The Influence of Secondary Stimulus Characteristics in the Assessment of Sexual Offenders Diagnosed with Intellectual Di...

MISSING IMAGE

Material Information

Title:
The Influence of Secondary Stimulus Characteristics in the Assessment of Sexual Offenders Diagnosed with Intellectual Disabilities
Physical Description:
1 online resource (81 p.)
Language:
english
Creator:
Walker, Stephen F
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology
Committee Chair:
Vollmer, Timothy Raymond
Committee Members:
Chambers, John Robert
Iwata, Brian A
Lanza Kaduce, Lonn M

Subjects

Subjects / Keywords:
assessment -- behavioral -- disabilities -- intellectual -- offenders -- penile -- plethysymograph -- sex
Psychology -- Dissertations, Academic -- UF
Genre:
Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
The present series of studies examined whether secondary stimulus characteristics can influence assessments of sexual interest of ID sex offenders. The majority of studies and clinical assessments on sexual interest of sex offenders have used stimulus materials that varied systematically by age and gender category, but not systematically across myriad other stimulus features such as body type, hair color or style, clothing type, and many others. We will refer to age and gender as primary stimulus characteristics (due to the emphasis in the literature and clinics), and we will refer to other stimulus features as secondary stimulus features. Study 1 used penile plethysmograph assessments to investigate whether arousal would occur differentially in the presence of stimuli within the same age and gender category that differed across secondary stimulus characteristics. Study 2 used paired stimulus preference assessments to investigate whether preference would be shown in the presence of stimuli within the same age and gender category that differed across secondary stimulus characteristics. The results of studies 1 and 2 show that secondary 10 stimulus characteristics must influence assessments of sexual interest because arousal occurred when age and gender were held constant. A limitation of Studies 1 and 2 was that, the specific stimulus characteristics influencing assessment results were unknown. Study 3 was an exploratory evaluation of the effects of specific stimulus features on outcomes of preference assessments evaluating sexual interest. Study 3A utilized paired stimulus preference assessments to evaluate the independent effects of four secondary stimulus characteristics (breast size, clothing type, waist-to-hip ratio, hair length). The results showed that breast size and clothing type influenced preference but hair length and waist-to-hip ratio did not. Because the results showed that preference was influenced by breast size and clothing type, it is possible that one of these stimulus characteristics might influence preference to a greater degree than the other. The purpose of Study 3B was to evaluate the relative preference of breast size and clothing type. Results showed that some secondary stimulus characteristics influence preference more than others; however, the effects of individual stimulus characteristics are idiosyncratic.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Stephen F Walker.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Vollmer, Timothy Raymond.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-05-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045313:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 THE INFLUENCE OF SECONDARY STIMULUS CHARACTERISTICS IN THE ASSESSMENT OF SEXUAL OFFENDERS DIAGNOSED WITH INTELLECTUAL DISABILITIES By STEPHEN F. WALKER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF F LORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 S tephen F. Walker

PAGE 3

3 To Nicole

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I would like to thank Tim Vollmer for the guidanc e, patience, and friendship that he has provided during my graduate training. I would also like to thank my committee members Drs. Brian Iwata, John Chambers, and Lonn Lanza Kaduce for their time and assistance on this project. I would also like to thank Ray Joslyn, Triton Ong, and Tony Oliver for their invaluable help on various aspects of this project. I would also like to thank Astrid Hall for her guidance and support. Finally, I would like to thank my wife and continued supporter, Nicole Zeug. You have made a great impression in my life, I would not be the person that I am today without you.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 9 Chapter 1 GENERAL INTRODUCTION ................................ ................................ .................. 11 Phallometric Assessments with Intellectually Disabled Offenders .......................... 12 Evaluating Deviant Sexual Interest via Preference Assessments ........................... 16 Current Studies ................................ ................................ ................................ ....... 17 2 STUDY 1 ................................ ................................ ................................ ................. 20 Method ................................ ................................ ................................ .................... 20 Results ................................ ................................ ................................ .................... 23 Discussion ................................ ................................ ................................ .............. 28 3 STUDY 2 ................................ ................................ ................................ ................. 48 Method ................................ ................................ ................................ .................... 49 Results ................................ ................................ ................................ .................... 50 Discussion ................................ ................................ ................................ .............. 52 4 STUDY 3 ................................ ................................ ................................ ................. 55 Study 3A ................................ ................................ ................................ ................. 56 Method ................................ ................................ ................................ ............. 56 Results and Discussion ................................ ................................ .................... 58 Study 3B ................................ ................................ ................................ ................. 61 Method ................................ ................................ ................................ ............. 61 Results and Discussion ................................ ................................ .................... 61 5 GENERAL DISCUSSION ................................ ................................ ....................... 73 LIST OF REFERENCES ................................ ................................ ............................... 77 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 80

PAGE 6

6 LIST OF TABLES Table page 1 1 Example secondary stimulus characteristics ................................ ...................... 19 2 1 Demographic characteristics for par ticipants ................................ ...................... 46 2 2 PPG Assessment results for Study 1. ................................ ................................ 47

PAGE 7

7 LIST OF FIGURES Figure page 2 1 7 assessment. ......................... 32 2 2 ........................ 33 2 3 ...................... 34 2 4 9 assessment. ............................ 35 2 5 .............. 36 2 6 ....................... 37 2 7 ....................... 38 2 8 9 assessment. ............................ 39 2 9 7 assessment. ............................ 40 2 10 ....................... 41 2 11 9 assessment. ......................... 42 2 12 n assessment. ........... 43 2 13 Assessment results for participants that showed arousal to all test stimuli within specific age and gender categories. ................................ ......................... 44 2 14 Representative age and gender assessments in which sexual arousal was not observed ................................ ................................ ................................ ....... 45 3 1 Preference assessment results for all participants.. ................................ ........... 54 4 1 Avatar stimuli used in the clothing type and breast size evaluations .................. 64 4 2 Avatar stimuli used in the waist to hip ratio and hair length evaluations ............ 65 4 3 Results of the clothing type evaluation. ................................ .............................. 66 4 4 Results of the breast size evaluation ................................ ................................ .. 67 4 5 Results of the waist to hip ratio evaluation ................................ ......................... 68 4 6 Results of the hair length evaluation ................................ ................................ ... 69 4 7 Stimuli used in study 3B ................................ ................................ ..................... 70

PAGE 8

8 4 8 Results of breast size and clothing type comparison grouped by breast size ..... 71 4 9 Results of breast size and clothing type comparison grouped by clothing type .. 72

PAGE 9

9 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Require ments for the Degree of Doctor of Philosophy THE INFLUENCE OF SECONDARY STIMULUS CHARACTERISTICS IN THE ASSESSMENT OF SEXUAL OFFENDERS DIAGNOSED WITH INTELLECTUAL DISABILITIES By Stephen F. Walker May 2013 Chair: Timothy R. Vollmer Major: Psychology Th e present series of studies examined whether secondary stimulus characteristics can influence assessments of sexual interest of intellectually disabled ( ID ) sex offenders. The majority of studies and clinical assessments on sexual interest of sex offenders have used stimulus materials that varied systematically by age and gender category, but not systematically across myriad other stimulus features such as body type, hair color or style, clothing type, and many others. We will refer to age and gender as pr imary stimulus characteristics (due to the emphasis in the literature and clinics), and we will refer to other stimulus features as secondary stimulus features. Study 1 u s ed penile plethysmograph assessments to investigate whether arousal w ould occur diffe rentially in the presence of stimuli within the same age and gender category that differed across secondary stimulus characteristics Study 2 u s ed paired stimulus preference assessments to investigate whether preference would be shown in the presence of st imuli within the same age and gender category that differed across secon dary stimulus characteristics. The results of studies 1 and 2 show that secondary

PAGE 10

10 stimulus characteristics must influence assessments of sexual interest because arousal occurred when a ge and gender were held constant A limitation of Studies 1 and 2 was that the specific stimulus characteristics influencing assessment results were unknown Study 3 was an exploratory evaluation of the effects of specific stimulus features on outcomes of preference assessments evaluating sexual interest. Study 3A utilized paired stimulus preference assessments to evaluate the independent effects of four secondary stimulus characteristics (breast size, clothing type, waist to hip ratio hair length). The re sults showed that breast size and clothing type influence d preference but hair length and waist to hip ratio did not Because t he results showed that preference was influenced by breast size and clothing type, it is possible that one of these stimulus char acteristics might influence preference to a greater degree than the other. The purpose of Study 3B was to evaluate the relative preference of breast size and clothing type. R esults showed that some secondary stimulus characteristics influence preference mo re than others; however, the effects of individual stimulus characteristics we re idiosyncratic.

PAGE 11

11 CHAPTER 1 GENERAL INTRODUCTION The assessment of deviant sexual arousal of child sex offenders has been the focus of a large body of research D eviant sexua l arousal (i.e., increased penile tumescence in the presence of pictures and videos of prepubescent children) has been shown to be a significant predictor of re off ense for child sex offenders (Hanson & Morton Bourgon, 2004; Lalumiere & Harris, 1998). R ela tively little research however, has evaluated common assessment techniques with child sex offenders diagnosed with intellectual disabilities (ID). The little research that does exist on the topic suggest s that common assessment techniques are viable for I D offenders ( Murphy et al. 1983 ; Quinsey et al.,1975; Reyes et al., 2006; Reyes, Vollmer, & Hall, 2011a; Reyes, Vollmer, & Hall, 2011b ) The initial pur pose of this dissertation (S tudy 1) with chil d sex offender s with ID, wa s to evaluate whether arousal will occur differentially in the presence of stimuli within the same age and gender group. The majority of studies and clinical assessments on sexual interest of sex offenders have used stimulus materials that varied systematically by age and gender catego ry, but not systematically across myriad other stimulus features such as body type, hair color or style, clothing type, and many others. In fact, virtually all marketed assessment stimulus sets are categorized by age and gender only ( e.g., Not Real People Visual Stimuli Set Pacific Psychological Assessment Corporation, Sex Offen der Assessment Series Northwest Media Inc.) and virtually all studies on the assessment of deviant arousal systematically control for age and gender only ( Laws, 2009; Vollmer, Reyes & Walker, 2011 ). We will refer to age and gender as primary stimulus characteristics (due to the emphasis in the literature and clinics), and

PAGE 12

12 we will refer to all other stimulus features that may vary within age and gender categories as secondary stimul us features. If differential arousal occurs within an age or gender category it suggests that factors other than the typically assumed age and gender secondary stimulus features, must be controlling arousal at least to some extent. A second purpose (Stu dy 2) wa s to use a less invasive method a paired stimulus preference assessment, to explore preference within stimuli of the same age and gender group. If preferences emerge, it suggests further that secondary stimulus features must be controlling prefer ence. Finally, because the first two studies only point to a possible influence of secondary stimulus features by exclusion, we used avatars in an effort to isolate some specific secondary stimulus features that may control preferences (S tudy 3) Phallome tric Assessments with Intellectually Disabled Offenders In sex offender research and practice, s exual arousal is commonly measured via penile plethysmograph (PPG), a device that measures changes in penile tumescence in real time. During the course of PPG a ssessments, participants are typically only exposed to test stimuli once or twice during an assessment (e.g., Laws & Osborn, 1983). Recent research has investigated the use of repeated measures with the aim of addressing the possibility of influence by ext raneous factors (e.g., health issues, masturbation prior to sessions) that might alter the outcomes of single session PPG assessments (Reyes et al., 2006; Reyes, Vollmer, & Hall, 2011a; Reyes, Vollmer, & Hall, 2011b). To date, research has shown that the u se of repeated measures can decrease the possibility of obtaining false negative outcomes due to extraneous factors. For example, Reyes et al. (2006) examined repeated measures by administering PPG assessments with 10 sex offenders with ID. The results sho wed that assessments

PAGE 13

13 including repeated measures provided a more complete assessment than shorter assessments because they show ed differences in arousal patterns across sessions as opposed to relative arousal levels on a single exposure to stimuli For exa mple, one participant showed no arousal to test stimuli the first time test stimuli were presented. However, when the participant was exposed to the test stimuli in subsequent sessions, high and differentiated levels of deviant arousal were observed. If th assessment had included only one exposure to the test stimuli, as is commonly seen in clinical practice, the assessment would have produced a false negative outcome. Other cases were reported showing possible false positive outcomes if onl y a first exposure had been tested. More r ecently, Reyes et al. ( 2011b ) examined the effects of pre session masturbation and pre session instructions to suppress arousal in an effort to systematically evaluate factors that might lead to false negative PPG assessment outcomes with three sex offenders with ID. During the pre session masturbation evaluation one participant was instructed to masturbate prior to each PPG session. The results indicated that pre session masturbation produced significantly lower le vels of penile arousal when compared to control conditions, suggesting that if an individual masturbat ed prior to PPG sessions, false negative assessment outcomes could be obtained. During the pre session instructions evaluation two participants were prov ided with a straightforward pre Results showed that, for one p articipant this simple instruction suppressed deviant and non deviant arousal However, for the seco nd participant this instruction only suppressed arousal to non deviant stimuli and a second

PAGE 14

14 instruction had little effect on levels of deviant sexual arousal. Additional factor s that can potentially influence the outcomes of PP G assessments surround the stimuli that are used in the assessments. For example, t he results of a meta analysis conducted by Lalumiere and Quinsey (1994) show that the discriminate validity of PPG assessments increases as a function of the number of test stimuli included in an assessment. Based on the results of this research Lalumiere and Harris (1998) recommend that PPG assessments include at least two test stimuli per age and gender category. One limitation of the studies conducted by Reyes et al. (200 6) and Reyes et al. (2011a) is that only one stimulus was included for each age and gender category during the PPG assessment. Combined, the studies included 13 participants. Of the 13 participants, five showed no arousal to test stimuli This outcome is surprising given that these individuals were charged as child sex offenders It is possible that if more than one stimulus was included for each age and gender category the number of participants who showed arousal to test stimuli would have increased O ther potential factor s that could influence the outcomes of PPG assessments are secondary stimulus features In the analyses that follow, we will consider age and gender to be primary stimulus features because they are the stimulus features usually contro lled in PPG research. Also in the analyses that follow, we will consider other variables that tend to change with stimulus presentations to be secondary stimulus features. For example, in most commercially available stimulus sets used for PPG assessments when the age and gender of the individual in a video clip changes, other stimulus features tend to change too (e.g., hair length, clothing type). These secondary

PAGE 15

15 stimulus features are not described in the literature as i t is commonly assumed that observe d arousal is a function of the age and gender of the presented stimulus (Laws, 2009). In fact, o ne limitation of studies conducted by Reyes et al. (2006) and Reyes et al. (2011a) is that they used stimuli that were designed to assess age and gender categor ies, but these stimuli also differed across a number of secondary stimulus features. We know this because we we re able to view the stimulus sets used in those studies ( T able 1 1 shows example s of some stimulus variations ) The use of stimulus sets with un controlled secondary stimulus features leaves open the possibility that s uch features in fluenced PPG outcomes For example, if an individual displayed high levels of arousal to an adult female who was wearing a two piece bathing suit and low levels of arou sal to a to a prepubescent female who was wearing a one piece bathing suit it is possible that that the type of bathing suit, rather than the age of the female was actually responsible for increased levels of arousal. Despite this general limitation, n o studies identified to date have evaluated the influence of secondary stimulus characteristics on PPG assessment outcomes. There is, however, a large body of research that has systematically evaluated the influence of physical characteristics on ratings o f physical attractiveness by non offenders. Specifically, the influence of waist to hip ratio (WHR) and breast size on ratings of female attractiveness has been replicated many times ( e.g., Furnham & Reeves, 2006; Furnham, Petrides, & Constantinides, 2005; Perilloux, Webster, & Gaulin, 2010; Singh, 1993; Singh & Young, 1995 ). The results of this research have shown that men generally rate women with a WHR of .7 and medium to large breasts as more attractive than women with higher a WHR (e.g., .9) and smalle r breasts (Singh, 1993; Signh &

PAGE 16

16 Young, 1995; Schutwoh l, 2006; Zelazniewicz, 2011). Very little research, however, exists on the influence of other secondary stimulus characteristics such as hair length and clothing type on rat ings of physical attractivene ss. A study conducted by Swami, Furnham, and Joshi (2008) is the only evaluation of the effects of hair length on ratings of female attractiveness, and their results showed that ratings of attractiveness were not influenced by hair length. No studies have evaluated the effects of clothing type on ratings of physical attractiveness. Further, it is not known whether typical features of and who happen to be ID. Eval uating Deviant Sexual Interest v ia Preference Assessments R eyes (2008) evaluated the use of paired choice preference assessments as a This approach was designed as an alternative to the presum ably more intrusive PPG method Reyes used pictures of adults and children captured from a commercially available video series designed for the assessment of sex offenders. A hierarchy of picture preference was obtained for each participant using a paired stimulus preference assessment (PSPA, Fisher et al., 1992), which is a method that is commonly used for individuals with ID to identify preferred leisure and edible items. R esults showed that for 4 out of 5 participants the preference assessment matched t he outcome of a previously conducted PPG assessment. That is, the higher stimulus preferences tended to correspond with stimuli producing the higher levels of arousal. Results of the Reyes study s uggest that preference assessments are a viable and less in vasive method of obtaining measures of deviant interest. Additionally, the results of the study showed that there was a high degree of correspondence between reflexive (as measured via PPG)

PAGE 17

17 and operant (as measured via preference assessment ) aspects of sex ual interest. The stimuli used in the study, again, however, were noted to be a limitation. The stimuli were designed to assess age and gender preferences, but they also differed across a number of secondary stimulus features Current Studies The purpose o f the current studies was to evaluat e whether secondary stimulus features influence assessments of sexual in terest of child sex offenders diagnosed with ID Study 1 evaluated whether secondary stimulus features influence PPG assessment outcom es by tes ting multiple stimuli for each age and gender category that differed across secondary stimulus characteristics. If secondary stimulus characteristics do not influence PPG outcomes it would be expected that levels of arousal would be undifferentiated across all members of an age and gender category that differed across other features (e.g., clothing type). However, if secondary stimulus features do influence PPG outcomes, it would be expected that arousal would be differentiated at least in some cases across mem bers of an age and gender category that differed across other features. Study 2 evaluated whether secondary stimulus features influenc e the outcomes of preference assessments using commercially available stimuli (still photos of stimuli used in Study 1) by tes ting multiple stimuli for each age and gender category that differed across secondary stimulus characteristics. If secondary stimulus characteristics do not influence the outcomes of preference assessments it would be expected that all stimuli withi n an age and gender category would be equally preferred. However, if stimulus features other than age and gender do influence preference assessment

PAGE 18

18 outcomes, it would be expected that hierarchies of preference would be shown within age and gender categorie s in at least some cases Study 3 was an exploratory two part study evaluating the effects of specific secondary stimulus characteristics using digitally rendered avatars as test stimuli. Study 3A was an evaluation of the independent effects of clothing t ype, waist to hip ratio, breast size, and hair length on preference. Study 3B evaluated the relative preference of stimulus features that were shown to produce hierarchies of preference in Study 3A.

PAGE 19

19 Tabl e 1 1 Example seconda ry stimulus characteristics F K Race Eye Color Hair Color Hair Length Complexion Clothing 1 Caucasian Light Brown Ponytail Shoulder Light Tank Top/Shorts 2 Hispanic Dark Dark Brown Pigtails Shoulder Medium 1 Piece Bathing Suit 3 African American Dark Da rk Brown Straight Shoulder Dark 2 Piece Bathing Suit

PAGE 20

20 CHAPTER 2 STUDY 1 Method Participants All participants reside d in a state funded treatment facility for ID offenders. All were accused of at least one sexual offen s e against a c hild and had been found incompetent to stand trial by a local area judge. Table 2 1 shows the demographic and offense related information for all participants. The procedures described are one component of an over arching assessment conducted by the facili ty and a certified PPG clinician supervised all assessments. All participants were assigned an individual identification number that allow ed for the review of data, while maintaining the individual s right to confidentiality. T analysis review committee approved the research. PPG Assessments. All PPG assessments were conducted using hardware Suite 10.0) developed by Limestone Technologies, penile strain gauges produced by D.M. Davis Inc., and commercially available stimuli designed for the assessment of sexual arousal (Sex Offen der Assessment Series Northwest Media I nc.). The video series included 31 videos: three videos for each of the 10 age and gender categories (male and female: kindergarten, 6 7, 8 9, teen, and adult) and one neutral stimulus (adult men fishing). Each video depicted an individual sitting in a cha ir reading a magazine, then moving to a set of stairs where the individual ate a piece of fruit, and then finally moving to a hot tub where the individual splashed water on themselves.

PAGE 21

21 The three videos for each age and gender category portrayed a differen t individual with different physical features and clothing. Assessments were conducted according to the s tandard assessment protocols for the facility Prior to the ir first session participants receive d a tour of the assessment space. During the tour, the technician explained the purpose of each piece of equipment to the participant. The assessment room contain ed a recliner with disposable absorbent pad, a computer monitor used to display assessment stimuli, a camera that provide d a live video feed of the participant from above the shoulders, and a metal lap tray. The purpose of the live video feed was to confirm that participants remain ed awake and orientated towards the computer screen. The purpose of the metal lap tray was to eliminate any visual feedbac k of arousal and decrease the likelihood that the participant interfere d with the gauge during the session. Additionally, the participant was asked to measure the circumference of his penis to ensure the appropriate gauge size wa s chosen for the assessment Before each session, the technician calibrate d the penile strain gauge with a calibration devise to ensure accurate measurement and instruct ed the participant to use the restroom. After the participant used the restroom the technician provide d the follo pull your pants and underwear around your ankles, sit in the recliner, place the gauge around the middle of your penis, place the metal lap tray over your lap, and if you feel yourself becoming aroused let it happ rticipant attached the gauge incorrectly or if the gauge broke an abnormal pattern of response was shown in the data stream, and the issue was resolved prior to continuing the session.

PAGE 22

22 Experimental Design and Procedure Video stimuli were presented in a multielement format with the order of the stimul us presentations randomized across sessions. Each age and gender category was evaluated independently (e.g., F emales in the age group 6 7 years would be one category, and this is abbreviated as F 6 7) and th e order in which each age and gender category was assessed was randomized across participants The three test stimuli and the neutral stimulus were presented in one of three predetermined orders with each stimulus being presented twice during each session for a total of 8 stimulus presentations per session. During each session, the technician monitor ed changes in penile tumescence in real time. Millimeter change (penile tumescence at the onset of the stimulus presentation subtracted from the peak level of p enile tumescence) serve d as the primary dependent variable and was calculated automatically by the software package (similar to Reyes et al., 2006) Following each stimulus presentation a d etumescence criterion of no changes greater than 5 mm for a period of 1 min had to be met before the next stimulus was presented. If this criterion wa s not met after 15 min, the session was terminated. One session was conducted per day, 3 to 5 days per week. Data a nalysis Five Board Certified Behavior Analysts (BCBAs) ex perienced in analyzing single subject data evaluated d ata for all assessments. This method was based on the procedure described by Kahng and Iwata (1999). Graphs were developed for each test stimulus evaluated. Specifically, each test stimulus (e.g., Fema le 6 7.1) and the neutral stimulus for that assessment were plotted in the same figure. T he evaluators reviewed each graph and reached a n agreement about which of the following three categories best described the data: (1) differentiated ( data paths have

PAGE 23

23 l ittle or no overlap and the difference between the data paths appear to be greater than 5mm) (2) differentiated low magnitude (data paths have little or no overlap and the difference between the data paths appear to be less than 5mm), or (3) undifferentia ted (data paths overlap a considerable amount). The reason for identifying low magnitude arousal is based on the recommendation of Lalumiere and Harris (1998) who advocate d for an arousal criterion that is based on the relative differences between control and test stimuli. Their criterion holds that any time test stimuli produce higher levels of arousal than control stimuli the assessment results are valid. We saw similarities in this logic with traditional behavior analytic interpretations of assessment a nd treatment outcomes. Additionally, the evaluators reached a n agreement on whether or not the data showed a downward trend. Cases in which an agreement was not reach ed (which was rare) were reviewed by the author and major professor and were assigned to one of the three categories. Results Figures 2 1 through 2 4 show the results of assessments when arousal was observed in the presence of one test stimulus and only one test stimulus for an age and gender category. A common feature of these figures is t hat the upper left panel depicts the results for all stimuli. The other three panels show test control comparisons for each stimulus versus neutral (for clearer visual inspection). All PPG data were plotted as millimeter change across successive stimulus presentations. Figure 2 1 shows the results of participant 6 6 7 assessment Arousal levels for the Female 6 7.1 stimulus (top right panel) were undifferentiated from neutral Arousal levels for the Female 6 7.2 stimulus (bottom left panel) wer e undifferentiated

PAGE 24

24 from neutral. A rousal levels for the Female 6 7.3 stimulus a lthough variable, were differentiated from the neutral Figure 2 2 shows the results een a ssessment. Arousal levels for the Female T een.1 stimulus ( top right panel) and F emale Teen.2 stimulus (bottom left panel) were undifferentiated from neutral A rousal levels for the F emale Teen.3 stimulus (bottom right panel) were differentiated from the neutral Figure 2 3 shows t emale a dult assessment Arousal levels for the F emale Adult.1 stimulus (top right panel) and Female Adult.2 (bottom left panel) were undifferentiated from neutral. A rousal levels for the F emale A dult.3 stimulus (bottom right panel ) were differentiated f rom neutral though at a low magnitude and with a downward trend w here the data paths converg ed at the end of the assessment. Figure 2 9 assessment. Arousal levels for the Male 8 9.1 stimulus (top right panel) w ere differentiated from neutral (low magnitude). A rousal levels for the Male 8 9.2 stimulus ( bottom left panel) and Male 8 9.3 stimulus (bottom right panel) were undifferentiated from neutral. A lthough arousal levels produced by the Male 8 9.3 st imulus may have been initially differentiated, they were on a downward trend and by the end of the assessment, differentiation was not seen Figures 2 5 through 2 12 show the results of assessments when arousal was generally observed in t he presence of t wo test stimuli for an age and gender category (with some less clear outcomes for the second or third age and gender category, however) Figure 2 5 shows th assessment. Arousal levels for the M ale K indergarten .1 stimulus (top right panel) were

PAGE 25

25 undifferentiated from neutral. A rousal levels for the Male Kindergarten.2 stimulus (bottom left panel) and Male Kindergarten.3 stimulus were d ifferentiated from neutral. Figure 2 6 f emale teen assessment. Arousal levels for the Female Teen.1 stimulus (top right panel) and Female Teen.3 (bottom right panel) were differentiated from neutral. A rousal levels for the Female Teen.2 stimulus (bottom left panel) were undifferentiated from neu tral and on a downward trend. Figure 2 7 Arousal levels for the Female Adult.1 stimulus (top right panel) were undifferentiated from neutral. A rousal levels for the Female Adult.2 stimulus (bo ttom left panel) and Female Adult.3 stimulus were differentiated from neutral. However, results for Adult.3 are somewhat ambiguous due to a high degree of variability. Figure 2 8 9 assessment. Arousal levels for the Male 8 9.1 (top right panel) and Male 8 9.3 stimuli (bottom right panel) were differentiated from neutral. A rousal levels for the Male 8 9.2 stimulus (bottom left panel) were conservatively judged as undifferentiated from neutral An argument can be made for calling these results differentiated too, but clinical exigencies for this participant required that the assessment end at after six stimulus exposures. Figure 2 9 7 assessment. Arousal levels for the Male 6 7.1 (top right panel) and Male 6 7.3 stimuli (bottom right panel) were differentiated from neutral. A rousal levels for the Male 6 7.2 stimulus (bottom left panel) were undifferentiated from neutral or possibly differentiated at a very low magnitude.

PAGE 26

26 Figure 2 10 levels for the Female Teen.1 stimulus (top right panel) were differentiated from neutral, and arousal levels Female Teen.3 stimulus (bottom right panel) were differentiated from neutral (low magnitude). A r ousal levels for the Female Teen.2 stimulus were undifferentiated from neutral or possibly differentiated at a very low magnitude Figure 2 11 9 assessment. Arousal levels for the Female 8 9.1 stimulus (top right panel) were differentiated from neutral, and arousal levels for the Female 8 9.3 stimulus (bottom right panel) were differentiated from neutral (low magnitude). A rousal levels for the Female 8 9.2 (bottom left panel) we re undifferentiated from neutral or possibly differentiated at a low magnitude Although arousal levels produced by the Female 8 9.2 stimulus may have been initially differentiated, they were on a downward trend and by the end of the assessment, differenti ation was not seen. Figure 2 Arousal lev els for the Female Kindergarten.1 stimulus (top right panel) were undifferentiated from neutral. A r ousal levels for the Female Kindergarten.2 (bottom left panel) and Female Kindergarten.3 (bottom right panel) stimuli were differentiated from neutral (low magnitude). Figure 2 13 shows the results of assessments for participant 1 ( female kindergarten, female 6 7, female 8 9, and female adult), p articipant 3 (female 6 7, and female adult), and participant 8 (female kindergarten, female 6 7, and female 8 9) in which all test stimuli produced differentiated arousal

PAGE 27

27 Representative age and gender assessments in which sexual arousal was not observed are displayed in Figure 2 14. T wo distinct patterns of no arousal were observed. One pattern was when both the test and neutral stimuli produced relatively low and undifferentiated levels of arousal. For example, the left panel in Figure 2 1 4 depicts the 9 assessment in which all test stimuli and the neutral stimulus produced low undifferentiated levels of arousal. This pattern of no arousal was observed in 34 other age and gender assessments. The second pattern was observed when both the test and neutral stimuli produced moderate and undifferentiated levels of arousal. For example, the second panel in Figure 2 1 4 depicts the r 6 7 assessment in which all test stimuli and the neutral stimulus produ ced moderate and undifferentiated levels of arousal. This pattern of no arousal was observed in 15 other age and gender assessments, but was only exhibited by participants 4 and 5. The summary of all PPG assessments are displayed in Table 2 2 Cases in w hich a stimulus produced a differentiated outcome are listed in the corresponding cell. For instance, Participant 1 showed differentiated arousal to the Female Teen.1 and Female Teen.3 stimuli, therefore the numbers 1 and 3 are listed in the corresponding cell across Cases in which a stimul us produced a differentiated low magnitude outcome are listed in the corresponding cell with an asterisk next to the stimulus number (e.g., 3*). Cases in which no stimuli produced arousal within an age and gender category were noted by a cell. A corresponding cell.

PAGE 28

28 Discussion The purpose of the study was to evaluate the poss ible influence of secondary stimulus features or those other than age and gender on PPG outcom es. This was accomplished by assessing three stimuli from specific age and gender categories that differed across secondary stimulus characteristics (e.g., clot hing type). O f the 70 age and gender assessments completed arousal was observe d in 21 assessments. In 12 cases of those 21 assessments differentiated arousal within age and gender categories was shown Specifically, in four of these cases only one stimul us produced arousal within an age and gender category (Participant 2: female teen; Participant 3: male 8 9 ; Participant 6: female 6 7: Parti cipant 7: female adult ), and in eight of these cases two stimuli produced arousal within an age and gender category (Participant 1: female teen; Participant 2: female adult; Participant 3: female kindergarten, fe male 8 9 female teen ; Participant 8: male kindergarten male 6 7, male 8 9 ). Even in cases when all stimuli produced arousal, there was sometimes greater arou sal to one stimulus o ver another (e.g., participant 3: female 6 7 ). The results of these 12 age and gender assessments show that secondary stimulus characteristics are likely to influence PPG assessment outcomes Furthermore these results show that second ary stimulus characteristics play an important r ole in sexual arousal, and need to be accounted for when evaluating PPG outcomes. For example, the result 7 assessments show that arousal was only produced by one of the three tes t stimuli used in that assessment (F 6 7.3). These results highlight the potential influence of secondary stimulus characteristics in PPG assessments, and show that if secondary stimulus characteristics are not accounted for, false negative assessment resu lts are possible. H owever, the methods utilized in the

PAGE 29

29 current study do not allow for the identification of which secondary stimulus characteristics exerted influence, as the age and gender of the presented stimuli were the only stimulus characteristics th at were held constant Also, in nine of the age and gender assessment s in which arousal was observed, all stimuli within an age and gender category were identified as producing either differentiated or differenti ated low magnitude arousal. Th is latter find ing suggest s that the age and the gender of the presented stimuli were the likely determinants of arousal in th o se cases. R esults of Study 1 also provide further empirical evidence to support the inclusion of multiple stimuli for each age and gender ca tegory evaluated in PPG. The 12 cases in which arousal was produced mainly by only one or two of the test stimuli in an age and gender category highlight the need to include multiple test stimuli for each age and gender category assessed. For example, the kindergarte n assessment show that arousal was produced by two stimuli (MK.2, MK.3). If this assessment had only included the MK.1 stimulus, the assessment would have produced a negative outcome for male kindergarten category In the current study a distinction was made between dif ferentiated and differentiated low magnitude patterns of arousal. One assessment produced results in which two test stimuli produced differentiated low magnitude arousal (Participa nt 3: Female Kind ergarten.2 and Female Kindergarten.3 ), and two assessments produced results in which onl y one test stimulus produced low magnitude differentiated arousal (Partic ipant 3: Male 8 9.1 ; Participant 7: Female Adult.3). There were cases, however, in which one o r more of the test stimuli produced differentiated arousal, and one or more of the test stimuli produced differentiated low

PAGE 30

30 magnitude showed that two stimuli produced differentiated arou sal (Female Kindergarten.2 and Female Kindergarten.3) and one stimulus produced differentiated low magnitude arousal (Female Kindergarten.1). Patterns of arousal similar to this were shown in one other age and gender assessments (Participant 3: female 8 9 ) Patterns of arousal such as these are similar to patterns of arousal in which only one or two test stimuli were identified as producing differentiated arousal, in that there are clear relative differences with in age and gender categories. One approach m ight be to compare the two data paths for test stimuli for ease of visual inspection, rather than just comparing a test stimulus to the control stimulus. If one test stimulus consistently produces higher levels of arousal than another test stimulus, it pr ovides further evidence of control by secondary stimulus features. Previous literature utilizing repeated measures in PPG has shown a downward trend, indicative of habituation or respondent extinction, in 2 out of 13 cases (Reyes et al., 2006; Reyes et al. 2011b). In the current study a downward trend was observed in 5 of the 21 assessments in which arousal was observed (Participant 1: female kindergarten and female teen ; Participant 3: female 8 9 and male 8 9; Participant 7 : female adult). Despite the a dvantages of repeated measurement, habituation, or extinction should be considered as possible confound s. It is important to note a nother potential limitation of the current study. Although it appears that secondary stimulus characteristics in fluence PPG assessment outcomes t he specific secondary stimulus characteristics responsible for producing these effects are unknown. Additionally, It is not known if secondary stimulus characteristics influence

PAGE 31

31 other assessments of deviant sexual interest (e.g., pref erence assessments of deviant sexual interest). Study 2 will evaluate preference within age and gender categories using paired stimulus preference assessments

PAGE 32

32 Figure 2 1. Assessment results for participant 6 s female 6 7 assessment Results are separ ated by test stimulus.

PAGE 33

33 Figure 2 2 Assessment een assessment Results are separated by test stimulus.

PAGE 34

34 Figure 2 3 Assessment adult assessment Re sults are separated by test stimulus.

PAGE 35

35 Figure 2 4 Assessment 9 assessment. Results are separated by test stimulus. The y axis for the top right, bottom left, and bottom right panels were modified to aid visual anal ysis.

PAGE 36

36 Figure 2 5 Assessment assessment. Results are separated by test stimulus.

PAGE 37

37 Figure 2 6 Assessment assessment. Results are sepa rated by test stimulus.

PAGE 38

38 Figure 2 7 Assessment assessment. Results are separated by test stimulus.

PAGE 39

39 Figure 2 8. Assessment 9 assessment. Results are sep arated by test stimulus.

PAGE 40

40 Figure 2 9 Assessment 7 assessment. Results are separated by test stimulus.

PAGE 41

41 Figure 2 10 Assessment assessment. Results are separated by t est stimulus.

PAGE 42

42 Figure 2 11 Assessment 9 assessment. Results are separated by test stimulus.

PAGE 43

43 Figure 2 12 Assessment assessment. Results are separated by test stimu lus.

PAGE 44

44 Figure 2 13 Assessment results for participants that showed arousal to all test stimuli within specific age and gender categories.

PAGE 45

45 Figure 2 1 4 Representative age and gender assessments in which sexual arousal was not observ ed

PAGE 46

46 Tabl e 2 1. Demographic characteristics for participants Participant Age IQ Diagnosis Offense 1 51 51 Mild MR Sexual battery on a child 6 years of age, Lewd/Lascivious assault on a child < 16 years old 2 34 62 Mild MR Sexual assault/molestation of a minor 3 41 56 Mild MR Sexual battery on a child under 12 years old 4 40 57 Mild MR 6 counts sexual battery on a child, 1 count sexual performance on a child 5 50 59 Mild MR Lewd and lascivious acts on a child 6 33 65 Mild MR Sexual battery on a child u nder 12 years old 7 28 55 Mild MR Sexual battery on a child under 12 years old 8 39 59 Mild MR Lewd and lascivious acts on a child, attempted kidnapping

PAGE 47

47 Table 2 2 PPG Assessment results for Study 1. Cases in which a stimulus produced a differentiate d outcome are listed in the corresponding cell (e.g., 3) Cases in which a stimulus produced a differentiated low magnitude outcome are listed in the corresponding cell with an asterisk next to the stimulus number (e.g., 3*). Cases in which no stimuli prod uced arousal within an age and gender cell. Female Stimuli Male Stimul i Participant K 6 7 8 9 Teen Adult K 6 7 8 9 Teen Adult 1 1*, 2, 3 1,2,3 1,2,3 1 3 1,2,3 NO NO NO NO / 2 NO NO NO 3 2,3 NO NO NO NO NO 3 2*, 3* 1,2,3 1,3 1,3* 1,2,3 NO NO 1* / / 4 NO NO NO NO NO / / / / / 5 NO NO NO NO NO NO N O NO NO NO 6 NO 3 NO NO NO NO NO NO NO / 7 NO NO NO NO 3* NO NO NO NO / 8 1,2,3 1,2,3 1,2,3 NO NO 2,3 1,3 1,3 NO NO

PAGE 48

48 CHAPTER 3 STUDY 2 Study 1 demonstrated that secondary stimulus characteristics must influence PPG assessment outcomes. However, the degree to which secondary stimulus characteristics influence outcomes of other assessments of deviant sexual interest is unknown. As previously noted, Reyes (2008) used a pa i red stimulus preference assessment (PSPA) to evaluate the age and ge nder preferences for five child sex offenders with ID R esults showed that for 4 out of 5 participants the preference assessment matched the results of a previously conducted PPG assessment. Though preliminary, the results suggest ed that preference assess ments are a viable and less invasive method of obtaining measures of sexual interest. Additionally, the results of the study showed that there was a high degree of correspondence between physiological (as measured via PPG) and operant (as measured via PSPA ) aspects of sexual interest. Reyes (2008) evaluated preference across age and gender categories. However, t he stimuli used in the study were noted to be a limitation. The stimuli were designed to assess age and gender preferences, but they also differed across a number of secondary stimulus features The purpose of the current study was to evaluate whether secondary stimulus features influence the outcomes of preference assessments using commercially available stimuli (still photos of stimuli used in Stud y 1). This was accomplished by evaluating multiple stimuli for each age and gender category that differed across secondary stimulus characteristics. If stimulus characteristics other than age and gender do not influence the outcomes of preference assessmen ts it would be expected that all stimuli within an age and gender category would be equally preferred.

PAGE 49

49 However, if secondary stimulus features do influence preference assessment outcomes, it would be expected that hierarchies of preference would be shown within age and gender categories. Method Participants. Participants 1, 2, and 3 were included. These individuals were selected because several of their PPG assessments in Study 1 showed differentiated arousal within age and gender categories. Participant 1 evaluation included assessments of female 6 7, female 8 9, and f emale teen. Stimuli Stimuli used in Study 2 were digital stills captured from the commercially available video series (Sex Offender Assessment Series, Northwest Media Inc.) used in Study 1. One picture was captured from each of the 30 test stimuli. Each picture measure d ed a standing model, facing forward. Paired Stimulus Preference Assessment The preference assessments were c onducted using a computer program written in Visual Basic. The programmed functionality was based on the descr iption provided by Reyes (2008). During each session, each stimulus was paired with every other stimulus twice in a randomized order. