examination of the data, this rarely occurr ed with George. 67

PAGE 68

There were several benefits to this investigation. First, se ssions were conducted in the environment caregivers indicate d the behavior was most problem atic (i.e., most frequent and most severe). Second, both maintenance and general ity of the procedures were assessed in a one month follow-up, with Georges teacher serving as the therapist in several of the sessions. Georges teacher reported that George had a history of attacking peers, making his behavior too severe for her to consistently ignore. His teacher also indicate d that the presence of four other children in the room limited the amount of attention she could deliver following Georges appropriate behavior This was a prime example of a case where EXT of problem behavior and regular attenti on following appropriate behavi or would be a beneficial recommendation, but likely implemented in the natu ral environment with integrity failures. The current procedure indicated an effective treatme nt option that did not require EXT of problem behavior or continuous reinforcem ent of appropriate behavior. In stead, Georges teacher briefly delivered a low quality of reinforcement after a short delay from some of Georges problem behavior. This attention ensured the safety of those around George. Contingent on some of Georges appropriate behavior, however, his teacher immediately delivered high quality attention for a relatively longer duration. Given a choice between these two consequences, George chose to engage in more appropriat e behavior than problem behavior. Given there is only one participant in the current analyses, however, additional research must be conducted in order to determine the generality of these results. 68

PAGE 69

George Sessions 51015202530354045Responses per minute 0 2 4 6 8 FA Attention VI VI BL Quality Duration Delay Analysis I VI VI BL Quality Duration Delay Analysis I 1 Month Follow-Up Quality Duration Delay Analysis I Teacher as therapist Problem Behavior Mand Figure 6-1. Georges overall response rates for problem and appropriate behavior. 69

PAGE 70

70 CHAPTER 7 GENERAL DISCUSSION The current experiments focused on evaluating several parameters of reinforcement and the effects on problem and appropriate behavior with the purpose of shedding light on manipulations that could be considered in the event EXT either cannot or will not be implemented. The experiments were designed to de termine if the manipulation of reinforcement parameters such as the quality, duration, or de lay of reinforcement could result in effective treatments for 6 individuals with developmenta l disorders and/or disabilities who engaged in severe problem behavior. In Experiment 1, functional analyses were conducted for problem behavior exhibited by five individuals diagnosed with developmental diso rders or disabilities to identify reinforcers for problem behavior. For Justin and Henry, probl em behavior was reinforced by escape from instructional demands. For Corey and Lana, prob lem behavior was reinforced by access to tangible items. For George, probl em behavior was reinforced by adult attention. For Kenneth, problem behavior was reinforced by adult attenti on and escape from highly aversive instructional demands. This experiment was a necessary prerequisite to Experiment II-V. The results of Experiment I provided a basis for each subseque nt experiment, during wh ich manipulations were made to the reinforcers delivered on concurre nt VI schedules of reinforcement following problem and appropriate behavior. Such anal yses would not have been possible without identifying the function(s) of problem behavior. In Experiment II-V, analyses of concurrent reinforcement schedules were conducted with independent reinforcement schedules in plac e for both problem beha vior and appropriate behavior. In Experiment II, we evaluated response allocation unde r a concurrent schedule of

PAGE 71

reinforcement when the reinforcement available following appropriate behavior was of a greater quality than that available fo llowing problem behavior. Analyses for Justin during the escape conditions and analyses for Kenneth under attention conditions showed the quality manipulations were effective in decreasing probl em behavior and increasing appr opriate behavior. This finding replicates the findings of prev ious investigations on the re lative effects of quality of reinforcement on choice responding (Conger & Killeen, 1974; Hoch, McComas, Johnson, Faranda, & Guenther, 2002; Mace, McCurdy, & Quigley, 1990; Martens & Houk, 1989; Neef, Mace, Shea, & Shade, 1992; Peck et al., 1996). In Experiment III, we evaluated the effect s of providing longer duration access to the reinforcer/s following appropriate behavior and shorter access following problem behavior. Analyses for Corey during the tangible condi tions and analyses for Henry under escape conditions showed the duration manipulations were effective in decreasing problem behavior and increasing appropriate behavior. Th is finding replicates the findings of previous investigations on the effects of reinforcement duration on c hoice responding (Catania, 1963; Ten Eyck, 1970). It is important to note that the manipulati on of duration of reinforcement can also be described as a manipulation of magnitude of rein forcement. Magnitude of reinforcement can take the form of intensity, number, or duration. For Jus tin in Experiment II, magnitude and quality of reinforcement were manipulated by increasing the number of high-quality toys available following appropriate behavior relative to a stable smaller number of low-quality toys available following problem behavior. In Experiment III, magnitude was manipulated by increasing and decreasing the duration of reinforcement ra nged from 5-s to 45-s across responses and participants, with durations favoring appropriate behavior relativ e to problem behavior. Findings from both investigations supported previous rese arch showing that manipulations of magnitude 71

