<%BANNER%>

Evaluations of the Overjustification Hypothesis

Permanent Link: http://ufdc.ufl.edu/UFE0045322/00001

Material Information

Title: Evaluations of the Overjustification Hypothesis
Physical Description: 1 online resource (56 p.)
Language: english
Creator: Peters, Kerri Patricia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

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

Notes

Abstract: The utility of reinforcement-based procedures has been well established in the behavior analysis literature. The overjustification effect is one commonly cited criticism of programs that usetangible rewards. Rewards are frequently delivered in school settings andinclude grades, stickers, social praise, and star charts. The overjustification hypothesis suggests that the delivery of an extrinsic (socially mediated) reward contingent on engagement with an activity that previously occurs at somelevel without apparent socially mediated reinforcement will result in a reduction in the amount of engagement in that activity from baseline levelswhen the reward phase is discontinued. This series of studies evaluated the effects of delivering tangible rewards contingent on engagement with age-appropriate leisure activities on the amount of engagement in the condition after reinforcement was discontinued. The Subjects in Studies 1-3 were children in first grade. In these Studies, tangible rewards were delivered for engaging with the preferred activities under different conditions, for example when one or several items was or were available at a time. The results did not support the overjustification hypothesis; however they suggest that extended exposure to a preferred item mayhave effects that look similar to the overjustification effect. Study 4 was a direct replication of a landmark study evaluating the overjustification hypothesis conducted by Deci (1971). The subjects in Study 4 were undergraduate students. Contrary to previous findings, both groups showed a decrease in engagement and individual data varied greatly, not providing support for the overjustification effect.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kerri Patricia Peters.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Vollmer, Timothy Raymond.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045322:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045322/00001

Material Information

Title: Evaluations of the Overjustification Hypothesis
Physical Description: 1 online resource (56 p.)
Language: english
Creator: Peters, Kerri Patricia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

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

Notes

Abstract: The utility of reinforcement-based procedures has been well established in the behavior analysis literature. The overjustification effect is one commonly cited criticism of programs that usetangible rewards. Rewards are frequently delivered in school settings andinclude grades, stickers, social praise, and star charts. The overjustification hypothesis suggests that the delivery of an extrinsic (socially mediated) reward contingent on engagement with an activity that previously occurs at somelevel without apparent socially mediated reinforcement will result in a reduction in the amount of engagement in that activity from baseline levelswhen the reward phase is discontinued. This series of studies evaluated the effects of delivering tangible rewards contingent on engagement with age-appropriate leisure activities on the amount of engagement in the condition after reinforcement was discontinued. The Subjects in Studies 1-3 were children in first grade. In these Studies, tangible rewards were delivered for engaging with the preferred activities under different conditions, for example when one or several items was or were available at a time. The results did not support the overjustification hypothesis; however they suggest that extended exposure to a preferred item mayhave effects that look similar to the overjustification effect. Study 4 was a direct replication of a landmark study evaluating the overjustification hypothesis conducted by Deci (1971). The subjects in Study 4 were undergraduate students. Contrary to previous findings, both groups showed a decrease in engagement and individual data varied greatly, not providing support for the overjustification effect.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kerri Patricia Peters.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Vollmer, Timothy Raymond.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045322:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVALUATIONS OF THE O VERJUSTIFICATION HYP OTHESIS By KERRI PATRICIA PETERS 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 2013

PAGE 2

2 2013 Kerri Patricia Peters

PAGE 3

3 To my loving family and friends who are like family

PAGE 4

4 ACKNOWLEDGMENTS I woul d like to acknowledge Timothy Vollmer for outstanding advising and for his mentoring both ins ide and outside the laboratory; I cannot provide enough thanks. Additionally, I would like to thank Brian Iwata, Scott Miller, and Maureen Co nroy for sitting on my defense committee and for comments on a previous draft of this paper. I would like to acknowledge Sean Peters, Michael Berard, and Allison Hamilton for their constant encouragement and support, and Kyle Peters for changing my life an d becoming my inspiration.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 2 GENE RAL METHOD, STUDIES 1 3 ................................ ................................ ....... 16 Subjects, Setting, and Response Definitions ................................ .......................... 16 Data Collection and Analysis ................................ ................................ .................. 16 General Procedure ................................ ................................ ................................ .. 17 Baseline ................................ ................................ ................................ ............ 17 Reward ................................ ................................ ................................ ............. 17 3 STUDY 1: THE EFFECTS OF CONTINGENT TANGIBLES ON ENGAGEMENT WITH A SINGLE ITEM ................................ ................................ ............................ 18 Method ................................ ................................ ................................ .................... 18 Subjects ................................ ................................ ................................ ............ 18 Design and Procedure ................................ ................................ ...................... 18 Interobserver Agreement ................................ ................................ .................. 19 Results and Discussion ................................ ................................ ........................... 19 4 STUD Y 2: THE EFFECTS OF CONTINGENT TANGIBLES ON ENGAGEMENT WITH MULTIPLE ITEMS ................................ ................................ ........................ 21 Method ................................ ................................ ................................ .................... 21 Subjects ................................ ................................ ................................ ............ 21 Design and Procedure ................................ ................................ ...................... 21 Interobs erver Agreement ................................ ................................ .................. 22 Results and Discussion ................................ ................................ ........................... 22 5 STUDY 3: THE EFFECTS OF EXTENDED EXPOSURE ON ACTIVITY ENGAGEMENT ................................ ................................ ................................ ...... 26 Study 3A: Evaluation of the Effects of an Extended Baseline with Multiple Activities ................................ ................................ ................................ .............. 28 Method ................................ ................................ ................................ ............. 28

PAGE 6

6 Subjects ................................ ................................ ................................ .... 28 Design and procedure. ................................ ................................ ............... 28 Interobserver agreement ................................ ................................ ............ 28 Resu lts and Discussion ................................ ................................ .................... 28 Study 3B: Evaluation of the Effects of Contingent Tangibles on Activity Engagement with an Extended Baseline Control ................................ ................. 30 Method ................................ ................................ ................................ ............. 32 Subjects ................................ ................................ ................................ ..... 32 Design and procedure ................................ ................................ ................ 32 Interobserver agreement ................................ ................................ ............ 33 Results and Discussion ................................ ................................ .................... 33 6 STUDY 4: REPLICATION OF DECI (1971) ................................ ............................ 39 Method ................................ ................................ ................................ .................... 41 Subjects and Setting ................................ ................................ ........................ 41 Design and Procedure ................................ ................................ ...................... 41 Interobserver Agreement. ................................ ................................ ................. 42 Results and Discussion ................................ ................................ ........................... 42 7 GENE RAL DISCUSSION ................................ ................................ ....................... 48 LIST OF REFERENCES ................................ ................................ ............................... 53 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 56

