Evaluations of Delayed Reinforcement in Children with Developmental Disabilities

MISSING IMAGE

Material Information

Title:
Evaluations of Delayed Reinforcement in Children with Developmental Disabilities
Physical Description:
1 online resource (93 p.)
Language:
english
Creator:
Sy,Jolene Rachel
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology
Committee Chair:
Vollmer, Timothy R
Committee Members:
Devine, Darragh P
Iwata, Brian
Jones, Hazel A

Subjects

Subjects / Keywords:
behavior -- delayed -- developmental -- parametric -- translational
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:
It is commonly recommended that reinforcers be provided immediately after behavior to establish or maintain responding (e.g., Miltenberger, 2008). These recommendations are usually provided in the absence of supporting empirical work, which is unfortunate because several laboratory experiments have found that delayed reinforcement can produce response acquisition and maintenance (e.g., Lattal & Gleeson, 1990). Relatively fewer studies with humans have isolated the effects of unsignaled, delayed reinforcement. Fewer still have evaluated unsignaled, delayed reinforcement in children with developmental disabilities (DD). Thus, the general purpose of the following experiments was to (a) examine whether delayed reinforcement could produce response maintenance in children with DD during an operant arrangement, (b) examine whether delayed reinforcement could produce discrimination acquisition in children with DD, (c) examine some of the variables that may affect the efficacy of delayed reinforcement in children with DD, and (d) examine one variable that may account for situations in which delayed reinforcement prevents acquisition. We found that delayed reinforcement produced response maintenance during three human operant arrangements. In addition, we found that 20-, 30-, and 40-s delays to reinforcement lead to discrimination acquisition for the majority of subjects. The availability of identical responses during the delay did not prevent response maintenance for 2 out of 3 subjects or discrimination acquisition for 2 out of 3 subjects. For 1 out of 2 subjects, discrimination acquisition was hindered when the number of responses targeted during each session was increased. These results suggest that delayed reinforcement can produce acquisition and maintenance under specific conditions. In addition, we found that longer intertrial intervals (time between delivery of a reinforcer and the presentation of the next trial) most likely do not account for cases in which discrimination acquisition fails to occur under conditions of delayed reinforcement.
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.
Statement of Responsibility:
by Jolene Rachel Sy.
Thesis:
Thesis (Ph.D.)--University of Florida, 2011.
Local:
Adviser: Vollmer, Timothy R.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0043092:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 EVALUATIONS OF DELAYED REINFORCEMENT IN CHILDREN WITH DEVELOPMENTAL DISABILITIES By JOLENE RACHEL SY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Jolene Rachel Sy

PAGE 3

3 To my parents, who taught me to value education and enjoy life

PAGE 4

4 ACKNOWLEDGMENTS I thank my mentor, Dr. Tim Vollmer, for his excellent suggestions, valuable feedback, and years of support. I would not be who I was today without his guidance. I also would like to thank my commit tee members: Dr. Brian Iwata, Dr. Darragh Devine, and Dr. Hazel Jones for their comments and suggestions on this project. Fina lly, I would like to thank Dr. Jesse Dallery for his feedbac k on this project I could not have gotten here without support from my dear family and friends. I thank them for the years and years of love and understanding.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ .......... 7 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Recomme ndations About the Necessity of Reinforcement Immediacy ................... 12 Laboratory Studies of Delayed Reinforcement ................................ ....................... 12 A Summary of Laboratory Studies ................................ ................................ .... 12 Innovative Methods Used to Evaluate Delayed Reinforcement ........................ 13 The Effects of Delayed Reinforcement on Human Behavior ................................ ... 16 Variables that May Influence Delayed Reinforcement ................................ ............ 17 Delay Duration ................................ ................................ ................................ .. 18 Rei nforcement Schedule ................................ ................................ .................. 18 Responses During the Delay ................................ ................................ ............ 18 Number of Alternatives Targeted at One Time ................................ ................. 20 Intertrial Intervals ................................ ................................ .............................. 21 General Purpose ................................ ................................ ................................ ..... 22 2 HUMAN OPERANT EVALUATIONS OF DELAYED REINFORCEMENT ............... 23 Purpose ................................ ................................ ................................ .................. 23 General Methods ................................ ................................ ................................ .... 23 Subjects and Setting ................................ ................................ ........................ 23 Target Responses ................................ ................................ ............................ 24 Procedure ................................ ................................ ................................ ......... 24 Experiment 1A: Procedure ................................ ................................ ...................... 27 Experiment 1A: Results and Discussion ................................ ................................ 28 Experiment 1B: Procedure ................................ ................................ ...................... 33 Experiment 1B: Results and Discussi on ................................ ................................ 35 Experiment 1C: Procedure ................................ ................................ ...................... 39 Experiment 1C: Results and Discussion ................................ ................................ 40 3 THE EFFECTS OF DELAYED REINFORCEMENT ON DISCRIMINATION ACQUISITION DURING CONDITIONAL DISCRIMINATION TRAINING ............... 63 Conditional Discrimination Trai ning ................................ ................................ ........ 63 Purpose ................................ ................................ ................................ .................. 63

PAGE 6

6 General Methods ................................ ................................ ................................ .... 64 Subjects and Setting ................................ ................................ ........................ 64 Target Responses and Data Collect ion ................................ ............................ 65 General Procedure ................................ ................................ ........................... 65 Preference assessment ................................ ................................ ............. 65 Baseline assessment ................................ ................................ ................. 66 Conditional discrimination training ................................ ............................. 66 Experiment 2A: Pr ocedure ................................ ................................ ...................... 67 Experiment 2A: Results and Discussion ................................ ................................ 67 Experiment 2B: Procedure ................................ ................................ ...................... 69 Experiment 2B: Results and Di scussion ................................ ................................ 70 Experiment 2C: Procedure ................................ ................................ ...................... 72 Experiment 2C: Results and Discussion ................................ ................................ 72 Experiment 2D: Procedure ................................ ................................ ...................... 73 Experiment 2D: Results and Discussion ................................ ................................ 74 4 GENERAL DISCUSSION ................................ ................................ ....................... 82 Overall Summary ................................ ................................ ................................ .... 82 Experiment 1 ................................ ................................ ................................ .... 82 Experiment 2 ................................ ................................ ................................ .... 84 Conclusions ................................ ................................ ................................ ............ 86 LIST OF REFERENCES ................................ ................................ ............................... 89 BIOGRAP HICAL SKETCH ................................ ................................ ............................ 93

PAGE 7

7 LIST OF FIGURES Figure page 2 1 and immediate reinforcement condition s of Experiment 1A. ................................ ....................... 46 2 2 responses as a function of delay, during Experiment 1A and hyperbolas depicting estimated relationship between these variables .. 47 2 3 during the delaye d reinforcement and immediate rein forcement conditions of Experiment 1A ................................ ...... 48 2 4 rei nforcement during Experiment 1A ................................ ................................ 49 2 5 Breakpoints obt ained during the delayed reinforce ment condition of Experiment 1A ................................ ................................ ................................ ... 50 2 6 Rat delaye d reinforc ement and immed iate reinforcement conditions of Experiment 1B ................................ ...................... 51 2 7 responses as a function of delay during Experiment 1B and hyperbola s depicting estimated relationship between these variables ....... 52 2 8 during delaye d reinforcement and immediate reinforcement conditions of Experiment 1B ................................ ...... 53 2 9 reinforceme nt during Experiment 1B ................................ ................................ 54 2 10 Breakpoints obtained during the del ayed reinforcement condition of Experiment 1B.. ................................ ................................ ................................ .. 55 2 11 ement and immediate reinfo rcement conditions of Experiment 1C ................................ ....................... 56 2 12 responses as a function of programmed delay during Experiment 1C and hyperbolas depicting relat ionship between variables ......... 57 2 13 responses as a function of obtained delay between .. 58 2 14 responses as a function of obtained delay between ...................... 59 2 15 delaye d reinforcement and immedi ate reinforcement conditions of Experiment 1C.. ................................ .... 60

PAGE 8

8 2 16 reinforcement in Experiment 1C.. ................................ ................................ ....... 61 2 17 Breakpoints obtained during the tandem VI PT schedule of reinforcement in the delayed reinforcement condition of Experiment 1C ................................ ..... 62 3 1 Percent cor rect across 20 s delaye d reinforcement and immedi ate reinforcement conditions in Experiment 2A.. ................................ ...................... 76 3 2 Percent correct across 30 s delaye d reinforcement and immedi ate reinforcement conditions during Experiment 2A.. ................................ ............... 77 3 3 Percent correct across 40 s delaye d reinforcement and immedi ate reinforcement conditions during Experiment 2A ................................ ................ 78 3 4 Percent correct across delaye d reinforcement and immedi ate reinforcement conditions of Experiment 2B. Stimuli available during delay. ............................. 79 3 5 Percent correct across delaye d reinforcement and immedi ate reinforcement co nditions of Experiment 2C. Four alternati ves were targeted per condition ..... 80 3 6 Percent correct across conditions in which either 0 s or 30 s ITI were programmed during Experiment 2D ................................ ................................ ... 81

PAGE 9

9 LIST OF ABBREVIATION S DD Developmental Disability; Characterized by below average scores on tests of mental ability and limited daily living skills (e.g., communication) DRO Diff erential Reinforcement of Other Behavior; Reinforcement provided contingent on the absence of problem behavior for a pre specified interval FI Fixed Interval; Reinforcement provided for the first response following a constant period of time FR Fixed Ratio; Reinforcement provided following the last of a constant number of responses IRI Interreinforcement Interval; Time between successive reinforcer deliveries ITI Intertrial Interval; Time between the consumption of the reinforcer and the start of the next tr ial S D Discriminative Stimulus; Stimulus that signals that a particular response will produce reinforcement S S Delta; Stimulus that signals that a particular response will not produce reinforcement PT Progressive Time; Schedule of reinforcement in which reinforcers are delivered following progressively increasing delays VI Variable Interval; Schedule of reinforcement in which reinforcement is provided for the first response following an average period of time VT Variable Time; Schedule of reinforcement in which reinforcement is provided, independent of responding, following an average interval of time

PAGE 10

10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the D egree of Doctor of Philosophy EVALUATIONS OF DELAYED REINFORCEMENT IN CHILDREN WITH DEVELOPMENTAL DISABILITIES By Jolene Rachel Sy August 2011 Chair: Timothy R. Vollmer Major: Psychology It is commonly recommended that reinforcers be provided immediat ely after behavior to establish or maintain responding (e.g., Miltenberger, 2008). These recommendations are usually provided in the absence of supporting empiric al work, which is unfortunate because several laboratory experiments have found that delayed r einforcement can produce response acquisition and maintenance ( e.g., Lattal & Gleeson, 1990). R elatively fewer studies with humans have isolated the effects of unsignaled, delayed reinforcement. Fewer still have evaluated unsignaled, delayed reinforcement in children with de velopmental disabilities (DD). Thus, the general purpose of the following experiments was to (a) examine whether delayed reinforcement could produce response maintenance in children with DD during an operant arrangement, (b) examine whet her delayed reinforcement could produce discrimination acquisitio n in children with DD (c) examine some of the variables that may affect the efficacy of delayed reinforcement in children with DD and (d) examine one variable that may account for situation s in which delayed reinforcement prevents acquisition. We found that delayed reinforcement produced response maintenance during three human operant arrangements. In addition, we found that 20 30 and 40 s

PAGE 11

11 delays to reinforcement lead to discrimination a cquisition for the majority of subjects. The availability of identical responses during the delay did not prevent response maintenance for 2 out of 3 subjects or discrimination acquisition for 2 out of 3 subjects For 1 out of 2 subjects, d iscrimination ac quisition was hindered when the number of responses targeted du ring each session was increased These results suggest that delayed reinforcement can produce acquisition and maintenance under specific conditions In addition, we found that longer intertrial intervals (time between delivery of a reinforcer and the presentation of the next trial) most likely do not account for cases in which discrimination acquisition fails to occur under conditions of delayed reinforcement.