PAGE 72

of reinforcement can produce shifts in response allocation to the response alternative that provides the greater magnitude of reinforcement, re lative to that concurrently available for other alternatives (Catania, 1963; Hoch et al., 2002; Lerman et al., 2002). In Experiment IV, we evaluated the effect s of delivering more immediate access to the reinforcer/s following appropriate behavior and delayed access following problem behavior. Analyses for Justin during the escape condi tions and analyses for Lana under attention conditions showed the delay manipulations were effective in decreasing problem behavior and increasing appropriate behavior. These results replicate the findings of previous investigations on the relative disruptive effects of unsignaled de lay to reinforcement (Sizemore & Lattal, 1978; Williams, 1976; Vollmer, Borrero, Lalli, & Daniel, 1999). The purpose of Experiments II-IV was to evaluate the effects of manipulating one feature of reinforcement at a time. This leads to the cl inical implication that all features could be manipulated in some way to favor appropriate behavior. Given that, in Experiment V we evaluated the effects of delivering more imme diate, longer duration access to high quality reinforcement following appropriate behavior and delayed, shorter du ration access to lower quality reinforcement following problem behavior Analyses for George during the attention condition showed these manipulations were e ffective in decreasing problem behavior and increasing appropriate behavior. In comparison to Experiments II-IV, Experiment V resulted in more rapid changes in responding. Overall, these results replicate the findings of previous investigations on the re lative effects of increas ing the value of rein forcement along several parameters following appropriate be havior relative to that follow ing problem behavior (Parrish, Cataldo, Kolko, Neef, & Egel, 1986; Piazza et al., 1997; Russo, Cataldo, & Cushing, 1981). 72

PAGE 73

Finally, at the conclusion of each individuals participation in the current investigations, a comprehensive treatment package was developed ( not outlined here), and his caregivers were trained in the treatment package as well as gi ven a set of general pa renting instructions. There are several implications of the result s of Experiment II-V. As has been noted previously, EXT is a highly effective compone nt of the DRA procedure and increases the efficacy of the DRA procedure as a treatment for problem behavior. Unfortunately, it is not always possible to implement EXT. Of prime concern are cases where treatment integrity failures with EXT may be inevitable. The present studies showed the effectiveness of differential reinforcement of appropriate behavior by provi ding some combination of more immediate, longer duration, or higher quality of reinfor cement for appropriate behavior relative to reinforcement for problem behavior. In cases wher e EXT is not feasible, the current studies offer a method of decreasing problem behavior and incr easing appropriate behavior without the use of EXT. For example, if problem behavior is so severe (e.g., severe aggression, head-banging on hard surfaces, etc.) that it is not possible to withhold reinforcement, it may be possible to manipulate other parameters of reinforcement such as duration or quality or reinforcement to favor appropriate behavior. If probl em behavior in the form of severe self-injury is maintained by attention, for example, a caregiv er could briefly attend to the se lf-injury to ensure the safety of the individual. The attention following problem could be of a lower quality than the attention available following appropriate behavior. Problem behavior coul d result in a s hort duration of social attention including monotone statements of concern, and brief phys ical contact to block the self-injury. Appropriate behavior could result in a longer duration of atten tion in the form of praise, smiles, laughter, and physic al attention such as hugs and tickling. Results of the current 73