PAGE 7

7 LIST OF TABLES Table page 4 1 Condition averages of percentage of session engaged with target item for all sessions in each condition and the last three sessions in each condition for each subject ................................ ................................ ................................ ....... 25 5 1 Condition averages of percentage of session engaged with Set 1 target item and Set 2 comparison item for each subject. Condition lengths for Set 2 comparison items were based on condition lengths for corres ponding Set 1 item. The top row for each subject displays the average for all sessions in each condition and the bottom row displays the average of the last 3 two session blocks for each condition. ................................ ................................ ...... 38 6 1 Number of Seconds Working on the Puzzle During the Free Choice Period ...... 45

PAGE 8

8 LIST OF FIGURES Figure page 3 1 Percentage of session engaged with the single item for Angel, Terrance, Georgette, and James. ................................ ................................ ....................... 20 4 1 Percentage of session enagaged for Sam and Angel (top two panels display all available items and bottom panels displays only target activity for Sam) ....... 24 5 1 Percentage of session engaged with all available items for Angel, Terrance, Georgette, and James ................................ ................................ ........................ 30 5 2 Percentage of session engaged with item with the initally highest percentage of engagement for Angel, Terrance, Georgette, and James ............................... 31 5 3 Per centage of session engaged with all items for Jack, Kayla, and Dennis for Set 1 and Set 2 (top and bottom panels, respectively). Sessions are combined in to two session blocks along the x axis. ................................ .......... 36 5 4 Percentage of session engaged with target items for Jack, Kayla, and Dennis for Set 1 and Set 2 (top and bottom panels, respectively). Sessions are combined in to two session bl ocks along the x axis. ................................ .......... 37 6 1 Number of seconds engaged with the puzzle during each break for individuals in the experime ntal group. ................................ ................................ 46 6 2 Number of seconds engaged with the puzzle during each break for individuals in the control group. ................................ ................................ .......... 47

PAGE 9

9 A bstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVALUATIONS OF THE O VERJUSTIFICATION HYP OTHESIS By Kerri Patricia Peters May 2013 Chair: Timothy R. Vollmer Major: Psychology The utility of reinforcement based procedures has been well established in the behavior analysis literature. However, t he overjustification effect is one commonly cited criticism of programs that use rewards. Rewards are frequently delivered in school settings and include grades, stic kers, social praise, and star charts. The overjustification hypothesis suggests that the delivery of an extrinsic (socially mediated) reward contingent on engagement with an activity that previously occurs at some level without apparent socially mediated r einforcement will result in a reduction in the amount of engagement in that activity from baseline levels when the reward phase is discontinued. This series of studies evaluated the effects of delivering tangible rewards contingent o n engagement with age a ppropriate leisure activities on the amount of engagement in the condition after rein forcement was discontinued. The s ubjects in Studies 1 3 were chi ldren in first grade. In these s tudies, t angible rewards were delivered for engaging with the preferred act ivities under different conditions, for example when one or several items was or were available at a time. The results did not support the overjustification hypothesis; however they suggest that extended exposure to a preferred item may have effects that l ook similar to the overjustification effect. Study 4 was a direct replication of

PAGE 10

10 a landmark study evaluating the overjustification hypothesis conducted by Deci (1971). The subjects in Study 4 were undergraduate students. Contrary to previous findings, both groups s howed a decrease in engagement across the study and individual data varied greatly, not providing support for the overjustification effect.

PAGE 11

11 CHAPTER 1 INTRODUCTION The utility of reinforcement based procedures has been well established in the behavior analysis literature. However, t he overjustification effect is one popular criticism of programs that use rewards (Deci, 1971; Kohn, 1993; Lepper & Greene, 1975; Lepper, Greene & Nisbett, 1973; Pink, 2009). The overjustification hypothesis states motivation to engage in the behavior that produced the reward (Roane, Fisher, & McDonough, 2003 and reinforcement ( Cameron & Pierce, 2004; Vasta, Andrews, Stirpe, & Comfort, 1978). The terms reward and reinforcer have been used interchangeably throughout the overjustification literature. A reinforcer is defined by its effect on behavior. Specifically a reinforcer is an event that increases the probability of the behavior it follows. A reward is a stimulus that is assumed to have a positive effect on behavior, but is not define d by its effects on behavior ; an increase in the probability of behavior has not necessarily been demonstrated (Cameron & Pierce). The commonly used definition of the overjustification effect does not require demonstration of a reinforcement effect only d elivery of an extrinsic reward (e.g., Deci, 1971; Lepper, Greene, & Nisbett, 1973). This is because the behavior of interest is already occurring at a high rate so increases in rate as a result of reinforcement are not necessary to see the effect (i.e., a decrease in rate across a pre reward phase to a post reward phase). A ccording to the overjustification hypothesis, the delivery of an extrinsic (socially mediated) reward contingent on engagement with an activity maintained by automatic reinforcement will when discontinued, result in a reduction in the amount of

PAGE 12

12 engagement with that activity below baseline levels. Based on the research showing this effect, many educational entities recommend against the use of rewards. For example, the Des Moines Public S chool ("Des Moines public," 2011) has a webpage backfire and reduce motivation and cited Lepper et al. (1973) and Deci (1971). In addition, one of the eight principles o f Montessori education is that extrinsic rewards have a negative impact on long term motivation and learning (Edwards, 2006). Edwards Deci (1971). Yet, the vast literature on the ove rjustification effect is riddled with methodological shortcomings, few efforts have been made to understand mechanisms accounting for the effect or lack thereof, and the primary study from which the effect is inferred has never been replicated despite the fact that several important measures were not reported. One of the most commonly cited series of studies with over 2,250 citations, on the overjustification effect was conducted by Deci (1971). Deci evaluated the effects of Deci concluded that intrinsic interest decreased when money was used as a reward and increased when verbal praise was used. Intrinsic interest was measured by recording the number of seconds the subjects worked on the puzzle task during a break period r was recorded covertly. Subsequently, there were several studies published in the 1970s addressing the issue of the detrimental effects of rewards (e.g., Lepper & Green, 1975; Lepper et al.,

PAGE 13

13 1973). Cameron, Banko and Pierce (2001) reported that in the sp an of 30 years over 100 studies were performed to evaluate the effects of extrinsic rewards. Tasks have varied across stud ies and include d dot to dot activities (Feingold & Mahoney, 1975), puzzles (Deci, 1971, 1972; Lepper & Green, 1975), coloring with mar kers (Lepper et al., 1973), listening to songs on a tape recorder (Reiss & Sushinsky, 1975), sorting silverware (Roane et al., 2003), and classroom rule following (Akin Little & Little, 2004), among many others. In addition to all of the empirical researc h on the topic, there have been literature reviews. First, Cameron and Peirce (1994) evaluated the results of 96 studies using group designs and 5 studies using single subject designs Based on the results of their statistical analyses, they concluded ove rall that rewards do not decrease intrinsic motivation. In 1999, Deci, Koestner and Ryan published a meta analysis claiming that and Pierce used inappropriate procedures an d made numerous errors. Based on the results of 128 studies, Deci et al. concluded that tangible rewards do have a substantial undermining effect on intrinsic interest. In 2001, Cameron, Banko and Pierce readdressed the issue, suggested that both of the p revious meta analyses had flaws, and attended to the issues brought up by Deci et al. The authors came to the conclusion that rewards overall do not have a negative effect, but under some conditions (e.g., tangible rewards delivered contingent on performa nce on high interest tasks) there might be some decrement in interest. The contradictory results from the meta analyses of the research on the overjustification effect may be due to the wide variation in methods used to study the