PAGE 12

12 CHAPTER 1 INTRODUCTION Recommenda tions About the Necessity of Reinforcement Immediacy Several notable textbooks recommend that reinforcers be provided immediately following a response to establish or maintain behavior (e.g., Miltenberger, 2008; ement which develops skills must be immediate Otherwise, the precision of the differential effect is lost Likewise, with both positive and negative reinforcement, immediate reinforcement is more effective than delayed reinf Given these types of recommendations, it is not surprising that the temporal proximity between a response and a reinforcing event has been considered to be one of the most important parameters of reinforcer efficacy (Williams, 1976). Neve rtheless, results from several experiments suggest that d elayed reinforcement can produce response acquisition and maintenance under some conditions Laboratory Studie s of Delayed Reinforcement A Summary of Laboratory Studie s Researchers have evaluated bo th signaled delays to reinforcement, which program a stimulus change immediately following a response, and unsignaled delays to reinforcement, which do not program a stimulus change following a response. Due to their pairing with delayed reinfo rcement, sti mulus changes may begin to function as conditioned reinforcers and supplement the effects of delayed reinforcement with immediate reinforcement. Laboratory studies have found that both signaled and unsignaled delays to reinforcement can produce response ac quisition (e.g., Critchfield & Lattal, 1993; Latta l & Gleeson, 1990) and maintenance (e.g., Gleeson & Lattal, 1 9 87;

PAGE 13

13 S c haal & Branch, 1988). Some of these studies program resetting delays, in which every response that occurs during the delay resets the dela y interval, ensuring that obtained delays to reinforcement match programmed delays. Resetting contingencies do not allow experimenters to isolate the effects of delayed reinforcement independent of a differential reinforcement of other behavior (DRO) conti ngency, which requires that a given amount of time elapse since the last response before a reinforcer can be delivered. Thus, many researchers program nonresetting delays in which responses that occur during the delay do not have programmed consequences. L at tal and Gleeson found t hat unsignaled, nonresetting 30 s and resetting 10 s delays produced response acquisition in pigeons, and that unsignaled resetting 10 s or 30 s delays produced response acquisition in rats. Innovati ve Methods Used to Evaluate De layed Reinforcement Given the wealth of basic research on delayed reinforcement, it is not surprising that several procedural advances have been made. These advances a llow researchers to isolate the effects of delayed reinforcement. Although concurrent arr angements, in which two or more responses are available and proportional responding is measured, can be used to study relative preference for immediate reinforcement over delayed reinforcement (e.g., Dixon, Horner, & Guercio, 2003) single operant arrangem ents in which only one response is available, provide information about the absolute effects of delaye d reinforcement, and several laboratory experiments have us ed such arrangement s (e.g., Critchfield & Lattal, 1993 ) Basic researchers have also created p rocedures that rule out the effects of (a) reinforcement rate, (b) immediate conditioned reinforcement, and (c) adventitious reinforcement. When delayed rei nforcement conditions are compared with immediate

PAGE 14

14 reinforcement condition s it is often the case tha t at least two variable s differ between conditions: (a) the time between a response and a reinforcer and (b) reinforcement rate (with delayed reinforcement sessions associated with lower rat es of reinforcement). T o rule out decreases in reinforcement rate as a confounding variable reinforcement rate must be equated across delayed reinforcement and immediate reinforcement conditions. Researchers have done this by yoking interreinforcement intervals (IRI ; intervals of time between successive reinforcer deli veries ) in each immediate reinforcement session to IRI obtained during each preceding delayed reinforceme nt session (e.g., Reilly & Lattal, 2004) Researchers have also eliminated t he effects of immediat e, conditioned reinforcement by programming unsignale d delays to reinforcement that do not include a stimulus change following a response (e.g., Lattal & Gleeson, 1990) Finally, the effects of adventitious reinforcement can be ru led out by imposing a re setting, or DRO, contingency so that responses cannot o ccur in temporal proximity to reinforcement (e.g., Lattal & Gleeson) In addition to these advances, a procedure used by Re illy and Lattal ( 2004) allows researchers to examine the effects of multiple delay s within a single session. Reilly and Lattal progr ammed chained schedules, in which a reinforcer was delivered following the successive completion of two component schedules, each of which operated in the presence of a different stimulus, and tandem schedules in which a reinforcer was delivered following the successive completion of two components schedules, each of which operated in the presence of the same stimulus. These schedules incorporated variable interval (VI) components, in which the first response after an average interval of time elapsed resul ted in reinforcement or the presentation of the next component

PAGE 15

15 schedule, fixed interval (FI) components, in which the first response after a fixed interval of time elapsed resulted in reinforcement or the presentation of the next component schedule, and pr ogressive time (PT) components, in which reinforcers were delivered following progressively increasing intervals of time, regardless of behavior occurring during that time. Progressive time intervals increased by 2 s following each reinforcer delivery. Mor e specifically Rei lly and Lattal programmed a tandem VI 30 s PT 2 s schedule of reinforcement, a tandem FI 30 s PT 2 s schedule of reinforcement, and a chain ed VI 30 s PT 2 s schedule of rein forcement. The s e arrangement allowed the researchers to ex amine response ra tes as a function of several different delay s within a single session Finally, laboratory s tudie s have evaluated the effects of delayed reinforcement on discriminated responding (i.e., differential responding that is the result of differ ent contingencies in place for different response alternatives) by comparing response rates on both reinforcement and no consequence operandi (e.g., Escobar & Bruner, 2007; Keely, Feola, & Lattal, 2007; Sutphin, Byrne, & Poling, 1998) For example, Sutphi n et al. examined the effects of unsignaled 8 16 32 and 64 s delays to reinforcement on the response distribution of eight rats when reinforcers were delivered on a fixed ratio ( FR) 1 schedule following every response The y found that all subjects wer e more likely to respond on the reinforcement lever relative to the no consequence lever when the delay to reinforcement was small (i.e., 0 or 8 s). However, rates of responding on each lever became more similar (i.e., disc riminated responding was lost) as delays to reinforcement increased.

PAGE 16

16 In summary, basic research has isolat ed the effects of delayed reinforcement and developed methods to control for a range of potentially confounding var iables. Unfortunately, there have been relatively fewer s tudie s wit h humans that isolate the effects of unsignaled, delayed reinforcement. The Effects of Delayed Reinforcement on Human Behavior Applied s tudie s of delayed reinforcement typically do not isolate delay to reinforcement as an independent variable. Instead, ap plied studies generally focus on identifying methods that can be used to shift response allocation to those responses that produce delayed reinforcement (e.g., Vollmer, Borrero, Lalli, & Daniel, 1999), or on methods that can be used to bridge the delay to reinforcement (e.g., Schwarz & Hawkins, 1970 ; Terrell, 1958 ). Schwarz and Hawkins used delayed reinforcement during a DRO procedure to reduce problematic behavior exhibited by a 12 year old gi rl. In this procedure, the subject viewed a video recording of h er behavior during a mathematics period that had t aken place 1 to 5 hr beforehand During this time, t okens, exchangeable for larger items, were delivered conti ngent on the absence of target responses. The researchers found that this procedure produced lon g lasting decreases in three target responses. Although these results suggest that delayed reinforcement is effective, it is unclear whether delayed reinforcement would have been effective had the subject not been allowed to view the video recording of her previous behavior. In addition, in studies in which delayed reinforcement is combined with instructions about the contingencies (e.g., Erickson & Lipsitt, 1960; Okouchi, 2009; Rieber, 1961) it is unclear whether delayed reinforcement would be effective i n the absence of instructions.

PAGE 17

17 In addition, it is difficult t o interpret the results of many applied s tudie s of delayed reinforcement, as these studies have several potential flaws. These potential flaws have included the lack of single subject data maki ng it difficult to evaluate the effects of delayed reinforcement on individual behavior (e.g., Brackbill & Kappy, 1962 ; Erickson & Lipsitt, 1960 ; Millar & Watson, 1979 ), the absence of any raw data (e.g., Terrell & Ware, 1961; Ware & Terrell, 1961), the in clusion of dependent variables that were not clearly defined (e.g., Brackbill & Kappy) brief analyses (e.g., Hockman & Lipsitt, 1961 ; Millar & Watson ) that may not have allowed response rates to stabilize (Reeve, Reeve, & Poulson, 1993) the use of group comparisons with small groups that may not be similar prior to the introduction of the independent variable ( e.g., Okouchi, 2009), the absence of a baseline condition (e.g., Rieber, 1961), and the inclusion of DRO contingencies ( e.g., Millar, 1972; Okouchi ) which can sometimes create conditions of partial reinforcement only in the delayed reinforcement condition (Millar & Watson). Variables that M ay Influence Delayed Reinforcement Evaluating the effects of delayed reinforcement on the behavior of individu als with developmental disabilities (DD) is difficult, as the effects of delayed reinforcement may depend on several variables, including (a) the duration of the delay, (b) the schedule of reinforcement, (c) whether responses occur during the delay, and (d ) the number of alternatives targeted a t one time In addition, it is possible that any disruption in responding associated with delayed reinforcement is due to increases in the duration of the intertrial interval (ITI ; time between consumption of a reinfo rcer and the presentation of the next trial ), rather than the delay.

PAGE 18

18 Delay Duration Response rate, speed of acquisition, an d whether acquisition occurs at all can all depend on the duration of the del ay to reinforcement. Several studies with non human ani mals have found a decreasing, negatively accelerating relationship between response rate and increases in the delay to reinforcement (e.g., Fowler & Trapold, 1962 ; Reilly & Lattal, 2004). In addition, speed of acquisition by non human s has been found to va ry as a f unction of delay duration (e.g., Williams, 1998). Finally, some researchers have found that larger (e.g., 64 s) delays impede response acquisition, while smaller (e.g., 32 s) delays do not (e.g., Dickinson, Watt, & Griffiths, 1992). Collectively, these results suggest that delay duration is an important variable to consider when evaluating the effects of delayed reinf orcement on response acquisition and maintenance. Little is known about the effects of different delay durations on responding of chi ldren with DD Reinforcement Schedule Reinforcement schedules have well documented effects on behavior (Ferster & Skinner, 1957). In addition, the delay to reinforcement has been found to interact with the reinforcement schedule In studies with rats, Ski nner (1938) found that delayed reinforcement interfered with response acquisition when a FI schedule of reinforcement was in place, but not when a continuous schedule of reinforcement was in place. Similar evaluations have yet to be conducted with children with DD. Responses During the Delay Responses that o ccur during the delay may influence the efficacy of delayed reinforcement. Sidman (1960) noted, "Whenever a delay occurs between the recorded behavior and its program m ed consequence, we can be sure that the period of delay is

PAGE 19

19 not empty. Some behavior is taking place all the time, and even though such behavior is unrecorded it may still play a vital role in mediating the effects of the delay." (p. 371) Watson (1917) noted that behavior that occurs during t he delay might become adventitiously reinforced by reinforcer delivery (i.e., the level of a response temporally contiguous with reinforcer delivery may increase even though a reinforcement contingency is not arranged for that response) Adventitious reinf orcement of non target responses may interfere with the acquisition or maintenance of target responses. Alternatively, Catania (1971) noted that, if a response is followed by a different response that is reinforced, the reinforcer might affect both respons es even though its delivery is only dependent on the second one. This possibility has practical implications: Namely, "teachers must be al ert for sequences in which a student's errors are followed by corrections, so that they don't strengthen incorrect res ponses along with the correct ones that they reinforce." ( Catania, 2007, p.177 ). Thus, responses that occur either before or after responses for which a reinforcement contingency is arranged may become adventitiously reinforced. However, the extent to whic h adventitious reinforcement of nontarget responses interferes with the acquisition of the target response may depend on the length of the analysis, with longer analyses offsetting the effects of any adventitious reinforcement contingencies (Reeve, Reeve, Brown, Brown, & Poulson, 1992). Indeed, res earchers have found that the effects of adventitious reinforcement are not necessarily stable over long periods of t ime because other responses may become adventitiously reinforced or the response that was previo usly adventitiously reinforced may no longer contact reinforcement (e.g., Morse & Skinner, 1957; Williams & Lattal,

PAGE 20

20 1999). Likewise, researchers (e.g., Lattal & Gleeson, 1990) have fo und that adventitious reinforcement of target responses that occur during the delay cannot account for response acquisition and maintenance under delayed reinforcement, a s responses can be acquired when DRO contingencies ensure that the target response does not occur in close temporal proximity to reinforcer delivery. Despite t he wealth of basic research on this phenomenon, less applied work has been done Thus, more research on the effects of adventiti ous reinforcement of nontarget responses on the rate of target responses exhibited by individuals with DD is needed. In additio n it is possible that the occurrence of intervening responses c ould weaken the contingency between the target response and the delayed consequence, regardless of whether intervening responses become adventitiously reinforced. Number of Alternatives Targe ted at One Time Delayed reinforcement may be associated with different outcomes, depending on the number of alternatives targeted within a single evaluation (e.g., Hockman & Lipsitt, 1961). Hockman and Lipsitt examined the effects of unsignaled 0 10 an d 30 s delays to reinforcement on rate of discrimination acquisition (i .e., engagement in a particular response in the presence of one stimulus and a different response in the presence of a different stimulus such that percent correct meets or exceeds a p re established criterion) of school aged children during conditional discrimination training (i.e., reinforcement is provided contingent on engagement in one response in the presence of a particula r stimu lus, and engagement in a differ e n t response in the p resence of a different stimulus ) when either two or three alternatives were targeted. Hockman and Lipsitt taught subjects to press a particular button in the presence of an orange light and to press a different button in the presence of a green light. The researchers did not find

PAGE 21

21 significant differences in either the rate of learning or the mean number of correct responses over learning trials when two alternatives were targeted and 0 10 or 30 s delays to reinforcement were programmed. However, when the number of alternatives was increased from two to three, conditions in which reinforcers were delivered following 10 or 30 s delays were associated with lower mean numbers of correct responses over learning trials and no acquisition. This study had severa l limitations: (a) data were averaged across subjects, (b) a maximum of 36 trials were conducted per s ubject, so it is unclear if additional trials would have facilitated acquisition, (c) mean number of correct responses, even though the total number of responses may have varied across subjects. Thus, further evaluations of the effects of the number of response alternatives on the efficacy of delayed reinforcement are warranted. Intertria l Intervals Some research suggests that increases in ITI are responsible for decreases in response rate associated with conditions in which reinforcement is delivered following a delay (e.g., Bilodeau & Bilodeau, 1958). This finding, however, is not robust Although Saltzman, Kanfer, and Greenspoon (1955) found that a delayed r einforcement condition that involved increases in the dur ation of the ITI relative to the immediate reinforcement condition did not produce d ifferences in the accuracy in which underg raduate students drew a 3 in line Denny, Allard, Hall, and Rokeach (1960) hypothesized that increases in ITI duration may hinder acquisition if subjects do not know what the target response is. In summary, the finding that longer ITI are associated with d ecreases in response rate is not robust, and more research is needed in this area.

