<%BANNER%>

An evaluation of reinforcer dimensions influencing food selection of individuals diagnosed with Prader-Willi Syndrome

University of Florida Institutional Repository

PAGE 1

1 AN EVALUATION OF REINFORCER DIMENSIONS INFLUENCING FOOD SELECTION OF INDIVIDUALS DIAGNOSED WITH PRADER WILLI SYNDROME By JESSICA L. THOMASON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN P ARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Jessica L. Thomason

PAGE 3

3 ACKNOWLEDGMENTS I would like to thank all of the individuals who encouraged my pursuit of higher education, both personally and professionally. I thank my family and friends for supporting me throughout school. I thank SungWoo Kahng for taking the time to nurture my interest in behavior analysis and encouraging me to attend graduate school. I could not have completed this work without the assistance of colleagues and research assistants who spent extra time and effort to help me design and conduct this s tudy, especially Claudia Dozier, Pamela Neidert, Zachariah Sims, and Brooke Jones. I also appreciate the time and advice of my committee members, Jennifer Elder, Neil Rowland, and Timothy Vollmer. I would like to express the deepest gratitude to my advi sor and committee chair Brian Iwata for his guidance and time, without which I would not have been able to accomplish this and other work

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ........... 3 LIST OF TABLES ................................ ................................ ................................ ...................... 5 LIST OF FIGURES ................................ ................................ ................................ .................... 6 ABSTRACT ................................ ................................ ................................ ............................... 7 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............... 9 2 GENERAL METHOD ................................ ................................ ................................ ....... 17 Participants and Setting ................................ ................................ ................................ ...... 17 Food Types Evaluated in Preference Assessment ................................ ............................... 17 Response Measurement and Reliability ................................ ................................ .............. 17 3 EXPERIMENT 1 ................................ ................................ ................................ ............... 25 Preference Assessment Procedure ................................ ................................ ...................... 26 Results and Discussion ................................ ................................ ................................ ....... 27 4 EXPERIMENT 2 ................................ ................................ ................................ ............... 48 Participants, Design and Proced ure ................................ ................................ .................... 49 Results and Discussion ................................ ................................ ................................ ....... 51 5 EXPERIMENT 3 ................................ ................................ ................................ ............... 56 Parti cipants, Design and Procedure ................................ ................................ .................... 57 Results and Discussion ................................ ................................ ................................ ....... 60 6 GENERAL DISCUSSION ................................ ................................ ................................ 63 REFERENCE LIST ................................ ................................ ................................ .................. 68 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ..... 73

PAGE 5

5 LIST OF TABLES Table page 2 1 Participant characteristics ................................ ................................ .............................. 20 2 2 Foods included in each preference assessment. ................................ .............................. 22 2 3 Reliability coefficients ................................ ................................ ................................ ... 23 3 1 Results of between category MSWO preference assessments for PWS gr oup. ............... 30 3 2 Results of between category PS preferen ce assessments for PWS group. ....................... 36 3 3 Results of between category MSWO preference assessments for control group. ............ 38 3 4 Results of between category PS preference assessments for control group. .................... 42

PAGE 6

6 LIST OF FIGURES Figure page 3 1 Percent of times a food was selected in between category MSWO preference assesws ments for 7 PWS participants. ................................ ................................ ........... 43 3 2 Percent of times a food was selected in between category MSWO pref erence asses sments for 6 PWS participants ................................ ................................ ............... 44 3 3 Percent of times a food was selected in between category PS preference asses sments for 5 PWS participants ................................ ................................ ................................ ... 45 3 4 Percent of times a food was selected in between category MSWO preference assessments for 6 con trol participants. ................................ ................................ ........... 46 3 5 Percent of times a food was selected in between category PS preference assessment s for 6 control participants. ................................ ................................ ............................... 47 4-1 Percent of times allocated to each response option during baseline and assessment........55 5 1 Percent of times allo cated to each response option during imme diacy and quality treatments. ................................ ................................ ................................ ..................... 62

PAGE 7

7 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AN EVALUATION OF REINFORCER DIMENSIONS INFLUENCING FOOD SELECTION OF INDIVIDUALS DIAG NOSED WITH PRADER WILLI SYNDROME By Jessica L. Thomason May 2007 Chair: Brian Iwata Major: Psychology Prader Willi Syndrome (PWS) is a genetic disorder associated with a variety of problem behaviors, including hyperphagia and food stealing, as well as a predisposition toward morbid obesity. Some reports have suggested that the food preferences of individuals with PWS differ from those of individuals with other developmental disabilities. The current study compared the relative influence of reinforce r characteristics such as quality, magnitude, and the delay to delivery on choices made by individuals diagnosed with PWS and individuals diagnosed with other developmental disabilities. First, preference assessments were conducted to identify foods that w ere of high quality. Second, an assessment was conducted to identify the reinforcer characteristic (quality, magnitude, or delay) that was most influential in determining choices among concurrently available vocational or academic tasks. Next, reinforcer characteristics were manipulated in an attempt to shift responding toward healthier food selections. For example, response allocation toward immediately available reinforcers was shifted by gradually increasing delays to reinforcer delivery. Results are discussed in terms of (a) similarities and differences

PAGE 8

8 among determinants of food preference in individuals with and without PWS, and (b) implications for dietary management and food related problem behaviors.

PAGE 9

9 CHAPTER 1 INTRODUCTION A common definition of p reference in research on learning is a choice from an array of reinforcers. Q uality, magnitude, and immediacy of delivery have been shown to be important determinants of preference, in that modifications to o ne parameter result in an altered response selection or allocation. This research focuses on determinants of food preference in individuals diagnosed with Prader Willi Syndrome and the extent to which preference can be modified through manipulation of rein forcer characteristics. Miller (1976) conducted an early investigation of the role of reinforcer quality as a determinant of choice. He arranged concurrent Variable Interval (VI) schedules in which pigeons were presented with two concurrently available re sponse options, each of which was associated with access a particular type of grain (hemp, buckwheat, and wheat were available in various paired combinations). Results showed that reinforcer quality (type of grain) influenced response allocation. Matthews and Temple (1979) also used concurrent VI schedules to assess dairy cows preference (response allocation) for two types of feed and showed that the cows exhibited a preference (response bias) for one type of feed over the other. Results of basic researc h on preference have been extended in the applied literature, most often with individuals who have severe developmental disabilities and, as a result, cannot readily communicate their preferences. Pace, Ivancic, Edwards, Iwata, and Page (1985) described a two stage procedure that has b ecome common in most preference assessment research. In the first stage, stimuli were presented singly while approach behavior was measured; in the second, stimuli that were frequently (or infrequently) approached were delivered as consequences in an oper ant learning task. In a variation of this procedure, Fisher et al. (1992) presented stimuli in all possible pairs, and p reference was determined based on the number of times a stimulus was

PAGE 10

10 selected given the number of times it was presented. In a subseq uent phase, preferred and nonpreferred items were delivered as consequences under a concurrent schedule arrangement in which each of two response options was associated with delivery of one of the items. Results showed that 3 of 4 participants allocate d most or all of their time to the response associated with the preferred item. Results of these and other studies on preference and performance in the developmentally disabled population (DeLeon et al., 2001; Graff, Gibson, & Galiatsatos, 2006; Roane, V ollmer, Ringdahl, & Marcus, 1998) indicate that qualitative aspects of stimuli exert control over choice. Amount of reinforcement also can influence choice and has been manipulated through changes in the rate of reinforcement (i.e., a schedule manipulatio n) or the magnitude (amount) of each reinforcer delivered. A large volume of literature exists on the effects of various schedules of reinforcement under concurrent schedule conditions. In short, interval schedules tend to produce response allocation amon g two alternatives that is roughly proportional to the amount of reinforcement available on each alternative, whereas ratio schedules tend to produce a strong bias in response allocation toward the denser schedule (Herrnstein, 1961). M anipulations of magn itude in basic research have included changes in the number of food reinforcers (Keesey & Kling, 1961; Reed, 1991; Schneider, 1973), the concentration of nutritive content in a solution (Lowe, Davey, and Harzem, 1974; Stebbins, 1962), and the duration of a ccess to reinforcers (Belke, 1997; Catania, 1963; Davison & Baum, 2003; Keller & Gollub, 1977; Neuringer, 1967). Although the results of many studies are inconsistent with respect to the influence of magnitude on responding (see Bonem & Crossman, 1988, an d Reed, 1991, for reviews), studies in which concurrent schedules were used have generally shown that animals

PAGE 11

11 allocate responding to the option that produces the larger reinforcer (Catania, 1963; Neuringer, 1967; Keller & Gollub, 1977). R elatively few app lied studies have examined the influence of reinforcer magnitude on responding. Results obtained when magnitude of reinforcement was manipulated for a single response option have been mixed. Some data show that larger magnitude reinforcers delivered contin gent on a response are correlated with slightly higher rates of responding (Lerman, Kelley, Vorndran, Kuhn, & LaRue, 2002), whereas other data sets show a reduction in overall rates of responding (Volkert, Lerman, & Vorndran, 2005). When reinforcers are d elivered noncontingently, data show that larger magnitude reinforcers produce better response suppression than smaller magnitude reinforcers (Roscoe, Iwata, & Rand, 2003). Only one applied study, Hoch, McComas, Johnson, Faranda, and Guenther (2002), has e xamined the effects of magnitude on response allocation. The dependent variable was the proportion of time allocated to playing in one of two areas; one area contained a peer, and the second area did not. When equal magnitudes (durations) of reinforcement (access to preferred items) were available for both options, the participant spent all available time in the area of the room not containing the peer. In a subsequent phase, a larger magnitude of reinforcement was available in the area of the room with t he peer versus the area without (90 and 15 s reinforcement durations, respectively); results showed that the participant began to allocate more time to the area of the room with the peer (and the larger magnitude of reinforcement). Finally, basic researc h on delay to reinforcement has shown that subjects select shorter over longer delays. Chung (1965) and Chung and Herrnstein (1967) arranged concurrent VI VI schedules in which the two response options resulted in differing delays to reinforcer delivery. Results showed that the animals tended to select the response option associated with the shorter

PAGE 12

12 delay to reinforcement. Few applied studies have examined the effects of reinforcement delay alone on response allocation (although numerous other studies ha ve compared the effects of magnitude and delay; e.g., Dixon et al., 1998; Schweitzer and Sulzer Azaroff, 1988; Vollmer, Borrero, Lalli, & Daniel, 1999). Horner and Day (1991) examined the effects of a delay to reinforcement (escape from tasks) on aggressi on and appropriate behavior. Escape was available either 20s or 1s after the participant requested a break, whereas aggression always resulted in immediate escape. Results showed that the participant allocated most responding to aggression when appropria te requesting resulted in a 20 s delay to escape, whereas the opposite occurred when requesting resulted in a 1 s delay to escape. Quality of reinforcement, amount of reinforcement, and delay to reinforcement in isolation influence behavior in predictable ways (subjects allocate responding to higher quality stimuli, larger amounts of stimuli, and immediately available stimuli); however, the effects of manipulating more than one variable simultaneously are more difficult to predict. Neef and colleagues have conducted an elegant series of experiments in which they examine the influence of several reinforcer and response requirement parameters on response allocation (Neef, Bicard, & Endo, 2001; Neef & Lutz, 2001; Neef, Mace, & Shade, 1993; Neef, Mace, Shea & Sh ade, 1992; Neef, Shade & Miller, 1994). In these studies, the authors used concurrent schedules to compare the relative influences of reinforcement frequency, immediacy, quality, and response effort on response allocation. For example, Neef et al. (2004) used this paradigm to provide an empirical evaluation of one of the diagnostic criteria for Attention Deficit Hyperactivity Disorder (ADHD), impulsivity. They observed that reinforcer immediacy exerted the greatest influence over participants choices amo ng work options and subsequently used this information to develop interventions to decrease impulsive choice. The assessment method presented by Neef

PAGE 13

13 et al. is an efficient way to isolate a number of variables that influence responding and may prove useful in application with a number of disorders in addition to ADHD. Although rare, some genetic disorders are correlated with specific behavioral characteristics. One example is Prader Willi Syndrome (PWS), which is diagnosed based on a combination of chromoso mal, physical, and behavioral traits (Holm, et al., 1993; Prader, Labhart, & Willi, 1956). Perhaps the most marked behavioral characteristic of PWS is hyperphagia, an insatiable appetite that typically results in morbid obesity. Intervention for this popul ation typically consists of environmental modifications, such as locking food cabinets and refrigerators, and restricting daily caloric intake (Butler & Thompson, 2000). Individuals with PWS have been reported to engage in a variety of problematic food rel ated behaviors, including property destruction to access food (Benjamin & Buot Smith, 1993), stealing (Donaldson, et al., 1995), and consumption of items not intended for human ingestion (e.g., food from trash cans, pet food, etc; Russell & Oliver, 2003). Assessment of the determinants of food preference may be particularly useful in identifying potential treatments for food related problems, such as teaching individuals to wait for delayed reinforcers when presented with immediately available ones, or by suggesting strategies that shift preference toward low calorie foods in lieu of higher calorie foods. Food preferences of individuals with PWS have been examined in several studies (Caldwell & Taylor, 1983; Glover, Maltzman, & Williams, 1996; Joseph, Egli, Koppekin, & Thompson, 2002; Rankin & Mattes, 1996; Taylor & Caldwell, 1985). Early results suggested that preferences were somewhat uniform. For example, Caldwell and Taylor (1983) suggested that individuals with PWS showed a pronounced preference for sw eets. Results of more recent

PAGE 14

14 studies have suggested preferences for other food characteristics (Glover et al., 1996; Rankin & Mattes, 1996) or particular macronutrients (Fieldstone, Zipf, Schwartz, & Berntson, 1997). S tudies of the interaction of quality, magnitude, and delay may provide even more information to be used in intervention. In a study by Glover et al. (1996), results of a series of comparisons among high preference foods, mixed preference foods (an array of both low and moderate preference fo ods), and low preference foods showed that participants diagnosed with PWS tended to make selections for larger magnitudes of mixed preference food over smaller magnitudes of high preference foods, in contrast to control participants, who tended to select higher preference foods in smaller quantity over mixed preference foods in larger quantity. In a second experiment, results showed that all groups (PWS and control) selected a smaller magnitude of a high preference item over a larger magnitude of low prefe rence items. The authors concluded that PWS participants had weaker taste preferences than the control participants and that the relative value of an item given other choices available influenced selection. A slightly different way to state these result s is that magnitude appeared to be a more influential variable than quality when items were ranked similarly (when quality was roughly equal across options), but that quality was more influential than magnitude when the difference among the qualities was h ighly disparate. Joseph et al. (2002) conducted a similar study in which the choices of PWS participants were compared to those of obese non PWS participants. In the first study, participants chose between one piece of food available immediately and 3 pi eces of food available after a delay (15, 30, or 60s). Data showed that PWS participants were more likely to select the larger quantity after a delay than the smaller quantity at no delay; obese comparison participants also tended to select the larger, de layed quantity slightly more than the smaller, immediate quantity but to a

PAGE 15

15 lesser extent than the PWS participants. In the second study, the same comparisons were made across four foods high in fat and carbohydrate content. First, the experimenters identif ied high preference and low preference foods by presenting an array of foods and instructing the participants to select their favorite food. During subsequent sessions, the participants were instructed to choose either a small magnitude reinforcer availabl e immediately or a large magnitude reinforcer available after a delay. This comparison (magnitude versus immediacy) was repeated across each of the identified high and low preference foods. Results were similar to the first experiment, in that PWS partic ipants tended to select (a) the larger, delayed quantity of food over the smaller, immediate option, and (b) the larger quantity more frequently than the control participants. Data also showed that when the food available in session was a high preference food, the participants selected the larger magnitude of food (at a delay) to a greater extent than when the food available was a low preference food. The authors concluded that individuals with PWS were more likely to make choices based on food magnitude. Although the Glover et al. (1996) and Joseph et al. (2002) studies provide some evidence of the variables influencing food selection, their presentation of aggregate data precludes inspection of individual differences in responding. That is, it is possi ble that some participants choices were closer to being indifferent, whereas others showed a moderate influence; thus, general statements based on the group averages may not accurately represent the performance of some individuals within the group. Furt her, although the authors conducted preference assessments, very few foods were included (4 and 10 foods, respectively). I nclusion of a broader array of foods would allow greater confidence to conclusions about relative preference among items. Finally, nei ther study included all possible comparisons (immediacy versus magnitude, magnitude versus quality, and immediacy versus quality).

PAGE 16

16 The purpose of the current research is to provide a preliminary analysis of variables influencing food choice in individual s with PWS and to increase healthy food choices when choice making appears maladaptive. The assessment methods used in the current study are based on those of Neef et al. (2001), with two of the variables adapted to the current situation. First, amount of reinforcement was manipulated by changing the number of reinforcers earned per response (magnitude) rather than reinforcement schedule. Given that the current population of participants was developmentally disabled (in many cases a diagnosis of mental re tardation was also present), it seemed that the visual cues inherent in a magnitude manipulation would likely be more salient than those available in a schedule manipulation. Second, although the series of studies by Neef and colleagues also assessed the effects of response effort, several of their data sets show that effort has limited, if any, effects relative to the other variables tested. Thus, the current experiment did not include an effort manipulation. The variables investigated in the current stu dy included immediacy of reinforcement delivery, magnitude of reinforcement, and quality of reinforcement.

PAGE 17

17 CHAPTER 2 GENERAL METHOD Participants and Setting Thirty individuals participated in Experiment 1, 11 individuals participated in Experiment 2, and 3 individuals participated in Experiment 3. Participants in Experiment 1 included 18 individuals diagnosed with PWS and 12 individuals diagnosed with mental retardation (MR) and/or developmental disabilities other than PWS. Participants in Experiments 2 and 3 were diagnosed with PWS (in addition to other diagnoses). Participant characteristics are listed in Ta ble 2 1. Sessions were conducted at either an adult vocational program or at a special education school. Session rooms included at least one table and chair, materials necessary to complete the target response, and reinforcers. Food Types Evaluated in Pr eference Assessment Six food groups (grain, dairy, meat, fruit, vegetable, and snack) were included, with 8 different foods per group. The foods included are listed in Table 2 2. The same foods were used across all participants assessments unless a part icipant was allergic to or intolerant of a food, in which case that particular food was omitted from the assessment. Each group of foods was compared separately (within category assessment), such that a hierarchy of preference within each category was gene rated. The top 3 choices from each food group were then included in a larger (18 item) between category preference assessment. Response Measurement and Reliability In Experiment 1, the t arget behavior consisted of reaching out and grasping a piece of foo d. Foods were presented either in pairs (Paired Stimulus Preference Assessment; PS) or in a larger array (Multiple Stimulus Without Replacement Preference Assessment; MSWO). In Experiments 2 and 3, target behaviors consisted of simple vocational or acade mic responses. The

PAGE 18

18 response for Participants JB, BK, AC, XD, KB, PM, MJ, DD, DV, and NR consisted of placing a washer on a screw and fastening a nut onto the screw such that the washer and nut remained in place. T he response for Participant MW consisted of placing an index card in an envelope, closing the envelope, and placing it in a manila envelope. Trained observers recorded food selections (PS and MSWO) and the order in which they were selected (MSWO) during Experiment 1. Data were collected on a form that listed all of the foods (MSWO assessment) and all possible combinations of foods (PS assessment). Reliability (interobserver agreement) for the preference assessments was assessed by having a second, independent observer collect data with the primary observer during an average of 50 % of sessions (range, 0% to 100%). Observers records were compared on a trial by trial basis, and reliability was calculated by dividing the number of scoring agreements by the number of agreements plus disagreements, and then multiplying by 100%. In Experiments 2 and 3, trained observers recorded the frequency of correct responses and incorrect responses, the duration of contact with session materials, and the delivery of reinforcers on handheld PDAs. A correct response f or the assembly task was defined as placing a washer onto a bolt, attaching a nut to the bolt, and placing the assembled piece into a bin. An incorrect response was scored for the assembly task if any of the components (washer, nut) were missing, or if th e pieces were detached upon placement in the bin. A correct response for the envelope stuffing task was defined as placing a card into an envelope, closing the envelope flap, and placing the envelope into a manila folder. An incorrect response for that tas k was scored if any components (card, envelope) were missing when an item was placed into the folder, or if the envelope flap was not closed when the item was placed into the folder. Duration measures of item contact were also recorded. T he on key for re sponse materials was pressed when a

PAGE 19

19 participant touched a set of materials. The off key was pressed when either (a) the materials for the other response were touched, or (b) the participant ceased contact with all materials for 3 consecutive seconds. Du ration measures were calculated by examining the raw data stream and determining the difference between the time (in seconds) that the on and off keys were scored for each response. Durations for response 1 and response 2 were summed, respectively. Rei nforcer delivery was scored when a reinforcer was placed on a participant plate or into a bag. A second observer independently recorded data during 32% (range, 22% to 57%) of the sessions for each participant. Reliability for frequency data was calculated by dividing each session into consecutive 10 s intervals, dividing the smaller number of responses scored in each interval by the larger number, averaging the results of these fractions, and multiplying by 100%. Reliability for response duration was calcu lated by dividing the number of intervals in which there was an agreement (for either occurrence or nonoccurrence) by the total number of intervals and multiplying by 100%. Reliability data for individual participants are listed in Table 2 3.