PAGE 74

analyses suggest these types of manipulations could decrease problem behavior and increase appropriate behavior. Another finding from Experiments II and V implicates the need for future research into history effects. In these experiments there was evidence of carry over from the functional analysis baseline. When concurrent VI schedul es were initially equated, response allocation favored problem behavior, presumably as a re sult of the preceding condition. Also, for both participants in Experiment III, and one in Experi ment IV, there was either a failure to replicate the findings from baseline conditions conducted earlier in the experi ments or a slower change in responding as the experiment progressed, such that responding began to favor appropriate behavior rather than problem behavior The slower change in re sponding observed throughout these experiments is expected because EXT was not in place. A proper comparison for parametric manipulations is conditions where EXT is not in place (i.e., ba seline conditions where problem behavior is reinforced). In such comp arisons, these gradual changes are expected, but represent positive results as problem behavior decreased despite an absence of the EXT component. It is unclear what caused the failure to replicate baseline conditions. There were, however, differences in the intervening histor y between each reversal. In these intervening conditions, the reinforcement contingencies favored more appropriate behavior. If history effects were a causal factor in this lack of experimental control, these history effects may have important implications for therapists and caregivers. If a treatment can be implemented with sufficient procedural integrity to produce a bias in respond ing toward appropriate behavior relative to problem behavior, subsequent failures in proced ural integrity may have less of an effect on behavior. This finding requires additional research. 74

PAGE 75

One potential contribution of the current series of experiments was methodological and common to the use of intermittent schedules of reinforcement. The use of intermittent schedules of reinforcement in the treatment of problem be havior had several benefits. For example, these schedules likely mimic to a degree the schedules of reinforcement in effect in the natural environment. It is unlikely that at home or sc hool, for example, each in stance of behavior is followed by reinforcement. It is likely, however, that variable amounts of appropriate and problem behavior are reinforced. Further, concur rent VI arrangements allow for comparisons to and translations from experiment al work in the matching law. Variable schedules are also less likely than fi xed schedules to force responding toward one response over another. Under fixe d schedules, reinforcer delivery is maximized by an individual favoring one response alternative. Under variable schedules, reinforcer delivery is maximized by an individual varying response al location across alternatives. If re sponding is observed to favor one response alternative over another under a variab le schedule, this woul d indicate a bias in responding that is independent from the schedule of reinforcement. Of particular in terest to the current study, most examinations of respondi ng under concurrent VI schedules show that response rates typically equal the distribution of reinforcements obtained from the schedules of reinforcement (Rider, 1981). This feature of concurrent VI schedules allowed changes in responding in the current experiments to be attributed to changes in the distribution of reinforcement obtained for responding. Another benefit to the use of intermittent rein forcement is its practicality for caregivers. In the natural environment, caregivers may not always implement EXT procedures accurately. They also may fail to implement reinforcement pr ocedures accurately (Shores et al., 1993). For example, Kenneths mom reported that while she could not always ignore problem behavior, she 75

PAGE 76

also could not always attend to appropriate beha vior. Georges teacher gave a similar report. Therefore, it may be important to identify a therapeutic differential reinforcement procedure that does not require constant implementation of EXT or reinforcement for problem or appropriate behavior. The use of concurrent VI schedules in the current experiments allowed for the examination of the effects of failure to ignore every problem behavior an d failure to reinforce every appropriate behavior in a highly controll ed analogue setting. It also allowed identification of the conditions sufficient to see therapeutic ch anges in behavior. While it is impractical to expect caregivers to implement VI schedules, these schedules give some information on the general effects of intermittent reinforcement de livery across problem a nd appropriate behavior. The current series of experiments also addr ess certain limitations inherent in previous investigations, such as control ling for histories of exposure to EXT and identification of the variables maintaining problem behavior. In addition, there is a relative paucity of research in the area of parametric manipulations of reinforcem ent during DRA without EXT procedures within a concurrent-operants arrangement. One limitation of these experiments is th e brevity of the c onditions. In a basic preparation, it may be possible to conduct conditions until meeting a stability criterion (e.g., a difference of less than 5% between data points), however, in a clinical setting, it was not always possible to bring each condition to stability before exposing behavi or to another condition (i.e., Corey and Kenneth). Therefore, the experimental control is not always as strong as would have been ideal. Ideally, for all participants, each condition would have been conducted until meeting stability criteria. A second potential limitation to the current experiment is the difference in obtained versus programmed schedules of reinforcement. Variable interval schedules of reinforcement 76