PAGE 22

22 General Purpose This series of experiments will determine whether the findings from basic research translate to the behavior of children with DD during an operant arrangeme nt and during conditional discrimination training. In addition, this series of studies will examine some factors that may affect the ef ficacy of delayed reinforcement during these arrangements. Both arrangements employed trial based procedures that allowed us to control and measure the types of responses that could occur during the delay (i.e., subjects could only make a single correct response or a single incorrect response).

PAGE 23

23 CHAPTER 2 HUMAN OPERANT EVALUA TION S OF DELAYED REINFORCE MENT Purpose The purpos e o f the following experiments was to evaluate whether the findings from basic research translate d to the behavior of school aged individuals with DD during an operant arrangement. With individuals with DD as subjects, w e sought to evaluate whether (a) pro gressively increasing delays to reinforcement ( i.e., delays that increased by a fixed amount followin g successive reinforcer deliveries ) would produce response ma intenance (b) response rates would decrease as delays to reinforcement increase d and (c) del ayed reinforcement would maintain discriminated r esponding In addition, we sought to evaluat e whether responses during the delay would disrupt responding. General Methods Subjects and Setting Subjects were 3 preschool aged children who were diagnosed with DD and who had limited verbal repertoires. These subjects were recruited from a local elementary school Vlade was a 4 ye ar old boy diagnosed with Autism Spectrum Disorder (ASD) Alice was a 4 year old girl d iagnosed as developmentally delayed, specifically in the area of language. Walden was a 4 year old boy diagnosed with ASD All subjects fell below the normal range of functioning in the areas of receptive and expressive communication on the Battelle Develo pmental Inventory which was administered by the pre Kindergarten school psychologist.

PAGE 24

24 subjects sat in a chair in front of a touch screen monitor that was placed on a table approxim ately 6 in in front of them. Target Response s (usually his or her hand) and one of two squares depicted on the touch screen monitor. During each session, responses on one square (the discriminative stimulus, or S D ; stimulus in the presence of which responses could be reinforced ) were considered to be (the s delta, or S ; stimulus in the presence of which responses were never reinforce d ) were considered A computer program designed using VB.NET responses in real time. Procedure During each experiment, a delayed reinforcement condition was alternated with an immediate reinforcement condition. During both conditions, subjects sat in front of a touch screen monitor that depicted two different squares. The color of the squares (Vlade and Alice) or the pictures displayed in the squares (Walden) varied from session to session and, at the start of each session, one color or picture was randomly designated as the S D while the other color or picture was randomly designated as the S Stimuli changed from session to session to eliminate the effects of history on responding During each session the location of the S D and S was randomized across trials. could result in the delivery of small edible items (e.g., half a fruit snack) or 30 s of access to an iPod, whereas never resul ted

PAGE 25

25 in reinforcement The experimenter sat next to the touch screen, and did not interact with the subject, unless an auditory sound from the touchscreen prompted the experimenter to deliver a reinforcer. The e xperimenter attempted to remain neutral throughout each session, minimi zing cues that might have indicated whether a Vlade and Alice received edible items, whereas Walden, due to severe food allergies, received access to an iPod. These items were identified as reinforcers during a previous reinforcer t est in which a reversal design was used to evaluate the reinforcing efficacy of the iPod or the edible items when a concurrent schedule of reinforcement was in place (i.e., two schedules of reinforcement operated simultaneously, but independently, for two different responses). During each session of the evaluation, a blue square and a white square were depicted on the monitor, and the position of the squares was randomized across trials. During one condition (A), every response to the blue square resulted i n the immediate delivery of either the iPod or a small edible item on a fixed ratio (FR) 1 schedule, whereas responses to the white square did not result in differential consequences. During the second condition (B) every response to the white square resu lted in the immediate delivery of either the iPod or a small edible item on a FR 1 schedule, whereas responses to the blue square did not result in differential consequences. Conditions were alternated in a ABAB reversal design. A reinforcement effect occu rred when subjects allocated more responding to the stimulus that produced reinforcement during a given condition. At the very beginning of Experiment 1A the experimenter prompted subject s to touch both squares to ensure they contacted the contingencies associated with either response. From that point on, additional prompting was not provided. However, a t the

PAGE 26

26 start of every session, subjects contacted ten forced exposure trials in which reinforcers were delivered immediately on a FR 1 schedule contingent on responses to the S D Responses to the S during these trials did not result in reinforcement. F orced exposure trials were programmed to ensure that subjects would acquire the discrimination prior to an assessment of response maintenance under either imm ediate or delayed reinforcement. Within session analyses the first five and last five trials) suggested that subjects generally acquired the discrimination during this period. Sessions in which reinforcers we re delivered following progressively increasing delays were alternated w ith yoked control sessions that controlled for reinforcement rate by delivering reinforcers immediately on either a ratio schedule (Experiment 1A) or an interval schedule (Experiments 1B and 1C) yoked to the obtained IRI of each previous delayed reinforcement session. A lthough the goal was to compare the effects of immediate reinforcement with delayed reinforcement while controlling reinforcement rate obtained reinforcement rate s were generally lower in the imm ediate reinforcement condition relative to the delayed reinforcement condition Delayed reinforcement s essions conti nued until a maximum delay of 120 s was reached or until subjects did not respond for three times the maximum paus e in responding observed during the previous ly described reinforcer test. This amounted to 84 s, 150 s, and 51 s for Vlade, Alice, and Walden, respectively. The duration of each immediate reinforcement session was yoked to the duration of each preceding de layed reinforcement session.

PAGE 27

27 One session was conducted per day, three to five days per week, depending on subject availability. Experiment 1A: Procedure During the delayed reinforcement condition of Experiment 1A, a chain ed FR 1 PT 5 or 10 s schedule of reinforcement was in effect. Each response to the S D initiated the PT terminal link, during which stimuli were removed from the screen. The duration of the terminal link increased by a 5 s (Vlade) or 10 s (Alice and Walden) following successive reinforc er deliveries, and reinforcers were delivered once the terminal link ended. A PT 5 s schedule was programmed for Vlade, the first subject in the Experiment. However, this schedule resulted in lengthy session durations of around 50 min, so a PT 10 s schedul e was programmed for Alice and Walden to decrease total session time. Each response to the S also initiated the terminal link, the duration of which was equal to the duration of the terminal link during the previous trial. However, a reinforcer was not pr esented at the end of the interval Following the terminal link, stimuli appeared on the screen followi ng a brief ITI During the immediate re inforcement condition of Experiment 1A, a chain ed variable time (VT) FR 1 schedule of reinforcement was program med. A VT schedule is a schedule in which reinforcers (or, in this case, access to the terminal link) are delivered following an interval of time that varies from one reinforcer delivery to the next, irrespective of behavior. This schedule was programmed t o keep the response requirement constant (FR 1) while making it more likely that reinforcement rate matched rates of reinforc ement obtained during each preceding delayed reinforcement session. The interval s in the initial link were yoked to the IRI of the previous delayed

PAGE 28

28 reinforcement session. Stimuli were only present during the term inal link. During the terminal link a response resulted in the immediate delivery of a small edible item or 30 s of access to an iPod and removal of the stim uli fro m the screen. An only resulted in the removal of stimuli from the screen. Experiment 1A: Results and Discussion suggested that discriminations were 100% of the time during the last five trials prior to each delayed reinforcement session and 92.73% of the time (range, 80% 100%) during the last five trials prior to each immediate reinforcement session. Alice 74.29% of the time (range, 20% 100%) during the last five trials prior to each delayed reinforcement session and 77.14% of the time (range, 20% 100%) during the last five trials prior to each immediate reinforcemen t session low during some of these first 10 trials, there was not a relationship between levels of percent correct during the last five trials and levels of percent correct during subsequen t delayed rein forcement (R 2 = .06) sessions However, there was somewhat of a positive relationship between levels of percent correct during the last five trials and levels of percent correct during subsequent immediate reinforcement sessions (R 2 = .66). Walden sele 100%) during the last five trials prior to each delayed reinforcement session and 93.85% of the time (range, 40% 100%) prior to each immediate reinforcement session. Once again, although percen relationship between percent correct during the last 5 trials and during subsequent delayed reinforcement (R 2 < .01) and immediate reinforcement sessions (R 2 = .03).

PAGE 29

29 Response rates were calculated for each delayed reinforcement and immediate reinforcement session. Given that the arrangement used in both conditions was trial based, subjects could n o t respond freely and responding might have been better categorized in terms of latency (time between the presentation of stimuli and the first response). However, this would have made it difficult to analyze del ay gradients, which are generated by analyzing the relationship between response rates and delays to reinforcement. In addition, th is would have made it difficult to compare the results from Experiment 1A to the results of Experiments 1B and 1C, in which subjects were able to make multiple response s during each trial. Thus, we chose to examine response rates, which, in Experiment 1A, were the reciprocals of response latencies. Response rates obtained during Experiment 1 A are depicted in Figure 2 1. During Experiment 1 A delayed reinforcement maintained lower response rates relative to immediate reinforcement. Response rates were only c alculated for the period of time following the first 10 forced exposure trials. Response rate was calculated by dividing the number of responses by the total time in which it was possible to make a response. Thus, time in which stimuli were remov ed from the screen was not included in the calculations Vlade engaged in an average of 12.57 responses per min during the immediate reinforcement condition (range, 10.60 15.29 responses per min) and an average of 4.58 responses per min in the delayed re inforcement condition (range, 3.62 5.89 responses per min) Alice engaged in an average of 10.29 responses per min during the immediate reinforcement condition (range, 3.50 21.80 responses per min) and an average of 3.04 responses per min in the delaye d reinforcement condition (range, 1.57 6.15 responses per min) Walden engaged in an average of 26.10 responses per min

PAGE 30

30 in the immediate reinforcement condition (range, 14.04 33.00 responses per min) and an average of 14.26 responses per min in the del ayed reinforcement condition (range, 2.86 24.83 responses per min) Figure 2 2 displays mean response rates and predicted curves for each subject as a function of each delay to reinforcement. During Experiment 1 A mean response rates decreased as de lays to reinforcement increased Mean res ponse rates for each delay were calculated by first computing the response rate during the time since the last trial This was done by dividing the number of correct responses (either zero or one in Experiment 1A) by the total duration in which stimuli were available (i.e., time between the presentation of the stimuli and their subsequent removal following a response) Thus, if stimuli were present on the screen for 5 s prior to a correct response, then response rate for that particular delay was considered to be 12 responses per min (one response divided by 5 s, multiplied by 60 s). Within each session, response rates for each d elay were averaged, even if reinforcers were not delivered at the end of the delay. Thus, would be equal for these responses, even though a reinforcer would only be delivered following the response. In these cases, response rates were calculated separately and then averaged. Response rates were then averaged across sessions a s a function of each delay It should be noted that more response rates were averaged into mean response rates during the smaller delay s. This is because subjects did not contact high delay s during sessions with low breakpoints. Thus, for Vlade, 11 sessions of data were averaged into each mean response rate for delays between 5 s and 6 0 s, 6 sessions of data were averaged into each mean response rate for delays between 65 s