PAGE 20

20 Table 2 1. Participant characteristics. Participant Age Diagnosis/sensory impairments Experiment J.B. 34 PWS, Mental retardation (level unspecified) 1, 2 K.B. 27 PWS 1, 2, 3 X.D. 31 PWS, Mild mental retardation 1, 2 M.J. 25 PWS 1, 2 D.D. 29 PWS 1, 2 D.W. 37 PWS Mild mental retardation 1 D.V. 36 PWS, Mental retardation (level unspecified), Seizure disorder 1, 2 P.M. 38 PWS, Mild mental retardation 1,2 B.K. 36 PWS, Mental retardation (level unspecified) 1, 2, 3 A.H. 37 PWS 1 T.C. 30 PWS, Mental retardation (level unspecified) 1 A.C. 18 PWS, Mental retardation (level unspecified) 1, 2 N.R. 31 PWS, Mental retardation (level unspecified) 1, 2, 3 M.W. 23 PWS, Mild mental retardation 1, 2 P.P. 56 PWS, Mental retardation (level unspecified) 1 P.B. 37 PWS, Men tal retardation (level unspecified) 1 D.K. 33 PWS, Mental retardation (level unspecified) 1 A.R. 35 PWS 1 C.C. 20 Trainable mentally handicapped* 1 B.C. 48 Mental retardation (level unspecified) 1 D.L. 19 Educable mentally handicapped, speech & langua ge impaired* 1 J.C. 18 Educable mentally handicapped, speech & language impaired* 1 L.J. 18 Educable mentally handicapped, speech & language impaired* 1 J.Sm. 27 Mental retardation (level unspecified), Hydrocephalus 1 J.Sp. 30 Mental retardation (level unspecified) 1

PAGE 21

21 Table 2 1. Continued Participant Age Diagnosis/sensory impairments Experiment M.T. 40 Mental retardation (level unspecified), Klinefelters syndrome 1 E.An. 17 Autism, Mental retardation (level unspecified) 1 E.Al. 18 Down syndrome, Mental retardation (level unspecified) 1 M.W. 18 Trainable mentally handicapped* 1 E.Ad. 50 Mental retardation (level unspecified), Autism 1 *Indicates that the diagnosis available was one assigned by the school system

PAGE 22

22 Table 2 2. Foods included in eac h category of preference assessment Category Foods Snack Jell y beans, Twizzler, M&M, Raisinet, oatmeal cookie, chocolate chip cookie, potato chips, Doritos Protein Chicken, turkey, roast beef, ham, sausage, bacon, pepperoni, tuna (canned) Grain Wheat bread, white bread, rice cake, oatmeal, Cheerios, corn flak es, Triscuit cracker, saltine cracker Dairy Vanilla yogurt, cottage cheese, cheddar cheese, mozzarella cheese, cream cheese, sour cream, skim milk, vanilla pudding Vegetable Broccoli, tomato, celery, carrot, cucumber, lettuce, squash, green bell pepper Fruit Orange, apple, banana, grapefruit, pear, grapes, pineapple, cantaloupe

PAGE 23

23 Table 2 3. Reliability coefficients Experiment Participant Percentage of trials/ sessions with reliability Mean Reliability percentage Range 1 J.B. 68.5 98.9 92.6 100 A.R. 24 100 N/A D.V. 33.3 100 N/A P.M. 46 100 N/A X.D. 42 100 N/A A.H. 55 100 N/A T.C. 53 100 N/A D.W. 66.7 98 91.6 100 B.K. 65 100 N/A M.J. 33.3 100 N/A K.B. 33.3 100 N/A D.D. 87 100 N/A A.C. 82.8 100 N/A N.R. 65.1 100 N/A M.W. 36.6 100 N/A D.K. 44.4 100 N/A P.P. 26.1 100 N/A P.B. 65.1 99.3 98 100 C.C. 54.5 100 N/A B.C. 37.6 100 N/A D.L. 58.6 100 N/A E.An. 41.4 100 N/A L.J. 45.5 100 N/A M.T. 78.8 100 N/A E.Al. 68.7 100 N/A J.C. 78.8 100 N/A M.W. 100 100 N/A

PAGE 24

24 Table 2 3. Continued Experiment Participant Percentage of sessions/ trials with reliability Mean reliability percentage Range 1 E.Ad. 26.1 98.8 96.4 100 J.Sm. 53.5 100 N/A J.Sp. 82.6 99.1 96.4 100 2 D.D. 22 97.7 81.7 100 K.B. 22 97.3 84.2 100 N.R. 40 98.8 88.3 100 M.J. 25 97.6 93.3 100 X.D. 30 97.1 86.7 100 P.M. 40 95.9 53.3 100 B.K. 11 97.3 85.8 100 D.V. 33 96.9 79.2 100 M.W. 33 98.8 89 100 A.C. 40 96.0 42.5 100 J.B. 45 97.9 83 .3 100 3 N.R. 37 98.0 78.3 100 B.K. 57 96.6 30 100 A.C. 30 98.6 86.7 100

PAGE 25

25 CHAPTER 3 EXPERIMENT 1: ASSESS MENT OF FOOD PREFERE NCES F ood preferences in the PWS population have been examined in several studies (Caldwell & Taylor, 1993; Fieldstone et al., 1997; Glover, Maltzman, & Williams, 1996; Rankin & Mattes, 1996; Taylor & Caldw ell, 1985). Research on qualitative aspects of food preference has included taste (sweet, salty, etc.; Caldwell & Taylor, 1985), macronutrient content (fat, carbohydrate, etc.; Fieldstone et al., 1997), and familiarity of foods (Rankin & Mattes, 1996). S ome results have shown group preferences for these qualitative variables (e.g., a preference for sweet foods; Caldwell & Taylor, 1983; Hinton, Holland, Gellatly, Soni, & Owen, 2006), whereas others report idiosyncratic preferences (e.g., Fieldstone et al., 1997). However, methodological features of several of these studies make data difficult to interpret. First, most of the studies presented aggregated data, which precluded analysis of individual sets of data. That is, it is unclear if the group average was an accurate representation of the responding of each individual within the group. Second, most of the studies assessed a very small number of stimuli (typically 4 or 9 items), which may limit the generality of their data. For example, it is not clear if a preference for the one sweet item from among 4 other items indicated a general preference for sweet foods versus a preference for that specific food relative to the other three foods included. Finally, some of these studies relied on verbal report (e .g., Hinton et al, 2006) rather than a food selection and consumption response. Data from Taylor and Caldwell (1985) and Glover et al. (1996) raise questions about the correspondence between verbal report and actual choices of PWS individuals during prefer ence assessments; thus, it is unclear to what extent the verbal report of preference is an actual indication of what an individual with PWS would choose to eat. The purpose of Experiment 1 was to determine what, if any, food preferences individuals with P WS exhibit. A secondary purpose was to determine if there were general characteristics of

PAGE 26

26 the preferences exhibited by the group (e.g., a preference for high calorie foods or for a particular food type). Finally, data were examined to determine if there we re any strong differences in the preferences across groups. Preference Assessment Procedures Preference assessments consisted of an experimenter presenting small samples of two or more foods and prompting the participant to select one. The serving size of each piece or portion of food was approximately 1 teaspoon during all sessions in Experiment 1. In Experiments 2 and 3, the portion size varied across conditions (ranging from 1 teaspoon to 15 teaspoons of food). All participants were exposed to the Mult iple Stimulus w ithout Replacement (MSWO) format (DeLeon & Iwata, 1996). If selections during that assessment suggested that participants were selecting based on stimulus position (e.g., selecting all items from the right hand side and progressing to the l eft), a Paired Stimulus Assessment (PS) was substituted (Fisher et al., 1992). During the MSWO assessment, all foods included in a particular assessment were presented in a grouped array. Participants were prompted to select their most preferred item and consume it, after which the array (minus the selected item) was rearranged and presented again. This procedure was continued until either (a) all items were consumed or (b) the participant did not make further selections. An assessment consisted of 3 rep eated presentations of the entire array, and a separate assessment was conducted for each food group. The top 3 choices from each food group were then included in a large (18 item) between category MSWO assessment. During the PS assessment, foods within a food group (see above) were presented in pairs, one pair per trial. Each trial consisted of two stimuli (foods) placed in front of the participant; approaches to one of the stimuli resulted in delivery of the food sample and removal of the other. Each ite m was presented in a pair with every other item until all possible pairs had been

PAGE 27

27 presented. The top 3 choices from each food group were then included in a large (18 item) between category PS assessment. Results and Discussion Tables 3 1 and 3 2 list the results of the between category preference assessments for the PWS participants, and T ab les 3 3 and 3 4 list the results of the between category preference assessments for the non PWS participants. Results from the within category assessments are not presented because those assessments were used only as a method of identifying high preference foods to be compared in the between category assessment. The tables list the top 3 items from each category that were included in the between category assessment, the percent of times those foods were selected in the between category assessment, the rank order of each food based on the percentage of times selected, and the rank for each category of food for each participant. Based on the category rankings, 9 out of 18 individuals in the PWS group tended to show a preference for the snack food category o ver other categories. Four individuals showed a preference for the protein category foods, 2 showed a preference for the fruit category foods, 1 showed a preference for the grain category foods, 1 showed a preference for the dairy category foods, and no in dividuals showed a preference for vegetable category. One individuals data showed that grain and protein category foods were ranked equally high. In the non PWS group, 4 out of 12 individuals showed a preference for the protein group, 3 showed a preferenc e for the dairy group, 2 showed a preference for the vegetable group, 2 showed a preference for the fruit group, and 1 showed a preference for the grain group. No control participants preferred the snack foods to other food groups. Figures 3 1, 3 2, and 3 3 graphically depict the results of the between category preference assessments for the PWS participants, and F igures 3 4 and 3 5 depict the results of the between category preference assessments for the non PWS participants. Results show that

PAGE 28

28 most indiv iduals showed preferences for specific foods, although a few individuals preferences appeared to be rather weak. Figure 3 1 depicts the results for 7 PWS participants who were exposed to the MSWO preference assessment. Participant KBs results show a st rong preference for 3 foods, all of which were snacks; these data suggest that KB not only showed preferences for a few specific foods, but also a general preference for snack foods. Data for DV show a similar pattern: B oth fruits and grains appeared to be more preferred than other types of foods. Data for PM, DK, AR, BK, and DW show strong preferences for particular foods but not a preference for foods within a category over other categories. Figure 3 2 depicts the data for the other 6 PWS participants ex posed to the MSWO preference assessment. AC, AH, MJ, XD, and TC showed weaker preferences than did participants whose data were shown in Figure 3 1, although some preferences for individual items can be identified. Data for DD show a fairly strong preferen ce for one item from the protein group (sausage) and a general preference for snack foods. Figure 3 3 depicts the data for the 5 PWS participants who were exposed to the PS preference assessment. JB showed strong preference for all of the snack foods. Par ticipants PB, PP, NR, and MW all appeared to prefer certain foods but did not show general preferences for a particular category of foods. Participant MWs data are interesting in that, although he did not appear to prefer any one category of foods, he di d appear to select away from one category of foods (vegetables). Figure 3 4 depicts the data for 6 of the non PWS participants exposed to the MSWO. Data for MT, CC, BC, and MW all show evidence of preference for particular categories in addition to some isolated items. MT showed a preference for dairy foods, CC and MW showed a preference for protein foods, and BC showed a preference for vegetables. Data for EAl and JC show

PAGE 29

29 preferences for a few items, although the preference appears somewhat weaker, and there does not appear to be a general category preference. The top four panels of Figure 3 5 depict the results for the other 4 non PWS participants exposed to the MSWO assessment. Data for LJ, EAn, DL, and JSm show weaker patterns of preference: A few ite ms were selected on a larger percentage of trials than others, but the differences in the rankings are slight compared to those of other participants (e.g., MT or CC). Data for the 2 non PWS participants who were exposed to the PS assessment are depicted i n the bottom two panels of Figure 3 5. Participants SP and EAd both showed preferences for particular items, but neither showed a general preference for one category of foods over all others. Overall, the data show a large amount of variation in preferenc es across participants; in addition, there was no striking difference in the preferences of PWS versus non PWS participants. Among those who showed strong preferences, the highly preferred items tended to be relatively high calorie for both PWS and non PW S participants, and very few people showed preferences for extremely low calorie foods (vegetables and fruits). The mean rankings showed some differences across groups in what categories were (on average) most preferred. In the PWS group, the snack catego ry was most often the highest ranked category (50% of individuals showed the highest ranking for the snack category), whereas the protein and dairy groups were ranked highest among the non PWS group. However, the snack food category didnt necessarily cont ain foods that were higher calorie than all other categories; the dairy and protein categories contained some of the most calorie dense foods (e.g., sausage, cheddar cheese). Thus, it is not the case that the differences observed represented preferences on the part of the PWS participants based simply on caloric value.

PAGE 30

30 Table 3 1. Results of between category MSWO preference assessments for PWS group

PAGE 31

3 1 Table 3 1. Continued

PAGE 32

32 Table 3 1. Continued

PAGE 33

33 Table 3 1. Continued

PAGE 34

34 Table 3 1. Continued

PAGE 35

35 Table 3 1. Continued

PAGE 36

36 Table 3 2. Between category paired stimulus preference assessments for PWS group

PAGE 37

37 Table 3 2. Continued

PAGE 38

38 Table 3 3. Results of between category MSWO preference assessments for control group

PAGE 39

39 Table 3 3. Continued

PAGE 40

40 Table 3 3. Continued

PAGE 41

41 Table 3 3. Continued

PAGE 42

42 Table 3 4. Results of between category paired stimulus preference assessments for control group

PAGE 43

43 Figure 3 1. Percent of times a food was selected in between category MSWO preference assessments for 7 PWS participants (Experiment 1).

PAGE 44

44 Figure 3 2. Percent of times a food was selected in between category MSWO preference assessments for 6 PWS participants (Experiment 1).

PAGE 45

45 Figure 3 3. Percent of times a food was selected in betwe en category Paired stimulus preference assessments for 5 PWS participants (Experiment 1).

PAGE 46

46 Figure 3 4. Percent of times a food was selected in between category MSWO preference assessments for 6 control participants (Experiment 1).

PAGE 47

47 Figure 3 5. Percent of times a food was selected in between category MSWO preference assessments and Paired Stimulus preference assessments for 6 control participants (Experiment 1).

PAGE 48

48 CHAPTER 4 EXPERIMENT 2: DETERM INANTS OF FOOD PREFE RENCE Studies on the relative influence of quality, magnitude, and delay have shown somewhat mixed results. In a two part experiment comparing the choices of individuals with PWS to developmentally disabled peers, Glover et al. (1996) showed that when foods were of relatively similar quality, magnitude appeared to be more influential than quality for individuals with PWS. By contrast, control participants tended to select based on food quality (rather than m agnitude). In the second experiment, all participants selected a smaller magnitude of the high preference item over larger magnitudes of low preference items. Joseph et al. (2002) compared the effects of magnitude and delay on foods of different qualities. Results showed that individuals in the PWS group were more likely to make choices based on magnitude than delay and did so to a greater extent than individuals in the obese comparison group. Results across different qualities of food showed that the hig her preferred foods were associated with a larger overall proportion of choices for the larger magnitude of food (at a delay), whereas magnitude appeared less influential when the foods were lower quality. The authors concluded that individuals with PWS we re more likely to make choices based on food magnitude than delay. The research to date suggests that both quality and magnitude may influence food choices. However, the influence of quality relative to both magnitude and delay has not been examined. Fu rther, no experiment has presented data in such a way that the influence of those variables on choice could be identified for individual participants. The purpose of the current experiment is to provide an analysis of the relative influence of quality, mag nitude, and immediacy on food choices of individuals with PWS.

PAGE 49

49 Participants, Design and Procedure Eleven individuals (all PWS) participated in Experiment 2. All sessions were conducted in either a sheltered workshop or a therapy room at a school. Targe t behaviors consisted of vocational tasks (piece work assembly and stuffing envelopes, defined previously). Two sets of materials were available, each associated with a different reinforcement contingency. Discriminative stimuli were included to signal the reinforcement contingency in place. The presence of a plate next to the participants session materials signaled that food on the plate would be available at the end of session (immediately), and the presence of a plastic bag signaled that food would be available after some delay. The food on the experimenters plate was visible, such that the participant was able to view the different foods associated with a response option (quality), and the number of pieces of food on the experimenters plate signaled the magnitude of food available for each response (1 versus 5 pieces). Prior to each session, participants were prompted to complete one response per set of materials, and the reinforcer associated with that set of materials was placed onto the participant s plate (or into a bag when delays to reinforcement were in effect). The participant was then told that you can do whatever you want to do during session, and sessions were begun. No further instructions or discussion occurred during session. Two meas ures were taken to ensure that participants did not consume an excessive number of calories. First, reinforcers were delivered according to a VI 30 s schedule during sessions (such that no more than an average of 20 pieces of food could be earned per sess ion). Second, no more than three, 10 min sessions were conducted per day. Participants were exposed to at least two sessions of each condition, and additional sessions were conducted if responding was inconsistent across those two sessions. Baseline. Durin g baseline, each parameter was evaluated singly to determine sensitivity to quantitative changes in it. When testing each parameter, the values of all other parameters were

PAGE 50

50 held constant. Foods were selected based on the between category preference assess ment (Experiment 1), and two high quality and two low quality items were selected (to reduce the chance of satiation to a particular item). The two items that were ranked first and second were selected as high quality foods. The two lowest ranked items we re selected as the low quality foods with the additional criterion that the low ranked foods must be low calorie (vegetable or fruit). During the immediacy (I) baseline, one response option resulted in reinforcer delivery at the end of the session, and th e other option resulted in reinforcer delivery one hour later. Magnitude and quality were held constant by using one piece of a high quality food across both options. During the magnitude (M) baseline, one option resulted in 5 pieces of food, and the oth er option resulted in 1 piece of food. Immediacy and quality were held constant by delivering the high quality food immediately after session across both options. During the quality (Q) baseline, one option resulted in delivery of one of the two highest ra nked foods, and the other option resulted in delivery of one of the two lower ranked foods. Foods of the same quality (high or low) were alternated across reinforcer deliveries. Magnitude and immediacy were held constant by delivering one piece of food imm ediately after session. Assessment of Competing Parameters. Disparate values of each parameter (Immediacy [I], Magnitude [M], and Quality [Q]) were compared, with only 2 of 3 parameters manipulated per session. The parameter that was not being compared w as at its optimal (quality, immediacy) or a moderate (magnitude) value across both response options. During Immediacy versus Magnitude (IvM) sessions, one response option was associated with an immediately available, small magnitude reinforcer, and the oth er option was associated

PAGE 51

51 with a delayed, large magnitude reinforcer. Because quality was not manipulated in this condition, the food item was the same (high quality) across both response options. Thus, the contingencies associated with response options we re (a) 1 piece of a high quality food available at the end of the session, or (b) 5 pieces of a high quality food available 50 min after the session was over. During the Immediacy versus Quality (IvQ) condition, one response option was associated with an immediately available, low quality reinforcer, and the other option was associated with a delayed, high quality reinforcer. Magnitude was not manipulated in this condition; therefore, the magnitude for each option was set at a moderate value (1 tablespoo n sized portion, which equals approximately 3 times the size of one teaspoon sized portion). The contingencies associated with response options were (a) 1 (tablespoon sized) piece of a low quality food available at the end of the session, or (b) 1 (tablesp oon sized) piece of a high quality food available one hour after the session was over. During the Magnitude versus Quality (MvQ) condition, one response option was associated with a large magnitude, low quality reinforcer, and the other option was associat ed with a small magnitude, high quality reinforcer. Immediacy was not manipulated in this condition; therefore, reinforcers were available at the end of session across each response option. The contingencies associated with response options were (a) 1 piec e of a high quality food available at the end of session, or (b) 5 pieces of a low quality food available at the end of session. Results and Discussion Results of the assessments are depicted in Figures 4 1 and 4 2. The figures show the final session of e ach condition, as this represents response allocation following a history of exposure to contingencies. That is, the first session per condition may not accurately represent behavior

PAGE 52

52 under the control of a particular set of contingencies because participan ts did not experience the (primary) reinforcement contingency until after the session. If behavior were to change as a result of exposure to the reinforcement contingency, this effect could not be observed until at least the second session per condition. Results show that for 10 of 11 participants, responding during baseline showed near 100% time allocation to what would traditionally be considered the favorable alternative: high quality over low quality, large magnitude over small magnitude, and immedia te over delayed alternatives. One participant, MJ, deviated from this pattern slightly in that she allocated most (76%) but not all of her time to the high quality option. The purpose of baseline was to demonstrate each participants sensitivity to the re inforcement parameters, as a prerequisite to participating in the subsequent assessment conditions. If a participant showed little or no sensitivity to one or more of the parameters of interest, then the subsequent assessment would not provide useful info rmation. Further, if a participant showed a preference for the unfavorable value of a parameter (e.g., allocating more time to the small magnitude food over the large magnitude food), then it was deemed likely that variables other than the parameters of i nterest might be influencing behavior (e.g., a history of instructions, a history of reinforcement for selecting smaller portions, etc.). Therefore, participants whose behavior showed indifference to the parametric manipulations (e.g., near chance levels of responding across both response options of a baseline condition) or a preference for the unfavorable option were excluded from the subsequent assessment. Three participants were exposed to baseline conditions but were excluded from assessment for this reason (data not presented here, but available from the authors upon request).

PAGE 53

53 Nine of 11 participants allocated most or all of their responding to the response option associated with the high quality food during assessment (Figure 4 1). For these 9 indiv iduals, data show that they selected the option associated with the high quality food, even when it produced a 50 min delay to that food (during the IvQ condition) and when it produced 1/5 the amount of food (during the MvQ condition). These results sugge st that quality was a more influential reinforcer parameter than immediacy or magnitude. Among those 9 participants, 5 participants showed that when immediacy and magnitude were compared, magnitude was the more influential parameter. Three participants s howed that immediacy was more influential than magnitude, and one participants responding was allocated equally among response options associated with immediately available versus larger magnitude reinforcers. Two of 11 participants allocated most or all of their responding to the response option associated with the reinforcer that was available immediately after session (Figure 4 2). These data show that responding was allocated to the immediately available reinforcers, even when it produced a lower qual ity food (during the IvQ condition) and when it produced a smaller magnitude of food (during the IvM condition). These results suggest that immediacy was a more influential parameter than quality or magnitude for these participants. In addition, data sho w that these two participants chose a higher quality reinforcer over a larger magnitude reinforcer (during the MvQ condition). No participants showed a pattern in which responding was influenced primarily by the magnitude of food. Magnitude appeared to be influential under some conditions, as observed for JB, KB, DD, DV, and PM. However, this effect was observed only when magnitude was compared with immediacy, and quality was the same across options. When quality was manipulated, magnitude was not a cont rolling variable.

PAGE 54

54 These results were somewhat unexpected. Numerous reports have documented that hyperphagia is a prevalent characteristic of PWS and that individuals with PWS tend to consume much larger quantities of food than control subjects (Holland, Treasure, Coskeran, & Dallow, 1995; Holland et al., 1993). Further, reports suggest that individuals with PWS are very slow to satiate, if satiation (reductions in speed of consumption) is noted at all (Lindgren, et al., 2000). The behavior (hyperphagia) itself seems to suggest that large quantities of food are valuable; therefore, it seems likely that magnitude would be an influential reinforcement characteristic. Instead, the results for all participants in the current show that the quality (9 particip ants) or the immediacy (2 participants) of the reinforcement was the most influential characteristic.