PAGE 77

involve delivery of a reinforcer for the first re sponse after an average le ngth of time has passed since the last reinforcer. Part icipants did not always respond immediately after the required length of time elapsed, resulting at times in a less dense reinforcement schedule than what was programmed. The differences in obtained vers us programmed reinforcement schedules were neither large nor consistent, however. Our study suggests several areas for future research. This experiment included concurrent schedules of VI 20 s reinforcement for problem and appropriate behavior Future research may conduct similar analyses using concurrent schedule arrange ments based on naturalistic observations. For example, descriptive anal yses (Bijou, Peterson, & Ault, 1968) could be conducted with caregivers and the results could be analyzed using rein forcers identified in a functional analysis (Iwata et al ., 1982) with procedures similar to those described by Borrero, Vollmer, Borrero, and Bourret (2005 ). For example, if descriptiv e analysis data showed that problem behavior was reinforced on average ever y 15 s and appropriate behavior was reinforced on average every 30 s, experimental analyses co uld be designed to mimic these reinforcement rates in an experimental context. The extent to which relative response allocation is similar under descriptive and experimental arra ngements may suggest values of reinforcement parameters that may increase both the acceptability and integrity of treatment implementation by caregivers. Investigations similar to the current experime nt could further explore the parameters of quality, duration, and delay more extensively and sy stematically. Future investigations could also further assess the generality of the matchi ng law in describing response allocation under concurrent VI schedules when th e target is problem behavior a nd parametric manipulations of reinforcement are conducted. 77

PAGE 78

In addition, investigations into the effect of concurrent ma nipulations of the parameters of reinforcement as treatment for problem behavior could be conducted. For example, it may be that the rate of reinforcement can continue to favor problem behavior if several parameters of reinforcement, such as magnitude, quality, and duration, favor appr opriate behavior. The implications for the treatment of severe problem behavior may be significant. This would be the case when problem behavior is so severe (e.g., h ead-banging on hard surfaces, severe aggression, etc.) that it is not possible to withhold reinforcement (i.e., extin ction) for any period of time. In conclusion, the present experiments focu sed on evaluating the treatment effects of providing differential reinforcement to appropr iate and problem beha vior by systematic manipulation of three parameters of reinforcemen t. It was designed to examine the effects of providing greater quality, longer duration, and mo re immediate reinforcement contingent on appropriate behavior relative to lower qualit y, shorter duration, and delayed reinforcement contingent on problem behavior. For 6 individuals with developmen tal delays or disabilities, an effective intervention was developed that reduc ed severe problem behavior and increased appropriate behavior to clinically acceptable levels. 78

PAGE 79

79 LIST OF REFERENCES Baum, W. M. (1974a). Choice in free-ranging wild pigeons. Science, 185 78-79. Baum, W. M. (1974b). On two types of devi ation from the matching law: Bias and undermatching. Journal of the Experimental Analysis of Behavior, 22 231-242. Baum, W. M. (1979). Matching, undermatchi ng, and overmatching in studies on choice. Journal of the Experimental An alysis of Behavior, 32 269-281. Belke, T. W., & Belliveau, J. (2001). The ge neralized matching law described choice on concurrent variable-interval schedu les of wheel running reinforcement. Journal of the Experimental Analysis of Behavior, 75 299-310. Bijou, S. W., Peterson, R. F., & Ault, M. H. (1968). A method to integrate descriptive and experimental field studies at the le vel of data and empirical concepts. Journal of Applied Behavior Analysis, 1 175-191. Borrero, C. S. W. (2006). Concurrent reinfo rcement schedules for problem behavior and appropriate behavior: Experimental applicatio ns of the matching law. Ph. D. dissertation, University of Florida, United StatesFlo rida. Retrieved August 7, 2007, from ProQuest Digital Dissertations database (Publication No. AAT 3248959) Borrero, C. S. W., Vollmer, T. R., Borrero, J. C., & Bourret, J. (2005). A method for evaluating the dimensions of reinforcement in parent-child interactions. Research in Developmental Disabilities, 26, 577-592. Borrero, J. C., & Vollmer, T. R. (2002). An application of the matching law to severe problem behavior. Journal of Applied Behavior Analysis, 35 13-27. Bostow, D.E., & Bailey, J.B. (1969). Modifica tion of severe disruptive and aggressive behavior using brief timeout and reinforcement procedures. Journal of Applied Behavior Analysis, 2, 31-37. Catania, A. C. (1963). Concurrent performan ces: A baseline for the study of reinforcement magnitude. Journal of the Experimental Analysis of Behavior, 6 299-300. Catania, A. C. (1966). Concurrent operants. In W. K. Honig (Ed.), Operant behavior: Areas of research and application (pp. 213-270). Englewood Cliffs, NJ: Prentice Hall. Conger, R., & Killeen, P. (1974). Use of conc urrent operants in small group research. Pacific Sociological Review, 17 399-416. Cooper, J.O., Heron, T.E., & Heward, W.L. (1987) Applied behavior analysis. Upper Saddle River, N.J: Prentice Hall.