PAGE 31

31 and 80 s, and 5 or fewer sessions of data were averaged into each mean response rate for delays between 90 s and 120 s. For Alice, 14 sessions of data were averaged into each mean response rate for delays between 10 s and 100 s, and 12 sessions of data were averaged into ea ch mean response rate for 110 s and 120 s delays. For Walden, 25 sessions of data were averaged into each mean response rate for delays between 10 s and 60 s, 24 sessions of data were averaged into each mean response rate for delays between 70 s and 80 s, 23 sessions of data were averaged into each mean response rate for 90 to 110 s delays, and 21 sessions of data were averaged into the mean response rate for the 120 s delay. The relationship between mean response rate and delay to reinforcement was quanti hyperbolic delay discounting equation: B = B I / (1 + kD ) Reilly and Lattal (2004) used this equation to predict response rate ( B ) a s a function of each delay ( D ). The equation contains two free parameters, B I and k which represent estimated rate of responding under immediate reinforcement and estimated degree of discounting due to increasing delays to reinforcement, respectively. B I and k were e stimated using Microsoft Excel, which minimized the sum of squar ed deviations between the obtained values and the predicted curve. data for all 3 su bjects are presented in Figure 2 3 Percentage of was calculated by dividing the number of responses by the total number of responses. Responses from the first 10 forced exposure trials were not included in these calculations 94.38 % during the delayed reinforcement condition (range, 91% 100%) and an

PAGE 32

32 average of 92.29% durin g the immediate reinforcement condition (range, 78% 100%) se 89.57% (range, 54.5% 100%) and 81.98% (range, 18.2% 100%) during the delayed reinforcement and immediate reinforcement conditions, respectively. Walden w during the immediate reinforcement condition. However, he eventually selected the ; range, 53.33% 100% ) and immediate reinforcement (98.66% ; range, 90.9% 100% ) conditions. Figure 2 increasing delays to reinforcement. Figure 2 5 presents breakpoints obtained during the chained FR 1 PT schedule in Experiment 1 A Breakpoints were defined as the last delay contacted prior to meeting the session termination criteria. The maximum possible delay that subjects could contact was 120 s. In general, breakpoints varied from session to sessio n mean breakpoint was 85.7 s (range, 60 (range, 110 120 s). For Alice a nd Walden there was a negative relationship between the k value estimated using mean breakpoi nt In other words greater degrees of discounting due to the delay were ass ociated with lower breakpoints, which suggests that k was an accurate representation of discounting due to the delay for these subjects In summary, Experiment 1 A evaluated the effects of delayed reinforcement on the behavior of children with DD using a chained schedu le of re inforcement. The removal of

PAGE 33

33 stimuli in the terminal component of the chained schedule ensured that programmed delays matched obtained delays because subjects could not emit the target response during the terminal component. Add itionally, the remova l of stimuli was programmed following both responses, thus de grading the relationship between the removal of stimuli and reinforcement. The results from Experiment 1A suggests that signaled, delayed reinforcement can maintain resp onding by individuals with DD even when the signal is degraded In addition, results suggest that adding a delay between a response a nd a reinforcer will not produce decreases the percent of made by the subjects R esults also suggest t hat individuals with DD will continue to respond, even as delays to reinforcement increase. However, these findings might be specific to the reinforcement schedule programmed (FR 1 schedule ). Thus, the purpose of Experiment 1B was to examine the effects of delayed re inforcement using a VI schedule. Experiment 1B: Procedure During Experiment 1B subjects were again exposed to two experimental conditions. During the delayed reinforcemen t condition, a chain ed VI 15 or 30 s PT schedule of reinforcement was in effect. Relative to the FR 1 schedule programmed in Experiment 1A, t his schedule was programmed to allow for greater response variability in the initial component. The VI 15 s schedule consisted of 20 intervals (range, 0.38 to 59.94 s) generated using a Fleshler and Hoffman (1962) progression, while the VI 30 s schedule consisted of 20 intervals (range, 0.76 to 119.87 s) generated using the same progression. All subjects were initially exposed to the chained VI 30 s PT schedule of reinforcement. Howeve appeared to be the r esult of interval strain, so for him the initial component was switched to a VI 15 s schedule for

PAGE 34

34 Experiments 1B an d 1C During the delayed reinforce ment condition the position of stimuli was randomize d following each response that occurred before the int erval had elapsed. S timuli were removed from the screen contingent on the first response after the inter val had elapsed and remained absent for the duration of the PT component, which increased by a fix ed amount (i.e., 5 or 10 s) following successive reinforcer deliveries. A t t he end of the terminal component a tone was sounded and reinforcers were delivered If this response had b een a tone was not sounded and reinforcers were not deliver ed. Following the comple tion of each terminal link, stimuli were represented on the screen following a 3 s or 30 s ITI. This ITI was programmed to allow subjects to consume the reinfor cer During the immediate reinforcement condition, a yoked VI schedule of reinforcement was in effect. A gain, this schedule was chosen to keep the response requirement relatively similar to that programmed in the delayed reinforcement condition, while at the same time equating reinforcement rate to that obtained during each previous delayed reinforcement session. In the yoked VI schedule, t he interval s were yoked to the obtained IRI of each previous delay session. The duration of each immediate reinforceme nt session was yoked to the duration of each previous delayed reinforcement session. During each session in the immediate reinforcement condition, responses that occurred prior to the end of the interval resulted in the randomized repositioning of stimuli on the screen, and the first response after the interval requirement elapsed either resulted in the immediate presentation of a tone and a

PAGE 35

35 reinforcer (if the response had been ) or an absence of differential consequences (i f the response had been ). Experiment 1B: Results and Discussion suggested that discriminations were acquired. Vlade sel 95.21 % (range, 60% 100%) of th e time during the last five trials prior to each delayed reinforcement session and 94.78% of the time (range, 4 0% 100%) during the last five trials prior to each immediate reinforcement session. Alice select responses 88 % of the time (ran ge, 20% 100%) during the last five trials prior to each delayed reinforcement session and 91% of the time (range, 60% 100%) during the last five trials prior to each immediate reinforcement session uring some of these sets of trials, there was not a relationship between levels of percent correct during the last five trials and levels of percent correct during subsequen t delayed reinforcement (R 2 = .02) or immediate reinforcement (R 2 = .03) sessio ns. Walden selec the time (range, 4 0% 100%) during the last five trials prior to each delayed reinf orcement session and 92.86% of the time (range, 6 0% 100%) prior to each immediate reinforcement session. For Walden, there was not a relationship between levels of percent correct during the last five trials and levels of percent correct during subsequent delayed reinforcement (R 2 = .02) or immediate reinforcement (R 2 = .01) sessions. Figure 2 6 displays response rat e data from Experiment 1B. Response rates were responses following the first 10 forced exposure trials divided by the total amount of

PAGE 36

36 time in which it was possible to make a response. R esponse rates were again slightly elevated during immediate reinforcement condition relative to the delayed reinforcement condition Vlade engaged in an average of 4.38 responses per min during the immediate reinforcement condition (range, 1.2 2 13.41 responses per min) and an average of 3.27 responses per min in the delayed reinforcement condition (range, 0.75 12.43 responses per min) Alice engaged in an average of 4.48 responses per min during the immediate reinforcement condition (range, 2.62 6.60 responses per min) and an average of 3.06 responses per min in the delayed reinforcement condition (range, 1.15 5.98 responses per min) Walden engaged in an average of 3.08 responses per min in the immediate reinforcement condition (range, 1.25 8.64 responses per min) and an average of 3.92 responses per min during the VI 30 s component of the delayed reinforcement condition (range, 1.65 8.57 responses per min) and 1.56 responses per min during the VI 15 s component of the delayed reinfo rcement condition (range, 1.00 2.62 responses per min) Figure 2 7 depicts mean response rates as a function of each delay and curves equation. Calculations of me an respons e rates were similar to those computed during Experiment 1A: T he number of correct responses made prior to each delay were divided by the total amount of time in which it was possible to make a response since the last trial Once again, more response rat es were averaged into mean response rates during the smaller delay s. For Vlade, at least 20 sessions of data were averaged into each mean response rate for delays between 5 s and 40 s, at least 10 sessions of data were averaged into each mean response rate for delays between 45 s and 60 s, and 7

PAGE 37

37 or fewer sessions of data were averaged into each mean response rate for delays between 65 s and 120 s. For Alice, at least 10 sessions of data were averaged into each mean response rate for delays between 10 s and 60 s, but 8 or fewer sessions of data were averaged into each mean response rate for delays between 70 s and 90 s. For Walden, 7 sessions of data were averaged into each mean response ra te for delays between 10 s and 3 0 s, but only 5 or fewer sessions of d ata were averaged into each mean re sponse rate for delays between 40 s and 7 0 s. from the VI 15 s component. For Vlade and Alice, mean response rates decreased as delays to reinforcement increased. For Walden, mean response rates remained fairly stable as delays to reinforcement increased. In general, did not differ across conditions. was forced exposure after the first 10 forced exposure trials. These data are depicted in Figure 2 8 Vlade an average of 82.45% during the delayed reinforcement condition (range, 40% 100%) and an average of 80.27% during the immediate reinforcement condition (range, 55.55% 100%) an average of 71.15% during the delayed reinforcement condition (range, 50% 91.18%) and an average of 67.81% during the immediate reinforcement condition (range, 41% 91.67%) an average of 64.72% during the VI 30 s component of the delayed reinforcement condition (range, 51.32% 88.24%) an average of 61.45% during the VI 15 s component of the delayed reinforcement condition (range, 48.57% 75%) and an average of 67.55% during the

PAGE 38

38 immediate reinforcement condition (range, 49.21% 82.76%) Figure 2 9 presents to reinforcement. For all 3 reinforcement. Figure 2 10 displays breakpoin ts obtained when a chained VI PT schedule of rein forcement was programmed. Mean breakpoints were 39.8 9 s (range, 10 s 85 s), 58.75 s (range, 20 s 80 s), and 38.57 s (range, 10 s 70 s) for Vlade, A lice, and Walden, respectively. obtained during the VI 30 s component (31.43 s) and the VI 15 s c omponent (45.71 s) Once again, for Alice and Walden, there was a negative relationship between mean breakpoints and the k delay discounting equation and data obtained during Experiment 2. Th us, greater degrees of discounting due to the delay were associated with lower breakpoints. In summary, the results of Experiment 1B suggest that delayed reinforcement will maintain responding, even when In addit ion, was again comparable across delayed reinforcement and im mediate reinforcement conditions, albeit low in both conditions. The fact that low levels of percent correct occurred in both conditions suggests that the VI schedule may have been unable to maintain discriminated responding, independent of the delay. In addition, i t appeared the change in the schedule of reinforcement in the initial component of the chained schedule (from FR 1 to VI 15 or 30 s) resulted in lower overa ll breakpoints, suggesting that, to some extent, the effects of delayed reinforcement may be influenced by the sc hedule of reinforcement

PAGE 39

39 One potential limitation of both Experiments 1A and 1B was that subjects could not engage in topographical ly similar responses during the delay. The trial based arrangement used in these experiments therefore differed from typical free operant situations in which subjects may engage in a variety of respo nses during the delay. It is possible that the occurrenc e of responses during the delay would disrupt responding maintained by delayed reinforcement. Therefore, the purpose of Experiment 1C was to evaluate whether delayed reinforcement would still maintain responding when subjects were free to engage in topograp hically similar responses during the delay. Experiment 1C: Procedure During the delayed reinforcement condition of Experiment 1C a tandem VI PT schedule of reinforcement was in effect. During the terminal link, stimuli continued to be presented on the screen and subjects could respond to these stimuli. During both the initial and terminal links, stimulus placement on the screen was randomized following each response Re inforcement contingencies were only arranged for the last response in the initial lin k (i.e., the first response after the interval requirement elapsed). If this response had been the duration of the terminal link increased by 5 o r 10 s across trials. If this the duratio n of the terminal link was e qual to the duration of the terminal link during the preceding trial. After the terminal link elapsed, a tone was sounded and reinforcers were delivered if the pre vious response Following all trials, stimuli were represe nted on the scre en following a 3 s or 30 s ITI. The immediate reinforcement condition was identical to that programmed in Experiment 1B ; a yoked VI schedule of reinforcement was in effect.