PAGE 14

14 effect (Cameron et al., 2 001; Cameron & Pierce, 1994; Deci, Koestner & Ry an, 1999; 2001). The majority of the research in the area used either a pre and post test paradigm without taking repeated measures of behavior (e.g., Anderson, Manoogian & Reznick, 1976; Deci, 1971; Lepper & Greene, 1975) or involved re peated measures but only reported averages (e.g., Mynatt Oakley, Arkkelin, Piccione, Margolis, & Arkkelin 1978; Lepper et al., 1973). In addition, aspects of the experimental sessions, including types of task, types of reward or reinforcer, and whether or not there are other items present duri ng the session differ across study. Some allow ed access to other activities during the experimental sessions (Deci, 1971; Feingold & Mahoney, 1975), others did not include an altern ative activity (Lepper & Greene, 1975; Roane et al., 2003; Bright & Penrod, 2009), and some studies did not specify (Anderson et al., 1976). A result of the mixed meta analytic outcomes and mixed methodology is that the overjustification effect is still c ommonly cited in psychology textbooks (e.g., Franzoi, 2010; Griggs, 2009; Lilienfield, Lynn, Namy, & Woolf, 2010 ) and in the popular media (e.g., Bronson & Merryman, 2009; Kohn, 1993; Pink, 2009) as a problem with the use of rewards. Also, no study has att Finally, no experiments have attempted to identify the conditions under which overjustification like effects would and would not occur. The current dissertation involves a series of studies evaluati ng the overjustification hypothesis. The purpose of Study 1 was to evaluate the effects of contingent tangible rewards on activity engagement when a single activity was available in order to evaluate whether the overjustification effect is observed when th ere are no other items available with which to engage. Study 2 examined the effects of contingent

PAGE 15

15 tangible rewards on activity engagement when multiple activities were available in order to evaluate whether the overjustification effect is observed when the subject has continuous opportunity to engage with other activities. Although on the surface Study 2 may seem to be a comparison to Study 1, this was not intended to be the case. Rather, Study 2 is simply designed to test for an overjustification effect w hen multiple items are available. The purpose of Studies 3A B was to evaluate the effects of extended exposure to activities when multiple activities were available. Study 3A was a descriptive study that assessed the effects of an extended baseline with mu ltiple activities. Study 3B was an extension of Mynatt et al. (1978) that compared the effects of contingent tangible rewards on activity engagement when multiple activi ties were available to a within subject control for exposure (i.e. extended baseline). Given that Deci (1971) is frequently cited in support of the overjustification effect and yet individual data were not presented and there has never been a replication of this study Study 4 was a direct replication of the first experiment of Deci (1971) with college age students in order to assess within subject measures.

PAGE 16

16 CHAPTER 2 GENERAL METHOD, STUDIES 1 3 Subjects, Setting, and Response Definitions Subjects were recruited from two first grade classrooms or from the afterschool program at a l ocal elementary school. All subjects were 6 or 7 years of age at the beginning of the study. Sessions took place at a large table in a quiet area of the school library or the school caf eteria. For all of the studies experimenters measured engagement with e ach activity; engagement was defined as touching the item or items corresponding with that activity with a three second onset and offset criteri on (i.e., if the dur ation of contact was less than 3 s engagement was not scored) Activities were chosen based on age appropriateness (listed on the packaging) and arbitrarily assigned to sets. Subjects could be engaged with multiple activities at the same time. Data Collection and Analysis Data were collected using handheld devices (HP iPAQ) with a behavior al recording program called Instant Data. The program allowed frequency and duration measures to be collected for multiple responses across time. For Studies 1 3A, a second, independent, observer collected data using a handheld stopwatch and for Study 3B t he second observer collected data using a second handheld device programed with Instant Data. Duration measures for activity engagement were converted to percentages by dividing the total duration of engagement by the session time and multiplying by 100.

PAGE 17

17 General Procedure Baseline During baseline conditions there was one or more (depending on the study) age appropriate leisure activities (e.g., puzzles, coloring books, and sticker books) within no programmed consequences for engagement with t he activities. Sessions lasted 5 min and one or two sessions were conducted per day. At the beginning of each baseline session the subject you like on the Reward Reward conditions were set up the same as the baseline sessions. At the beginning of each reward session the subject was asked to choose from an array of small edible items. One s mall piece was placed on a plate in front of the subject contingent on 30 s continuous engagement with the target activity; the target activity was determined from the baseline levels of engagement. An activity with somewhat high (>30% ) engagement in basel ine was selected as the target activity in the reward condition. The 30% value was chosen to ensure that there was some level of responding to the activity, or intrinsic interest, prior t o the contingent reward. At the beginning of each reward session the You may play with whatever you like on the table. You will only get (preferred edible) if

PAGE 18

18 CHAPTER 3 STUDY 1: THE EFFECTS OF CONTINGENT TANGIBLES ON ENGAGEMENT WITH A SINGLE ITEM The availability of other activities for the child to engage with during the baseline and experimental sessions is either not reported or inconsistent across studies evaluating the overjustif ication effect. Several studies d id not include an alternative activity during any of the sessions (Bright & Penrod, 2009; Lepper & Greene, 1975; Roane et al., 2003 ) and several studies specified that alternative activities were available throughout the sessions (Deci, 1971; Feingold & Mahoney, 1975). Anderson et al. (1976) told the subjects that if they did not want to engage in the target activity, drawing, they co uld do something else in the room, but whether there were other activities in the room was not reported. The availability of alternative activities could have direct effects on the level of engagement with a target activity. The purpose of Study 1 was to e valuate the effects of contingent tangible rewards on activity engagement when a single item was available and to assess whether overjustification like effects are seen when only one item is available Method Subjects Subjects were four typically develop ing first grade children recruited from two classrooms in the same local elementary school. There were two male and two female subjects; Terrance, James, Angel, and Georgette. Design and Procedure An ABA d esign was used to evaluate the effects of continge nt tangibles on activity engagement where A was no tangible reward and B was tangible reward There was a single age appropriate leisure act on the table in

PAGE 19

19 front of the subject. Baseline and reward conditions were as descri bed in the General Method. Interobserver Agreement I nterobserver agreement (IOA) was collected by a second observer simultaneously but independently collecting data on 33%, 40%, 32%, and 52% of sessions for Angel, Terrance, Georgette, and James, respectively T otal agreement scores were calculat ed by dividing the smaller duration of observed engagement by the larger duration of observed engagement in each session and multiplying that value by 100 Averag e agreement scores were 99.6% (range, 99 % to 100%), 99.2% (range, 96.7% to 100%), 99.6% (range, 98.7% to 100%), and 99.2% (range, 98.9% to 99.3%) for the four subjects respectively. Results and Discussion Figure 3 1 shows the results of Study 1 for Angel, Terrance, Georgette, and James. None of the four subjects showed a sustained detrimental effect of contingent tangibles on levels of activity engagement. In fact, three of the four subjects were engaged with the activity over 98 % of the session for all baseline sessions, contingent reward and return to baseline sessions. James wa s engaged with the activity 100% of the session for all baseline and contingen t reward sessions and above 70 % of the session for all return to baseline se ssions. Thus, the overjustification effect was not observed when only on e activity choice was available.

PAGE 20

20 Figure 3 1 Percentage of session engaged with the single item for Angel, Terrance, Georgette, and Jame s.

PAGE 21

21 CHAPTER 4 STUDY 2: THE EFFECTS OF CONTINGENT TANGIBLES ON ENGAGEMENT WITH MULTIPLE ITEMS The results of Study 1 appear to debunk the overjustification hypothesis However, a follow up question is whether the overjustification effect would be obtained if alternative activities were available. Thus, the purpose of Study 2 was to examine the effects of contingent tangible rewards on activity engagement when multiple activities were available and to assess whether overjustification like effects were observed under these conditions. Method Subjects Three subjects were originally enrolled for Study 2. One of the subjects was excluded from the study because of a diagnosis of Autism Spectrum Disorder (ASD). His data have been included i n a follow up study with subjects with ASD (not reported here The remaining subjects were Sam and Angel. Sam was a typically developing male in first grade. Angel was a typically developing female and was also a subject in Study 1. Design and Procedure An ABA design was used to evaluate the effects of contingent rewards on activity engagement. There were more than one age appropriate leisure activities on the table in front of the subject. Sam had five items available and Angel had two. Baseline and reward conditions were as described in the General Method. In the reward condition a small edible item was delivered contingent on 30 s of continuous engagement with the target activity.

PAGE 22

22 The ta rget activity was determined based on the levels of engagement observed during the baseline sessions. The activity was chosen as the target activity if the average percent age engagement with the activity was above 30% for the last three sessions The 30% v alue was chosen to ensure that there was some level of responding to the activity, or intrinsic interest, prior to the contingent reward. Interobserver Agreement Agreement was collected and calculated in the same manner as Study 1. Agreement data were collected during 56% and 30% of sessions, and average agreement scores were 95.1% (range, 82% to 100%) and 99.2% (range, 96.7% to 100%) for Sam and Angel, respectively. Results and Discussion Figures 4 1 display the results of Study 2 for Sam and Angel. T he top two panels of Figure 4 1 show the patterns of responding for all available items. Stamps were chosen as the target activity for Sam, and b enders were chosen as the target activity for Angel. Due to the large number of data paths some of the effect s may be obscured, so the bottom panel of Figure 4 1 displays only the percentage of session engaged with the target activity for Sam For both subjects there was a decrease in activity engagement from the first to the second baseline condition During the r eward condition for both subjects, engagement with the t arget activity increased to 100% of the session. For Sam, engagement with th e stickers was above 80% of the session at the end of the first baseline and dropped to zero after the second session of the second baseline. Angel enga ged with the benders during 100% of the session for five consecutive sessions in the first baseline Immediately, engagement with b enders dropped to 0% of the session during the second baseline. Table 4 1 shows the average percentage of

PAGE 23

23 session engaged with the target activity in ea ch condition for each subject. In addition the average of the last three sessio ns in each condition is shown for both subjects From the first baseline to the second ba seline there was a 56.68 and 71.43 decrease in percentage of session engaged for Sam and Angel, respectively. The decrease s in engagement with the target activities for both subjects could provide support for the overjustification hypothesis or could be a result of repeated exposure to the activities ( one possibility would be satiation). The purpose of Studies 1 and 2 was not to evaluate the effects of single versus multiple available activities on the probabil ity of post reward condition decrements, rather to assess separately whether the overjustification like effects would be observed given thes e two conditions. However, post reward decrements were only observed when multiple activities were available so future research could focus on a within subject com parison of these two conditions.

PAGE 24

24 Figure 4 1. Percentage of session enagaged for Sam and Angel (top two panels display all available items and bottom panels displays only target activity for Sam)

PAGE 25

25 Table 4 1. Condition averages of percentage of session engaged with target item for all sessions in each condition and the last three sessions in each condition for each subject Subject BL 1 Reward BL 2 BL 2 BL1 Sam All 76.58 75.33 19.9 56.68 Last 3 68.78 100 0 68.78 Angel All 71.42 96.00 0 71.42 Last 3 100.00 100.00 0 100.00

PAGE 26

26 CHAPTER 5 STUDY 3: THE EFFECTS OF EXTENDED EXPOSURE ON ACTIVITY ENGAGEMENT Mynatt et al. (1978) criticized the use of a single trial to evaluate the overjustification effect and included conditions with several baseline and reinforcement trials in their study. Mynatt et al. included 9 days in the baseline condition, 11 days in the rein forcement condition, and 9 in the second baseline. In addition to using repeated measures within subject s Mynatt et al. used a group design to compare the experimental group to a group with no reinforcement trials designed to control for amount of exposur e to the activities. The results showed a high interest level in the first baseline and reinforcement conditions and a decrease to levels that match the control group in the second baseline. Interest was defined as number of times a subject chose a game di vided by the number of games chosen in each session. Based on these results, the authors concluded that the decline in interest in the high base rate condition was due to satiation, with the rate of satiation not affected b y the reinforcement condition. Un fortunately, the authors averaged the sessions together for each condition so within condition changes cannot be seen. Also, the 6 and 7 year old subjects collected the data and interobserver agreement measures were not reported. Thus, there has been ver y limited research with children to date that ha s used repeated measures and acceptable observation methods when multiple activities are available. Controls for amount of exposure have typically been across groups and there have been no studies to date tha t have used a within subject control. The purpose of Study 3 was to evaluate the effects of an extended exposure on activity engagement when multiple activities were available and provide an extension to Mynatt et al There were t wo phases to Study 3. The purpose of Study 3A was to

PAGE 27

27 observe the effects of an extended baseline with multiple activities and no programmed consequences The purpose of Stud y 3B was to compare the effects of a contingent tangible reinforcer on activity engagement to a within subject control for exposure condition. Study 3B involve d an extended baseline as a control

PAGE 28

28 Study 3A: Evaluation of the Effects of an Extended Baseline with M ultiple Activities Method Subjects Subjects were four first grade students recruited from two classrooms in a local elementary school. All subjects included in this study were also involved in Study 1. Design and pr ocedure. There were five age appropriate leisure activ on the table in front of the subject. Subjects were exposed to between eight and fifteen 5 min sessions with no programmed consequences for activity engagement. One to two sessions were conducte d per day, two to five days per week. Interobserver a greement Agreement was collected and calculated in the same manner as Studies 1 and 2. Agreement data were collected on 50% of sessions for all subjects and average agreement scores were 90.3% (range, 74.3% to 100%), 90.5% (64.7% to 100%), 97.9% (96.4% to 99.6%) and 98.1% (94.8% to 99.7%) for Angel, Terrance, Georgette, and James, respectively. Results and Discussion Figure 5 1 displays the percentage of session engaged with each of the activities and Figure 5 2 displays the percentage of session engaged with the activity with the highest initial engagement for Angel, Terrance, Georgette, and James. For all four subjects, activity engagement was initially stable; engagement with one of the activitie s re mained at approximately 100% of the session for between three to five sessions. Then, for all subjects, engagement with the activity that was at the highest percent age

PAGE 29

29 of the session dropped to zero and subsequently became variable as the subjects sampled the other activities. This finding supports the notion that the presumed overjustification effect may b e a satiation effect as described by Mynatt et al By reinforcing engagement (in a typical overjustification arrangement), the subject spends more time t han they ordinarily would have with the item and the decrease in engagement is therefore not entirely surprising.

PAGE 30

30 Figure 5 1 Percentage of session engaged with all available items for Angel, Terrance, Georgette, and James

PAGE 31

31 Figure 5 2 Percentage of session engaged with item with the initally highest percentage of engagement for Angel, Terrance, Georgette, and James

PAGE 32

32 Study 3B: Evaluation of t he Effects of Contingent Tangibles on Activity Engagement with an Extended Baseline Control The results of Study 3A show that if activity engagement with a single activity occurs at a high rate when other activities are available over time engagement with that activity will decrease, become more variable, or both. As Mynatt et al. (1978) suggested, reinforcement in the reward condition may simply maintain temporarily the level of engagement and once the reward condition is discontinued the level of engagement will drop to approximately the levels observed in an extende d baseline. Thus, the purpose of Study 3B is to extend on Mynatt et al. b y conducting a within subject comparison by i ncluding two sets of activities implementing a reward condition for a target activity in one set while keeping the other set in baseline. Method Subjects Subjects were three typically developing first grade students two males and one female. Jack and Kayla were recruited from a classroom in a local elementary school Design and p rocedure A combined ABA and multiple baseline design was used to compare the effects of a contingent tangible on activity engagement to an extended baseline condition. Subjects were e xposed to two sets of materials; one set of materi als in each 5 mi n session with two sessions per day. There were five age appropriate leisure activities in each set. The activities in each set were pl on the table in front of the subject. The presentation of sets was alternate d each session

PAGE 33

33 For one set of toys, an ABA design was used to evaluate the effects of contingent tangibles for activity engagement. This manipulation was the same as Study 2. Baseline and Reward conditions were as described in the General Method. In the Reward condition a sm all edible item was delivered contingent on 30 s of continuous engagement with the target activity. The target activity was defined in the same manner as in Study 2. The second set of toys remain ed in an extended baseline as described in Study 3A. Interob server a greement A second observer collected data during 29%, 43%, and 30% of sessions for Jack, Kayla, and Dennis, respectively. P ercentage agreement scores were computed by dividing the session length into consecutive 10 s intervals. The smaller duration of observed responses was divided by the larger duration of observed responses in each interval and multiplied by 100. These values were averaged across the sess ion. Average agreement scores were 99.0% (range, 94.5% to 100%), 99.2% (range, 96.74% to 100%), and 98.2% (range, 87.4% to 100%) for Jack, Kayla and Dennis, respectively. Results and Discussion Figures 5 3 and 5 4 display the results for all three subjec ts. For each subject, Set 1 is displayed on the top in a n ABA design and Set 2 is displayed on the bottom in an extended baseline. Sessions are combined in two session blocks along the x axis for Figures 5 3 and 5 4 In addition, due to the large number of data paths some of the effect s may be obscured, so Figure 5 4 displays only the percentage of session eng aged with the target activity. The results from all three subjects are consistent with the results reported by Mynatt et al. (1978) Table 5 1 shows t he average percentage of session engage d with the Set 1 target item and Set 2 comparison item for each subject.

PAGE 34

34 Condition lengths for comparison items were based on the condition lengths for corresponding Set 1 item There was an average decrease in percentage of session engaged with Set 1 items from the first baseline condition to the second baseline of 45.65, 55.22, and 43.53 for Jack, Kayla, and Dennis, respectively. The average decrease in percentage of session engaged with the Set 2 comparison it em was 50.56 and 54.79 for Jack and Kayla. For Dennis, there was an increase in percentage of session engaged with the Set 2 comparison item of 7.97, but when only the last 3 two session blocks of each condition were averaged together there was a decrease of 28.22. These decreases are similar to those observed with the Set 1 target activity. For all three subjects t here was an increase in the level of engagement with the target activity in the set of toys exposed to the contingent tangible reward conditio n and once the condition was discontinued, the level of engagement dropped to approximately the levels observed in the other set o f toys in the extended baseline as seen in Figure 5 3. This finding supports the n otion that the presumed overjustification e ffect may b e a satiation effect or at least an effect of repeated exposure, and that the reward condition may simply prolong the amount of engagement with the target item and, hence, provoke the repeated exposure effect. In the current study the control condition was an extended baseline, consistent with the procedures used by Mynatt et al. (1978). In future research, a different control condition for this type of analysis would be to restrict access to all other activitie s. Based on the results of Study 1, this would increase time allocation to the target activity. If a decrease in activity engagement follows a phase of only the target activity available

PAGE 35

35 then there would be further evidence to support that the decrement in responding is not due to the reward.

PAGE 36

36 Figure 5 3 Percentage of session engaged with all items for Jack, Kayla, and Dennis for Set 1 and Set 2 (top and bottom panels, respectively). Sessions are combined in to two session blocks along the x axis.

PAGE 37

37 Figure 5 4 Percentage of session engaged with target items for Jack, Kayla, and Dennis for Set 1 and Set 2 (top and bottom panels, respectively). Sessions are combined in to two session blocks along the x axis.

PAGE 38

38 Table 5 1. Condition averages of percentage of session engaged with Set 1 target item and Set 2 comparison item for each subject. Condition lengths for Set 2 comparison items were based on condition lengths for corresponding Set 1 item. The top row for each subject displays the average for all sessi ons in each condition and the bottom row displays the average of the last 3 two session blocks for each condition. BL1 Reward BL 2 BL 2 BL 1 Set # 1 2 1 2 1 2 1 2 Jack All 46.82 52.81 66.97 33.92 1.17 2.25 45.65 50.56 Last 3 40.23 48.11 91.28 31.11 1.17 2.25 39.57 45.86 Kayla All 55.22 87.96 99.29 69.42 0 33.17 55.22 54.79 Last 3 56.96 84.50 99.22 63.34 0 33.17 56.96 51.32 Dennis All 47.00 34.84 88.20 89.25 3.47 42.81 43.53 7.97 Last 3 66.28 33.22 91.90 99.80 5.78 5.01 60.50 28.22

PAGE 39

39 CHAPTER 6 STUDY 4: REPLICATION OF DECI (1971) The results from S tudies 1 3 do not provide evidence to support the overjustification hypothesis. The results do, however, suggest that decrements in activity engagement under conditions in which a reward has been previously delivered for engagement could be due to repeated exposure t o that particular activity. The discrepancy between the present results and the results reported by Deci (1971) could be due to the analytic methods used to evaluate the effect. used a group design with 12 subjects in the experimental group and 12 subjects in the control group. The experimental group was exposed to an ABA design. The experimental phase involved the delivery of one dollar contingent on successful puzzle completion. The control group was exposed to three baseline conditions. For both groups, each condition was a single hour long session. The subjects were asked to complete puzzle configurations that matched a drawing presented by the experimenter F our puzzle config urations were presented in each phase. The dependent measure was the number of seconds that the subject was working on the puzzles during an 8 min break occurring in the middle of each phase (i.e., after the second configuration of that phase had been comp leted). During the break, the experimenter left the room and the behavior of the subject was being monitored through a one way window. Deci subtracted the number of seconds that the subject was engaged with the puzzle during the third break from the seco nds engaged in the first break. These values were compared across groups. Based on a high alpha value of .10, Deci concluded that there was a significant difference between groups and that these results support the

PAGE 40

40 overjustification hypothesis. The differe nce was not significant at the .05 level, although that level is commonly used in statistical tests. Besides the high alpha value, there were several other limitations with the study. First, there were no individual or repeated measures of behavior. The individual differences could not be seen based on the averages and the amount of engagement during the sessions in which the experimenter was present was not reported. Second, the amount of engagement with other available activities was not reported. Third, the number of times each subject contacted the extrinsic reward was not reported. At the extreme, it is possible that a subject never completed a puzzle and, therefore, never contacted the extrinsic reward The overjustification hypothesis, by definition, requires that a subject contact an extrinsic reward, so it would be inappropriate to include the dat a from a subject who did not contact re ward in the experimental group average. In analysis, only 4 of the 24 total subjects showed a decrease in engagement with the puzzle from break 1 to break 3. In addition, individual patterns d id not necessarily indicate a high level of intrinsic interest in that seven subjects in each group engage d with the puzzle for less than 240 s during break 1. Still, how often the subjects completed the puzzle was not reported and how often the subjects in the experimental group earned the money was not reported. The purpose of Study 4 was to replicate the methods used by Deci and address the limitations by including individual measures of puzzle engage ment during sessions and breaks and by reporting how often the subjects completed the puzzle and earned the reward.

PAGE 41

41 Method Subjects and Setting Subjects wer e 24 undergraduate students in psychology recruited via the General Psychology (PSY2012) Research Participation Pool. There were 12 subjects in the experimental group and 12 in the control group. All Study 4 subjects were enrolled in and compensated with PSY2012 class credi t in accordance with the class protocol (i.e., one credit per ever y half hour spent in a study). Sessions for Study 4 took place at a table in an experimental room in the Psychology building on the University of Florida campus. The expe rimenter measured engagement with available activities. E ngagement was defined as touching the item or items c orresponding with that activity. E ngagement also include d orienting toward and looking directly at the materials. Subjects could be engaged with m ultiple activities at the same time. Design and Procedure A combined ABA and group design was used to evaluate the effects of money contingent on puzzle completion. One group was exposed to all three phases as described below; the second group was expo sed to three baseline sessions. Baseline The subjects were asked to reproduce a shape printed on a piece of paper using block puzzle pieces. There were four puzzle configurations in this condition. The subjects had 13 min to complete the puzzle before the next configuration wa s presented. If they complete d the puzzle before the 13 min wa s over they were presented with the next configuration. After the second configuration in this phase there was a break. Procedures for the break are described below.

PAGE 42

42 Experimental Phase Procedures for this phase were the same as baseline with one exception; if the subject complete d the configuration within 13 min they were given one dollar. Break There were three breaks during the entire study, one in the middle of each phase of the study. The first occur red after the second puzzle configuration, the second occur red after the sixth puzzle configuration, and the third occur red after the tenth puzzle configuration. Before each break t he experimenter xcuse me, I have to leave the room for a few minutes to determine the next configuration. You may were a newspaper a nd a few up to date magazines. The subject s were observed through a one way mirror and data were collected on puzzle engagement as well as engagement with other items that were available. Interobserver Agreement. I nterobserver agreement (IOA) was collected by a second observer simultaneously b ut independently collecting data on 23% of all breaks T otal agreement scores were calculat ed by dividing the smaller duration of observed engagement during each break by the larger duration of engagement and multiplying by 100 Average agreement scores f or all subjects were 99.5% (range, 97.5 % to 100%), 98.9% (range, 95.5% to 100%), and 99.2% (range, 95.48% to 100) for breaks 1, 2, and 3, respectively. Results and Discussion Table 6 1 shows the number of seconds engaged with the puzzle during each break f or all subject s a nd the means for both groups. For both groups, the number in parenthesis indicates the number of puzzles completed during the corresponding phase

PAGE 43

43 of the study for that group. If there is no number, then there were no puzzles completed during that phase for that subject. For example, subject 5 in the experimental group completed zero puzzles during the first phase, one puzzle during the second phase (experimental phase), thus ea rning the money for that configuration, and one puzzle during the third phase with no money delivered following completion of that configuration. Interestingly, of the 288 configurations that were presented to both of the groups, only 22 were completed in less than 13 min and furthermore, only four subjects in the experimental group actually earned money (subjects 5, 6, 7 and 10) This means that eight subjects never contacted the reward so any changes in puzzle engagement could not be attributed to recei ving the reward for these subjects In addition subjects 3, 9, and 11 showed a decrease i n puzzle engagement from break 1 to break 3 but did not contact the reward during phase 2 of the study. Contrary to Deci, b oth of the groups in the current study sho wed an average decrease from break 1 to b reak 3 with the control group show ing a larger decrease Figure s 6 1 and 6 2 show the individual times engaged with the puzzle for the experimental and control group s, respectively There were three general patterns of individual results in the experimental group : 1) five subjects showed a decrease in the am ount of puzzle engagement from b rea k 1 to b reak 2 and a further decrease from break 2 to b reak 3 (subjects 3, 5, 9, 10 and 11) 2) t wo subjects showed a n increase from b reak 1 to b reak 3 (subjects 1 and 8), and 3) five subjects did not engage with the puz zles during any of the breaks (subjects 2, 4, 6 and 7) For the experimental group, w hen the amount of puzzle engagement during each break is averaged acro ss subject s the results are consiste nt with those reported by Deci. In the current study, i ndividual

PAGE 44

44 patterns of engagement did not show a high level of intrins ic interest meaning that there was little engagement with the puzzle during the breaks when th e subject was left alone Several of the subjects across both groups had less than 240 s of engagement during b reak 1, and eight subjects did not engage with the puzzle at all during the entire study. Based on MaWhinney ) analysis, more than half of the subjects in the original study engaged with the puzzle for less than 240 s during the first break Future evaluations should include a task with a higher level of engageme nt during the first break, or implement an exclusion criterion for subjects with a low level of engagement during the first break.