PAGE 80

Crowley, M. A., & Donahoe, J. W. (2004). Ma tching: Its acquisition and generalization. Journal of the Experimental An alysis of Behavior, 82 143-159. Day, R. M., Rea, J.A., Schussler, N. G., Larse n, S. E., & Johnson, W. L. (1988). A functionally based approach to the treatment of self-injurious behavior. Behavior Modifications, 12 565-589. Deitz, D.E.D. & Repp, A.C. (1983). Reduc ing behavior through reinforcement. Exceptional Education Quarterly, 3, 34-46. DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of a multiple-stimulus presentation format for assessing reinforcer preferences. Journal of Applied Behavior Analysis, 29, 519-532. Ferster, C. B., & Skinner, B. F. (1957). Schedules of reinforcement New York: Appleton-Century-Crofts. Fisher, W., Piazza, C. C., Bowman, L. G., Hagopi an, L. P., Owens, J. C. & Slevin, I. (1992). A comparison of two approaches for identifying reinforcers for person with severe and profound disabilities. Journal of Applied Behavior Analysis, 25 491-498. Fisher, W., Piazza, C. C., Cataldo, M. F., Harrell, R., Jefferson, G., & Conner, R. (1993). Functional communication training with and without extinction and punishment. Journal of Applied Behavior Analysis, 26, 23-36. Goh, H., & Iwata, B. A. (1994). Behavioral pers istence and variability during extinction of SIB maintained by escape. Journal of Applied Behavior Analysis, 27 173-174. Hagopian, L. P., Fisher, W., Sullivan, M. T., Acquisto, J., & LeBlanc, L. A. (1998). Effectiveness of functional communication training with and without extinction and punishment: A summary of 21 inpatient cases. Hanley, G. P., Iwata, B. A., & McCord, B. E. (2003). Functional analysis of problem behavior: A review. Journal of Applied Behavior Analysis, 36 147-185. Herrnstein, R. J. (1961). Relative and absolute stre ngth of response as a function of frequency of reinforcement. Journal of the Experimental Analysis of Behavior, 4 563-573. Herrnstein, R. J., & Loveland, D. H. (1975). Maximizing and matching on concurrent ratio schedules. Journal of the Experimental Analysis of Behavior, 24 107-116. Hoch, H., McComas, J. J., Johnson, L., Faranda, N., & Guenther, S. L. (2002). The effects of magnitude and quality of reinforcement on choice responding duri ng play activities. Journal of Applied Behavior Analysis, 35 171-181. 80

PAGE 81

Hoch, H., McComas, J. J., Thompson, A. L., & Paone, D. (2002). Concurrent reinforcement schedules: Behavior chance and maintenance without extinction. Journal of Applied Behavior Analysis, 35 155-169. Horner, R. H., & Day, H. M. (1991). The eff ects of response efficiency on functionally equivalent competing behaviors. Journal of Applied Behavior Analysis, 24 719-732. Horner, R. H., & Keilitz, I. (1975). Training mental ly retarded adolescents to brush their teeth. Journal of Applied Behavior Analysis, 8 301-309. Iwata, B. A., Dorsey, M. F., Slifer, K. J., Ba uman, K. E., & Richman, G. S. (1982). Toward a functional analysis of self-injury. Analysis and Intervention in Developmental Disabilities, 3 1-20. Iwata, B. A., Pace, G. M., Cowdery, G. E., & Mi ltenberger, R. G. (1994). What makes extinction work: An analysis of procedural form and function. Journal of Applied Behavior Analysis, 27 131-144. Lalli, J. S., & Casey, S. D. (1996). Treatment of multiply controlled problem behavior. Journal of Applied Behavior Analysis, 29, 391-395. Lalli, J. S., Vollmer, T. R., Progar, P. R., Wr ight, C., Borrero, J., Daniel, D., et al. (1999). Competition between positive and negative reinforcement in the treatment of escape behavior. Journal of Applied Behavior Analysis, 32 285-296. Lerman, D. C., Kelley, M. E., Vorndran, C. M ., Kuhn, S. A. C., & LaRue, R. H. J. (2002). Reinforcement magnitude and respondi ng during treatment with differential reinforcement. Journal of Applied Behavior Analysis, 35 29-48. MacDonall, J. S. (1988). Concurrent variable-ra tio schedules: Implications for the generalized matching law. Journal of the Experimental Analysis of Behavior, 50 55-64. Mace, F. C., McCurdy, B., & Quigley, E. A. (1990) A collateral effect of reward predicted by matching theory. Journal of Applied Behavior Analysis, 23 197-205. Mace, F. C., Neef, N. A., Shade, D., & Mauro, B. C. (1994). Limited matching on concurrentschedule reinforcement of academic behavior. Journal of Applied Behavior Analysis, 24 719-732. Mace, F. C., & Roberts, M. L. (1993). Factors aff ecting selection of behavi oral interventions. In J. Reichle & D. P. Wacker (Eds.), Communicative alternatives to challenging behavior: Integrating functiona l assessment and intervention strategies (pp. 113-134). Baltimore: Paul H. Brooks. 81

PAGE 82

Martens, B. K., & Houk, J. L. (1989). The applicati on of Herrnstein's law of effect to disruptive and on-task behavior of a retarded adolescent girl. Journal of the Experimental Analysis of Behavior, 51, 17-27. Mazur, J. E. (1994). Learning and behavior. Englewood Cliffs, NJ: Prentice Hall. McSweeney, F. K., Farmer, V. A., Dougan, J. D ., & Whipple, J. E. (1986). The generalized matching law as a description of multiple-schedule responding. Journal of the Experimental Analysis of Behavior, 45 83-101. Neef, N. A., Mace, F. C., & Shade, D. (1993). Impulsivity in students with serious emotional disturbance: The interactive effects of reinforcer rate, delay, and quality. Journal of Applied Behavior Analysis, 26 37-52. Neef, N. A., Mace, F. C., Shea, M. C., & Shad e, D. (1992). Effects of reinforcer rate and reinforcer quality on time allocation: Exte nsions of matching theory to educational settings. Journal of Applied Behavior Analysis, 25 691-699. Neef, N. A., Shade, D., & Miller, M. S. (1994). Assessing influential dime nsions of reinforcers on choice in students with se rious emotional disturbance. Journal of Applied Behavior Analysis, 27 575-583. Oliver, C., Hall, S., & Nixon, J. (1999). A molecular to molar analysis of co mmunicative and problem behavior. Research in Developmental Disabilities, 20 197-213. Parrish, J. M., Cataldo, M. F., Kolko, D. J., N eef, N. A., & Egel, A. L. (1986). Experimental analysis of response covariation among compliant and inappropriate behaviors. Journal of Applied Behavior Analysis, 19, 241-254. Peck, S. M., Wacker, D. P., Berg, W. K., Cooper, L. J., Brown, K. A., Richman, D., et al. (1996). Choice-making treatment of young children's severe behavior problems. Journal of Applied Behavior Analysis, 29 263-290. Piazza, C. C., Bowman, L. G., Contrucci, S. A., Delia, M. D., Adelinis, J. D., & Goh, H. (1999). An evaluation of the properties of atten tion as reinforcement for destructive and appropriate behavior. Journal of Applied Behavior Analysis, 32 437-449. Piazza, C. C., Fisher, W. W., Hanley, G. P., Remi ck, M. L., Contrucci, S. A., & Aitken, T. L. (1997). The use of positive and negative reinforcement in the treatment of escapemaintained destructive behavior. Journal of Applied Behavior Analysis, 30 279-298. Rider, D. P. (1981). Concurrent fixed-interval variable-ratio schedules and the matching relation. Journal of the Experimental Analysis of Behavior, 36 317-328. 82