PAGE 40

40 Experiment 1C: Results and Discussion ponses during the first 10 trials generally 97.65% (range, 80% 100%) of the time during the last five trials prior to each delayed reinforcement session and 96.47% of the time (range, 80% 100%) during the last five responses 93% of the time (range, 40% 100%) during the last five trials prior to each delayed reinforcement session and 93% of the time (range, 80% 100%) during the last prior to delayed reinforcement sessions, there was only a minimal rela tionship between levels of percent correct during the last five trials and levels of percent correct during subsequent delayed reinforcement (R 2 time (range, 60% 100%) during the last five trials prior to each delayed reinforcement session and 98% of the time (range, 80% 100%) prior to each immediate reinforcement session. For Walden, delayed reinforcement sessions, there was only a mi nimal positive relationship between levels of percent correct during the last five trials and levels of percent correct during subsequent delayed reinforcement (R 2 = .29) sessions. Response rate data obtained during Experimen t 1 C are depicted in Figur e 2 11 W e compared response rates across both components of a tandem VI PT schedule of reinforcement with response rates during the immediate reinforcement condition. D uring responses made following the first 10 forced exposure trials by the total amount of time

PAGE 41

41 in which it was possible to make a response. Vlade engaged in an average of 3.82 responses per min during the immediate reinforcement condition (range, 1.56 7.6 8 responses per min) and an average of 2.70 responses per min in the delayed reinforcement condition (range, 1.0 0 5.07 responses per min) Alice engaged in an average of 4.64 responses per min during the immediate reinforcement condition (range, 1.32 8 .21 responses per min) and an average of 2.55 responses per min in the delayed reinforcement condition (range, 1.18 6.91 responses per min). Walden engaged in an average of 3.00 responses per min in the immediate reinforcement condition (range, 1.44 5. 77 responses per min) and an average of 3.93 responses per min in the delayed reinforcement condition (range, 2.19 6.50 responses per min) Figure 2 12 displays mean response rates as a fun ction of each programmed delay. Unlike Experiments 1A and 1B mea n response rates were calculated across both initial and terminal components of the tandem VI PT schedule as a function of each programmed delay Thus, the total both initial and terminal links was divided by the total duration of the initial and terminal link, f or each delay A gain, within each session, response rates for identical delays were calculated separately and then averaged. Response rates were then averaged across sessions as a function of each delay. O nce again, more values were averaged into mean response rates for smaller delays because subjects did not contact high delays during sessions with low breakpoints For Vlade, at least eight values were averaged into mean response rates for delays between 5 s and 30 s. However, only three or fewer values were averaged into mean response rates for delays between 35 and 55 s. For Alice, at least 17 values were averaged into mean response rates for 10 and 20 s

PAGE 42

42 delays, whereas nine or fewer values were averaged into mean response rates for 30 and 40 s delays. For Walden, at least nine values were averaged into mean response rates for 10 and 20 s delays, but only six or fewer values were averaged into mean response rates for 30 40 50 and 60 s delays. For Vlade and Alice, mean response rates decreased as delays to reinforcement increased. For Walden, response rates remained relatively stable across increasing delays. We fit the modified version of did not do so for Alice because we obtained so few measures of response r ates under different delay s. Figure 2 13 displays mean response rates as a function of obtained delays to reinforcement, when obtained delays were defined as the time between the last Response rates were calculated for each obtained delay, and then averaged (both within and across sessions ) for each identical delay (in bi ns of 1 s) In general, response rates decreased as obtained delays to reinforcement increased. We fit the data to the It should be noted that mean resp onse rates and obtained delays we re not completely independent because lower rates of responding could produce longer obtained delays to reinforcement. Figure 2 14 displays mean response rates as a function of obtained delays to reinforcement, when obtained delays were defined as the t ime between the responses occurred between the last A gain, response rates were calculated for each obtained delay, and then avera ged across

PAGE 43

43 identical delays. Mean response rates decreased as obtained delays to reinforcement increased. Once again, it should be noted that mean response rates and obtained delays were not completely independent of one another. Figure 2 15 displays perc data for all subjects. A gain, was calculated using data obtained following the first 10 forced exposure trials of each session. Overall, generally did not differ between immediate reinforcement and delayed responses during the delayed reinforcement condition were correct (range, 53.85% 83.87%) and an average of 71.75% of his responses during the immediate reinforcement condition were correct (ra nge, 55.81% 100%). Alice selected the an average of 57.23% during the delayed reinforcement condition (range, 35.7% 85.19%) and an average of 59.17% during the immediate reinforcement condition (45.71% 81.81%) Walden selected the an average of 61.49% during the delayed reinforcement condition (range, 42.86% 79.17%) and an average of 60.26% (range, 45.71% 75%) during the immediate reinforcement condition. Th ability to di scriminate between stimuli deteriorated during both conditions of this evaluation. There was no noticeable effect of the delay data as a function of each delay to reinforcement. For all 3 su remained fairly stable across increasing delays to reinforcement, with only a slight decreasing trend for Vlade. Figure 2 1 7 displays breakpoints obtained during the tandem VI PT schedule of reinforcement. Breakpoints were lowe st during this evaluation, and varied across

PAGE 44

44 breakpoint was 23.5 s (range, 0 (range, 10 60 s). For Vlade and Walden, the re was a negative relationship between the k equation and breakpoints, such that greater degrees of discounting due to the delay were associated with lower breakpoints. As not The methods used in the third evaluation allowed us to measure the effects of at least two possible i ntervening responses (responses to the S D and responses to the S ). R esults from Experiment 1C sug gest that delayed reinforcement will maintain responding when subject s can engage in topographically similar res ponses during the delay. R esponses that occur during the delay may increase the efficacy of delayed reinforcement by making obtained delays (i.e., the delay betwe en the last response and reinforcer delivery) shorter than programmed delays (i.e., the delay between the response that initiated the delay and reinforcer delivery) Howev er, this arrangement may instead decrease the effica cy of delayed reinforcement if responses. In deed, f or Vlade, 42% (range, 0% 100%) of reinforcer deliveries were preceded by a For Alice, 44% (range, 0% 100%) of reinforcer deliveries were preceded by a For Walden, 25.3% (range, 0% 100%) of reinforcer deliveries were preceded by a Nevertheless, this arrangement still resulted in maintenance of the target response. However, this

PAGE 45

45 arrangement was less likely to maintain discriminated responding, relative to Experiment 1A.

PAGE 46

46 F igure 2 1. responses during the first component of the chained FR 1 PT schedule of reinforcement in the delayed reinforcement condition (closed circles) and the second component of the chained VT FR 1 schedule of reinforcement in the immediate reinforcement condi tion (open circles) during Experiment 1A A) Data for Vlade. A chained FR 1 PT 5 s schedule of reinforcement wa s in place during the delayed reinforcement condition. B) Data for Alice. A chained FR 1 PT 10 s schedule of reinforcement w as in place in th e delayed reinforcement condition C) Data for Walden. A chained FR 1 PT 10 s schedule of reinforcement w as in place in the delayed reinforcement condition.

PAGE 47

4 7 Figure 2 2. as a function of each delay during the first comp onent of the chained FR 1 PT schedule of reinforcement in the delayed reinforcement condition of Experiment 1A and a hyperbola depicting the estimated relationship between mean response rate and delay to reinforcement B I and k are fitted parameters tha t represent estimated response rate under immediate reinforcement and estimated degree of discounting, respectively. A ) Data for Vlade. A chained FR 1 PT 5 s schedule of reinforcem ent was in place. B) Data for Alice. A chained FR 1 PT 10 s schedule of reinforcement w as in place C) Data for Walden. A chained FR 1 PT 10 s schedule of reinforcement was in place.

PAGE 48

48 F igure 2 3. Percentage of across consecutive delayed reinforcement (closed circles) and immediate reinforcement (open circles) sessions during Experiment 1A. A) Data for Vlade. A chained FR 1 PT 5 s schedule of reinforcement wa s in place during the delayed reinforcement condition. B) Data for Alice. A chained FR 1 PT 10 s schedule of reinforcement w as in place during the delayed reinforcement condition C) Data for Walden. A chained FR 1 PT 10 s schedule of reinforcement w as in place during the delayed reinforcement condition.

PAGE 49

49 Figure 2 delays to reinf orcement during Experiment 1A. A) Data for Vlade. A chained FR 1 PT 5 s schedule of reinforcement was in place. B) Data for Alice. A chained FR 1 PT 10 s schedule of reinforcement was in place. C) Data for Walden. A chained FR 1 PT 10 s schedule of re inforcement was in place.

PAGE 50

50 F igure 2 5 Breakpoints obtained during the chained FR 1 PT schedule of reinforcement in the delayed reinforcement condition of Experiment 1A A) Data for Vlade. A chained FR 1 PT 5 s schedule of reinforcement was in place B) Data for Alice. A chained FR 1 PT 10 s schedule of reinf orcement was in place C) Data for Walden. A chained FR 1 PT 10 s schedule of reinforcement was in place

PAGE 51

51 F igur e 2 6 during t he fi rst component of the chained VI PT schedule of reinforcement in the delayed reinforcement condition (closed circles) and during the yoked VI schedule of reinforcement in the immediate reinforcement condition (open circles) of Experiment 1B A) Data for Vlade. A chained VI 30 PT 5 s s chedule of reinforcement wa s in place during the delayed reinforcement condition. B) Data for Alice. A chained VI 30 PT 10 s schedule of reinforcement w as in place during the delayed reinforcement condition C) Data for Walden. Prior to the phase change line, a chained VI 30 PT 10 s schedule of reinforcement w as in place during the delayed reinforcement condition After the ph ase change line, a chained VI 15 PT 10 s schedule of reinforcement was in place during this condition.

PAGE 52

52 Figure 2 7 Mean rate a s a function of each delay during the fi rst compone nt of the chained VI PT schedule of reinforcement in Experiment 1B a nd a hyperbola depicting the estimated relationship between mean response rate and delay to reinforcement B I and k are fitted parameters that represent estimated response rate under immediate reinforcement and estimated degree of discounting, respectively. A) Data for Vlade. A chained VI 30 PT 5 s schedule of reinforcement wa s in place. B) Data for Alice. A ch ained VI 30 PT 10 s schedule of reinforcement w as in place C) Data for Walden. A chained VI 15 PT 10 s schedule of reinforcement wa s in place

PAGE 53

53 Figure 2 8 Percentage of responses across consecutive delayed reinforcement (closed circles) and immediate reinforcement (open circles) session s during Experiment 1B A) Data for Vlade. A chained VI 30 PT 5 s schedule of reinforcement was in place in the delayed reinforcement condition. B) Data for Alice. A chained VI 30 PT 10 s schedule of reinf orcement was in place in the delayed reinforcement condition C) Data for Walden. Prior to the phase change line, a chained VI 30 PT 10 s schedul e of reinforcement was in place in the delayed reinforcement condition.

PAGE 54

54 Figure 2 9. Mean percentage of across consecutive delays to reinforcement during Experiment 1B. A) Data for Vlade. A chained VI 30 PT 5 s schedule of reinforcement w as in place B) Data for Alice. A chained VI 30 PT 10 s schedule of reinforce ment was in place C) Data for Walden. A chained VI 15 PT 10 s schedule of reinforcement was in p lace

PAGE 55

55 Fi gure 2 10 Breakpoints obtained d uring the delayed reinforcement condition of Experiment 1B. A) Data for Vlade. A chained VI 30 PT 5 s schedule of reinforcement was i n place. B) Data for Alice. A chained VI 30 PT 10 s schedule of reinforcement was in place C) Data for Walden. Prior to the phase change line, a chained VI 30 PT 10 s schedul e of reinforcement was in place.

PAGE 56

56 Figure 2 11 R ate d uring both components of the tandem VI PT schedule of reinforcement in the delayed reinforcement condition (closed circles) and during the yoked VI schedule of reinforcement in the immediate reinforcement condition (open circl es) of E xperiment 1C A) Dat a for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place in the delayed reinforcement condition. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement was in place in the delayed reinforcement condition. C) Data for Walde n. A tandem VI 15 PT 10 s schedule of reinforcement was in place in the delayed reinforcement condition.

PAGE 57

57 Figure 2 12 Mean rate as a function of each delay during both components of the tandem VI PT schedule of reinforcement in Experiment 1C and hyperbolas depicting the estimated relationship between these variables B I and k are fitted parameters that represent estimated response rate under immediate reinforcement and estimated degree of discounting, respectively. A) Data f or Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement w as in place. C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place.

PAGE 58

58 Figure 2 13 obtained delay, during both components of the tandem VI PT schedule of reinforcement in Experiment 1C and hyperbolas depicting the estimated relationship between these variables Obtained delay s defined as time between the last response B I and k are fitted parameters that represent estimated response rate under immediate reinforcement and estimated degree of discounting, respectively. A) Data for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement was in place. C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place.

PAGE 59

59 F igure 2 14 obtained delay, during both components of the tandem VI PT schedule of reinforcement in Experiment 1C and hyperbolic curve depicting the estimated relationship between these variables Obtained delays defined as time between the last B I and k are fitted parameters that represent estimated response rate under immediate reinforcement and estimated degree of discounting, respectively. A) Data for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement was in place. C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place.

PAGE 60

60 Figure 2 15 Percentage of responses across consecutive delayed reinforcement (closed circles) and immediate reinforcement (open circles) sessions during Experiment 1C A) Data for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place in the delayed reinforcement condition. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement was in place in the delayed reinforcement condition C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place in the delayed r einforcement condition.

PAGE 61

61 Figure 2 reinforcement in Experiment 1C A) Data for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place in the delayed reinforcement con dition. B) Data for Alice. A tandem VI 30 PT 10 s schedule of reinforcement was in place in the delayed reinforcement condition. C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place in the delayed reinforcement condition.