PAGE 55

55 Figure 4 1. Percentage of time allocated to each response option for the final session of each baseline and assessment condition. Open portions of ba rs during assessment indicate choice of an unfavorable dimension (Experiment 2).

PAGE 56

56 CHAPTER 5 EXPERIMENT 3: PARAME TRIC MANIPULATIONS T O SHIFT CHOICE The results of Experiment 2 showed that 9 of 11 participants choices were determined by food quality. Selection based on food quality is not a problem per se; however, many of these partici pants highly preferred foods were relatively high calorie (top ranked foods included pork breakfast sausage, Raisinets, cheddar cheese, pepperoni, and cookies). Therefore, many individuals who selected foods based on quality were selecting extremely high calorie foods. This type of responding may be problematic, as many individuals diagnosed with PWS typically are on restricted calorie diets. Interventions that teach them to select lower calorie options may have therapeutic benefits. Two of 11 participa nts in Experiment 2 selected foods based on the immediacy of their delivery. Food selections based primarily on immediacy may present challenges to controlled eating in this population. Individuals may choose to eat extremely high calorie or dangerous (s poiled, contaminated) foods because they are immediately available, which may result in weight gain or illness. Teaching these individuals to wait for delayed alternatives may allow for a number of therapeutic interventions to control weight (e.g., reinfo rcement contingencies for the absence of inappropriate food consumption, reinforcement for surrendering found foods, etc.). A variety of interventions have been used to decrease impulsive choices, including the use of signals to indicate reinforcement will be delivered at a later time (Vollmer, Borrero, Lalli, & Daniel, 1999), introducing activities during the delay (Dixon, Rehfeldt, & Randich, 2003), and initially decreasing and then gradually increasing delays (Schweitzer & Sulzer Azaroff, 1988). Schweit zer and Sulzer Azaroff identified 5 participants who tended to select smaller, immediate reinforcers over larger, delayed reinforcers. After establishing a baseline in which the participants consistently selected the smaller, immediate reinforcer, the expe rimenters introduced

PAGE 57

57 training. Initially, delays to both the small and the large reinforcers were set at equal values (0s); during subsequent sessions, the delay to the large reinforcer was increased in 5 s increments. Following training, 4 of 5 participa nts consistently selected the larger, delayed reinforcer over the smaller, immediate reinforcer. Neef et al. (1994) used a similar procedure to teach 3 participants diagnosed with ADHD to tolerate delays to reinforcers. During assessment, each individual allocated most or all responding to the immediately available reinforcers. Self control training consisted of initially decreasing and then gradually increasing the delays to a higher quality reinforcer, while the lower quality, no delay reinforcer conti nued to be available. Results showed that the participants continued to allocate most of their responding to the delayed reinforcer, even when delays were increased up to 24 hours. Results of post assessment sessions showed that some amount of generalizati on occurred across untrained conditions. The purpose of Experiment 3 was to shift participant responding to either a lower calorie option (BK and AC) or to a delayed option (NR). BK and AC were exposed to a variety of parametric manipulations of immediacy and magnitude, and a subsequent replacement of the higher calorie foods with lower calorie versions. NR was exposed to delay and magnitude fading. Participants, Design and Procedure Three individuals from Experiment 2 participated in Experiment 3. NR sel ected foods that were immediately available (control by immediacy), even when that resulted in a smaller amount of food or a lower quality food. BK always selected the high quality food (control by quality) in Experiment 2, even when that resulted in a sma ller amount of food or a delay to food availability. Further, her high quality foods were relatively high calorie (pork sausage and Raisinets candy). AC also selected only high quality foods (control by quality); her high quality foods were pepperoni and oatmeal raisin cookies, which were also high calorie.

PAGE 58

58 The experimental arrangement, target behaviors, pre session exposure to contingencies, and reinforcement schedules were identical to Experiment 2. The reinforcement contingencies associated with each r esponse varied according to the desired change in responding (i.e., selection of a lower calorie food, or selecting foods that would be available after a delay). Immediacy treatment The purpose of treatment was to teach NR to tolerate (i.e., choose) del ays to food availability. A combination of magnitude and delay fading was used. The sessions were run in the context of the IvM condition, with the exception that lower calorie foods were substituted. Throughout treatment sessions, one response optio n continued to be associated with one piece of food available immediately after the session (the option to which she allocated all responding during the IvM assessment sessions). The contingencies associated with the second response option changed across s essions. Initially, both the magnitude of the food delivered per reinforcer was increased (to 15 pieces), and the delay to that foods availability was decreased (to the end of session). During subsequent sessions, the delay to foods delivery was gradua lly increased. After the delay was increased to one hour, magnitude fading was initiated by decreasing the number of pieces by one per session. By the final sessions of delay and magnitude fading, the second response option was for 5 pieces of food avail able after one hour. Thus, the options available were the same as those available during the original IvM assessment sessions. Quality treatments The purpose of the quality treatment was to arrange conditions such that BK and AC would select lower calori e foods over higher calorie foods. Several parametric manipulations were introduced in an attempt to reduce caloric intake. During all quality treatment sessions, the response options were associated with two different sets of foods. BKs high quality fo ods were

PAGE 59

59 sausage and cheddar cheese, her low quality foods were lettuce and apple, and her high quality substitutes were vegetable sausage and fat free cheddar cheese. ACs high quality foods were pork pepperoni and oatmeal raisin cookies, her low quality foods were tomatoes and carrots, and her high quality substitutes were turkey pepperoni and South Beach Diet chocolate chip oatmeal cookies. Delay manipulation During these sessions, the delay to the high quality food was increased from one hour (durin g assessment) to 3 hours, and the magnitude was the same across both options (1 piece of food). Thus, one response option was associated with one piece of a high quality food available after 3 hours, and the second option was associated with one piece of a low quality food available at the end of the session. Magnitude manipulation During these sessions, the magnitude of the low quality food was increased from 5 pieces (during assessment) to 15 pieces, and the delay was equal across both options (end of session). One response option was associated with one piece of a high quality food at the end of the session, and the second option was associated with 15 pieces of a low quality food available at the end of the session. Delay plus magnitude manipulatio n Both the delay to the high quality food and the magnitude of the low quality food were manipulated in these sessions. One response option was associated with one piece of a high quality food available after 3 hours, and the second option was associated with 15 pieces of a low quality food available at the end of session. Quality substitution The low quality foods were replaced with reduced calorie versions of the high quality foods during these sessions, and both magnitude and delay values were equa l across both options (1 piece and end of session, respectively). Thus, one response option was associated with one piece of the high quality food available at the end of the session,

PAGE 60

60 and the second option was associated with one piece of the high quality substitute available at the end of the session. Results and Discussion Figure 5 2 depicts the results of the immediacy (top panel) and quality (bottom panel) treatments. Results of NRs assessment showed that she selected the immediately available reinfor cer whenever delay was manipulated. When the delay and magnitude fading treatment was introduced, she allocated all of her responding to the option associated with the larger magnitude reinforcer that was delivered after gradually increasing delays. She continued to allocate all responding to the delayed option throughout all increases in the delay and decreases in the magnitude. During the final four sessions, she continued to allocate all of her time to the response option associated with the delayed ( 60 min), larger magnitude (5 pieces) reinforcer. This reflects a shift from the assessment condition, during which she allocated all of her time to the option associated with the immediately available, smaller reinforcer. Results of the quality treatment for BK are depicted in the center panel of Figure 5 1. During assessment, she allocated all of her time to the option associated with the higher quality foods (sausage and cheddar) whenever quality was manipulated. Throughout the delay, magnitude, and de lay plus magnitude manipulations, BK continued to allocate all of her responding to the higher quality foods. When high quality substitutes were introduced, she selected the high quality foods during the initial sessions, but in subsequent sessions alloca ted all of her time to the substitutes. Results for AC (depicted in the bottom panel of Figure 5 1) are similar; the delay, magnitude, and delay plus magnitude manipulations had no influence on her response allocation. Following the introduction of the h igh quality substitutes, she allocated all of her time to the high quality substitutes without further manipulation.

PAGE 61

61 The results of NRs treatment suggest that simple fading manipulations may be sufficient to teach individuals to tolerate delays to reinf orcers. This finding may be especially important in the PWS population because these individuals have a history of hyperphagia; teaching them not to immediately consume foods at each opportunity may be a necessary skill for achieving re integration into s ituations in which foods are available but should not be consumed. The results of BKs and ACs treatments were somewhat surprising; quality (as an independent variable) was sufficiently strong to prevent shifts despite long delays (3 hours), very dispa rate magnitudes (1 versus 15 pieces of food), and a combination of the increased delay and magnitude. The shift to the lower calorie version of the high quality foods was unexpected, given the extent to which quality had previously controlled responding. Their data suggest that the high quality substitutes may have been even higher quality than the high quality foods identified in the preference assessment.

PAGE 62

62 Figure 5 1. Percentage of time allocated to each response option during immediacy (top panel) and quality (bottom panel) treatments (Experiment 3).

PAGE 63

63 CHAPTER 6 GENERAL DISCUSSION The prevalence of hyperphagia and other food related problem behaviors in the PWS population suggest that food is an unusually potent reinforcer for this group. Hyperphagia is listed as one of the major clinical features of th e disorder, and researchers have noted the occurrence of other food related problems such as property destruction to access food (Benjamin & Buot Smith, 1993), stealing (Donaldson, et al., 1995), hoarding food (Dykens, Leckman, & Cassidy, 1996), and consum ption of substances such as pet food or food from trash cans. (Russell & Oliver, 2003). Unrestricted access to food also poses a danger for this population because individuals with PWS are more likely than those in the general population to die from choki ng (Stevenson et al., 2006), gastric necrosis (Wharton, Wang, Graeme Cook, Briggs, & Cole, 1997), and obesity related diseases such as diabetes or heart failure (Schrander Stumpel et al., 2004). Given the behavioral and physical problems presented by food in this population, an assessment of variables that might influence food selection (and subsequent consumption) seems relevant to a comprehensive treatment for PWS. Most studies on reinforcers characteristics that influence food choice in PWS have focused on preferences for food type. Results of some studies suggested that individuals with PWS prefer sweet foods to all others (Caldwell & Taylor, 1983; Glover et al., 1996; Hinton et al., 2006; Taylor & Caldwell, 1985), whereas others suggest a variety of ta stes and types of food are preferred (Fieldstone et al., 1997; Rankin & Mattes, 1996). However, the small number of stimuli presented in all studies limits the extent to which generalizations about food preferences can be made. Further, the presentation of data in an aggregated format obscures individual differences in preference. The results of Experiment 1 represent a departure from previous studies on food preference in the population. First, several different samples and types of food

PAGE 64

64 were used (48 fo ods were assessed across 6 types of food). Therefore, conclusions about preferences for various types of food may have more generality than those of previous studies. Second, each individuals data showed slightly different preferences, and it did not appe ar to be the case that (a) preferences were uniform across individuals within a group, or (b) there were important differences in the caloric value of choices across groups. The data from Experiment 1 suggest that the preferences of individuals with PWS a re not drastically different than those of individuals with developmental disabilities other than PWS. The results of Experiment 2 showed that quality was most often the variable determining choice. These results are consistent with one previous study tha t suggested that individuals with PWS sometimes choose smaller quantities of preferred foods over larger quantities of less preferred food (Caldwell & Taylor, 1983). However, other studies have shown that magnitude may occasionally be more influential than quality (Glover et al., 1996) or delay (Joseph et al., 2002) in food selections of individuals with PWS. Several factors may have contributed to the emergence of quality as such an influential variable in the current study. First, it is possible that e xtra experimental factors influenced responding. Most individuals diagnosed with PWS are prescribed therapeutic (calorie restricted) diets, which may result in the restriction or total elimination of higher calorie foods. Thus, high calorie foods are typ ically unavailable, which may increase their value in the context of the study. Future studies may wish to rule out this influence by restricting the foods used to those typically available to individuals with PWS (i.e., an array of all low calorie foods) Second, it is possible that the quality parameter was stronger in the current study than it has been in previous studies. Previous studies sampled a small number of stimuli, which may have resulted in the identification of foods that were of roughly si milar value. The wide array of

PAGE 65

65 foods included in the current study may have allowed for the identification of a set of stimuli with more disparate reinforcing potencies, such that the lower preferred foods had little value relative to the highly preferred foods. (However, it should be noted that all foods used as low preferred foods were in fact highly preferred during the within category preference assessments). Data presented by Glover et al. (1996) support this hypothesis: Their results showed that mag nitude influenced selection among similarly preferred foods (foods ranked first in a preference assessment versus those ranked second and third), but that quality was more influential than magnitude when the comparison was among highly preferred and non pr eferred foods. The large number and broad variety of items assessed in the current study may have ultimately increased the strength of the quality variable, relative to the immediacy and magnitude variables. Third, it is possible that the differences in t he magnitude and/or the delay values constituted a fairly weak independent variable. The variables were of sufficient strength to influence responding during the baseline sessions (i.e., all subjects allocated most or all of their time to the favorable va lues); however, those variables exerted little control over most participants responding during the subsequent assessment. It is unclear to what extent those values would have to be changed such that they were as powerful as the quality variable. Data f rom BK and ACs treatment phase showed that three fold increases in the delay and magnitude values did not produce a shift in responding. Thus, it may be the case that much larger differences in the delay and magnitude variables (e.g., 5 or 10 fold increa ses) are necessary to make those variables more potent. Another alternative may be to change the magnitude or delay values of the higher quality food (e.g., decrease the size of the preferred food or the delay to the

PAGE 66

66 lower preferred food). These changes w ere not attempted in the current study, but may prove useful in future research. The implications of quality as an influential variable may provide some insight into food related problem behaviors in individuals with PWS. Early researchers had suggested t hat the unusual consumptive behavior (i.e., ingestion of contaminated or unusual foods) was an indication of indiscriminate eating or a lack of preference (Pipes & Holm, 1973). However, data from Experiment 1 and 2 suggest that this is not the case. The unusual consumptive behavior may not be due to a lack of food preference; rather, variables such as restricted access to foods may create a situation in which any ingested material is sufficiently reinforcing to maintain such behavior, despite preferences among foods. Results of a study by Roscoe, Iwata, and Kahng (1999) showed that although participants tended to allocate responding to a preferred food over a non preferred food under a concurrent reinforcement schedule, contingent delivery of the non prefe rred food was sufficient to maintain responding during a single schedule. Thus, consumption of unusual or contaminated food by individuals with PWS does not necessarily indicate a lack of preference; rather, these foods may simply be sufficiently reinforc ing to consume when other foods are unavailable, despite preferences for other foods. Food appears to be a particularly powerful reinforcer for the PWS population, making information on variables that influence that reinforcer relevant to management of eat ing in the PWS. This information may allow for the development of more highly individualized treatments that better address each individuals problematic patterns of eating. For instance, individuals who choose immediately available foods may be appropriat e for delay fading, as illustrated in Experiment 3. Or, interventions in which appropriate (low calorie) foods are constantly (immediately) available may mitigate inappropriate food consumption. Individuals who show a

PAGE 67

67 strong tendency to select high qualit y items may be good candidates for research on shifting preference to lower calorie items. If substitution interventions such as those described in the current study are not effective, then interventions such as simultaneous delivery of preferred and non p referred foods (Piazza et al., 2002) or differential reinforcement contingencies in which a very small portion of preferred (high calorie) food is available contingent on consumption of larger amounts of low calorie foods may prove useful in managing the f ood consumption of this population. Despite the potential contribution of antecedent manipulations, it is likely that differential consequences will be necessary to suppress inappropriate food consumption. Studies on food stealing in PWS have shown that i nterventions including differential reinforcement of other behavior (DRO; Page, Finney, Parrish, & Iwata, 1983), self monitoring combined with contingency contracting and punishment (Altman, Bondy, & Hirsch, 1978), and verbal reprimands combined with stimu lus control interventions (Maglieri, DeLeon, Rodriguez Catter, and Sevin, 2000) may be successful in reducing inappropriate eating; however, limited evidence of generalization is available. Both Page et al. (1983) and Maglieri et al. (2000) evaluated thei r interventions in a multiple baseline across settings design, and results of both studies showed that the treatment did not generalize to untrained settings. Thus, comprehensive treatments for food related problem behaviors in PWS will likely require both antecedent and consequences manipulations, as well as explicit programming for generalization.

PAGE 68

68 REFERENCES Bekle, T. W. (1997). Running and responding reinforced by the opportunity to run: Effect of reinforcer duration. Journal of the Experimental Analysis of Behavior, 67, 337 351. Benjamin, E., & Buot Smith, T. (1993). Naltrexone and fluoxetine in Prader Willi syndrome. Journal of the American Academy of Child and Adolescent Psychiatry, 32 870 873. Berntson, G.G., Zipf, W.B., O'Dorisiod, T.M., Hoffmane, J.A., & Chancee, R.E. (1993). Pancreatic polypeptide infusions reduce food intake in Prader Wi lli syndrome. Peptides, 14, 497 503. Butler, M. G., & Thompson, T. (2000). Prader Willi syndrome: Clinical and genetic findings. Endocrinologist, 10, 3S 16S. Caldwell, M. L., & Taylor, R. L. (1983). A clinical note on food preference of individuals with Prader Willi syndrome The need for empirical research. Journal Of Mental Deficiency Research, 27, 45 49. Caldwell, M. L., Taylor, R. L., & Bloom, S. R. (1986). An investigation of the use of high preference and low preference food as a reinforcer for i ncreased activity of individuals with Prader Willi syndrome. Journal Of Mental Deficiency Research, 30 347 354. Catania, A. C. (1963). Concurrent performances: A baseline for the study of reinforcement magnitude. Journal of the Experimental Analysis of B ehavior, 6, 299 300. Chung, S. H. (1965). Effects of delayed reinforcement in a concurrent situation. Journal of the Experimental Analysis of Behavior, 8 439 444. Chung, S. H., & Herrnstein, R. J. (1967). Choice and delay of reinforcement. Journal of the Experimental Analysis of Behavior, 10 67 74. Davison, M. & Baum, W. M. (2003). Every reinforcer counts: Reinforcer magnitude and local preference. Journal of the Experimental Analysis of Behavior, 80 95 129. DeLeon, I. G., Fisher, W. W., Rodriguez Catter, V., Maglieri, K., Herman, K., & Marhefka, J. M. (2001). Examination of relative reinforcement effects of stimuli identified through pretreatment and daily brief preference assessments. Journal of Applied Behavior Analysis, 34 463 473. DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of a multiple stimulus presentation format for assessing reinforcer preferences. Journal of Applied Behavior Analysis, 29 519 533. Dixon, M. R. & Cummings, A. (2001). Self control in children with autism: Response a llocation during delays to reinforcement. Journal of Applied Behavior Analysis, 34 491 495. Dixon, M. R., Rehfeldt, R. A., & Randich, L. (2003). Enhancing tolerance to delayed reinforcers: The role of intervening activities. Journal of Applied Behavior A nalysis, 36 263 266.

PAGE 69

69 Donaldson, M.D.C., Chu, C.E., Cooke, A., Wilson, A., Greene, S.A., & Stephenson, J.B.P. (1994). The Prader Willi syndrome. Archives of Disease in Childhood, 70, 58 63. Dykens, E. M., Leckman, J. F., & Cassidy, S. B. (1996). Obsessi ons and compulsions in Prader Willi syndrome. Journal of Child Psychology and Psychiatry, 27, 995 1002. Fieldstone, A., Zipf, W. B., Schwartz, H. C., & Berntson, G. C. (1997). Food preferences in Prader Willi syndrome, normal weight and obese controls. I nternational Journal of Obesity, 21, 1046 1052. Glover, D., I. Maltzman, I., & Williams, C. (1996). Food preferences among individuals with and without Prader Willi syndrome. American Journal on Mental Retardation, 101 195 205. Graff, R. B., Gibson, L., & Galiatsatos, G. T. (2006). The impact of high and low preference stimuli on vocational and academic performances of youths with severe disabilities. Journal of Applied Behavior Analysis, 39 131 135. Herrnstein, R. J. (1961). Relative and absolute str ength of response as a function of frequency of reinforcement. Journal of the Experimental Analysis of Behavior, 4, 267 272. Hinton, E. C., Holland, A. J., Gellatly, M. S. N., Soni, S., & Owen, A. M. (2006). An investigation into food preference and the n eural basis of food related incentive motivation in Prader Willi syndrome. Journal of Intellectual Disability Research, 50, 633 642. Hoch, H., McComas, J. J., Johnson, L., Faranda, N., & Guenther, S. L. (2002). The effects of magnitude and quality of re inforcement on choice responding during play activities. Journal of Applied Behavior Analysis, 35 177 181. Holland, A.J., Treasure, J., Coskeran, P., Dallow, J., Milton, N., Hillhouse, E. (1993). Measurement of excessive appetite and metabolic changes in Prader Willi syndrome. International Journal of Obesity and Related Metabolic Disorders, 17, 527 532. Holland, A.J., Treasure, J., Coskeran, P., Dallow, J. (1995). Characteristics of the eating disorder in Prader Willi syndrome: Implications for treatmen t. Journal of Intellectual Disability Research, 39 373 81. Holm, V.A., Cassidy, S.B., Butler, M.G., Hanchett, J.M., Greenswag, L.R., Whitman, B.Y., Greenberg, F. (1993). Prader Willi syndrome: Consensus diagnostic criteria. Pediatrics, 91, 398 402. Jos eph, B., Egli, M., Koppekin, A., & Thompson, T. (2002). Food choice in people with Prader Willi syndrome: Quantity and relative preference. American Journal on Mental Retardation, 107, 128 35.

PAGE 70

70 Keesey, R. E., & Kling, J. W. (1961). Amount of reinforcement and free operant responding. Journal of the Experimental Analysis of Behavior, 4 125 132. Keller, J. V., & Gollub, L. R. (1977). Duration and rate of reinforcement as determinants of concurrent responding Journal of the Experimental Analysis of Behavi or, 28 145 153. Landon, J. Davison, M. & Elliffe, D. (2003). Concurrent schedules: Reinforcer magnitude effects. Journal of the Experimental Analysis of Behavior, 79 351 365. Lerman, D. C., Kelley, M. E., Van Camp, C. M., & Roane, H. S. (1999). Effects of reinforcement magnitude on spontaneous recovery. Journal of Applied Behavior Analysis, 32 197 200. Lindgren, A.C., Barkeling, B., Hagg, A., Ritzen, E.M., Marcus, C., & Rossner, S. (2000). Eating behavior in Prader Willi syndrome, normal weight, and o bese control groups. Journal of Pediatrics, 137 50 5. Lowe, C. F., Davey, G. C. L., & Harzem, P. (1974). Effects of reinforcement magnitude on interval and ratio schedules. Journal of the Experimental Analysis of Behavior, 22, 553 560. Maglieri, K. A., DeLeon, I. G., Rodriguez Catter, V., & Savin, B. M. (2000). Treatment of covert food stealing in an individual with Prader Willi syndrome. Journal of Applied Behavior Analysis, 33, 615 618. Matthews, L. R., & Temple, W. (1979). Concurrent schedule assess ment of food preference in cows. Journal of the Experimental Analysis of Behavior, 32, 245 254. McLean, A. P. & Blampied, N. M. (2001). Sensitivity to relative reinforcer rate in concurrent schedules: Independence from relative and absolute reinforcer dur ation. Journal of the Experimental Analysis of Behavior, 75 25 42. Miller, H. L. (1975). Matching based hedonic scaling in the pigeon. Journal of the Experimental Analysis of Behavior, 26, 335 347. Neef, N. A., Bicard, D. F., & Endo, S. (2001). Assessme nt of impulsivity and the development of self control in students with attention deficit hyperactivity disorder. Journal of Applied Behavior Analysis 34, 397 408. Neef, N. A., & Lutz, M. N. (2001). A brief computer based assessment of reinforcer dimension s affecting choice. Journal of Applied Behavior Analysis 34 57 60. Neef, N. A., Mace, F. C., & Shade, D. (1993). Impulsivity in students with severe emotional disturbance: The interactive effects of reinforcer rate, delay, and quality. Journal of Applie d Behavior Analysis, 26 37 52. Neef, N. A., Mace, F. C. Shea, M.C., & Shade, D. (1992). Effects of reinforcer rate and

PAGE 71

71 reinforcer quality on time allocation: Extensions of matching theory to educational settings. Journal of Applied Behavior Analysis, 2 5, 691 699. Neef, N. A., Shade, D, & Miller, M.S. (1994). Assessing influential dimensions of reinforcers on choice in students with serious emotional disturbance. Journal of Applied Behavior Analysis, 27 575 583. Neuringer, A. J. (1967). Effects of rein forcement magnitude on choice and rate of responding. Journal of the Experimental Analysis of Behavior, 10, 417 424. Pace, G. M., Ivancic, M. T., Edwards, G. L., Iwata, B. A., & Page, T. J. (1985). Assessment of stimulus preference and reinforcer value w ith profoundly retarded individuals. Journal of Applied Behavior Analysis, 18 249 255. Page, T. J., Finney, J. W., Parrish, J. M., & Iwata, B. A. (1983). Assessment and reduction of food stealing in Prader Willi children. Applied Research in Mental Retar dation, 4, 219 228. Piazza, C. C., Patel, M. R., Santana, C. M., Goh, H., Delia, M. D., & Lancaster, B. M. (2002). An evaluation of simultaneous and sequential presentation of preferred and nonpreferred food to treat food selectivity. Journal of Applied B ehavior Analysis, 35, 259 269. Pipes, P., & Holm, V. (1973). Weight control of children with Prader Willi syndrome Journal of the American Dietetic Association, 62, 520 524. Prader, A., Labhart, A., & Willi, H. (1956). Ein syndrom von adipositas, kleinw ucs, kryptorchismus und oligophrenie nach myatonicartigem zustand im neugeborenalter. Schweiz Med Wochenchr, 86, 1260 1261. Rankin, K. M., & Mattes, R. D. (1996). Role of food familiarity and taste quality in food preferences of individuals with Prader Willi syndrome. International Journal on Obesity and Related Metabolic Disorders, 20, 759 62. Reed, P. (1991). Multiple determinants of the effects of reinforcement magnitude on free operant response rates. Journal of the Experimental Analysis of Behavior 55 109 123. Roane, H. S., Vollmer, T. R., Ringdahl, J. E., & Marcus, B. A. (1998). Evaluation of a brief stimulus preference assessment. Journal of Applied Behavior Analysis, 31 605 620. Roscoe, E. M., Iwata, B. A., & Rand, M. S. (2003). Effects of reinforcer consumption and magnitude on response rates during noncontingent reinforcement. Journal of Applied Behavior Analysis, 36 525 539. Russell, H., & Oliver, C. (2003). The assessment of food related problems in individuals with Prader Willi syndro me. British Journal of Clinical Psychology, 42, 379 392. Schneider, J. W. (1973). Reinforcer effectiveness as a function of reinforcer rate and magnitude:

PAGE 72

72 A comparison of concurrent performances. Journal of the Experimental Analysis of Behavior, 20 461 471. Schrander Stumpel, C. T., Curfs, L. M., Sastrowijoto, P., Cassidy, S. B., Schrander, J. J., Fryns, J.P. (2004). Prader Willi syndrome: Causes of death in an international series of 27 cases. American Journal of Medical Genetics A, 124 333 338. Schweitzer, J. B., & Sulzer Azaroff, B. (1988). Self control: Teaching tolerance for delay in impulsive children. Journal of the Experimental Analysis of Behavior, 50, 173 186. Stebbins, W. C. (1962). Response latency as a function of amount of reinforcem ent. Journal of the Experimental Analysis of Behavior, 5 305 307. Stevenson, D. A., Heinemann, J., Angulo, M., Butler, M.G., Loker, J., Rupe, N., Kendell, P., Clericuzio, C.L., Scheimann, A. O. (2006). Deaths due to choking in Prader Willi syndrome. Ame rican Journal of Medical Genetics, 143A, 484 487. Taylor, R. L., & Caldwell, M. L. (1985). Type and strength of food preferences of individuals with Prader Willi syndrome. Journal Of Mental Deficiency Research, 29 109 112. Todorov, J. C. Hanna, E. S. & de SA, M. C. N. B. (1984). Frequency versus magnitude of reinforcement: New data with a different procedure. Journal of the Experimental Analysis of Behavior. 41 157 167. Tustin, R. D. (1994). Preference for reinforcers under varying schedule arrangem ents: A behavioral economic analysis. Journal of Applied Behavior Analysis, 27 597 606. Volkert, V. M., Lerman, D. C., & Vorndran, C. M. (2005). The effects of reinforcement magnitude on functional analysis outcomes. Journal of Applied Behavior Analysi s, 38 147 162. 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. Wharton, R.H., Wang, T., Graeme Cook F., Briggs, S., Cole, R.E. (1997). Acute idiopathic gastric dilation with gastric necrosis in individuals with Prader Willi syndrome. American Journal of Medical Genetics 73 437 41. Zipf, W.B., Berntson, G.G. (1987). Characteristics of abnormal food i ntake patterns in children with Prader Willi syndrome and study of effects of naloxone. American Journal of Clinical Nutrition, 46 277 81.

PAGE 73

73 BIOGRAPHICAL SKETCH Jessica Thomasons career in applied behavior analysis (ABA) began at the University of Florida, where she volunteered at the Florida Center on Self Injury under the supervision of Dr. Brian Iwata. After receiving her B.S. (1998), she accepted a position at the Kennedy Krieger Institute, where she coordinated the assessment and treatment of behavior disorders for inpatients diagnosed with developmental disabilities. In August of 2001, she began graduate school in behavior analysis at U niversity of Florida. While in graduate school, Je ssica conducted research on a variety of topics, including the assessment and treatment of problem behavior and reinforcer identification and assessment. She also provided clinical services in several dif ferent settings, including a school for individuals with developmental disabilities, an outpatient clinic for children diagnosed with autism, and a vocational and residential program for individuals diagnosed with developmental disabilities. Jessica served as the coordinator of two clinics (the school and autism clinics), as a teaching assistant for an introductory class and lab in ABA, and as instructor of a class in ABA. Jessica received her Ph.D. in May of 2007. Jessica hopes to continue on to a career in which she conducts clinical research on behavior disorders and skill acquisition.


Permanent Link: http://ufdc.ufl.edu/UFE0019203/00001

Material Information

Title: An evaluation of reinforcer dimensions influencing food selection of individuals diagnosed with Prader-Willi Syndrome
Physical Description: Mixed Material
Language: English
Creator: Thomason, Jessica L. ( Dissertant )
Iwata, Brian ( Thesis advisor )
Elder, Jennifer ( Reviewer )
Rowland, Neil ( Reviewer )
Vollmer, Timothy ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007
Copyright Date: 2007

Subjects

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

Notes

Abstract: Prader-Willi Syndrome (PWS) is a genetic disorder associated with a variety of problem behaviors, including hyperphagia and food stealing, as well as a predisposition toward morbid obesity. Some reports have suggested that the food preferences of individuals with PWS differ from those of individuals with other developmental disabilities. The current study compared the relative influence of reinforcer characteristics such as quality, magnitude, and the delay to delivery on choices made by individuals diagnosed with PWS and individuals diagnosed with other developmental disabilities. First, preference assessments were conducted to identify foods that were of “high quality.” Second, an assessment was conducted to identify the reinforcer characteristic (quality, magnitude, or delay) that was most influential in determining choices among concurrently-available vocational or academic tasks. Next, reinforcer characteristics were manipulated in an attempt to shift responding toward healthier food selections. For example, response allocation toward immediately-available reinforcers was shifted by gradually increasing delays to reinforcer delivery. Results are discussed in terms of (a) similarities and differences among determinants of food preference in individuals with and without PWS, and (b) implications for dietary management and food-related problem behaviors.
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 73 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0019203:00001

Permanent Link: http://ufdc.ufl.edu/UFE0019203/00001

Material Information

Title: An evaluation of reinforcer dimensions influencing food selection of individuals diagnosed with Prader-Willi Syndrome
Physical Description: Mixed Material
Language: English
Creator: Thomason, Jessica L. ( Dissertant )
Iwata, Brian ( Thesis advisor )
Elder, Jennifer ( Reviewer )
Rowland, Neil ( Reviewer )
Vollmer, Timothy ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007
Copyright Date: 2007

Subjects

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

Notes

Abstract: Prader-Willi Syndrome (PWS) is a genetic disorder associated with a variety of problem behaviors, including hyperphagia and food stealing, as well as a predisposition toward morbid obesity. Some reports have suggested that the food preferences of individuals with PWS differ from those of individuals with other developmental disabilities. The current study compared the relative influence of reinforcer characteristics such as quality, magnitude, and the delay to delivery on choices made by individuals diagnosed with PWS and individuals diagnosed with other developmental disabilities. First, preference assessments were conducted to identify foods that were of “high quality.” Second, an assessment was conducted to identify the reinforcer characteristic (quality, magnitude, or delay) that was most influential in determining choices among concurrently-available vocational or academic tasks. Next, reinforcer characteristics were manipulated in an attempt to shift responding toward healthier food selections. For example, response allocation toward immediately-available reinforcers was shifted by gradually increasing delays to reinforcer delivery. Results are discussed in terms of (a) similarities and differences among determinants of food preference in individuals with and without PWS, and (b) implications for dietary management and food-related problem behaviors.
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 73 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0019203:00001


This item has the following downloads:


Full Text





AN EVALUATION OF REINFORCER DIMENSIONS INFLUENCING FOOD
SELECTION OF INDIVIDUALS DIAGNOSED WITH PRADER-WILLI SYNDROME




















By

JESSICA L. THOMASON


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

2007

































O 2007 Jessica L. Thomason









ACKNOWLEDGMENTS

I would like to thank all of the individuals who encouraged my pursuit of higher education,

both personally and professionally. I thank my family and friends for supporting me throughout

school. I thank SungWoo Kahng for taking the time to nurture my interest in behavior analysis

and encouraging me to attend graduate school. I could not have completed this work without

the assistance of colleagues and research assistants who spent extra time and effort to help me

design and conduct this study, especially Claudia Dozier, Pamela Neidert, Zachariah Sims, and

Brooke Jones. I also appreciate the time and advice of my committee members, Jennifer Elder,

Neil Rowland, and Timothy Vollmer. I would like to express the deepest gratitude to my

advisor and committee chair Brian Iwata for his guidance and time, without which I would not

have been able to accomplish this and other work.












TABLE OF CONTENTS


page

ACKNOWLEDGMENT S ............. ......___ .............. 3....


LIST OF TABLES ............. ...... __ ..............5....


LIST OF FIGURES ............. ...... __ ..............6....


ABSTRACT ............. ...... __ ..............7.....


CHAPTER


1 INTRODUCTION ................. ..............9........... .....


2 GENERAL METHOD ................. ................. 17......... ...


Participants and Setting ................ ........... ................ 17....
Food Types Evaluated in Preference Assessment ................. ................. 17............
Response Measurement and Reliability ................. .......... ......... ..........1

3 EXPERIMENT 1 ................ ..............25. .......... ....


Preference Assessment Procedure ................. ..............26................
Results and Discussion ................ ................. 27.............


4 EXPERIMENT 2 ................ ..............48. .......... ....


Participants, Design and Procedure ................. ......... ..............49. ....
Results and Discussion.............. ............... 51


5 EXPERIMENT 3 ................. ................. 56......... ...


Participants, Design and Procedure ................. ......... ......... ...........5
Results and Discussion ................ ..............60. ..............


6 GENERAL DISCUS SION ................ ..............63. ..............


REFERENCE LIST ................. ..............68.......... ......


BIOGRAPHICAL SKETCH. ................ ..............73. ...............










LIST OF TABLES


Table page

2-1 Participant characteristics ................. ................. 20......... ...

2-2 Foods included in each preference assessment. ................ ..............22. .......... ..

2-3 Reliability coefficients. ................ ..............23. .......... ....

3-1 Results of between-category MSWO preference assessments for PWS group. ...............30

3-2 Results of between-category PS preference assessments for PWS group. ................... .... 36

3-3 Results of between-category MSWO preference assessments for control group............. 38

3-4 Results of between-category PS preference assessments for control group. ................... .42










LIST OF FIGURES


Figure page

3-1 Percent of times a food was selected in between-category MSWO preference
assessments for 7 PWS participants. ................ ..............43. .......... ...

3-2 Percent of times a food was selected in between-category MSWO preference
assessments for 6 PWS participants ................ ..............44. .......... ...

3-3 Percent of times a food was selected in between-category PS preference assessments
for 5 PW S participants. ................ ................. 45.............

3-4 Percent of times a food was selected in between-category MSWO preference
assessments for 6 control participants. ................ ..............46. .......... ...

3-5 Percent of times a food was selected in between-category PS preference assessments
for 6 control participants. ................ ................. 47.............

4-1 Percent of times allocated to each response option during baseline and assessment........ 55

5-1 Percent of times allocated to each response option during immediacy and quality
treatments ................ .............62..................









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

AN EVALUATION OF REINFORCER DIMENSIONS INFLUENCING FOOD
SELECTION OF INDIVIDUALS DIAGNOSED WITH PRADER-WILLI SYNDROME


By

Jessica L. Thomason

May 2007

Chair: Brian Iwata
Major: Psychology


Prader-Willi Syndrome (PWS) is a genetic disorder associated with a variety of problem

behaviors, including hyperphagia and food stealing, as well as a predisposition toward morbid

obesity. Some reports have suggested that the food preferences of individuals with PWS differ

from those of individuals with other developmental disabilities. The current study compared the

relative influence of reinforcer characteristics such as quality, magnitude, and the delay to

delivery on choices made by individuals diagnosed with PWS and individuals diagnosed with

other developmental disabilities. First, preference assessments were conducted to identify foods

that were of "high quality." Second, an assessment was conducted to identify the reinforcer

characteristic (quality, magnitude, or delay) that was most influential in determining choices

among concurrently-available vocational or academic tasks. Next, reinforcer characteristics were

manipulated in an attempt to shift responding toward healthier food selections. For example,

response allocation toward immediately-available reinforcers was shifted by gradually increasing

delays to reinforcer delivery. Results are discussed in terms of (a) similarities and differences









among determinants of food preference in individuals with and without PWS, and (b)

implications for dietary management and food-related problem behaviors.









CHAPTER 1
INTRODUCTION

A common definition of preference in research on learning is a choice from an array of

reinforcers. Quality, magnitude, and immediacy of delivery have been shown to be important

determinants of preference, in that modifications to one parameter result in an altered response

selection or allocation. This research focuses on determinants of food preference in individuals

diagnosed with Prader-Willi Syndrome and the extent to which preference can be modified

through manipulation of reinforcer characteristics.

Miller (1976) conducted an early investigation of the role of reinforcer quality as a

determinant of choice. He arranged concurrent Variable-Interval (VI) schedules in which

pigeons were presented with two concurrently available response options, each of which was

associated with access a particular type of grain (hemp, buckwheat, and wheat were available in

various paired combinations). Results showed that reinforcer quality (type of grain) influenced

response allocation. Matthews and Temple (1979) also used concurrent VI schedules to assess

dairy cows' preference (response allocation) for two types of feed and showed that the cows

exhibited a preference (response bias) for one type of feed over the other.

Results of basic research on preference have been extended in the applied literature, most

often with individuals who have severe developmental disabilities and, as a result, cannot readily

communicate their preferences. Pace, Ivancic, Edwards, Iwata, and Page (1985) described a

two-stage procedure that has become common in most preference-assessment research. In the

first stage, stimuli were presented singly while approach behavior was measured; in the second,

stimuli that were frequently (or infrequently) approached were delivered as consequences in an

operant-learning task. In a variation of this procedure, Fisher et al. (1992) presented stimuli in

all possible pairs, and preference was determined based on the number of times a stimulus was










selected given the number of times it was presented. In a subsequent phase, "preferred" and

"nonpreferred" items were delivered as consequences under a concurrent-schedule arrangement

in which each of two response options was associated with delivery of one of the items. Results

showed that 3 of 4 participants allocated most or all of their time to the response associated with

the preferred item. Results of these and other studies on preference and performance in the

developmentally disabled population (DeLeon et al., 2001; Graff, Gibson, & Galiatsatos, 2006;

Roane, Vollmer, Ringdahl, & Marcus, 1998) indicate that qualitative aspects of stimuli exert

control over choice.

Amount of reinforcement also can influence choice and has been manipulated through

changes in the rate of reinforcement (i.e., a schedule manipulation) or the magnitude (amount) of

each reinforcer delivered. A large volume of literature exists on the effects of various schedules

of reinforcement under concurrent-schedule conditions. In short, interval schedules tend to

produce response allocation among two alternatives that is roughly proportional to the amount of

reinforcement available on each alternative, whereas ratio schedules tend to produce a strong

bias in response allocation toward the denser schedule (Herrnstein, 1961).

Manipulations of magnitude in basic research have included changes in the number of food

reinforcers (Keesey & Kling, 1961; Reed, 1991; Schneider, 1973), the concentration of nutritive

content in a solution (Lowe, Davey, and Harzem, 1974; Stebbins, 1962), and the duration of

access to reinforcers (Belke, 1997; Catania, 1963; Davison & Baum, 2003; Keller & Gollub,

1977; Neuringer, 1967). Although the results of many studies are inconsistent with respect to the

influence of magnitude on responding (see Bonem & Crossman, 1988, and Reed, 1991, for

reviews), studies in which concurrent schedules were used have generally shown that animals









allocate responding to the option that produces the larger reinforcer (Catania, 1963; Neuringer,

1967; Keller & Gollub, 1977).

Relatively few applied studies have examined the influence of reinforcer magnitude on

responding. Results obtained when magnitude of reinforcement was manipulated for a single

response option have been mixed. Some data show that larger magnitude reinforcers delivered

contingent on a response are correlated with slightly higher rates of responding (Lerman, Kelley,

Vorndran, Kuhn, & LaRue, 2002), whereas other data sets show a reduction in overall rates of

responding (Volkert, Lerman, & Vorndran, 2005). When reinforcers are delivered

noncontingently, data show that larger magnitude reinforcers produce better response

suppression than smaller magnitude reinforcers (Roscoe, Iwata, & Rand, 2003). Only one

applied study, Hoch, McComas, Johnson, Faranda, and Guenther (2002), has examined the

effects of magnitude on response allocation. The dependent variable was the proportion of time

allocated to playing in one of two areas; one area contained a peer, and the second area did not.

When equal magnitudes (durations) of reinforcement (access to preferred items) were available

for both options, the participant spent all available time in the area of the room not containing the

peer. In a subsequent phase, a larger magnitude of reinforcement was available in the area of the

room with the peer versus the area without (90- and 15-s reinforcement durations, respectively);

results showed that the participant began to allocate more time to the area of the room with the

peer (and the larger magnitude of reinforcement).

Finally, basic research on delay to reinforcement has shown that subj ects select shorter

over longer delays. Chung (1965) and Chung and Herrnstein (1967) arranged concurrent VI VI

schedules in which the two response options resulted in differing delays to reinforcer delivery.

Results showed that the animals tended to select the response option associated with the shorter










delay to reinforcement. Few applied studies have examined the effects of reinforcement delay

alone on response allocation (although numerous other studies have compared the effects of

magnitude and delay; e.g., Dixon et al., 1998; Schweitzer and Sulzer-Azaroff, 1988; Vollmer,

Borrero, Lalli, & Daniel, 1999). Horner and Day (1991) examined the effects of a delay to

reinforcement (escape from tasks) on aggression and appropriate behavior. Escape was available

either 20s or 1s after the participant requested a break, whereas aggression always resulted in

immediate escape. Results showed that the participant allocated most responding to aggression

when appropriate requesting resulted in a 20-s delay to escape, whereas the opposite occurred

when requesting resulted in a 1-s delay to escape.

Quality of reinforcement, amount of reinforcement, and delay to reinforcement in isolation

influence behavior in predictable ways (subj ects allocate responding to higher-quality stimuli,

larger amounts of stimuli, and immediately available stimuli); however, the effects of

manipulating more than one variable simultaneously are more difficult to predict. Neef and

colleagues have conducted an elegant series of experiments in which they examine the influence

of several reinforcer and response-requirement parameters on response allocation (Neef, Bicard,

& Endo, 2001; Neef & Lutz, 2001; Neef, Mace, & Shade, 1993; Neef, Mace, Shea & Shade,

1992; Neef, Shade & Miller, 1994). In these studies, the authors used concurrent schedules to

compare the relative influences of reinforcement frequency, immediacy, quality, and response

effort on response allocation. For example, Neef et al. (2004) used this paradigm to provide an

empirical evaluation of one of the diagnostic criteria for Attention-Deficit Hyperactivity

Disorder (ADHD), impulsivity. They observed that reinforcer immediacy exerted the greatest

influence over participants' choices among work options and subsequently used this information

to develop interventions to decrease impulsive choice. The assessment method presented by Neef









et al. is an efficient way to isolate a number of variables that influence responding and may prove

useful in application with a number of disorders in addition to ADHD.

Although rare, some genetic disorders are correlated with specific behavioral

characteristics. One example is Prader-Willi Syndrome (PWS), which is diagnosed based on a

combination of chromosomal, physical, and behavioral traits (Holm, et al., 1993; Prader,

Labhart, & Willi, 1956). Perhaps the most marked behavioral characteristic of PWS is

hyperphagia, an insatiable appetite that typically results in morbid obesity. Intervention for this

population typically consists of environmental modifications, such as locking food cabinets and

refrigerators, and restricting daily caloric intake (Butler & Thompson, 2000). Individuals with

PWS have been reported to engage in a variety of problematic food-related behaviors, including

property destruction to access food (Benjamin & Buot-Smith, 1993), stealing (Donaldson, et al.,

1995), and consumption of items not intended for human ingestion (e.g., food from trash cans,

pet food, etc; Russell & Oliver, 2003). Assessment of the determinants of food preference may

be particularly useful in identifying potential treatments for food-related problems, such as

teaching individuals to wait for delayed reinforcers when presented with immediately available

ones, or by suggesting strategies that shift preference toward low-calorie foods in lieu of higher-

calorie foods.

Food preferences of individuals with PWS have been examined in several studies

(Caldwell & Taylor, 1983; Glover, Maltzman, & Williams, 1996; Joseph, Egli, Koppekin, &

Thompson, 2002; Rankin & Mattes, 1996; Taylor & Caldwell, 1985). Early results suggested

that preferences were somewhat uniform. For example, Caldwell and Taylor (1983) suggested

that individuals with PWS showed a pronounced preference for sweets. Results of more recent









studies have suggested preferences for other food characteristics (Glover et al., 1996; Rankin &

Mattes, 1996) or particular macronutrients (Fieldstone, Zipf, Schwartz, & Berntson, 1997).

Studies of the interaction of quality, magnitude, and delay may provide even more

information to be used in intervention. In a study by Glover et al. (1996), results of a series of

comparisons among high-preference foods, mixed-preference foods (an array of both low- and

moderate-preference foods), and low-preference foods showed that participants diagnosed with

PWS tended to make selections for larger magnitudes of mixed-preference food over smaller

magnitudes of high-preference foods, in contrast to control participants, who tended to select

higher-preference foods in smaller quantity over mixed-preference foods in larger quantity. In a

second experiment, results showed that all groups (PWS and control) selected a smaller

magnitude of a high-preference item over a larger magnitude of low-preference items. The

authors concluded that PWS participants had weaker taste preferences than the control

participants and that the relative "value" of an item given other choices available influenced

selection. A slightly different way to state these results is that magnitude appeared to be a more

influential variable than quality when items were ranked similarly (when quality was roughly

equal across options), but that quality was more influential than magnitude when the difference

among the qualities was highly disparate.

Joseph et al. (2002) conducted a similar study in which the choices of PWS participants

were compared to those of obese non-PWS participants. In the first study, participants chose

between one piece of food available immediately and 3 pieces of food available after a delay (15,

30, or 60s). Data showed that PWS participants were more likely to select the larger quantity

after a delay than the smaller quantity at no delay; obese comparison participants also tended to

select the larger, delayed quantity slightly more than the smaller, immediate quantity but to a









lesser extent than the PWS participants. In the second study, the same comparisons were made

across four foods high in fat and carbohydrate content. First, the experimenters identified high-

preference and low-preference foods by presenting an array of foods and instructing the

participants to select their favorite food. During subsequent sessions, the participants were

instructed to choose either a small magnitude reinforcer available immediately or a large

magnitude reinforcer available after a delay. This comparison (magnitude versus immediacy)

was repeated across each of the identified high- and low-preference foods. Results were similar

to the first experiment, in that PWS participants tended to select (a) the larger, delayed quantity

of food over the smaller, immediate option, and (b) the larger quantity more frequently than the

control participants. Data also showed that when the food available in session was a high-

preference food, the participants selected the larger magnitude of food (at a delay) to a greater

extent than when the food available was a low-preference food. The authors concluded that

individuals with PWS were more likely to make choices based on food magnitude.

Although the Glover et al. (1996) and Joseph et al. (2002) studies provide some evidence

of the variables influencing food selection, their presentation of aggregate data precludes

inspection of individual differences in responding. That is, it is possible that some participants'

choices were closer to being indifferent, whereas others' showed a moderate influence; thus,

general statements based on the group averages may not accurately represent the performance of

some individuals within the group. Further, although the authors conducted preference

assessments, very few foods were included (4 and 10 foods, respectively). Inclusion of a broader

array of foods would allow greater confidence to conclusions about relative preference among

items. Finally, neither study included all possible comparisons (immediacy versus magnitude,

magnitude versus quality, and immediacy versus quality).










The purpose of the current research is to provide a preliminary analysis of variables

influencing food choice in individuals with PWS and to increase healthy food choices when

choice-making appears maladaptive. The assessment methods used in the current study are

based on those of Neef et al. (2001), with two of the variables adapted to the current situation.

First, amount of reinforcement was manipulated by changing the number of reinforcers earned

per response (magnitude) rather than reinforcement schedule. Given that the current population

of participants was developmentally disabled (in many cases a diagnosis of mental retardation

was also present), it seemed that the visual cues inherent in a magnitude manipulation would

likely be more salient than those available in a schedule manipulation. Second, although the

series of studies by Neef and colleagues also assessed the effects of response effort, several of

their data sets show that effort has limited, if any, effects relative to the other variables tested.

Thus, the current experiment did not include an effort manipulation. The variables investigated

in the current study included immediacy of reinforcement delivery, magnitude of reinforcement,

and quality of reinforcement.









CHAPTER 2
GENERAL METHOD

Participants and Setting

Thirty individuals participated in Experiment 1, 11 individuals participated in Experiment

2, and 3 individuals participated in Experiment 3. Participants in Experiment 1 included 18

individuals diagnosed with PWS and 12 individuals diagnosed with mental retardation (MR)

and/or developmental disabilities other than PWS. Participants in Experiments 2 and 3 were

diagnosed with PWS (in addition to other diagnoses). Participant characteristics are listed in

Table 2-1. Sessions were conducted at either an adult vocational program or at a special

education school. Session rooms included at least one table and chair, materials necessary to

complete the target response, and reinforcers.

Food Types Evaluated in Preference Assessment

Six food groups (grain, dairy, meat, fruit, vegetable, and snack) were included, with 8

different foods per group. The foods included are listed in Table 2-2. The same foods were used

across all participants' assessments unless a participant was allergic to or intolerant of a food, in

which case that particular food was omitted from the assessment. Each group of foods was

compared separately (within-category assessment), such that a hierarchy of preference within

each category was generated. The top 3 choices from each food group were then included in a

larger (18-item) between-category preference assessment.

Response Measurement and Reliability

In Experiment 1, the target behavior consisted of reaching out and grasping a piece of

food. Foods were presented either in pairs (Paired Stimulus Preference Assessment; PS) or in a

larger array (Multiple Stimulus Without Replacement Preference Assessment; MSWO). In

Experiments 2 and 3, target behaviors consisted of simple vocational or academic responses. The










response for Participants JB, BK, AC, XD, KB, PM, MJ, DD, DV, and NR consisted of placing a

washer on a screw and fastening a nut onto the screw such that the washer and nut remained in

place. The response for Participant MW consisted of placing an index card in an envelope,

closing the envelope, and placing it in a manila envelope.

Trained observers recorded food selections (PS and MSWO) and the order in which they

were selected (MSWO) during Experiment 1. Data were collected on a form that listed all of the

foods (MSWO assessment) and all possible combinations of foods (PS assessment). Reliability

(interobserver agreement) for the preference assessments was assessed by having a second,

independent observer collect data with the primary observer during an average of 50% of

sessions (range, 0% to 100%). Observers' records were compared on a trial-by-trial basis, and

reliability was calculated by dividing the number of scoring agreements by the number of

agreements plus disagreements, and then multiplying by 100%.

In Experiments 2 and 3, trained observers recorded the frequency of correct responses and

incorrect responses, the duration of contact with session materials, and the delivery of reinforcers

on handheld PDAs. A correct response for the assembly task was defined as placing a washer

onto a bolt, attaching a nut to the bolt, and placing the assembled piece into a bin. An incorrect

response was scored for the assembly task if any of the components (washer, nut) were missing,

or if the pieces were detached upon placement in the bin. A correct response for the envelope-

stuffing task was defined as placing a card into an envelope, closing the envelope flap, and

placing the envelope into a manila folder. An incorrect response for that task was scored if any

components (card, envelope) were missing when an item was placed into the folder, or if the

envelope flap was not closed when the item was placed into the folder. Duration measures of

item contact were also recorded. The "on" key for response materials was pressed when a










participant touched a set of materials. The "off' key was pressed when either (a) the materials

for the other response were touched, or (b) the participant ceased contact with all materials for 3

consecutive seconds. Duration measures were calculated by examining the raw data stream and

determining the difference between the time (in seconds) that the "on" and "off' keys were

scored for each response. Durations for response 1 and response 2 were summed, respectively.

Reinforcer delivery was scored when a reinforcer was placed on a participant plate or into a bag.

A second observer independently recorded data during 32% (range, 22% to 57%) of the

sessions for each participant. Reliability for frequency data was calculated by dividing each

session into consecutive 10-s intervals, dividing the smaller number of responses scored in each

interval by the larger number, averaging the results of these fractions, and multiplying by 100%.

Reliability for response duration was calculated by dividing the number of intervals in which

there was an agreement (for either occurrence or nonoccurrence) by the total number of intervals

and multiplying by 100%. Reliability data for individual participants are listed in Table 2-3.










Table 2-1. Participant characteristics.


Participant

J.B.

K.B.

X.D.

M.J.

D.D.

D.W.

D.V.

P.M.

B.K.

A.H.

T.C.

A.C.

N.R.

M.W.

P.P.

P.B.

D.K.

A.R.

C.C.

B.C.

D.L.

J.C.


Age Diagnosis/sensory impairments

34 PWS, Mental retardation (level unspecified)

27 PWS

31 PWS, Mild mental retardation

25 PWS

29 PWS

37 PWS, Mild mental retardation
PWS, Mental retardation (level unspecified), Seizure
36 disorder

38 PWS, Mild mental retardation

36 PWS, Mental retardation (level unspecified)

37 PWS

30 PWS, Mental retardation (level unspecified)

18 PWS, Mental retardation (level unspecified)

31 PWS, Mental retardation (level unspecified)

23 PWS, Mild mental retardation

56 PWS, Mental retardation (level unspecified)

37 PWS, Mental retardation (level unspecified)

33 PWS, Mental retardation (level unspecified)

35 PWS

20 Trainable mentally handicapped*

48 Mental retardation (level unspecified)
Educable mentally handicapped, speech & language
19 impaired*
Educable mentally handicapped, speech & language
18 impaired*
Educable mentally handicapped, speech & language
18 impaired*

27 Mental retardation (level unspecified), Hydrocephalus

30 Mental retardation (level unspecified)


Experiment

1, 2

1, 2, 3

1, 2

1, 2

1, 2

1

1, 2

1,2

1, 2, 3

1

1

1, 2

1, 2, 3

1, 2

1

1

1

1

1

1

1

1

1

1

1


L.J.

J. Sm.

J. Sp.













Participant Age Diagnosis/sensory impairments Experiment
Mental retardation (level unspecified), Klinefelter's
M.T. 40 syndrome 1

E.An. 17 Autism, Mental retardation (level unspecified) 1

E.Al. 18 Down syndrome, Mental retardation (level unspecified) 1

M.W. 18 Trainable mentally handicapped* 1

E.Ad. 50 Mental retardation (level unspecified), Autism 1
*Indicates that the diagnosis available was one assigned by the school system


Table 2-1. Continued










Table 2-2. Foods included in each category of preference assessment

Category Foods
Jelly beans, TwizzlerTM, M&MTM, RaisinetTM, oatmeal cookie, chocolate chip
Snack cookie, potato chips, DoritosTM

Protein Chicken, turkey, roast beef, ham, sausage, bacon, pepperoni, tuna (canned)
Wheat bread, white bread, rice cake, oatmeal, CheeriosTM, corn flakes, Triscuit
Grain crackerTM, saltine cracker
Vanilla yogurt, cottage cheese, cheddar cheese, mozzarella cheese, cream cheese,
Dairy sour cream, skim milk, vanilla pudding

Vegetable Broccoli, tomato, celery, carrot, cucumber, lettuce, squash, green bell pepper
Fruit Orange, apple, banana, grapefruit, pear, grapes, pineapple, cantaloupe










Table 2-3. Reliability coeffieients
Percentage of trials/
Experiment Participant sessions with reliability
1 J.B. 68.5

A.R. 24

D.V. 33.3

P.M. 46

X.D. 42

A.H. 55

T.C. 53

D.W. 66.7

B.K. 65

M.J. 33.3

K.B. 33.3

D.D. 87

A.C. 82.8

N.R. 65.1

M.W. 36.6

D.K. 44.4

P.P. 26.1

P.B. 65.1

C.C. 54.5

B.C. 37.6

D.L. 58.6

E.An. 41.4

L.J. 45.5

M.T. 78.8

E.Al. 68.7

J.C. 78.8

M.W. 100


Mean Reliability
percentage
98.9

100

100

100

100

100

100

98

100

100

100

100

100

100

100

100

100

99.3

100

100

100

100

100

100

100

100

100


Range
92.6 10

N/A

N/A

N/A

N/A

N/A

N/A

91.6 -10

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

98 100

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A










Table 2-3. Continued


Percentage of sessions/
trials with reliability

26.1

53.5

82.6

22

22

40

25

30

40

11

33

33

40

45

37

57

30


Mean reliability
percentage

98.8

100

99.1

97.7

97.3

98.8

97.6

97.1

95.9

97.3

96.9

98.8

96.0

97.9

98.0

96.6

98.6


Range


Experiment

1





2


Participant

E.Ad.

J. Sm.

J. Sp.

D.D.

K.B.

N.R.

M.J.

X.D.

P.M.

B.K.

D.V.

M.W.

A.C.

J.B.

N.R.

B.K.

A.C.


96.4

N/A

96.4

81.7

84.2

88.3

93.3

86.7

53.3

85.8

79.2

89 -

42.5

83.3

78.3

30 -

86.7


- 100



- 100

- 100

- 100

- 100

- 100

- 100

- 100

- 100

- 100

100

- 100

- 100

- 100

100

- 100









CHAPTER 3
EXPERIMENT 1: ASSESSMENT OF FOOD PREFERENCES

Food preferences in the PWS population have been examined in several studies (Caldwell

& Taylor, 1993; Fieldstone et al., 1997; Glover, Maltzman, & Williams, 1996; Rankin & Mattes,

1996; Taylor & Caldwell, 1985). Research on qualitative aspects of food preference has

included taste (sweet, salty, etc.; Caldwell & Taylor, 1985), macronutrient content (fat,

carbohydrate, etc.; Fieldstone et al., 1997), and familiarity of foods (Rankin & Mattes, 1996).

Some results have shown group preferences for these qualitative variables (e.g., a preference for

sweet foods; Caldwell & Taylor, 1983; Hinton, Holland, Gellatly, Soni, & Owen, 2006), whereas

others report idiosyncratic preferences (e.g., Fieldstone et al., 1997). However, methodological

features of several of these studies make data difficult to interpret. First, most of the studies

presented aggregated data, which precluded analysis of individual sets of data. That is, it is

unclear if the group average was an accurate representation of the responding of each individual

within the group. Second, most of the studies assessed a very small number of stimuli (typically

4 or 9 items), which may limit the generality of their data. For example, it is not clear if a

preference for the one sweet item from among 4 other items indicated a general preference for

sweet foods versus a preference for that specific food relative to the other three foods included.

Finally, some of these studies relied on verbal report (e.g., Hinton et al, 2006) rather than a food

selection and consumption response. Data from Taylor and Caldwell (1985) and Glover et al.

(1996) raise questions about the correspondence between verbal report and actual choices of

PWS individuals during preference assessments; thus, it is unclear to what extent the verbal

report of preference is an actual indication of what an individual with PWS would choose to eat.

The purpose of Experiment 1 was to determine what, if any, food preferences individuals

with PWS exhibit. A secondary purpose was to determine if there were general characteristics of










the preferences exhibited by the group (e.g., a preference for high-calorie foods or for a

particular food type). Finally, data were examined to determine if there were any strong

differences in the preferences across groups.

Preference Assessment Procedures

Preference assessments consisted of an experimenter presenting small samples of two or

more foods and prompting the participant to select one. The serving size of each piece or portion

of food was approximately 1 teaspoon during all sessions in Experiment 1. In Experiments 2 and

3, the portion size varied across conditions (ranging from 1 teaspoon to 15 teaspoons of food).

All participants were exposed to the Multiple Stimulus without Replacement (MSWO) format

(DeLeon & Iwata, 1996). If selections during that assessment suggested that participants were

selecting based on stimulus position (e.g., selecting all items from the right-hand side and

progressing to the left), a Paired Stimulus Assessment (PS) was substituted (Fisher et al., 1992).

During the MSWO assessment, all foods included in a particular assessment were

presented in a grouped array. Participants were prompted to select their most-preferred item and

consume it, after which the array (minus the selected item) was rearranged and presented again.

This procedure was continued until either (a) all items were consumed or (b) the participant did

not make further selections. An assessment consisted of 3 repeated presentations of the entire

array, and a separate assessment was conducted for each food group. The top 3 choices from

each food group were then included in a large (18-item) between-category MSWO assessment.

During the PS assessment, foods within a food group (see above) were presented in pairs,

one pair per trial. Each trial consisted of two stimuli (foods) placed in front of the participant;

approaches to one of the stimuli resulted in delivery of the food sample and removal of the other.

Each item was presented in a pair with every other item until all possible pairs had been










presented. The top 3 choices from each food group were then included in a large (18-item)

between-category PS assessment.

Results and Discussion

Tables 3-1 and 3-2 list the results of the between-category preference assessments for the

PWS participants, and Tables 3-3 and 3-4 list the results of the between-category preference

assessments for the non-PWS participants. Results from the within-category assessments are not

presented because those assessments were used only as a method of identifying high-preference

foods to be compared in the between-category assessment. The tables list the top 3 items from

each category that were included in the between-category assessment, the percent of times those

foods were selected in the between-category assessment, the rank order of each food based on

the percentage of times selected, and the rank for each category of food for each participant.

Based on the category rankings, 9 out of 18 individuals in the PWS group tended to show

a preference for the snack food category over other categories. Four individuals showed a

preference for the protein category foods, 2 showed a preference for the fruit category foods, 1

showed a preference for the grain category foods, 1 showed a preference for the dairy category

foods, and no individuals showed a preference for vegetable category. One individual's data

showed that grain and protein category foods were ranked equally high. In the non-PWS group, 4

out of 12 individuals showed a preference for the protein group, 3 showed a preference for the

dairy group, 2 showed a preference for the vegetable group, 2 showed a preference for the fruit

group, and 1 showed a preference for the grain group. No control participants preferred the

snack foods to other food groups.

Figures 3-1, 3-2, and 3-3 graphically depict the results of the between-category

preference assessments for the PWS participants, and Figures 3-4 and 3-5 depict the results of

the between-category preference assessments for the non-PWS participants. Results show that










most individuals showed preferences for specific foods, although a few individuals' preferences

appeared to be rather weak.

Figure 3-1 depicts the results for 7 PWS participants who were exposed to the MSWO

preference assessment. Participant KB's results show a strong preference for 3 foods, all of

which were snacks; these data suggest that KB not only showed preferences for a few specific

foods, but also a general preference for snack foods. Data for DV show a similar pattern: Both

fruits and grains appeared to be more preferred than other types of foods. Data for PM, DK, AR,

BK, and DW show strong preferences for particular foods but not a preference for foods within a

category over other categories. Figure 3-2 depicts the data for the other 6 PWS participants

exposed to the MSWO preference assessment. AC, AH, MJ, XD, and TC showed weaker

preferences than did participants whose data were shown in Figure 3-1, although some

preferences for individual items can be identified. Data for DD show a fairly strong preference

for one item from the protein group (sausage) and a general preference for snack foods. Figure

3-3 depicts the data for the 5 PWS participants who were exposed to the PS preference

assessment. JB showed strong preference for all of the snack foods. Participants PB, PP, NR, and

MW all appeared to prefer certain foods but did not show general preferences for a particular

category of foods. Participant MW' s data are interesting in that, although he did not appear to

prefer any one category of foods, he did appear to select away from one category of foods

(vegetables).

Figure 3-4 depicts the data for 6 of the non-PWS participants exposed to the MSWO. Data

for MT, CC, BC, and MW all show evidence of preference for particular categories in addition to

some isolated items. MT showed a preference for dairy foods, CC and MW showed a preference

for protein foods, and BC showed a preference for vegetables. Data for EAl and JC show










preferences for a few items, although the preference appears somewhat weaker, and there does

not appear to be a general category preference. The top four panels of Figure 3-5 depict the

results for the other 4 non-PWS participants exposed to the MSWO assessment. Data for LJ,

EAn, DL, and JSm show weaker patterns of preference: A few items were selected on a larger

percentage of trials than others, but the differences in the rankings are slight compared to those

of other participants (e.g., MT or CC). Data for the 2 non-PWS participants who were exposed to

the PS assessment are depicted in the bottom two panels of Figure 3-5. Participants SP and EAd

both showed preferences for particular items, but neither showed a general preference for one

category of foods over all others.

Overall, the data show a large amount of variation in preferences across participants; in

addition, there was no striking difference in the preferences of PWS versus non-PWS

participants. Among those who showed strong preferences, the highly-preferred items tended to

be relatively high-calorie for both PWS and non-PWS participants, and very few people showed

preferences for extremely-low calorie foods (vegetables and fruits). The mean rankings showed

some differences across groups in what categories were (on average) most preferred. In the PWS

group, the snack category was most often the highest-ranked category (50% of individuals

showed the highest ranking for the snack category), whereas the protein and dairy groups were

ranked highest among the non-PWS group. However, the snack food category didn't necessarily

contain foods that were higher-calorie than all other categories; the dairy and protein categories

contained some of the most calorie-dense foods (e.g., sausage, cheddar cheese). Thus, it is not

the case that the differences observed represented preferences on the part of the PWS participants

based simply on caloric value.











Table 3-1. Results of between-category MSWO preference assessments for PWS group

Participant Category Food Percent selected Rank Category rank
KB Snack Twizzler 33.3 3 1
Raisinet 75 1
Jelly bean 60 2
Grain Triscuit 17.7 5.5 6
Cheerio 6.8 14.5
Rice cake 5.6 18
Dainy Cream cheese 12.5 8 2
Pudding 13.6 7
Cheddar cheese 11.5 9
Protein Sausage 23.0 4 3
Pepperoni 17.7 5.5
Roast beef 7.7 13
Fruit Pear 10.3 5 5
Pineapple 8.3 12
Banana 9.4 11
Vegetable Bell pepper 6.8 14.5 4
Tomato 6.4 16
Cucumber 5.9 17
PM Snack Raisinet 33.3 1 2
Twizzler 13.0 5.5
Chocolate chip cookie 7 16.5
Grain Triscuit 11.5 9 3
Corn flakes 10.7 10.5
Oatmeal 13.0 5.5
Dainy Cheddar 17.7 4 5
Mozzarella 7.3 13.5
Pudding 10.7 10.5
Protein Sausage 27.3 2.5 1
Pepperoni 10 12
Chicken 12.5 7
Fruit Orange 27.3 2.5 4
Banana 12 8
Pear 7 16.5
Vegetable Cucumber 7.3 13.5 6
Squash 7.1 15
Tomato 6.3 18
DK Snack M&M 10.7 11.5 5
Potato Chip 6.8 15
Doritos 11.1 9
Grain Triscuit 21.5 2 1
Cheerio 10.7 11.5
Oatmeal 42.8 1
Dairy Cottage cheese 10.3 13 4
Sour cream 3.7 18
Mozzarella 18.8 3.5













Participant Category Food Percent selected Rank Category rank
DK Protein Chicken 16.6 5 3
Continued Bacon 14.2 6
Tuna 6.3 16
Fruit Grapes 13 7 2
Cantaloupe 11.1 9
Banana 18.8 3.5
Vegetable Squash 11.1 9 6
Bell pepper 7.6 14
Tomato 6.1 17
AR Snack Oatmeal cookie 16.7 6.5 1
Chocolate chip cookie 33.3 2
Raisinet 23 3.5
Grain Oatmeal 6.5 17 4
Wheat bread 15 8
Cheerios 13 9
Dainy Cheddar 10 13 6
Mozzarella 10.7 12
Yogurt 11 11
Protein Sausage 23 3.5 2
Roast beef 8.8 14
Chicken 21.4 5
Fruit Banana 16.7 6.5 3
Pineapple 11.5 10
Grapes 6.9 15
Vegetable Broccoli 100 1 5
Celeny 6.2 18
Bell pepper 6.8 16
BK Snack Jelly bean 6.5 18 6
Chocolate chip cookie 10.7 8
M&M 6.7 17
Grain Oatmeal 9.4 11.5 5
White bread 7.1 16
Cheerios 8.3 13.5
Dainy Mozzarella 13.0 4.5 2
Cheddar 30 2
Pudding 8.3 13.5
Protein Pepperoni 12.5 6.5 1
Sausage 60 1
Turkey 23.1 3
Fruit Grapes 9.4 11.5 4
Orange 13.0 4.5
Apple 8.1 15
Vegetable Broccoli 9.7 10 3
Bell pepper 12.5 6.5
Lettuce 10.3 9


Table 3-1. Continued













Participant Category Food Percent selected Rank Category rank
DV Snack Chocolate chip cookie 20 7 3
Twizzler 10.3 9.5
M &M 21.4 6
Grain Triscuit 12.5 8 2
Corn flakes 75 1
Oatmeal 25 4.5
Dairy Yogurt 6 17 6
Pudding 6.1 15
Cottage cheese 6 17
Protein Tuna 10.3 9.5 4
Sausage 8.6 12
Chicken 9.4 11
Fruit Pineapple 27.3 3 1
Apple 25 4.5
Orange 30 2
Vegetable Carrot 7.5 13 5
Tomato 6.3 14
Cucumber 6 17
DW Snack Chocolate chip cookie 12 7 1
M&M 18 4
Raisinet 75 1
Grain Wheat bread 12 7 5
Cheerios 1 18
Rice cake 8 13
Dairy Cream cheese 7 15 6
Pudding 6 16.5
Mozzarella 12 7
Protein Roast beef 10 9.5 3
Sausage 60 2
Pepperoni 6 16.5
Fruit Pineapple 9 11 2
Pear 20 3
Banana 8 13
Vegetable Tomato 10 9.5 4
Bell pepper 16 5
Squash 8 13
AC Snack Twizzler 12.5 6.5 1
Jelly bean 11.1 10
Oatmeal cookie 23.1 1
Grain Corn flakes 14.3 5 5
Triscuit 7.1 15
Rice cake 7 16
Dairy Milk 10.7 11.5 3
Mozzarella 20 3
Cheddar 12 8.5


Table 3-1. Continued













Participant Category Food Percent selected Rank Category rank
AC Protein Sausage 12 8.5 2
Continued Chicken 10.7 11.5
Pepperoni 21.4 2
Fruit Banana 12.5 6.5 4
Apple 15.8 4
Orange 9.7 13
Vegetable Carrot 6.5 17.5 6
Tomato 6.5 17.5
Broccoli 8.1 14
AH Snack Jelly bean 7.5 15 6
Twizzler 7.5 15
Raisinet 7.5 15
Grain Rice cake 9.4 11 4
Cheerio 8.8 13
Oatmeal 15 6
Dairy Cottage cheese 17.7 4.5 2
Pudding 33.3 1
Yogurt 10 10
Protein Chicken 20 3 1
Tuna 23.1 2
Roast beef 17.7 4.5
Fruit Grapefruit 9.1 12 3
Cantaloupe 13.04 8
Grape 13.7 7
Vegetable Carrot 10.3 9 5
Celery 6.5 17
Cucumber 6.4 18
MJ Snack Jelly bean 12 7 5
Licorice 10 12
Raisinet 7.7 16
Grain Rice cake 13.6 5 3
Cheerio 11.1 10
Oatmeal 9.6 13
Dairy Cottage cheese 15 4 1
Pudding 15.8 3
Yogurt 16.7 1.5
Protein Chicken 13 6 4
Tuna 7.9 15
Roast beef 11.5 8.5
Fruit Grapefruit 16.7 1.5 2
Cantaloupe 10.3 11
Grape 11.5 8.5
Vegetable Carrot 8.3 14 6
Celery 7.5 17
Cucumber 6.5 18


Table 3-1. Continued














Participant Category Food Percent selected Rank Category rank
XD Snack Potato chip 10.7 9.5 4
Jelly bean 12 7
Chocolate chip cookie 6.7 18
Grain Oatmeal 9.7 12 6
Corn flakes 7. 1 16
Wheat bread 6.8 17
Dairy Sour cream 11.5 8 2
Cream cheese 33.3 1
Milk 13.6 6
Protein Roast beef 8.3 14 5
Pepperoni 7.3 15
Bacon 9.1 13
Fruit Apple 10.7 9.5 1
Grapefruit 21.4 2
Grape 16.7 3
Vegetable Broccoli 14.3 5 3
Squash 15 4
Tomato 10 11
TC Snack Twizzler 9.1 11 6
Raisinet 9.1 11
Oatmeal cookie 8.6 14
Grain Wheat bread 33.3 1 1.5
Oatmeal 14.3 5
Triscuit 7.9 16
Dairy Cottage cheese 16.7 4 4.5
Sour cream 9.1 11
Milk 8.1 15
Protein Chicken 6.7 18 1.5
Bacon 17.7 2
Tuna 17.7 2
Fruit Pineapple 10 9 3
Grapefruit 10.3 8
Grapes 9.7 10
Vegetable Tomato 7.7 17 4.5
Cucumber 13.6 6
Bell pepper 11.5 7
DD Snack Chocolate chip cookie 27.3 3 1
Oatmeal cookie 25 4.5
Raisinet 30 2
Grain Triscuit 13.6 7 3
Oatmeal 10.3 9.5
Wheat bread 21.4 6
Dairy Cottage cheese 10.3 9.5 4
Pudding 8.6 12
Sour cream 9.4 11


Table 3-1. Continued

















Sausage 75 1
Fruit Banana 7.5 13 5
Orange 6.3 15
Grapes 7.1 14
Vegetable Lettuce 6 17.5 6
Broccoli 6.1 16
Tomato 6 17.5


Table 3-1. Continued


Participant
DD
Continued


Category Food
Protein Tuna
Chicken


Percent selected
25
12.5


Rank
4.5
8


Category rank










Table 3-2. Between-category paired stimulus preference assessments for PWS group

Participant Category Food Percent selected Rank Category rank
PB Snack Twizzler 71 3 1
Oatmeal cookie 59 7
Raisinet 59 7
Grain Oatmeal 71 3 5
Corn flakes 41 13
Rice cakes 24 16
Dairy Yogurt 65 5 3.5
Pudding 47 10.5
Mozzarella 47 10.5
Protein Sausage 82 1 3.5
Tuna 35 14.5
Bacon 47 10.5
Fruit Pear 47 10.5 2
Grapefruit 71 3
Banana 59 7
Vegetable Tomato 23 17 6
Squash 18 18
Celeny 35 14.5
PP Snack Oatmeal cookie 82 2 3
Raisinet 53 9
M&M 58.8 7
Grain Triscuit 41 11 5
Saltine 41 11
Oatmeal 35 14
Dainy Cream cheese 70.5 3 2
Cottage cheese 58.8 7
Pudding 64.7 6
Protein Roast beef 82.3 1 1
Ham 47 10
Tuna 70.5 3
Fruit Pineapple 35 14 4
Pear 23.5 17
Grape 70.5 3
Vegetable Celery 0 18 6
Tomato 41 11
Squash 29.4 16
JB Snack Twizzler 82 5 1
Jelly bean 94 1.5
Raisinet 94 1.5
Grain Triscuit 24 14.5 5.5
Corn flake 24 14.5
Cheerio 12 17
Dainy Mozzarella 71 6 2.5
Cheddar 88 3.5
Pudding 41 10













Participant Category Food Percent selected Rank Category rank
JB Protein Sausage 59 7 2.5
Continued Pepperoni 88 3.5
Roast beef 47 9
Fruit Pineapple 53 8 4
Orange 29 12
Banana 35 11
Vegetable Cucumber 12 17 5.5
Bell pepper 24 14.5
Squash 24 14.5
NR Snack Raisinet 70.6 3 1
Jelly beans 76.5 2
M&M 64.7 4.5
Grain Wheat bread 58.8 6.5 5
White bread 29.4 17
Oatmeal 35.3 14.5
Dairy Mozzarella 58.8 6.5 2
Cheddar 64.7 4.5
Pudding 52.9 8
Protein Bacon 35.3 14.5 3
Chicken 41.2 11.5
Sausage 82.4 1
Fruit Pineapple 41.2 11.5 4
Cantaloupe 47.1 9
Grapefruit 41.2 11.5
Vegetable Cucumber 29.4 17 6
Bell pepper 41.2 11.5
Squash 29.4 17
MW Snack Raisinet 82.4 3 1
Jelly bean 88.2 1.5
Chocolate chip cookie 64.7 6.5
Grain Triscuit 58.8 9 4
Saltine 17.7 15.5
Oatmeal 58.8 9
Dairy Yogurt 64.7 6.5 2
Cream cheese 88.2 1.5
Pudding 76.5 4.5
Protein Roast beef 47.1 11.5 3
Sausage 76.5 4.5
Ham 47.1 11.5
Fruit Orange 23.5 13.5 5
Pineapple 58.8 9
Pear 23.5 13.5
Vegetable Cucumber 5.9 17 6
Carrot 17.7 15.5
Tomato 0 18


Table 3-2. Continued










Table 3-3. Results of between-category MSWO preference assessments for control group

Participant Category Food Percent selected Rank Category rank
EAl Snack Raisinet 20 4.5 4
Oatmeal raisin cookie 10.7 9
Twizzler 9.7 13
Grain Saltine 15 6 3
White bread 8.8 15
Oatmeal 20 4.5
Dainy Cheddar 37.5 1.5 1
Cottage cheese 37.5 1.5
Cream cheese 9.4 14
Protein Tuna 23.1 3 2
Chicken 10 11
Ham 13.6 8
Fruit Grape 14.3 7 5
Grapefruit 10 11
Orange 6.4 18
Vegetable Lettuce 7.9 16 6
Cucumber 6.8 17
Bell pepper 10 11
MT Snack Twizzler 9.4 9.5 3
Oatmeal cookie 6.5 15.5
Potato chips 16.7 6
Grain Oatmeal 12.5 8 4
Corn flakes 9.4 9.5
Triscuit 6.8 14
Dairy Yogurt 21.4 4 1
Cottage cheese 27.3 3
Milk 33.3 2
Protein Tuna 100 1 2
Chicken 13 7
Sausage 20 5
Fruit Pineapple 6.1 17 6
Cantaloupe 8.6 11
Banana 6.5 15.5
Vegetable Bell pepper 8.3 12 5
Cucumber 7.1 13
Celeny 6 18
CC Snack Chocolate chip cookie 13.6 5 2
Potato chip 13.6 5
M&M 7.5 14
Grain Wheat bread 7.5 14 6
White bread 8.8 12
Corn flakes 5.9 18
Dainy Cheddar 13.6 5 4
Mozzarella 11.5 9
Milk 6.1 16













Participant Category Food Percent selected Rank Category rank
CC Protein Bacon 25 3 1
Continued Sausage 60 1
Ham 60 1
Fruit Pear 7.9 13 5
Banana 11.5 9
Orange 9.4 11
Vegetable Tomato 13 8 3
Carrot 15 4
Squash 6.1 16
BC Snack Doritos 9 11.5 4
Potato chips 10 9.5
Chocolate chip cookie 8 13.5
Grain Oatmeal 20 6 2
Wheat bread 23 5
Rice cake 14 7.5
Dairy N/A
N/A
N/A
Protein Tuna 25 3.5 3
Roast beef 14 7.5
Pepperoni 10 9.5
Fruit Grapefruit 9 11.5 5
Cantaloupe 8 13.5
Banana 7 15
Vegetable Cucumber 25 3.5 1
Celery 43 1
Squash 33 2
JC Snack Potato chip 12 9 3
Oatmeal cookie 8.8 14.5
Raisinet 13.6 5
Grain Corn flakes 8.8 14.5 2
Saltine 27 2
Triscuit 12.5 7.5
Dairy Milk 12.5 7.5 6
Cream cheese 11 10
Pudding 6 18
Protein Tuna 6.4 17 5
Ham 9.7 12
Roast beef 23 3
Fruit Apple 7 16 1
Banana 15.8 4
Grapes 33 1
Vegetable Broccoli 9.1 13 4
Celery 10 11
Bell pepper 13 6


Table 3-3. Continued













Participant Category Food Percent selected Rank Category rank
MW Snack Raisinet 7 15 5
Potato chip 16.7 5.5
Doritos 6.8 16
Grain Rice cake 14.3 8 4
Saltine 5.9 17
Triscuit 8.3 11
Dairy Cheddar 37.5 3 3
Cottage Cheese 7.3 14
Cream cheese 7.7 13
Protein Roast beef 42.9 1.5 1
Sausage 42.9 1.5
Ham 23.1 4
Fruit Pineapple 15.8 7 2
Grapes 16.7 5.5
Orange 10.3 9.5
Vegetable Celery 10.3 9.5 6
Cucumber 7.9 12
Carrot 5.7 18
L J Snack Chocolate chip cookie 9 14 4
Oatmeal cookie 10.7 10
M &M 15 4.5
Grain Saltine 8.6 15 2
Triscuit 12 7
Rice cake 18.6 1
Dairy Cheddar 10 12 6
Yogurt 7.9 16
Pudding 6.7 18
Protein Chicken 10.7 10 5
Tuna 11 8
Bacon 6.8 17
Fruit Grape 15.8 2.5 1
Orange 15.8 2.5
Banana 13 6
Vegetable Celery 9.3 13 3
Broccoli 15 4.5
Cucumber 10.7 10
EA Snack Jelly bean 16 3 2
Potato chips 11 9.5
Oatmeal cookie 11 9.5
Grain Triscuit 8 15 5
White bread 8 15
Oatmeal 12 7
Dairy Mozzarella 7 17.5 6
Yogurt 12 7
Vanilla pudding 8 15


Table 3-3. Continued













Participant Category Food Percent selected Rank Category rank
EA Protein Sausage 12 7 3
Continued Roast beef 9 12.5
Bacon 10 11
Fruit Pineapple 9 12.5 4
Grapefruit 7 17.5
Grape 13 4.5
Vegetable Celery 13 4.5 1
Squash 18 2
Broccoli 21 1
DL Snack Raisinet 8.3 14 5
Jelly Bean 7.9 15
Doritos 12.5 8
Grain Triscuit 8.8 13 3
Rice Cake 10.7 9
Saltine 15 5
Dairy Yogurt 13.6 6 2
Pudding 6.3 18
Mozzarella 18.8 2
Protein Roast Beef 13 7 1
Sausage 23.1 1
Turkey 18.8 2
Fruit Banana 10.7 9 4
Grape 17.6 4
Apple 7.9 15
Vegetable Celery 10.3 9.5 6
Cucumber 7.9 12
Carrot 5.7 18
JSm Snack Twizzler 16.7 7.5 5
Raisinet 9.7 15
Doritos 7.3 18
Grain Rice cake 18.8 4 1
Saltine 11.5 10.5
Triscuit 21.4 2.5
Dairy Cheddar 16.7 7.5 6
Mozzarella 27.3 1
Milk 10 14
Protein Tuna 9.1 16 3
Bacon 10.3 13
Turkey 21.4 2.5
Fruit Grapes 11.5 10.5 4
Pineapple 7.7 17
Banana 16.7 7.5
Vegetable Carrot 17.7 5 2
Squash 16.7 7.5
Broccoli 10.7 12


Table 3-3. Continued










Table 3-4. Results of between-category paired stimulus preference assessments for control group

Participant Category Food Percent selected Rank Category rank
EAd Snack Raisinet 53.9 9.5 3
Potato Chip 56 8
Doritos 58.8 6.5
Grain Saltine 47 11.5 4.5
Triscuit 53.9 9.5
White bread 47 11.5
Dairy Cheddar 76.4 2 2
Pudding 58.8 6.5
Mozzarella 64.7 5
Protein Chicken 70.5 3.5 1
Sausage 82.3 1
Pepperoni 70.6 3.5
Fruit Grapefruit 35.2 13 6
Apple 35.7 14
Banana 29.4 15.5
Vegetable Celery 11.7 18 4.5
Carrot 29.4 15.5
Bell pepper 17.6 17
JSp Snack Doritos 81 2 4
Potato Chips 50 9.5
Oatmeal Cookie 38 12.5
Grain Oatmeal 63 5 5
Triscuit 38 12.5
Rice Cake 56 7
Dairy Pudding 69 4 1
Yogurt 56 7
N/A
Protein Sausage 88 1 2
Chicken 56 7
Bacon 31 14
Fruit Orange 50 9.5 3
Apple 75 3
Pineapple 50 11
Vegetable Bell Pepper 19 16 6
Cucumber 6 17
Celery 25 15












PM








m,~,,, o
~~~ST. =:~a
a ~N m 88"
.~ 3 8 e ~ 9 E
c;i O N r~
CL: ~ e, ~ EOUN de ya a
98 c~


AR






nnnnn,,,nnn,,


100_



50-


V


u STIMULI

Figure 3-1. Percent of times a food was selected in between-category MSWO preference
assessments for 7 PWS participants (Experiment 1).


691 "g
o ~e j ,
'j: o o m
~ ~~ ~~
"ru
CL: 3
8
u


OT~n



BK m~~~g




oS3ce 9~
100 g
e 8 DW
O eo c~ M75- M
c~ClrUO,50 -
n 25-


~FI"~"~"~~m"~-~~
O
eg e ~ oa,,~ ~E o m
EO e,~ ~ ~M
c;i "ohl~ ~ O O
OC~iDUN r~ U M
$U 3 O g
CL:
L
O


n In,,,,,,nnn,,,


m ~ m ~ 5 ~a ~ ~ ~ g ~ ~ ~b 6 B b
~~~-a~s~r:j-e$e~
8'5 ~a e, b0 Y
C~reO uo
gr o ~b a u
u a
% u


,nnnnnnnnn
















AH bYc2 W





a co










DD


100~

S751

S501





- 25





50 ~
7-1

25
2-1


100-


a~~ 0


STIMULI

Figure 3-2. Percent of times a food was selected in between-category MSWO preference
assessments for 6 PWS participants (Experiment 1).


nnnn,,nnnnnnnnn,,n
= $ ~ B ~a W ~ L c~ e
cce ym
;3c O m ~ ~B
N O m Ye ~ $ E~
m WbrUo e
'~30 ." Er- mo "m O r
a~ CL: U& m
,~s


hnnnnnnnnnnnn,,,
ccce
W 6 "~ ~ ~ s Y oa o W ~
o~
~-"~cb"~'r.c'~ c
O
ra e0 L,
gL U r~0 c~
~F U a c
E~ u m


nnnnnnnnnnnnnnnnnn
~~Y 0~~~~0~~
$sc~~~, o as
"6%~~ E
M c t~O
e ~"Y &
9~ I li C~
c~ ro
3
r ra O *
CL~O O
6 u


aaa~a~aa
""~~~~
a E,,
8?;a-c" ~~~r
c CL: c U
9 ~ m
E ~
ea
o u

























































~8 0B~~";


STIMULI

Figure 3-3. Percent of times a food was selected in between-category Paired stimulus preference
assessments for 5 PWS participants (Experiment 1).


TTT T TTTT




O U


100


4~pr
'~ a 's
~eu
o
u


I I


ool M


m~ c~ 0~
alt~~'~~~?"~e,~~_
c
N g o E rd Oa" $ ~ ~:~; W E
oSS
"~ 8>~~, t~m r
r~CL: OCL: S


m$$~
o ~ ~ ~b W
e o W W j
~,~c~M "'Il
~oo o
acL~" 8
o


~ "" $ a ~ o ~~ ~~ ~ $~
e Ec I
c;;~I: MO "at~
~: W =
O c
U M
















100-
hIT
75-

50-



205





o a i aY ):

BC


75-

so-

25-


100


75

50
2-







0-





Wo-
25
S7-

5-




S2-



W -




10-





0-


~ .s $
g a~
~ 8 '5
CL:
m
E
o


12~c~8~~$myog
gp ~~~~~~78~)Bg
o~,~$Lm~
i;; r~
~Li~~e ~ m
~; ~~U C~


O U


mme
s~y
'e o
o
"8
as
ai~
o
o
u


MW


"m~~
a~o~
" 8 'j:

d ~ae
as U
o


u =


a "" d


o u


STIMULI


Figure 3-4. Percent of times a food was selected in between-category MSWO preference
assessments for 6 control participants (Experiment 1).


nnnnnnnnnnnnnnnnn


O
aE~$ g g ~ ~ g:~ ~o$
o~1 O g r E P OWY g
~ o~
a $

a ~e ~1 m


nn,,nnni II II Innnnn,


P
~~ E ~~ ra ra ra o e r ~ ~ ~ ~`O
o aaa cm m o W gb
~"
Oc~0 Um
"CL: ~o 9
F cL~a
u


nnnnnnnn,,nnnnnnn


,cy~~" gm
- W
po;3~a
'e cb
1""~~C PIE
P m
CL:
~1
u


u
EU



















hnnnnnnnnnnnnnnnnn
g~,- gca i L
~~C~D ~~ o,
o o oa E
oo,~~ o orm' o
L, ifl.~ Q ~~ ~ O M ~ ~ m ~, o
~" u m
F1 U
oi~i
%o
u


IIIIIIIIIIIIIIIIII
~ g ~s c g oa r rr, ~, o o ~b~' "L~~
aao,
.. r. E :: U; ~C~"P~"O
Uo oo~Y~~~~aa~ U
C1 o U i~i." g ~ ~ C1 m o M L ~ ~ ~
Oc~a
a~U
e!
,E


IIIIIIIIIIIIIIIIII
C CL
a ~c~~~qo~ C,~~g
9~gr3 e!Bu e ~I gi ,.
e a
O ~9 b ~~ -o ~.i! Y g ~, ~ m ~,
m e oma~ MUUa
m ifl r y U a g u ifl a, o
B$"a h


100-


Paired Stimulus Preference Assessments


STIMULI

Figure 3-5. Percent of times a food was selected in between-category MSWO preference
assessments and Paired Stimulus preference assessments for 6 control participants
(Experiment 1).


n,,nnnnnnnnnnnnnnn
~ y m ~ ~ C ~ m ~I ~O o o B
O~~am=m o
B.9 o ~~ e ~~ L ~(
.Y LmmUifl g
rma~~ru~ o
FIFI ~j a m
L,


nnnnnnnnnnnnnnnnnn
Clolo~r~ "~h~~~~~~
~, m ,~~ o ~ o~m ,~,~~ O a
e ~ "o ~ My 5 O a~ E
on m~~Ug
'~x~'"~~~"~~mrM


nnnnn









CHAPTER 4
EXPERIMENT 2: DETERMINANTS OF FOOD PREFERENCE

Studies on the relative influence of quality, magnitude, and delay have shown somewhat

mixed results. In a two-part experiment comparing the choices of individuals with PWS to

developmentally disabled peers, Glover et al. (1996) showed that when foods were of relatively

similar quality, magnitude appeared to be more influential than quality for individuals with PWS.

By contrast, control participants tended to select based on food quality (rather than magnitude).

In the second experiment, all participants selected a smaller magnitude of the high-preference

item over larger magnitudes of low-preference items. Joseph et al. (2002) compared the effects

of magnitude and delay on foods of different qualities. Results showed that individuals in the

PWS group were more likely to make choices based on magnitude than delay and did so to a

greater extent than individuals in the obese comparison group. Results across different qualities

of food showed that the higher-preferred foods were associated with a larger overall proportion

of choices for the larger magnitude of food (at a delay), whereas magnitude appeared less

influential when the foods were lower quality. The authors concluded that individuals with PWS

were more likely to make choices based on food magnitude than delay.

The research to date suggests that both quality and magnitude may influence food

choices. However, the influence of quality relative to both magnitude and delay has not been

examined. Further, no experiment has presented data in such a way that the influence of those

variables on choice could be identified for individual participants. The purpose of the current

experiment is to provide an analysis of the relative influence of quality, magnitude, and

immediacy on food choices of individuals with PWS.










Participants, Design and Procedure

Eleven individuals (all PWS) participated in Experiment 2. All sessions were conducted in

either a sheltered workshop or a therapy room at a school. Target behaviors consisted of

vocational tasks (piece-work assembly and stuffing envelopes, defined previously). Two sets of

materials were available, each associated with a different reinforcement contingency.

Discriminative stimuli were included to signal the reinforcement contingency in place. The

presence of a plate next to the participants' session materials signaled that food on the plate

would be available at the end of session (immediately), and the presence of a plastic bag signaled

that food would be available after some delay. The food on the experimenter' s plate was visible,

such that the participant was able to view the different foods associated with a response option

(quality), and the number of pieces of food on the experimenter' s plate signaled the magnitude of

food available for each response (1 versus 5 pieces). Prior to each session, participants were

prompted to complete one response per set of materials, and the reinforcer associated with that

set of materials was placed onto the participants' plate (or into a bag when delays to

reinforcement were in effect). The participant was then told that "you can do whatever you want

to do during session," and sessions were begun. No further instructions or discussion occurred

during session. Two measures were taken to ensure that participants did not consume an

excessive number of calories. First, reinforcers were delivered according to a VI 30-s schedule

during sessions (such that no more than an average of 20 pieces of food could be earned per

session). Second, no more than three, 10-min sessions were conducted per day. Participants were

exposed to at least two sessions of each condition, and additional sessions were conducted if

responding was inconsistent across those two sessions.

Baseline. During baseline, each parameter was evaluated singly to determine sensitivity to

quantitative changes in it. When testing each parameter, the values of all other parameters were









held constant. Foods were selected based on the between-category preference assessment

(Experiment 1), and two high-quality and two low-quality items were selected (to reduce the

chance of satiation to a particular item). The two items that were ranked first and second were

selected as high-quality foods. The two lowest-ranked items were selected as the low-quality

foods with the additional criterion that the low-ranked foods must be low-calorie (vegetable or

fruit).

During the immediacy (I) baseline, one response option resulted in reinforcer delivery at

the end of the session, and the other option resulted in reinforcer delivery one hour later.

Magnitude and quality were held constant by using one piece of a high-quality food across both

options.

During the magnitude (M) baseline, one option resulted in 5 pieces of food, and the other

option resulted in 1 piece of food. Immediacy and quality were held constant by delivering the

high-quality food immediately after session across both options.

During the quality (Q) baseline, one option resulted in delivery of one of the two highest-

ranked foods, and the other option resulted in delivery of one of the two lower-ranked foods.

Foods of the same quality (high or low) were alternated across reinforcer deliveries. Magnitude

and immediacy were held constant by delivering one piece of food immediately after session.

Assessment of Competing Parameters. Disparate values of each parameter (Immediacy

[I], Magnitude [M], and Quality [Q]) were compared, with only 2 of 3 parameters manipulated

per session. The parameter that was not being compared was at its optimal (quality, immediacy)

or a moderate (magnitude) value across both response options.

During Immediacy versus Magnitude (IvM) sessions, one response option was associated

with an immediately available, small-magnitude reinforcer, and the other option was associated










with a delayed, large-magnitude reinforcer. Because quality was not manipulated in this

condition, the food item was the same (high-quality) across both response options. Thus, the

contingencies associated with response options were (a) 1 piece of a high-quality food available

at the end of the session, or (b) 5 pieces of a high-quality food available 50 min after the session

was over.

During the Immediacy versus Quality (IvQ) condition, one response option was associated

with an immediately available, low-quality reinforcer, and the other option was associated with a

delayed, high-quality reinforcer. Magnitude was not manipulated in this condition; therefore, the

magnitude for each option was set at a moderate value (1 tablespoon-sized portion, which equals

approximately 3 times the size of one teaspoon-sized portion). The contingencies associated with

response options were (a) 1 (tablespoon-sized) piece of a low-quality food available at the end of

the session, or (b) 1 (tablespoon-sized) piece of a high-quality food available one hour after the

session was over.

During the Magnitude versus Quality (MyQ) condition, one response option was

associated with a large-magnitude, low-quality reinforcer, and the other option was associated

with a small-magnitude, high-quality reinforcer. Immediacy was not manipulated in this

condition; therefore, reinforcers were available at the end of session across each response option.

The contingencies associated with response options were (a) 1 piece of a high-quality food

available at the end of session, or (b) 5 pieces of a low-quality food available at the end of

session.

Results and Discussion

Results of the assessments are depicted in Figures 4-1 and 4-2. The figures show the final

session of each condition, as this represents response allocation following a history of exposure

to contingencies. That is, the first session per condition may not accurately represent behavior










under the control of a particular set of contingencies because participants did not experience the

(primary) reinforcement contingency until after the session. If behavior were to change as a

result of exposure to the reinforcement contingency, this effect could not be observed until at

least the second session per condition.

Results show that for 10 of 11 participants, responding during baseline showed near-100%

time allocation to what would traditionally be considered the favorable alternative: high quality

over low quality, large magnitude over small magnitude, and immediate over delayed

alternatives. One participant, MJ, deviated from this pattern slightly in that she allocated most

(76%) but not all of her time to the high-quality option. The purpose of baseline was to

demonstrate each participant's sensitivity to the reinforcement parameters, as a prerequisite to

participating in the subsequent assessment conditions. If a participant showed little or no

sensitivity to one or more of the parameters of interest, then the subsequent assessment would

not provide useful information. Further, if a participant showed a preference for the unfavorable

value of a parameter (e.g., allocating more time to the small-magnitude food over the large-

magnitude food), then it was deemed likely that variables other than the parameters of interest

might be influencing behavior (e.g., a history of instructions, a history of reinforcement for

selecting smaller portions, etc.). Therefore, participants whose behavior showed indifference to

the parametric manipulations (e.g., near-chance levels of responding across both response

options of a baseline condition) or a preference for the unfavorable option were excluded from

the subsequent assessment. Three participants were exposed to baseline conditions but were

excluded from assessment for this reason (data not presented here, but available from the authors

upon request).









Nine of 11 participants allocated most or all of their responding to the response option

associated with the high-quality food during assessment (Figure 4-1). For these 9 individuals,

data show that they selected the option associated with the high-quality food, even when it

produced a 50-min delay to that food (during the IvQ condition) and when it produced 1/5 the

amount of food (during the MyQ condition). These results suggest that quality was a more

influential reinforcer parameter than immediacy or magnitude. Among those 9 participants, 5

participants showed that when immediacy and magnitude were compared, magnitude was the

more influential parameter. Three participants showed that immediacy was more influential than

magnitude, and one participant' s responding was allocated equally among response options

associated with immediately-available versus larger-magnitude reinforcers.

Two of 11 participants allocated most or all of their responding to the response option

associated with the reinforcer that was available immediately after session (Figure 4-2). These

data show that responding was allocated to the immediately-available reinforcers, even when it

produced a lower-quality food (during the IvQ condition) and when it produced a smaller

magnitude of food (during the IvM condition). These results suggest that immediacy was a more

influential parameter than quality or magnitude for these participants. In addition, data show that

these two participants chose a higher-quality reinforcer over a larger-magnitude reinforcer

(during the MyQ condition).

No participants showed a pattern in which responding was influenced primarily by the

magnitude of food. Magnitude appeared to be influential under some conditions, as observed for

JB, KB, DD, DV, and PM. However, this effect was observed only when magnitude was

compared with immediacy, and quality was the same across options. When quality was

manipulated, magnitude was not a controlling variable.









These results were somewhat unexpected. Numerous reports have documented that

hyperphagia is a prevalent characteristic of PWS and that individuals with PWS tend to consume

much larger quantities of food than control subj ects (Holland, Treasure, Coskeran, & Dallow,

1995; Holland et al., 1993). Further, reports suggest that individuals with PWS are very slow to

satiate, if satiation (reductions in speed of consumption) is noted at all (Lindgren, et al., 2000).

The behavior (hyperphagia) itself seems to suggest that large quantities of food are valuable;

therefore, it seems likely that magnitude would be an influential reinforcement characteristic.

Instead, the results for all participants in the current show that the quality (9 participants) or the

immediacy (2 participants) of the reinforcement was the most influential characteristic.














AC





XD


Baseline Assessment
BK

.

I I I I I


Baseline Assessment


Qa it
SMagnitude
I Inunediac


PM


NR VIVIV


MW




QI VIVII I


QM IMyQ IvM IvQ


SESSIONS
Figure 4-1. Percentage of time allocated to each response option for the final session of each
baseline and assessment condition. Open portions of bars during assessment indicate
choice of an unfavorable dimension (Experiment 2).









CHAPTER 5
EXPERIMENT 3: PARAMETRIC MANIPULATIONS TO SHIFT CHOICE

The results of Experiment 2 showed that 9 of 11 participants' choices were determined by

food quality. Selection based on food quality is not a problem per se; however, many of these

participants' highly-preferred foods were relatively high-calorie (top-ranked foods included pork

breakfast sausage, Raisinets, cheddar cheese, pepperoni, and cookies). Therefore, many

individuals who selected foods based on quality were selecting extremely high-calorie foods.

This type of responding may be problematic, as many individuals diagnosed with PWS typically

are on restricted calorie diets. Interventions that teach them to select lower-calorie options may

have therapeutic benefits.

Two of 11 participants in Experiment 2 selected foods based on the immediacy of their

delivery. Food selections based primarily on immediacy may present challenges to controlled

eating in this population. Individuals may choose to eat extremely high-calorie or dangerous

(spoiled, contaminated) foods because they are immediately available, which may result in

weight gain or illness. Teaching these individuals to wait for delayed alternatives may allow for

a number of therapeutic interventions to control weight (e.g., reinforcement contingencies for the

absence of inappropriate food consumption, reinforcement for surrendering found foods, etc.).

A variety of interventions have been used to decrease impulsive choices, including the use

of signals to indicate reinforcement will be delivered at a later time (Vollmer, Borrero, Lalli, &

Daniel, 1999), introducing activities during the delay (Dixon, Rehfeldt, & Randich, 2003), and

initially decreasing and then gradually increasing delays (Schweitzer & Sulzer-Azaroff, 1988).

Schweitzer and Sulzer-Azaroff identified 5 participants who tended to select smaller, immediate

reinforcers over larger, delayed reinforcers. After establishing a baseline in which the

participants consistently selected the smaller, immediate reinforcer, the experimenters introduced









training. Initially, delays to both the small and the large reinforcers were set at equal values (Os);

during subsequent sessions, the delay to the large reinforcer was increased in 5-s increments.

Following training, 4 of 5 participants consistently selected the larger, delayed reinforcer over

the smaller, immediate reinforcer. Neef et al. (1994) used a similar procedure to teach 3

participants diagnosed with ADHD to tolerate delays to reinforcers. During assessment, each

individual allocated most or all responding to the immediately available reinforcers. Self-control

training consisted of initially decreasing and then gradually increasing the delays to a higher-

quality reinforcer, while the lower-quality, no-delay reinforcer continued to be available. Results

showed that the participants continued to allocate most of their responding to the delayed

reinforcer, even when delays were increased up to 24 hours. Results of post-assessment sessions

showed that some amount of generalization occurred across untrained conditions.

The purpose of Experiment 3 was to shift participant responding to either a lower-calorie

option (BK and AC) or to a delayed option (NR). BK and AC were exposed to a variety of

parametric manipulations of immediacy and magnitude, and a subsequent replacement of the

higher-calorie foods with lower-calorie versions. NR was exposed to delay and magnitude

fading.

Participants, Design and Procedure

Three individuals from Experiment 2 participated in Experiment 3. NR selected foods that

were immediately available (control by immediacy), even when that resulted in a smaller amount

of food or a lower-quality food. BK always selected the high-quality food (control by quality) in

Experiment 2, even when that resulted in a smaller amount of food or a delay to food

availability. Further, her high-quality foods were relatively high-calorie (pork sausage and

RaisinetsTM candy). AC also selected only high-quality foods (control by quality); her high-

quality foods were pepperoni and oatmeal raisin cookies, which were also high-calorie.










The experimental arrangement, target behaviors, pre-session exposure to contingencies,

and reinforcement schedules were identical to Experiment 2. The reinforcement contingencies

associated with each response varied according to the desired change in responding (i.e.,

selection of a lower-calorie food, or selecting foods that would be available after a delay).

Immediacy treatment

The purpose of treatment was to teach NR to tolerate (i.e., choose) delays to food

availability. A combination of magnitude and delay fading was used. The sessions were run in

the context of the IvM condition, with the exception that lower-calorie foods were substituted.

Throughout treatment sessions, one response option continued to be associated with one

piece of food available immediately after the session (the option to which she allocated all

responding during the IvM assessment sessions). The contingencies associated with the second

response option changed across sessions. Initially, both the magnitude of the food delivered per

reinforcer was increased (to 15 pieces), and the delay to that food's availability was decreased (to

the end of session). During subsequent sessions, the delay to food's delivery was gradually

increased. After the delay was increased to one hour, magnitude fading was initiated by

decreasing the number of pieces by one per session. By the final sessions of delay and

magnitude fading, the second response option was for 5 pieces of food available after one hour.

Thus, the options available were the same as those available during the original IvM assessment

sessions.

Quality treatments

The purpose of the quality treatment was to arrange conditions such that BK and AC

would select lower-calorie foods over higher-calorie foods. Several parametric manipulations

were introduced in an attempt to reduce caloric intake. During all quality treatment sessions, the

response options were associated with two different sets of foods. BK's high-quality foods were









sausage and cheddar cheese, her low-quality foods were lettuce and apple, and her high-quality

substitutes were vegetable sausage and fat-free cheddar cheese. AC's high-quality foods were

pork pepperoni and oatmeal raisin cookies, her low-quality foods were tomatoes and carrots, and

her high-quality substitutes were turkey pepperoni and South Beach DietTM chocolate chip

oatmeal cookies.

Delay manipulation. During these sessions, the delay to the high-quality food was

increased from one hour (during assessment) to 3 hours, and the magnitude was the same across

both options (1 piece of food). Thus, one response option was associated with one piece of a

high-quality food available after 3 hours, and the second option was associated with one piece of

a low-quality food available at the end of the session.

Magnitude manipulation. During these sessions, the magnitude of the low-quality food

was increased from 5 pieces (during assessment) to 15 pieces, and the delay was equal across

both options (end of session). One response option was associated with one piece of a high-

quality food at the end of the session, and the second option was associated with 15 pieces of a

low-quality food available at the end of the session.

Delay plus magnitude manipulation. Both the delay to the high-quality food and the

magnitude of the low-quality food were manipulated in these sessions. One response option was

associated with one piece of a high-quality food available after 3 hours, and the second option

was associated with 15 pieces of a low-quality food available at the end of session.

Quality substitution. The low-quality foods were replaced with reduced-calorie

versions of the high-quality foods during these sessions, and both magnitude and delay values

were equal across both options (1 piece and end of session, respectively). Thus, one response

option was associated with one piece of the high-quality food available at the end of the session,









and the second option was associated with one piece of the high-quality substitute available at

the end of the session.

Results and Discussion

Figure 5-2 depicts the results of the immediacy (top panel) and quality (bottom panel)

treatments. Results of NR' s assessment showed that she selected the immediately available

reinforcer whenever delay was manipulated. When the delay and magnitude fading treatment

was introduced, she allocated all of her responding to the option associated with the larger

magnitude reinforcer that was delivered after gradually increasing delays. She continued to

allocate all responding to the delayed option throughout all increases in the delay and decreases

in the magnitude. During the final four sessions, she continued to allocate all of her time to the

response option associated with the delayed (60-min), larger magnitude (5 pieces) reinforcer.

This reflects a shift from the assessment condition, during which she allocated all of her time to

the option associated with the immediately-available, smaller reinforcer.

Results of the quality treatment for BK are depicted in the center panel of Figure 5-1.

During assessment, she allocated all of her time to the option associated with the higher-quality

foods (sausage and cheddar) whenever quality was manipulated. Throughout the delay,

magnitude, and delay plus magnitude manipulations, BK continued to allocate all of her

responding to the higher-quality foods. When high-quality substitutes were introduced, she

selected the high-quality foods during the initial sessions, but in subsequent sessions allocated all

of her time to the substitutes. Results for AC (depicted in the bottom panel of Figure 5-1) are

similar; the delay, magnitude, and delay plus magnitude manipulations had no influence on her

response allocation. Following the introduction of the high-quality substitutes, she allocated all

of her time to the high-quality substitutes without further manipulation.









The results of NR' s treatment suggest that simple fading manipulations may be sufficient

to teach individuals to tolerate delays to reinforcers. This finding may be especially important in

the PWS population because these individuals have a history of hyperphagia; teaching them not

to immediately consume foods at each opportunity may be a necessary skill for achieving re-

integration into situations in which foods are available but should not be consumed.

The results of BK' s and AC's treatments were somewhat surprising; quality (as an

independent variable) was sufficiently strong to prevent shifts despite long delays (3 hours), very

disparate magnitudes (1 versus 15 pieces of food), and a combination of the increased delay and

magnitude. The shift to the lower-calorie version of the high-quality foods was unexpected,

given the extent to which quality had previously controlled responding. Their data suggest that

the high-quality substitutes may have been even higher-quality than the high-quality foods

identified in the preference assessment.






























MyQIvMIvQ IvM

~3High Quality (HQ) [] HQ Substitute

~aImmediate High Magnitude


Assessment


Delay Fading + Magnitude Fading


1LQ @ end v. 15LQ @ (n time)
~ minutes a
0 02 OR 1 2 5 10 20 70 45 60


1LQ @ end v. n LQ @ 60 min
I magnitude
17 11 9 7 5 5


S100.



60.



C40


Assessment


Treatment


MyQ IvM IvQ IvQ MyQ I+MyQ


HQ v HQ sub


SMy Q IvM Iv Q IvQ MyQ I+MyQ HQ v HQ sub
SESSIONS



Figure 5-1. Percentage of time allocated to each response option during immediacy (top panel)
and quality (bottom panel) treatments (Experiment 3).









CHAPTER 6
GENERAL DISCUSSION

The prevalence of hyperphagia and other food-related problem behaviors in the PWS

population suggest that food is an unusually potent reinforcer for this group. Hyperphagia is

listed as one of the maj or clinical features of the disorder, and researchers have noted the

occurrence of other food-related problems such as property destruction to access food (Benj amin

& Buot-Smith, 1993), stealing (Donaldson, et al., 1995), hoarding food (Dykens, Leckman, &

Cassidy, 1996), and consumption of substances such as pet food or food from trash cans.

(Russell & Oliver, 2003). Unrestricted access to food also poses a danger for this population

because individuals with PWS are more likely than those in the general population to die from

choking (Stevenson et al., 2006), gastric necrosis (Wharton, Wang, Graeme-Cook, Briggs, &

Cole, 1997), and obesity-related diseases such as diabetes or heart failure (Schrander-Stumpel et

al., 2004). Given the behavioral and physical problems presented by food in this population, an

assessment of variables that might influence food selection (and subsequent consumption) seems

relevant to a comprehensive treatment for PWS.

Most studies on reinforcers characteristics that influence food choice in PWS have focused

on preferences for food type. Results of some studies suggested that individuals with PWS

prefer sweet foods to all others (Caldwell & Taylor, 1983; Glover et al., 1996; Hinton et al.,

2006; Taylor & Caldwell, 1985), whereas others suggest a variety of tastes and types of food are

preferred (Fieldstone et al., 1997; Rankin & Mattes, 1996). However, the small number of

stimuli presented in all studies limits the extent to which generalizations about food preferences

can be made. Further, the presentation of data in an aggregated format obscures individual

differences in preference. The results of Experiment 1 represent a departure from previous

studies on food preference in the population. First, several different samples and types of food









were used (48 foods were assessed across 6 types of food). Therefore, conclusions about

preferences for various types of food may have more generality than those of previous studies.

Second, each individual's data showed slightly different preferences, and it did not appear to be

the case that (a) preferences were uniform across individuals within a group, or (b) there were

important differences in the caloric value of choices across groups. The data from Experiment 1

suggest that the preferences of individuals with PWS are not drastically different than those of

individuals with developmental disabilities other than PWS.

The results of Experiment 2 showed that quality was most often the variable determining

choice. These results are consistent with one previous study that suggested that individuals with

PWS sometimes choose smaller quantities of preferred foods over larger quantities of less-

preferred food (Caldwell & Taylor, 1983). However, other studies have shown that magnitude

may occasionally be more influential than quality (Glover et al., 1996) or delay (Joseph et al.,

2002) in food selections of individuals with PWS.

Several factors may have contributed to the emergence of quality as such an influential

variable in the current study. First, it is possible that extra-experimental factors influenced

responding. Most individuals diagnosed with PWS are prescribed therapeutic (calorie-restricted)

diets, which may result in the restriction or total elimination of higher-calorie foods. Thus, high-

calorie foods are typically unavailable, which may increase their value in the context of the

study. Future studies may wish to rule out this influence by restricting the foods used to those

typically available to individuals with PWS (i.e., an array of all low-calorie foods).

Second, it is possible that the quality parameter was stronger in the current study than it

has been in previous studies. Previous studies sampled a small number of stimuli, which may

have resulted in the identification of foods that were of roughly similar value. The wide array of









foods included in the current study may have allowed for the identification of a set of stimuli

with more disparate reinforcing potencies, such that the lower-preferred foods had little value

relative to the highly-preferred foods. (However, it should be noted that all foods used as low-

preferred foods were in fact highly-preferred during the within-category preference assessments).

Data presented by Glover et al. (1996) support this hypothesis: Their results showed that

magnitude influenced selection among similarly preferred foods (foods ranked first in a

preference assessment versus those ranked second and third), but that quality was more

influential than magnitude when the comparison was among highly-preferred and non-preferred

foods. The large number and broad variety of items assessed in the current study may have

ultimately increased the strength of the quality variable, relative to the immediacy and magnitude

variables.

Third, it is possible that the differences in the magnitude and/or the delay values

constituted a fairly weak independent variable. The variables were of sufficient strength to

influence responding during the baseline sessions (i.e., all subj ects allocated most or all of their

time to the favorable values); however, those variables exerted little control over most

participants' responding during the subsequent assessment. It is unclear to what extent those

values would have to be changed such that they were as powerful as the quality variable. Data

from BK and AC's treatment phase showed that three-fold increases in the delay and magnitude

values did not produce a shift in responding. Thus, it may be the case that much larger

differences in the delay and magnitude variables (e.g., 5- or 10-fold increases) are necessary to

make those variables more potent. Another alternative may be to change the magnitude or delay

values of the higher-quality food (e.g., decrease the size of the preferred food or the delay to the









lower-preferred food). These changes were not attempted in the current study, but may prove

useful in future research.

The implications of quality as an influential variable may provide some insight into food-

related problem behaviors in individuals with PWS. Early researchers had suggested that the

unusual consumptive behavior (i.e., ingestion of contaminated or unusual foods) was an

indication of indiscriminate eating or a lack of preference (Pipes & Holm, 1973). However, data

from Experiment 1 and 2 suggest that this is not the case. The unusual consumptive behavior

may not be due to a lack of food preference; rather, variables such as restricted access to foods

may create a situation in which any ingested material is sufficiently reinforcing to maintain such

behavior, despite preferences among foods. Results of a study by Roscoe, Iwata, and Kahng

(1999) showed that although participants tended to allocate responding to a preferred food over a

non-preferred food under a concurrent-reinforcement schedule, contingent delivery of the non-

preferred food was sufficient to maintain responding during a single schedule. Thus,

consumption of unusual or contaminated food by individuals with PWS does not necessarily

indicate a lack of preference; rather, these foods may simply be sufficiently reinforcing to

consume when other foods are unavailable, despite preferences for other foods.

Food appears to be a particularly powerful reinforcer for the PWS population, making

information on variables that influence that reinforcer relevant to management of eating in the

PWS. This information may allow for the development of more highly individualized treatments

that better address each individual's problematic patterns of eating. For instance, individuals who

choose immediately available foods may be appropriate for delay fading, as illustrated in

Experiment 3. Or, interventions in which appropriate (low-calorie) foods are constantly

(immediately) available may mitigate inappropriate food consumption. Individuals who show a









strong tendency to select high-quality items may be good candidates for research on shifting

preference to lower-calorie items. If substitution interventions such as those described in the

current study are not effective, then interventions such as simultaneous delivery of preferred and

non-preferred foods (Piazza et al., 2002) or differential reinforcement contingencies in which a

very small portion of preferred (high-calorie) food is available contingent on consumption of

larger amounts of low-calorie foods may prove useful in managing the food consumption of this

population.

Despite the potential contribution of antecedent manipulations, it is likely that differential

consequences will be necessary to suppress inappropriate food consumption. Studies on food

stealing in PWS have shown that interventions including differential reinforcement of other

behavior (DRO; Page, Finney, Parrish, & Iwata, 1983), self-monitoring combined with

contingency contracting and punishment (Altman, Bondy, & Hirsch, 1978), and verbal

reprimands combined with stimulus control interventions (Maglieri, DeLeon, Rodriguez-Catter,

and Sevin, 2000) may be successful in reducing inappropriate eating; however, limited evidence

of generalization is available. Both Page et al. (1983) and Maglieri et al. (2000) evaluated their

interventions in a multiple baseline across settings design, and results of both studies showed that

the treatment did not generalize to untrained settings. Thus, comprehensive treatments for food-

related problem behaviors in PWS will likely require both antecedent and consequences

manipulations, as well as explicit programming for generalization.










REFERENCES


Bekle, T. W. (1997). Running and responding reinforced by the opportunity to run: Effect of
reinforcer duration. Journal of the Experimental Analysis ofBehavior, 67, 337-351.

Benjamin, E., & Buot-Smith, T. (1993). Naltrexone and fluoxetine in Prader-Willi syndrome.
Journal of the American Academy of Child and Adolescent Psychiatry, 32, 870-873.

Berntson, G.G., Zipf, W.B., O'Dorisiod, T.M., Hoffmane, J.A., & Chancee, R.E. (1993).
Pancreatic polypeptide infusions reduce food intake in Prader-Willi syndrome. Peptides,
14, 497-503.

Butler, M. G., & Thompson, T. (2000). Prader-Willi syndrome: Clinical and genetic findings.
Endocrinologist, 10, 3S-16S.

Caldwell, M. L., & Taylor, R. L. (1983). A clinical note on food preference of individuals with
Prader-Willi syndrome The need for empirical research. Journal Of~entalDeficiency
Research, 27, 45-49.

Caldwell, M. L., Taylor, R. L., & Bloom, S. R. (1986). An investigation of the use of high-
preference and low-preference food as a reinforcer for increased activity of individuals
with Prader-Willi syndrome. Journal Of2~ental Deficiency Research, 30, 347-3 54.

Catania, A. C. (1963). Concurrent performances: A baseline for the study of reinforcement
magnitude. Journal of the Experimental Analysis ofBehavior, 6, 299-300.

Chung, S. H. (1965). Effects of delayed reinforcement in a concurrent situation. Journal of the
ExperimentalAnalysis ofBehavior, 8, 439-444.

Chung, S. H., & Herrnstein, R. J. (1967). Choice and delay of reinforcement. Journal of the
ExperimentalAnalysis ofBehavior, 10, 67-74.

Davison, M. & Baum, W. M. (2003). Every reinforcer counts: Reinforcer magnitude and local
preference. Journal of the Experimental Analysis ofBehavior, 80, 95-129.

DeLeon, I. G., Fisher, W. W., Rodriguez-Catter, V., Maglieri, K., Herman, K., & Marhefka, J. M.
(2001). Examination of relative reinforcement effects of stimuli identified through pretreatment
and daily brief preference assessments. Journal ofAppliedBehavior Analysis, 34, 463-473.

DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of a multiple-stimulus presentation format for assessing
reinforcer preferences. Journal of Applied Behavior Analysis, 29, 519-533.

Dixon, M. R. & Cummings, A. (2001). Self-control in children with autism: Response allocation
during delays to reinforcement. Journal ofApplied Behavior Analysis, 34, 491-495.

Dixon, M. R., Rehfeldt, R. A., & Randich, L. (2003). Enhancing tolerance to delayed reinforcers: The
role of intervening activities. Journal ofApplied Behavior Analysis, 36, 263-266.











Donaldson, M.D.C., Chu, C.E., Cooke, A., Wilson, A., Greene, S.A., & Stephenson, J.B.P.
(1994). The Prader-Willi syndrome. Archives ofDisease in Childhood, 70, 58-63.

Dykens, E. M., Leckman, J. F., & Cassidy, S. B. (1996). Obsessions and compulsions in Prader-
Willi syndrome. Journal of Child Psychology and Psychiatry, 27, 995-1002.

Fieldstone, A., Zipf, W. B., Schwartz, H. C., & Berntson, G. C. (1997). Food preferences in
Prader-Willi syndrome, normal weight and obese controls. International Journal of
Obesity, 21, 1046-1052.

Glover, D., I. Maltzman, I., & Williams, C. (1996). Food preferences among individuals with
and without Prader-Willi syndrome. American Journal on M~entalRetardation, 101, 195-
205.

Graff, R. B., Gibson, L., & Galiatsatos, G. T. (2006). The impact of high- and low-preference stimuli
on vocational and academic performances of youths with severe disabilities. Journal ofApplied
Behavior Analysis, 39, 13 1-13 5.

Herrnstein, R. J. (1961). Relative and absolute strength of response as a function of frequency of
reinforcement. Journal of the Experimental Analysis ofBehavior, 4, 267-272.

Hinton, E. C., Holland, A. J., Gellatly, M. S. N., Soni, S., & Owen, A. M. (2006). An
investigation into food preference and the neural basis of food-related incentive
motivation in Prader-Willi syndrome. Journal oflntellectual Disability Research, 50,
633-642.

Hoch, H., McComas, J. J., Johnson, L., Faranda, N., & Guenther, S. L. (2002). The effects of magnitude
and quality of reinforcement on choice responding during play activities. Journal ofApplied
Behavior Analysis, 35, 177-18 1.

Holland, A.J., Treasure, J., Coskeran, P., Dallow, J., Milton, N., Hillhouse, E. (1993).
Measurement of excessive appetite and metabolic changes in Prader-Willi syndrome.
International Journal of Obesity andRelated Metabolici Disorders, 1 7, 527-532.

Holland, A.J., Treasure, J., Coskeran, P., Dallow, J. (1995). Characteristics of the eating disorder
in Prader-Willi syndrome: Implications for treatment. Journal oflntellectual Disability
Research, 39, 373-8 1.

Holm, V.A., Cassidy, S.B., Butler, M.G., Hanchett, J.M., Greenswag, L.R., Whitman, B.Y.,
Greenberg, F. (1993). Prader-Willi syndrome: Consensus diagnostic criteria. Pediatrics,
91, 398-402.

Joseph, B., Egli, M., Koppekin, A., & Thompson, T. (2002). Food choice in people with Prader-
Willi syndrome: Quantity and relative preference. American Journal on M~ental
Retardation, 107, 128-35.











Keesey, R. E., & Kling, J. W. (1961). Amount of reinforcement and free-operant responding.
Journal of the Experimental Analysis ofBehavior, 4, 125-132.

Keller, J. V., & Gollub, L. R. (1977). Duration and rate of reinforcement as determinants of
concurrent responding. Journal of the Experimental Analysis ofBehavior, 28, 145-153.

Landon, J. Davison, M. & Elliffe, D. (2003). Concurrent schedules: Reinforcer magnitude
effects. Journal of the Experimental Ana~lysis of Behavior, 79, 351-365.

Lerman, D. C., Kelley, M. E., Van Camp, C. M., & Roane, H. S. (1999). Effects of reinforcement
magnitude on spontaneous recovery. Journal of Applied Behavior Analysis, 32, 197-200.

Lindgren, A.C., Barkeling, B., Hagg, A., Ritzen, E.M., Marcus, C., & Rossner, S. (2000). Eating
behavior in Prader-Willi syndrome, normal weight, and obese control groups. Journal of
Pediatrics, 137, 50-5.

Lowe, C. F., Davey, G. C. L., & Harzem, P. (1974). Effects of reinforcement magnitude on
interval and ratio schedules. Journal of the Experimental Analysis ofBehavior, 22, 553-
560.

Maglieri, K. A., DeLeon, I. G., Rodriguez-Catter, V., & Savin, B. M. (2000). Treatment of
covert food stealing in an individual with Prader-Willi syndrome. Journal ofApplied
Behavior Analysis, 33, 615-618.

Matthews, L. R., & Temple, W. (1979). Concurrent schedule assessment of food preference in
cows. Journal of the Experimental Analysis ofBehavior, 32, 245-254.

McLean, A. P. & Blampied, N. M. (2001). Sensitivity to relative reinforcer rate in concurrent
schedules: Independence from relative and absolute reinforcer duration. Journal of the
ExperimentalAnalysis ofBehavior, 75, 25-42.

Miller, H. L. (1975). Matching-based hedonic scaling in the pigeon. Journal of the Experimental
Analysis of Behavior, 26, 3 35-3 47.

Neef, N. A., Bicard, D. F., & Endo, S. (2001). Assessment of impulsivity and the development of
self-control in students with attention deficit hyperactivity disorder. Journal ofApplied
Behavior Analysis 34, 397-408.

Neef, N. A., & Lutz, M. N. (2001). A brief computer-based assessment of reinforcer dimensions
affecting choice. Journal ofApplied Behavior Analysis 34, 57-60.

Neef, N. A., Mace, F. C., & Shade, D. (1993). Impulsivity in students with severe emotional
di sturb ance: The interactive effects of reinforcer rate, delay, and quality. Journal of
Applied Behavior Analysis, 26, 37-52.

Neef, N. A., Mace, F. C. Shea, M.C., & Shade, D. (1992). Effects of reinforcer rate and










reinforcer quality on time allocation: Extensions of matching theory to educational
settings. Journal ofApplied Behavior Analysis, 25, 691-699.

Neef, N. A., Shade, D, & Miller, M. S. (1994). Assessing influential dimensions of reinforcers on
choice in students with serious emotional disturbance. Journal ofAppliedBehavior
Analysis, 27, 575-583.

Neuringer, A. J. (1967). Effects of reinforcement magnitude on choice and rate of responding.
Journal of the Experimental Analysis of Behavior, 10, 417-424.

Pace, G. M., Ivancic, M. T., Edwards, G. L., Iwata, B. A., & Page, T. J. (1985). Assessment of
stimulus preference and reinforcer value with profoundly retarded individuals. Journal of
Applied Behavior Analysis, 18, 249-255.

Page, T. J., Finney, J. W., Parrish, J. M., & Iwata, B. A. (1983). Assessment and reduction of
food stealing in Prader-Willi children. Applied Research in M~entalRetardation, 4, 219-
228.

Piazza, C. C., Patel, M. R., Santana, C. M., Goh, H., Delia, M. D., & Lancaster, B. M. (2002). An evaluation
of simultaneous and sequential presentation of preferred and nonpreferred food to treat food
selectivity. Journal of Applied Behavior Analysis, 35, 259-269.

Pipes, P., & Holm, V. (1973). Weight control of children with Prader-Willi syndrome. Journal of the
American Dietetic Association, 62, 520-524.

Prader, A., Labhart, A., & Willi, H. (1956). Ein syndrome von adipositas, kleinwucs,
kryptorchismus und oligophrenie nach myatonicartigem zustand im neugeborenalter.
Scinveiz Med Wochenchr, 86, 1260 -1261.

Rankin, K. M., & Mattes, R. D. (1996). Role of food familiarity and taste quality in food
preferences of individuals with Prader-Willi syndrome. International Journal on Obesity
andRelated Metabolic Disorders, 20, 759-62.

Reed, P. (1991). Multiple determinants of the effects of reinforcement magnitude on free-
operant response rates. Journal of the Experimental Analysis ofBehavior, 55, 1 09-123.

Roane, H. S., Vollmer, T. R., Ringdahl, J. E., & Marcus, B. A. (1998). Evaluation of a brief stimulus
preference assessment. Journal ofApplied Behavior Analysis, 31, 605-620.

Roscoe, E. M., Iwata, B. A., & Rand, M. S. (2003). Effects of reinforcer consumption and magnitude on
response rates during noncontingent reinforcement. Journal of Applied Behavior Analysis, 36,
525-539.

Russell, H., & Oliver, C. (2003). The assessment of food-related problems in individuals with
Prader-Willi syndrome. British Journal of Clinical Psychology, 42, 379 392.

Schneider, J. W. (1973). Reinforcer effectiveness as a function of reinforcer rate and magnitude:










A comparison of concurrent performances. Journal of the Experimental Analysis of
Behavior, 20, 461-471.

Schrander-Stumpel, C. T., Curfs, L. M., Sastrowijoto, P., Cassidy, S. B., Schrander, J. J., Fryns,
J.P. (2004). Prader-Willi syndrome: Causes of death in an international series of 27
cases. American Journal of2~edical Genetics A, 124, 333 338.

Schweitzer, J. B., & Sulzer-Azaroff, B. (1988). Self-control: Teaching tolerance for delay in
impulsive children. Journal of the Experimental Analysis ofBehavior, 50, 173-186.

Stebbins, W. C. (1962). Response latency as a function of amount of reinforcement. Journal of
the Experimental Analysis ofBehavior, 5, 305-307.

Stevenson, D. A., Heinemann, J., Angulo, M., Butler, M.G., Loker, J., Rupe, N., Kendell, P.,
Clericuzio, C.L., Scheimann, A. O. (2006). Deaths due to choking in Prader-Willi
syndrome. American Journal of2~edical Genetics, 143A, 484 487.

Taylor, R. L., & Caldwell, M. L. (1985). Type and strength of food preferences of individuals
with Prader-Willi syndrome. Journal Of2~ental Deficiency Research, 29, 109-112.

Todorov, J. C. Hanna, E. S. & de SA, M. C. N. B. (1984). Frequency versus magnitude of
reinforcement: New data with a different procedure. Journal of the Experimental Analysis
ofBehavior. 41, 157-167.

Tustin, R. D. (1994). Preference for reinforcers under varying schedule arrangements: A
behavioral economic analysis. Journal ofApplied Behavior Analysis, 27, 597-606.

Volkert, V. M., Lerman, D. C., & Vorndran, C. M. (2005). The effects of reinforcement magnitude on
functional analysis outcomes. Journal of Applied Behavior Analysis, 38, 147-162.

Vollmer, T. R., Borrero, J. C., Lalli, J. S., & Daniel, D. (1999). Evaluating self-control and impulsivity
in children with severe behavior disorders. Journal ofAppliedBehavior Analysis, 32, 451-466.

Wharton, R.H., Wang, T., Graeme-Cook, F., Briggs, S., Cole, R.E. (1997). Acute idiopathic gastric
dilation with gastric necrosis in individuals with Prader-Willi syndrome. American Journal of
Medical Genetics, 73, 437-41.

Zipf, W.B., Berntson, G.G. (1987). Characteristics of abnormal food-intake patterns in children with
Prader-Willi syndrome and study of effects of naloxone. American Journal of Clinical Nutrition,
46, 277-81.









BIOGRAPHICAL SKETCH

Jessica Thomason's career in applied behavior analysis (ABA) began at the University of

Florida, where she volunteered at the Florida Center on Self-Injury under the supervision of Dr.

Brian Iwata. After receiving her B.S. (1998), she accepted a position at the Kennedy Krieger

Institute, where she coordinated the assessment and treatment of behavior disorders for inpatients

diagnosed with developmental disabilities. In August of 2001, she began graduate school in

behavior analysis at University of Florida. While in graduate school, Jessica conducted research

on a variety of topics, including the assessment and treatment of problem behavior and reinforcer

identification and assessment. She also provided clinical services in several different settings,

including a school for individuals with developmental disabilities, an outpatient clinic for

children diagnosed with autism, and a vocational and residential program for individuals

diagnosed with developmental disabilities. Jessica served as the coordinator of two clinics (the

school and autism clinics), as a teaching assistant for an introductory class and lab in ABA, and

as instructor of a class in ABA. Jessica received her Ph.D. in May of 2007. Jessica hopes to

continue on to a career in which she conducts clinical research on behavior disorders and skill

acquit sition.