PAGE 83

Russo, D. C., Cataldo, M. F., & Cushing, P. J. (1981). Compliance training and behavioral covariation in the treatment of multiple behavior problems. Journal of Applied Behavior Analysis, 14 209-222. Shirley, M. J., Iwata, B. A., Kahng, S., Mazaleski, J. L., & Lerman, D. C. (1997). Does functional communication training co mpete with ongoing co ntingencies of reinforcement? An analysis during response acquisition and maintenance. Journal of Applied Behavior Analysis, 30 93-104. Shores, R. E., Jack, S. L., Gunter, P. L., Ellis, D. N., DeBriere, T. J., & Wehby, J. H. (1993). Classroom interactions of child ren with behavior disorders. Journal of Emotional and Behavioral Disorders, 1 27-39. Sizemore, O. J., & Lattal, K. A. (1978). Unsignall ed delay of reinforcement in variable-interval schedules. Journal of the Experimental Analysis of Behavior, 30 169-175. Symons, F. J., Hoch, J., Dahl, N. A., & McComas, J. J. (2003). Sequential matching analyses of self-injurious behavior: A case of ove rmatching in the natural environment. Journal of Applied Behavior Analysis, 36 267-270. Ten Eyck, R. L. J. (1970). Effects of rate of reinforcement-time upon concurrent operant performance. Journal of the Experimental Analysis of Behavior, 14 269-274. Volkert, V. M., Lerman, D. C., & Vorndran, C. M. (2005). The effects of reinforcement magnitude on functional analysis outcomes. Journal of Applied Behavior Analysis, 38 147-162. Vollmer, T. R., Borrero, J. C., Lallie, J. S., & Daniel, D. (1999). Evaluating self-control and impulsivity in children with severe behavior disorders. Journal of Applied Behavior Analysis, 32 451-466. Vollmer, T. R., & Bourret, J. (2000). An application of the matching law to evaluate the allocation of twoand three-point shots by college basketball players. Journal of Applied Behavior Analysis, 33 137-150. Vollmer, T. R., & Iwata, B. A. (1992). Different ial reinforcement as treatment for behavior disorders: Procedural and functional variations. Research in Developmental Disabilities, 13, 393-417. Vollmer, T. R., Roane, H. S., Ringdahl, J. E ., & Marcus, B. A. (1999). Evaluating treatment challenges with differential reinforcement of alternative behavior. Journal of Applied Behavior Analysis, 32 9-23. Williams, B. A. (1976). The effects of unsignalled delayed reinforcement. Journal of the Experimental Analysis of Behavior, 26 441-449. 83

PAGE 84

Worsdell, A. S., Iwata, B. A., Hanley, G. P ., Thompson, R. T., & Kahng, S. W. (2000). Effects of continuous and intermittent reinforcem ent for problem behavior during functional communication training. Journal of Applied Behavior Analysis, 33 167-179. 84

PAGE 85

85 BIOGRAPHICAL SKETCH Elizabeth Athens has been interested in beha vior analysis since vol unteering as a research assistant her first year as an undergraduate at the University of Nevada, Reno. She began her graduate studies in behavior analysis at the Un iversity of Florida in 2002, under the supervision of Timothy Vollmer. Elizabeths research interests center around the development and implementation of interventions for children w ho have fallen behind academically or exhibit problem behavior. Elizabeth is sp ecifically interested in the vari ables affecting learning under a discrete trial format, parametric evaluations of common behavioral treatments, and the development of verbal behavior in y oung or developmentally delayed children. During her graduate school career, Elizabeth conducted research in several diverse settings, including the laboratory, clinic, schools, and homes. She is currently studying the effects of treatment integrity failures on acqui sition of complex discri minations, acquisition of verbal behavior, and methods to increase appropr iate behavior while tr eating problem behavior. Following graduation, Elizabeth will seek a position allowing her to contribute to the field of behavior analysis through research, teaching, and clinical application.