PAGE 62

62 Figure 2 17 Breakpoints obtained during the tandem VI PT schedule of reinforcement in the delayed reinforcement condition of Experiment 1C A) Data for Vlade. A tandem VI 30 PT 5 s schedule of reinforcement was in place. B) Data for Alice. A tandem V I 30 PT 10 s schedule of reinforcement was in place. C) Data for Walden. A tandem VI 15 PT 10 s schedule of reinforcement was in place.

PAGE 63

63 CHAPTER 3 THE EFFECTS OF D ELAYED REINFORCEMENT ON DISCRIMINATION ACQUISITION DURING CONDITIONAL D ISCRIMINATION TR AINING Conditional Discrimination Training To date, no research has evaluated whether unsignaled delayed reinforcement alone can lead to discrimination acquisition by individuals with DD. Grindle and Remington (2002) conducted an initial evaluation of the effects of reinforcement delay on discrimination training and found that 3 children diagnosed with ASD could be taught to receptively identify one new picture at a time when reinforcement followed a signaled 5 s delay and an error correction procedure was implemented following incorrect responses As noted by the authors, it might be important to evaluate delayed reinforcement when a signal is not provided following the response. Additionally, it may be useful to evaluate the effects of longer delays and t o evaluate whether delayed reinforcement can lead to conditional discrimination acquisition when more than one alternative is targeted at a time and an error correction procedure is not implemented Finally, it is important to identify the conditions under which delayed reinforcement will produce conditional discrimination acquisition. Purpose The purpose of Experiment 2A wa s to exami ne whether conditional discriminations could be acquired by children with DD during conditional discrimina tion training when reinforcers we re delivered following unsignaled delays to reinforcement The purpose of Experiment 2B wa s to evaluate how the availability of re sponses during the delay affected discrimination acquisition The purpose of Experiment 2C wa s to evaluate the effects of delayed reinforcement when the number of alternative s targeted during each condition wa s increased from two to four The purpose of Experiment 2D

PAGE 64

64 wa s to evaluate wh ether immediate reinforcement would produce discrimination acquisition when ITI w e re matched to ITI programmed during one of the delayed reinforcement condition s. General Methods Subjects and Setting Subjects were preschool age and school aged individuals diagnosed with DD All subjects had limited verbal repertoires, as indexed by th eir below average performance on the expressive and receptive communication components of the Battelle Developmental Inventory administered by the pre Kindergarten school psychologist. Jorma was a 4 ye ar old boy diagnosed with ASD Alice was a 5 year old girl diagnosed as developmental ly delayed Victor was a 4 year old boy diagnosed with ASD Vlade was a 5 year old boy diagnosed with ASD Jade was a 3 year old girl diagnosed as developmental ly delayed Morgan was a 7 year old girl diagnosed with Down Synd rome. Amira was a n 8 yea r old girl diagnosed with an Intellectual disability Mara was a 4 year old girl diagnosed as developmentally delayed All subjects scored in the below average range in the areas of receptive and expressive communication on the Batt elle Developmental Inventory. Alice and V lade had previously served as subjects in Experiments 1A, 1B, and 1C. Alice, Amira, Jade, Jorma, Mara, Morgan, Victor, and Vlade, served as subjects in Experiment 2A. Jade, Mara, and Victor served as subjects in Exp eriment 2B. Jade and Jorma served as subjects in Experiment 2C Amira, Jade, and Morg an served as subjects in Experiment 2D

PAGE 65

65 Target Response s and Data Collection The target responses di ffered for each subject, but generally consisted of academic discrimination tasks (e.g., receptive identification of a num eral from a field of stimuli fo llowing a verbal prompt ). Independent observers collected data on correct responses, incorrect respons es, and the absence of responding using handheld computers. Responses made within 5 s of the instruction that corresponded with the instruction were counted as correct. Responses that did not correspond with the instruction were counted as incorrect. Respo nses were counted as incorrect if the subject pi cked both response options (i.e. if the subject placed both stimuli in the experimenters hand ) or if the subj ect did something with the stimuli other than place them in the respons Interobserver agreement (IOA) data were collected during 29.40 % of all sessions during Experiment 2A, 20.91 % of all se ssions during Experi ment 2B, 53.85 % of all sessio ns du ring Experiment 2C, and 32 .55 % of all sessions during Experiment 2 D IOA was calculated by dividing each session into 10 s intervals, calculating agreement for each interval, and averaging across intervals. Across subjects, IOA averaged 95.53% (range, 78.1 3 % to 100%) for Experiment 2A, 96.66% (range, 80.15 % to 100%) for Experiment 2B, 97.87% (range, 89.17% % t o 100%) for Experiment 2C, and 96.96% (range, 80 % to 100 %) for Experiment 2 D. General Procedure Preference assessment Prior to each discrimination trai ning session, subjects were allowed to choose from an array of edible items and toys. The item chosen first was subsequently used as a

PAGE 66

66 reinforcer duri ng the remainder of the session, unless the subject indicated that a new item became more preferred. Bas eline assessment During baseline, consequences were never p rovided for correct or incorrect responses Throughout all evaluations, either two or four alternatives were targeted per condition. During each session, a fixed number of trials of each instructio n w as delivered The order in which each instruction was presented was randomize d across trials. front of the subject, and then randomly alternated both the position of the stimu li and the t side of field. Instructions associated with low percentages of correct respon ses were targeted during conditional discrimination training. Conditional d iscrimination training During conditional discrimination training, sessions were identical to baseline, e xcept that reinforcers were delivered following correct responses. Incorrec t responses did not result in differential consequences. During the im mediate reinforcement condition a correct response resulted in the immediate delivery of a small edible item or 30 s of access to a preferred toy on a FR 1 schedule. During the delayed reinforcement condition, either a chained (Experiments 2A and 2C) or tandem (Experiment 2B) FR 1 FT schedule of reinforcement was in effect. During the chained schedule, stimuli were

PAGE 67

67 removed during the delay During the tandem schedule, stimuli remained in front of the subject during the delay. Preferred items were delivered following the completion of the FT component. Experiment 2A: Procedure In Experiment 2A, a combined nonconcurrent multiple baseline multielement design was used to evaluate whether the delivery of a preferred item (a) immediately or (b) following an unsignaled delay could produce discrimination acquisition. Two alternatives were targeted during each condition. The duration of the unsignaled delay varied across evaluati ons and ranged from 20 to 40 s. For all subjects except for Amira, a 20 s delay was initially programmed in the delayed reinforcement condition. If this delay led to discrimination acquisition, a 30 s delay to reinforcement was subsequently evaluated. If the 30 s delay to reinforcement led to discrimination acquisition, a 40 s delay to reinforcement was subsequently evaluated. Immediate reinforcement sessions were alternated with delayed reinforcement s essions. For all subjects except Jade, a total of 20 trials (10 trial s of each alternative) were administe red per session. For Jade, with the exception of the 20 s delayed reinforcement versus immediate reinforcement c omparison and the first 24 session s of the 30 s delayed reinforcement versus immediate reinforcement compar ison, a total of 10 trials (5 trials of each alternative) were administered per session. This was done because within session analyses revealed that Jade performed better during the first 10 trials. Experiment 2A: Results and Discussion Results from Exper iment 2A are depicted in Figures 3 1, 3 2, and 3 3 During the baseline conditions of all evaluations of Experiment 2A percent correct was at or below leve ls expected by chance. Discrimination acquisition was defined as

PAGE 68

68 percent correct equ al to or greater than 85% across three consecutive sessions. During the first reinforcement condition immediate reinforcement led to discrimin ation acquisition for 6 out of 7 subjects and 20 s delays to reinforcement led to discrimin ation acquisition for 4 out of 7 subjects. Results were replicated with new sets of stimuli for Jorma and Alice. For the 4 subjects who acquired discriminations under delayed reinforcement rates of discrimination acquisition were comparable across immediate reinforcement and 20 s delayed reinforcement conditions. For Mara, the subject who did not acquire discriminations under either immediate reinforcement or delayed reinforcement we subsequently increased response ef fort by having the subject stand up and walk 3 ft to select a card following each instruction This manipulation resulted in further separation between the delayed reinforcement and immediate reinforcement condition during most sessions but did not lead to discrimination acquisition in either condition. These resu lt s are depicted in Figure 3 1. W e subsequently found that 30 s delays to reinforcement led to discrimination acquisition for 4 out of 4 subjects However, for Jade discriminations were acquired slower under the 30 s delay ed reinforcement condition relative to the immediate reinforcement condition Jade only acquired discriminations under 3 0 s delays to reinforcement when the total number of trials was reduced from 20 to 10. These data are displayed in Figure 3 2 We also found tha t 40 s delays to reinforcem ent led to discrimination acq uisition for 3 out of 3 subjects. These data are displayed in Figure 3 3. For J orma, we replicated the effect with a new set of stimuli. During this evaluation, rates of discrimination acquisition were comparable across immedia te reinforcement and 40 s delayed reinforcement conditions for Jorma

PAGE 69

69 s delayed reinforcement condition relative to the immediate reinforcement condition. In summary, t hese results suggest that brief (i.e., 20 to 40 s) delays to reinforcement can produce conditional discrimination acquisition for some children with DD. These findings extend the results of previous research (Grindle & Remington, 2002) by demonstrating th at delayed reinforcement can produce discrimination acquisition (a) in the absence of a prompting procedure, (b) when a signal is not programmed between a response and a reinfor cer, (c) when larger delay s are programmed, and (d) when more than one response a lternative is targeted in each condition One limitat ion of the current study was that the subject was unable to make either correct or incorrect responses during the delay. The efficacy of delayed reinforcement may depend on the type of responses that occur during the delay. In addition, t he removal of stimuli following a response, even though not directly correlated with rein forcement, may have functioned as a signal and made the preceding response more salient. Thus, the purpose of Experiment 2B was t o examine whether procedures used in Experiment 2A would produce discrimination acquisition when stimuli were available during the delay, and the subject was free to make either correct or incorrect responses during the delay. A trial based procedure was p rogrammed to limit the number and type of responses that could occur during the delay (i.e., subjects were limited to making only correct or incorrect responses during the delay). Experiment 2B: Procedure In Experiment 2B, a combined nonconcurrent multipl e baseline multielement design was used to evaluate whether the delivery of a preferred item (a) immediately or

PAGE 70

70 (b) following an unsignaled delay in which response alternatives were available during the delay, w ould produce discrimination acquisition. Tw o alternatives were targeted during each condition. The duration of the unsignaled delay was 20 s (Mara) 30 s (Jade), or 40 s (Victor) Immediate reinforcement sessions were alternated with delayed reinforcement s essions. For Mara and Victor, a total of 2 0 trials (10 trials of each alternative) were administered per session. For Jade, a total of 10 trials (5 trials of each alternative) were administered per session. During each session, the experimenter randomly alternated between two instructions. During the delayed reinforcement condition, a tandem FR 1 FT schedule of reinforcement was in effect. During both components of the tandem schedule, stimuli were left out in front of the subject. Thus, the subject could continue to make either correct or incor rect responses during the delay. Following any of these responses, the experimenter simply replaced the cards back in front of the subject. Ten trials of each of two instructions were delivered per session. Experiment 2B: Results and Discussion Figure 3 4 displays percent correct as a function of consecutive sessions during both delayed reinforcement and immediate reinforcement conditions of Experiment 2B. The availabi lity of stimuli during the delay did not interfere with discrimi nation acquisition for Ja de or Victor In addition, for Victor discriminations were acquired rapidly across delayed reinforcement and immediate reinforcement conditions. For Jade, the rate of discrimination acquisition was initially slower under the delayed reinforcement conditio n relative to the immediate reinforcement condition. However, when a replication was ran with new sets of stimuli, rates of discrimination acquisition were comparable across

PAGE 71

71 condition s For Mara the availability of stimuli during the delay interfere d with dis crimination acquisition Differences between subjects may have been due to the fact that Victor did not respond during the delay J ade generally made correct responses during the delay and Mara made both correct and incorrect responses during the del ay. Thus, for Victor, reinforcers were never preceded by incorrect responses during the delay. On average, 81.37 (range, 53.33% 100%) In addition, for Jade, only 17.27 % (range, 0% 71.43%) of reinforce rs were preceded by an incorrect response and 47.17 % (range, 0% 100%) of final correct responses were not followed by reinforcement. O n the other hand, o n average, only 45.82% (range, 31.82% addition, for Mara, 52.71 % (range, 28.57% 70%) of reinforcer s were preceded by an incorrect response and 43.49 % (range, 7.69% 85.71%) of final correct responses were not followed by reinforcement. Final responses were defined as the last response made before a trial ended. These responses could have consisted of the initial response (the response for which a reinforcement contingency was programmed) or a response that occurred during the delay ( a r esponse for which reinforcement contingencies were not programmed). These results suggest that the ability to respond during the delay may prevent discrimination acquisition if subjects often make incorrect responses during the delay and if reinforcers are often preceded by incorrect responses Less is known a bout whether discrimination acquisition will occur when reinforcement is delayed and more than two alternatives are targeted at one time. Although Hockman and Lipsitt (1961)

PAGE 72

72 found that increasing the number of alternatives targeted at one time made it less likely that typically functioning school age children would acquire discriminations under delayed reinforcement, th e study had several limitations (noted previously). In addition, to date, no study has examined the combined effects of delayed reinforcemen t and an increase in the number of alte rnatives targeted on discrimination acquisition by school aged individuals with DD. Thus, the purpose of Experiment 2C was to evaluate whether discrimination acquisition would still occur when the number of alternativ es was increased from two to four. Experiment 2C: Procedure In Experiment 2C, a combined nonconcurrent multiple baseline multielement design was used to evaluate whether the delivery of a preferred item (a) immediately or (b) following an unsignaled del ay could produce discrimination acquisition. Unlike Experiments 2A and 2B in which two alternatives were targeted during each condition, four alternatives were targeted during each condition. The duration of the unsignaled delay varied across subjects and evaluati ons and ranged from 10 to 40 s. Immediate reinforcement sessions were alternated with delayed reinforcement sessions. The experimenter randomly selected one of the four instructions during each trial until 5 trials of each instruction had been d elivered. Experiment 2C: Results and Discussion Figure 3 5 displays results from Experimen t 2C For Jade, a child for whom 30 s delays to reinforcement had previously produced discrimination acquisition when two alternatives were targeted, 30 s delays to r einforcement produce d discrimination acquisition when four alternatives were targeted. In addition, discrimination acquisition occurred in the immediate reinforcement condition when four alternatives were targeted.

PAGE 73

73 The rate of discrimination acquisition wa s comparable across conditions. For Jorma, a child for whom 40 s delays to reinforcement had previously produced discrimination acquisition when two alternatives were targeted, 40 s delays to reinforcement did not produce discrimination acquisition when fo ur alternatives were targeted. In addition, Jorma began to protest when he observed the stimuli associated with the delayed reinforcement condition on the table. Howe ver, immediate reinforcement produce d discrimination acquisition when four alternatives we re targeted. We then replicate d the evaluatio n using different stimuli and brief er (i.e., 10 s) delay to reinforcement (third and fourth conditions). A gain, immediate reinforcement produced discrimination acquisition when four alternatives w ere targeted. A lthough 10 s delay s to reinforcement started to produce discrimination acquisition, discrimination acquisition wa s not conclusively demonstrate d because Jorma left the school before the evaluation could be completed In summ ary, these results suggest that for some subjects, the efficacy of delayed reinforcement may depend on the number of alternatives targeted at one time, with a negative relationship between the number of responses targeted and the length of the delay associated with discrimination acqui sition. In cases in which discrimination acquisition does not occur under delayed reinforcement, it is difficult to evaluate whether the delay to reinforcement, or the longer ITI associated with delayed reinforcement relative to immediate reinforcement, a re responsible for the lack of discrimination acquisition. Thus, the purpose of Experiment 2D was to compare the effects of different ITI lengths on discrimination acquisition. Experiment 2 D : Procedure In Experiment 2D a combined nonconcurrent multiple b aseline multielement design was used to compare the occurrence and speed of discrimination acquisition

PAGE 74

74 during conditions in which either 0 s or 30 s ITI were programmed In both conditions, reinforcers were delivered immediately on a FR 1 schedule contin gent on correct responses. Two response alternatives were targeted during each condition. Either five (Jade) or ten (all other subjects) trials of each of two instructions were delivered per session. The experimenter randomly alternated between the instruc tion s Experiment 2 D : Results and Discussion The percentage of correct responses made during both 0 s ITI and 30 s ITI conditions in Experiment 2 D are depicted in Figure 3 6 For all 3 subjects discrimination acquisition occurred in the 30 s ITI conditio n. For Jade (top panel), discrimination acquis i tion initially fa iled under the 0 s ITI condition However, for Jade, discrimination acquisition had previously occurred under immediate reinforcement and 0 s ITI ( see bottom panels of Figures 3 1 and 3 2 mid dle panel of Figure 3 4, and top panel of Figure 3 5 ) Thus we hypothesized that something about the alternatives targeted (the numerals 13 and 14) hindered discrimination acquisition independent of ITI length and we compared 0 s and 30 s ITI conditions with two new set s of stimuli. During the second and third evaluations discrimination acquisition occurred in both 0 s and 30 s ITI conditions For Amira discrimination acquisition occurred slightly more rapidly in the 30 s ITI condition relative to the 0 s ITI condition For Morgan, the rate of discrimination acquisition was comparable across conditions. In addition, discrimination acquisition occurred rapidly in each condition. These results are contrary to those of previous resea rchers (e.g., Bilodeau & Bilo deau 1959) and suggest that longer ITI may not always have detrimental effects on behavior. In addition, these results suggest that cases in which delayed reinforcement fails to produce discrimination acquisition likely

PAGE 75

75 cannot be accounted for by lon ger ITI typically associated with the delayed reinforcement condition.

PAGE 76

76 Figure 3 1. Percent c orrect across 20 s delayed reinforcement (closed circles) and immediate reinforcement (open circles) conditions in Experiment 2A A) Data for Jorma. Findings were replicated with a new set of stimuli. B) Data for Alice. Findings were replicated with a new set of stimuli. C) Data for Mara. Response effort was increased following the phase change line by requiring Mara to walk 3 ft to select a response. D) Data f or Morgan. E) Data for Victor. F) Data for Vlade. G) Data for Jade.

PAGE 77

77 Figure 3 2. Percent correct across 30 s delayed reinforcement (closed circles) and immediate reinforcement (open circles) conditions during Experiment 2A. Unless otherwise noted, 20 trials were completed per session. A) Data for Jorma. B) Data for Vlade. C) Data for Victor. D) Data for Jade. Following the second phase change line, 10 trials were completed per session.

PAGE 78

78 Figure 3 3. Percent c orrect across 40 s delayed reinforcement (c losed circles) and immediate reinforcement (open circles) conditions during Experiment 2A A) Results for Jorma Findings were replicated with a new set of stimuli. B) Results for Victor C) Results for Amira

PAGE 79

79 Figure 3 4. Percent correct across delayed reinforcement (closed circles) and immediate reinforcement (open circles) conditions of Experiment 2B. A) Data for Mara. A 20 s delay was programmed in the delayed reinforcement condition. B) Data for Jade A 30 s delay was programmed in the delayed reinfo rcement condition. Findings were replicated with a new set of stimuli. C) Data for Victor. A 40 s delay was programmed in the delayed reinforcement condition. Findings were replicated with a new set of stimuli.

PAGE 80

80 Figure 3 5. Percent correct across delayed reinforcement (closed circles) and immediate reinforcemen t (open circles) conditions of Experiment 2C when 4 alternatives were targeted per condition A) Data for Jade. A 30 s delay was programmed in the delayed reinforcement condition. B) Data for Jorma. In the first delayed reinforcement condition, a 40 s delay was programmed. In the second delayed reinforcement condition, a 10 s delay was programmed.

PAGE 81

81 Figure 3 6 Percent correct across conditions in which either 0 s (open circles) or 30 s (closed ci rcles) ITI were programmed during Experiment 2D During baseline, reinforcers were not delivered. In the reinforcement phase, reinforcers were delivered immediately contingent on correct responses. A) Data for Jade. Two replications were conducted with new sets of stimuli. B) Data for Amira. C) Data for Morgan.

PAGE 82

82 CHAPTER 4 GENERAL DISCUSSION Overall Summary Experiment 1 evaluat ed whether arrangements similar to those used in the laboratory could be used to evaluate the behavior of children with DD under co nditions of delayed reinforcement, and whether the findings from nonhum an laboratories would translate to th is population Experiment 2 evaluated (a) whether delayed reinforcement could produce discrimination acquisition in children with DD, (b) some of th e variables that might affect the efficacy of delay ed reinforcement, and (c) whether longer ITI prevent ed discrimination acquisition Experiment 1 Overall, the results of Experiment 1 suggest that reinforcement immediacy may not be essential under some te aching arrangements when delays to reinforcement are progressively increa sed following successive reinforcer deliveries and the goal is simply to maintain already established responding The results also suggest that delayed reinforcement may be more likel y to maintain responding when a stimulus change (e.g., the removal of squares dep icted on a touch screen) follows a response, even if the stimulus change is not exclusively paired with reinforcement. This was indexed by higher response rates during Experim ents 1A and 1B (which programmed stimulus removal during the terminal component) r elative to Experiment 1C (which did not program stimulus removal during the terminal component). This finding is contrary to that of previous research (i.e., Ferster, 1953). Ferster found that a VI schedule of reinforcement with response contingent blackouts not correlated with reinforcement did not produce differences in responding relative to a VI schedule that did not include

PAGE 83

83 blackouts and a VI schedule that included respon se independent blackouts. However, these discrepancies may be due to the fact that Ferster did not compare conditions in which blackouts were and were not correlated with later reinforcer delivery. (evidence o f discrimination) was lower during Experiments 1B and 1C relative to the Experiment 1A One interpretation of these results is that delayed reinforcement may be less likely to maintain discri minated responding under an intermittent schedule of reinforcemen t (i.e., a VI 30 s schedule relative to a FR 1 schedule responses may simply have been due to interval strain, independent of the delay. Previous evaluations of delayed reinforcement typically establish stead y state responding under an interval schedule prior to introducing a delay to reinforcement (e.g., Gleeson & Lattal, 1987). In the present experiments, this was not done. Instead, responding was briefly established at the start of each session by exposing subjects to 10 trials in which reinforcers were delivered immediately on a FR 1 schedule contingent Although discriminations were typically established during t his initial exposure period they quickly deteriorated once the chain ed or tandem VI PT schedule was introduced. Thus although these procedures were specifically designed to increase the speed in which we could evaluate response rates under various delays to reinforcement they may not have been sufficient to examine the ef fects of delayed reinforcement on the maintenance of discriminated responding However this arrangement may be analogous to classroom situations in which a teacher demonstrates how to complete a problem, reinforces the correct completion of a few problems and then moves away to assist other students, only to come back at a later

PAGE 84

84 he methods used in Experiments 1B and 1C in some way resemble s how delayed reinforcement contingencies operate in the natural environ ment. M ean levels of percentage correct generally did not decrease as delays to reinfor cement increased, which is contrary to the results of some experiments (e.g., Escobar & Bruner, 2007; LeSage, Byrne, & Polin g, 1996) Escobar and Bruner compared resp onding on both an operative lever and six nonoperative levers across seven groups of three rats exposed to 0 1 2 4 8 16 or 32 s delays to reinforcement. The researchers found that rats exposed to 16 or 32 s delays to reinforcement allocated mo re of their responses to the nonoperative levers relative to rats exposed to 0 or 2 s delays to reinforcement The responses to the nonoperative levers competed with the targe responses Differences betwe present experiments may be due to several procedural differences between the two studie s (e.g., within subjects versus between subjects comparisons). Lattal (2010) noted that delayed reinforcement is associated with a variety of outcomes, depending on the circumstanc es in which i t is evaluated. Experiment 2 The purpose of Experiment 2 was to examine the effects of delayed reinforcement on discrimination acquisition in children with DD and to evalu at e some of the variables that may affect the effi cacy of delayed reinforcement. The results of Experiment 2A demonstrate d that for some subject s, discriminations c ould be acquired when reinfo rcers were delivered following a 20 to 40 s unsignaled delay. Th e results of Experiment 2B demonstrate d that delayed reinforcement could produce discrimination

PAGE 85

85 acquisition when responses occur during the delay, as long as incorrect responses do not occur often The results of Experiment 2C demonstrated that, for some i ndividuals, the efficacy of delayed reinf orcement was dependent on the number of response alternati ves targeted per condition ITI length (which was longer in all previous delayed reinforcement conditions in Experiment 2 relative to immediate reinforcement conditions) was not found to produce an effect on the rapidity or occurrence of discrimination acquisition. Findings from Experiment 2 have several implications for practitioners who use discrimination training to teach individuals with DD new skills. Fi rst, findings suggest that, in many cases, practitioners can complete other necess ary tasks (e.g., data recording blocking of dangerous behavior) prior to delivering a reinforcer as long as the reinforcer is delivered within a reasonable amount of time an d as long as there is a contingency arranged between a correct response and reinforcement. Second, results suggest that, in cases in which practical limitations necessitate the delay of reinforcement, it may be necessary to limit the number of alternatives targeted at one time for some individuals Third, results suggest that it may not be necessary to remove stimuli immediately following a response. Although these implications may seem minor in isolation collectively, they may help practitioners better ma nage their time, and they may make procedures appear less daunting to new practitioners in the field. It should be noted, however, that evaluations similar to those conducted in Experiments 1 and 2 should be conducted for each individual to determine the e ffects of delayed reinforcement on an individual basis. In addition, results do not suggest that reinforcement should be intentionally delayed. Instead, results simply suggest that there

PAGE 86

86 may be more leeway in the extent to which reinforcers must be provide d immediately under some controlled circumstances. Conclusions The m ethods used in Experiments 1 and 2 extend the methods used in basic research (e.g., Reilly & Lattal, 2004) to children with DD As noted by Mace and Critchfield (2010), more translational research is needed to bridge the gap between basic and applied research. Replications of effects seen with nonhumans provide important empirical support for the notion of interspecies generality of operant principles and suggest that the questions asked b y basic researchers have some applied value (Mace & Critchfield). Skinner was a proponent of translational research, arguing that the importance of a science of behav ior derives largely from the possibility of an eventual Experiment 1 replicated the results of basic, laboratory experiments, in that increases in the delay to reinforcement were associated with decrease s in response rate (e.g., Reilly & Lattal, 2004). This relationship was nicely described by the modified subjects ranking of the parameter k which represents estimated degree of d iscounting due to the delay, did not remain consistent across Experiments 1A, 1B, and 1C. For example, k k value in Expe riments 1A and 1B, but lower tha k value in Experiment 1C. These results suggest that r elative ranking of k values may not remain consistent across time. Experiment 2 extended the results of basic, laboratory experiments to an arrangement commonly used to teach children with DD new skills. It is unknown whether delays to reinforcement are li kely to occur under

PAGE 87

87 this arrangement in the natural environment. However, this arrangement was designed as one way to carefully control the number and type of responses that could occur during the delay. Future evaluations will first examine the effects of delayed reinforcement on free operant responding under controlled, laboratory conditions, next examine the effects of delayed reinforcement on both problem behavior and appropriate behavior in analogue settings, and finally examine the effects of delayed reinforcement on both problem behavior and appropriate behavior in the natural environment. There has been some debate about the mechanisms responsible for response acquisition and maintenance under delayed reinforcement (see Lattal, 2010, for a review) For example, it has been suggested that responding can be maintained when reinforcement (Keller & Schoenfeld, 1950). Indeed, Ferster (1953) found that each subject consist ently engaged in a particular response (e.g., turning in a circle with the head stretched high) during a 60 s delay interval. These responses then continued to occur throughout an approximately 100 hr experimental period. In the present investigation all subjects engaged in alternative responses during the delay interval. These responses varied over time and across subjects, ranging from stereotypy (Walden) to drinking water (Vlade and Amira ). The lack of consistency suggests that the efficacy of delayed r einforcement was not occurring during the delay. Instead, these responses could have either been adjunctive responses induced by the reinforcement schedule ( e.g., water consumption, Falk, 1961) or responses that were briefly adventitiously rein forced. Reeve et al. (1993) noted that responses that are adventitiously reinforced by temporally contiguous reinforcement

PAGE 88

88 may vary and fade over time as the overall contingency (reinforcement for the target response) begins to c ontrol behavior. Assuming that the procedures used in the current investigation accurately estimate response rates under delayed reinforcement the results of this investigation may have implications for skill acquisition and maintenance of pro blem behavi or. I t may be the case that teachers need not provide reinforcers immediately to maintain an appropriate response, even if problem behavior occurs following those appropriate responses. However, it may also be the case that the delivery of reinforcers for appropriate responses that occur some time after problem behavior may actually serve to maintain problem behavior. However, it is difficult to make a jump from the systematic controlled arrangement set up in the current experiments to the more complex cont ingencies operating in the natural environment. It is hoped that the current series of studies will set the stage for more research on delayed reinforcement in complex environments.

PAGE 89

89 LIST OF REFERENCES Bilodeau E. A., & Bilodeau, I. M. (1958). Variation of temporal intervals among critical events in five studies of knowledge of results. Journal of Experimental Psychology, 55, 603 612. Brackbill Y., & Kappy, M. S., (1962). Delay of reinforcement and retention. Journal of Comparative and Physiological Ps ychology, 55, 14 18. Catania A. C. (1971). Reinforcement schedules: The role of responses preceding the one that produces the reinforcer. Journal of the Experimental Analysis of Behavior, 15, 271 287. Catania A. C. (2007). Learning (4 th ed.). New York Sloan Publishing Critchfield, T. S., & Lattal, K. A. (1993). Acquisition of a spatially defined operant with delayed reinforcement. Journal of the Experimental Analysis of Behavior, 59, 373 387. Denny, M. A., Allard, M., Hall, E., & Rokeach, M. (1960) Supplementary report: Delay of knowledge of results, knowledge of task, and intertrial interval. Journal of Experimental Psychology, 60 327 Dews, P. B. (1960). Free operant behavior under conditions of delayed reinforcement. I. CRF type schedules. Jou rnal of Experimental Psychology, 45, 27 45. Dickinson, A., Watt, A., & Griffiths, W. J. H. (1992). Free operant acquisition with delayed reinforcement. The Quarterly Journal of Experimental Psychology, 45, 241 258. Dixon, M. R., Horner, M. J., & Guercio, J. (2003). Self control and the preference for delayed reinforcement: An example in brain injury. Journal of Applied Behavior Analysis, 36, 371 374. Erickson, M. T., & Lipsitt, L. P. (1960). Effects of delayed reward on simultaneous and successive discri mination learning in children. Journal of Comparative & Physiological Psychology, 53, 256 260. Escobar, R., & Bruner, C. A. (2007). Response induction during the acquisition and maintenance of lever pressing with delayed reinforcement. Journal of the Expe rimental Analysis of Behavior, 88, 29 49. Falk J. L. (1961). Production of polydipsia in normal rats by an intermittent food schedule. Science, 133, 195 196. Ferster, C. B. (1953). Sustained behavior under delayed reinforcement. Journal of Experimental Psychology, 45, 27 45.

PAGE 90

90 Ferster C. B., & Skinner B.F. (1957) Schedules of reinforcement New York: Appleton Century Crofts. Fleshler, M., & Hoffman, H. S. (1962). A progression for generating variable interval schedules. Journal of the Experimental Analy sis of Behavior, 5, 529 530. Fowler H., & Trapold M. A. (1962) Escape performance as a function of delay of reinforcement. Journal of Experimental Psychology, 63, 464 467. Gleeson, S., & Lattal, K. A. (1987). Response reinforcer relations and the main tenance of behavior. Journal of the Experimental Analysis of Behavior, 48, 383 393. Grindle, C. F., & Remington, B. (2002). Discrete trial training for autistic children when reward is delayed: A comparison of conditioned cue value and response marking. J ournal of Applied Behavior Analysis, 35, 187 190. Herrnstein R. J. (1970) On the law of effect. Journal of the Experimental Analysis of Behavior, 13, 243 266. Hockman C. H., & Lipsitt L. P. (1961). Delay of reward gradients in discrimination learning with children for two levels of difficulty. Journal of Comparative Physiological Psychology, 54, 24 27. Keely, J., Feola, T., & Lattal, K. A. (2007). Contingency tracking during unsignaled delayed reinforcement. Journal of the Experimental Analysis of B ehavior, 88, 229 247. Keller, F. S., & Schoenfeld, W. N. (1950). Principles of Psychology New York: Appleton Century. Lattal, K. A. (2010). Delayed reinforcement in operant behavior. Journal of the Experimental Analysis of Behavior, 93, 129 139. Lattal K. A., & Gleeson, S. (1990). Response acquisition with delayed reinforcement. Journal of Experimental Psychology, 16, 27 39. LeSage, M. G., Byrne, T., & Poling, A. (1996). Effects of d amphetamine on response acquisition with immediate and delayed reinf orcement. Journal of the Experimental Analysis of Behavior, 66, 349 367. Mace, F. C., & Critchfield, T. S. (2010). Translational research in behavior analysis: Historical traditions and imperative for the future. Journal of the Experimental Analysis of Be havior, 93, 293 312. Mazur, J. E. (1987). An adjusting procedure for studying delays of reinforcement. In M. L. Commons, J. E. Mazur, J. A. Nevin, & H. Rachlin (Eds.), Quantitative analyses of behavior, Vol 5: The effects of delays and of intervening even ts on reinforcement value (pp. 55 73). Hillsdale, N: Erlbaum.

PAGE 91

91 Meichenbaum, D. H., Bowers, K. S., & Ross, R. R. (1968). Modification of classroom behavior of institutionalized female adolescent offenders. Behavior Research & Therapy, 6, 343 353. Millar W S. (1990). Span of integration for delayed reward contingency learning in 6 to 8 month old infants. Annals of the New York Academy of Sciences, 608, 239 266. Millar, W. S., & Watson, J. S. (1979). The effect of delayed feedback on infant learning reexa mined. Child Development, 50, 747 751. Miltenberger, R. (2008). Behavior modification (4 th ed.). California, Thomson Wadsworth. Morse W. H., & Skinner B. F. (1957) A second type of superstition in the pigeon. The American Journal of Psychology, 70, 308 311. Okouchi, H. (2009). Response acquisition by humans with delayed reinforcement. Journal of the Experimental Analysis of Behavior, 91, 377 390. Reeve, L., Reeve, K. F., & Poulson C. L. (1993). A parametric variation of delayed reinforcement in infan ts. Journal of the Experimental Analysis of Behavior, 60, 515 527. Reeve, L., Reeve, K. F., Brown, A. K., Brown, J. L., & Poulson C. L. (1992). Effects of delayed reinforcement on infant vocalization rate. Journal of the Experimental Analysis of Behavior 58, 1 8. Reilly, M. P., & and Lattal, K. A. (2004). Within session delay of reinforcement gradients. Journal of the Experimental Analysis of Behavior, 82 21 35. Renner K. E. (1964). Delay of reinforcement: A historical review. Psychological Bulletin 61, 341 361. Rieber M. (1961). The effect of CS presence during delay of reward on the speed of an instrumental response. Journal of Experimental Psychology, 61, 290 294. Saltzman, I. J., Kanfer, F. H., & Greenspoon J. (1955) Delay of reward and hum an motor learning. Psychological Reports, 1, 139 142. Schaal, D. W., & Branch, M. N. (1988). Responding of pigeons under variable interval schedules of unsignaled, briefly signaled, and completely signaled delays to reinforcement. Journal of the Experiment al Analysis of Behavior, 50, 33 54. Schwarz, M. L., & Hawkins, R. P. (1970). Application of delayed reinforcement procedures to the behavior of an elementary school child. Journal of Applied Behavior Analysis, 3, 85 96. Sidman, M. (1960). Tactics of scie ntific research. New York: Basic Books.

PAGE 92

92 Skinner, B. F. (1938). The behavior of organisms: An experimental analysis. New York: Appleton Century Crofts. Skinner, B. F. (1953). Science and human behavior. New York: The Free Press. Spence K. W. (1956) Beha vior theory and conditioning. New Haven: Yale University Press. Sutphin, G., Byrne, T., and Poling, A. (1998). Response acquisition with delayed reinforcement: A comparison of two lever procedures. Journal of the Experimental Analysis of Behavior, 69, 17 28. Terrell G. (1958) The role of incentive in discrimination learning in children. Child Development, 29, 231 236. Terr ell, G., & Ware R. (1961) Role of delay of reward in speed of size and form discrimination learning in childhood. Child Developmen t, 32, 409 415. Vollmer, T. R., Borrero, J. C., Lalli, J. S., & Daniel, D. (1999). Evaluating self control and impulsivity in children with severe behavior disorders. Journal of Applied Behavior Analysis, 32, 451 466. Ware, R., & Terrell, G. (1961). Eff ects of delayed reinforcement on associative and incentive factors. Child Development, 32, 789 793. Watson, J. B. (1917). The effects of delayed feeding upon learning. Psychobiology, 1, 51 59. Williams, B. A. (1976). The effects of unsignaled delayed rei nforcement. Journal of the Experimental Analysis of Behavior, 26, 44 449. Williams B. A. (1998) Relative time and delay of reinforcement. Learning and Motivation, 29, 236 248. Williams A. M., & Lattal K. A. (1999) The role of the response reinforcer relation in delay of reinforcement effects. Journal of the Experimental Analysis of Behavior, 71, 187 194.

PAGE 93

93 BIOGRAPHICAL SKETCH Jolene Rachel Sy was born in Sacramento, CA in 1981 and was raised in the Sacramento area. In 1998, Jolene moved to Santa Cruz CA to attend the University of California, Santa Cruz. After earning a Bachelor of Arts in l anguage s tudies, Jolene began work as a behavior analyst working with children with ASD It was at this time that she became interested in behavior analysis. In 2 005, Jolene moved to Stockton, CA to attend the University of the Pacific (UOP). At UOP, Jolene served as a research earned a m aster s degree under his supervision. In 2007, Jolene entered the doctoral program in behavior analysis under the supervision of Dr. Tim Vollmer. Jolene earned a doctor of philosophy in 2011, and will relocate to Saint Louis to begin an assistant professor position at Saint Louis University