PAGE 45

45 Table 6 1. Number of Seconds Working on the Puzzle During the Free Choice Period Experimental Group Control Group Break 1 Break 2 Break 3 Break 3 Break 1 1 0 0 85.92 85.92 2 0 0 0 0 3 479.04 437.76 408.48 70.56 4 0 0 0 0 5 472.99 386.02 (1) 183.98 (1) 289.01 6 0 0 (1) 0 0 7 0 0 (1) 0 0 8 70.99 0 81.98 10.99 9 478.99 480.00 28.80 450.19 10 166.99 (1) 0 (1) 0 (1) 166.99 11 470.40 427.20 379.20 (1) 91.20 12 0 0 0 0 1 472.80 425.33 354.58 118.22 2 470.40 480.00 460.80 9.60 3 474.00 316.99 26.02 447.98 4 0 0 (1) 0 0 5 459.98 (1) 392.02 330.00 129.98 6 0 0 (1) 0 (1) 0 7 345.98 (1) 0 0 345.98 8 0 0 0 0 9 272.02 0 0 272.02 10 480.00 (2) 480.00 (2) 480.00 (1) 0 11 201.60 338.02 33.02 168.58 12 459.02 75.98 (1) 0 (3) 459.02 Mean 178.28 144.25 97.36 80.92 Mean 302.98 209.03 140.37 162.62

PAGE 46

46 Figure 6 1. Number of seconds engaged with the puzzle during each break for individuals in the experimental group.

PAGE 47

47 Figure 6 2. Number of seconds engaged with the puzzle during each break for individuals in the control group.

PAGE 48

48 CHAPTER 7 GENERAL DISCUSSION The results of Studies 1 3 do not provide evidence to support the overjustificati on hypothesis. Rather, the results demonstrate that extended exposure to the preferred activities may account for a decrease in activity engagement from the first to the second baseline condition, when a decrease is observed. In Study 1, none of the four s ubjects showed a sustained detrimental effect of contingent rewards o n levels of activity engageme nt and, with the exception of James, the subjects were engaged with the available activity over 98% of the session for all sessions. James showed a slight dec rease in activity engagement from the first to the second baseline, but remained engaged o ver 70% of the session for all sessions in the second baseline. The overjustification effect was not observed when only one item was available and levels of engagemen t remained high and stable for all four subjects. Contrary to the results of Study 1, in Study 2, there was a decrease in activity engagement from the first to the second baseline for both subjects. For Sam, a drastic decrease to 0% engagement with the ta rget item, stickers, occurred after the second session of the second baseline. For Angel, the same decrease occurred immediately in the first session of the second baseline. Both subjects were informed of the contingency at the beginning of each session an d thus were told at the beginning of the second baseline that they would no longer receive edibles for playing with the target item. These effects are in alignment with the overjustification hypothesis and could provide support for the overjustification ef fect. However, due to the number of sessions of exposure to the same activities, effects that look similar to the overjustification effect could, instead, be a result of repeated exposure This possibility was evaluated in Study

PAGE 49

49 3. It is also important to note that in Studies 1 and 2 the effects of contingent rewards differed depending on whether there are other items available with which the child coulf engage In previous research, this aspect of the experimental sessions has varied across studies, and in some cases was not even specified (e.g., Anderson et al., 1976). Based on these findings, in future research it will be important to note whether other activity choices are available to the children during the experimental sessions. In addition, within su bject comparisons of one versus many items available should be conducted. The results of Studies 3A and 3B provide evidence to suggest that the putative overjustification effect may in fact be a result of extended repeated exposure to the activities. Study 3A demonstrates that extended exposure to one set of activities will resu lt in decreases in engagement with a single activity, if engagement with that activity was initially high. The results from all four subjects showed this effect. The results of Study 3B demonstrate that if a contingent reward condition is included and subs equently removed for a target activity in one set of activities, over time, engagement with the target activity will match the levels observed in a set of activities that remained in baseline. These findings are consistent with those reported by Mynatt et al. (1978). Future research should evaluate the effects of other control conditions that would increase time allocation to the target activity without the use of a reward condition for example a condition with only the target item available. Although these results provide support for the notion that the o verjustification effect is simply a result of extended exposure to the activities, it is still not conclusive. If the effects of repeated exposure to the activi ties is minimized or eliminated and if th e

PAGE 50

50 overjustification effect is in fact a result of repeated exposure, then activity engagement should remain at high levels throughout the assessment. If the overjustification effect is not an effect of repeated exposure, then when these effects are minimiz ed, activity engagement should remain high for the activities that remain in baseline, but drop below baseline levels for those that have been exposed to a contingent reward. One possible explanation for the effect of repeated exposure is satiation. The re has been ample research showing that varied stimulus presentation will reduce the effects of satiation (e.g., Dunlap & Koegel, 1980; Egel, 1980,1981; Bowman, Piazza, Dun lap and Koegel (1980) evaluated the differential effectiveness of two methods of task presentation. In one condition a single task was presented throughout the training session; in the other condition, the target task was interspersed among a variety of o ther tasks during a single training session. The results showed higher and more stable levels of correct responding during the varied task condition for both subjects. Egel (1980) found that the presentation of varied reinforcers produced higher rates of bar pressing for ten individuals with autism. Subsequently, Egel (1981) replicated those findings with a population of children diagnosed with autism spectrum disorder or other developmental disabilities in a classroom setting. There were decreasing trends in correct responding and on task behavior for all three subjects during the constant reinforcer condition; however, correct responding returned to high levels and maintained for 125 trials during the varied reinforcement condition. Varying the presentati on of the reinforcers minimized the effects of satiation. Thus, possible future direction would be to minimize the effects of extended exposure to the activities,

PAGE 51

51 perhaps by varying the presentation of sets of activities across sessions, and evaluate the e ffects of contingent rewards on a target activity in each set. T he terms reward and reinforcer have been used interchangeably throughout the literature (Cameron & Pierce, 2004). The term reward was used throughout the current study because the effectiven ess of the edibles as reinforcers was not directly assessed. However, with the exception of Study 1, there was an increase in the level of activity engagement in all reward conditions for all subjects, so it is probably safe to say the edibles were reinfor cers given that the baseline lengths were staggered overall providing a context similar to a nonconcurrent multiple baseline design In Study 1, all subjects were engaged with the items for 100% of th e session in the first baseline therefore due to a cei ling effect increases could not be observed. Bright and Penrod (2009) directly evaluated the effects of verbal praise, non reinforcing stimuli, and reinforcers as identified by a reinforcer assessment and found no evidence to support the overjustification hypothesis in any of the conditions. Continued research on possible negative effects of reward is warranted. The results of Study 4, a replication of one of the most commonly cited studies, do not provide evidence to support the overjustification hypothe sis. Both groups in the current replication showed an average decrease in puzzle engagement from break 1 to break 3 and analysis of the individual data showed a variety of patterns of responding in both groups. F uture research should continue to identify the conditions under which overjustification like effects will and will not occur. Some variations may include evaluations with different types of rewards, (e.g., praise instead of edibles) and different contingen cies, (e.g., reward contingent on task completion instead of engagement).

PAGE 52

52 Finally, it is possible that different populations of children and adults would respond differently to the conditions of this study. The children in this study were all typically de veloping. In some preliminary follow up work, we are finding that some children with autism are less sensitive to repeated exposure following a reward condition. Thus, we are in the process of evaluating variations of overjustification research in our curr ent studies.

PAGE 53

53 LIST OF REFERENCES Akin Little, K. A., & Little, S. G. (2004) Re Examining the o verjustification ef fect. Journal of Behavioral Education, 13, 179 192 Anderson, R., Manoogian, S. T., & Reznick, J. S. (1976). The undermining and enhancing of intrinsic motivation in preschool children. Journal of Personality and Social Psychology, 34, 915 922. Bowman, L. G., Piazza, C. C., Fisher, W. W., Hagopian, L. P., & Kogan, J. S. (1997). Assessment of preference for varied versus constant reinforcers. Jo urnal of Applied Behavior Analysis, 30 451 458. Bronson, P. & Merryman, A. (2009) NurtureShock: n ew t hinking a bout c hildren. New York: Twelve. Bright, C. N. & Penrod, B. (2009). An evaluation of the overjustification effect across multiple contingency arrangements Behavioral Interventions, 24, 185 194. Cameron, J., Banko, K. M., & Pierce, W. D. (2001). Pervasive negative effects of rewards on Intrinsic motivation: The myth continues. The Behavior Analyst, 24 1 44. Cameron, J., & Pierce, W. D. (1994) Reinforcement, reward and intrinsic motivation: A meta analysis. Review of Educational Research, 64 363 423. Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation J ournal of Personality and Social Psychology, 18, 105 115. Deci, E. L. (1972). Intrinsic motivation, extrinsic reinforcement, and inequity. Journal of Personality and Social Psychology, 22, 113 120. Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta analytic review of experiments examining the effects of ex trinsic rewards on intrinsic motivation. Psychological Bulletin,125, 627 668. Deci, E. L., Koestner, R., & Ryan, R. M. (2001) Extrinsic rewards and intrinsic m otivation in education: Reconsidered once again. Review of Educational Research, 71, 1 27. Des Moines public schools art department (2011). Retrieved from http://www.dmpsmart.org/instruction.cfm?subpage=156587 Dunlap, G., & Koegel, R. L. (1980). Motivating autistic children through stimulus variation. Journal of Applied Behavior Analysis, 13, 619 627.

PAGE 54

54 Edwards, C. P., (2006). Montessori education and its scientific basis. Applied Developmental Psychology, 27 183 187. Egel, A. L. (1980). The effects of constant vs. varied reinforcer presentation on responding by autistic children. Journal of Experimental Child Psychology, 30, 455 463. Egel, A. L. (1981). Reinforcer variation: Implications for motivating developmentally disabled children. Journal of Applied Behavior Analysis, 14, 345 350. Feingold, B. D., & Mahoney, M. J. (1975). Reinforceme nt effects on intrinsic interest: Undermining the overjustification hypothesis. Behavior Therapy, 6, 357 377. Fisher, W. W., Piazza, C. C., Bowman, L. G., Hagopian, L. P., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying r einforcers for persons with severe to profound disabilities. Journal of Applied Behavior Analysis, 5 491 498 Franzoi, S. L. (2002). Psychology: A journey of discovery Cincinnati, OH: Atomic Dog Publishing. Griggs, R. A. (2009). Psychology: A concise introduction. New York, NY: Worth Publishers. Effects of stimulus variation on the reinforcing capability of nonpreferred stimuli. Journal of App lied Behavior Analysis, 38 469 484. Kohn, A. (1993) praise, and other bribes. Boston: Houghton Mifflin. Lepper, M. R., & Greene, D. (1975). Turning play into work: Effects of adu lt Surveillance and extrinsic rewards on children's intrinsic motivation. Journal of Personality and Social Psychology, 31, 479 486. Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children's intrinsic interest with extrinsic reward: A tes t of the "overjustification" hypothesis. Journal of Personality and Social Psychology, 28, 129 137. Levine, F. M., & Fasnacht, G. Token rewards may lead to token learning. American Psychologist 1974, 29, 816 820. Lilienfeld, S.O., Lynn, S. J., Namy L. L., & Woolf, N. J. (2009). Psychology: a framework for everyday thinking United States of America: Allyn & Bacon.

PAGE 55

55 MaWhinney, T. C. (1979). Intrinsic X extrinsic work motivation: perspectives from behaviorism. Organizational Behavior and Human Performance, 24 411 440. Mynatt, C., Oakley, T., Arkkelin, D., Piccione, A., Margolis, R., & Arkkelin, J. (1978). An examination of overjustification under conditions of extended observation and multiple reinforcement: Overjustification or boredom? Cogni tive Therapy and Research, 2, 171 177. Pink, H, P. (2009). Drive: the surprising truth about what motivates us United States of America: Riverhead Books (Hardcover) Reiss, S., & Sushinsky, L. W. (1975). Overjustification, competing responses and the acq uisition of intrinsic interest. Journal of Personality and Social Psychology, 31, 1116 1125. Roane, H. S., Fisher, W. W., & McDonough, E. M. (2003) Progressing from programmatic to discovery research: A case example with the overjustification effect. Jou rnal of Applied Behavior Analysis, 36, 35 46. Vasta, R., Andrews, D. E., McLaughlin, A. M., Stirpe, L. A., & Comfort, C. (1978). Reinforcement effects on intrinsic interest: A classroom analog. Journal of School Psychology, 16, 161 168.

PAGE 56

56 BIOGRAPHICAL SK ETCH Kerri P. Peters graduated from the University of Florida with a Bachelor of S cience in psychology and a minor in e ducation in 2004. She began working in the field of behavior analysis in late 2002 as an in home therapist for a 3 year old girl diagnos ed with autism. From there, her interests in the field grew. She enrolled in all of the classes in the area that were available. In addition to classes, she gained research experience by volunteering in multiple labs. These experiences prepared her for the next step in her academic career: a Master of Science from the Department of Behavior Analysis a t the University of North Texas. She continued her graduate studies at the University of Florida in 2007. Since beginning graduate school, Kerri has ha d the opportunity to work a variety of settings with multiple populations, including individuals with intellectual disabilities, parents at risk for maltreatment, foster parents and children, and children with behavior disorders. Following graduation, Kerr i intends to pursue a career in applied behavior analysis, with her goal being to teach and continue to conduct research in clinical settings.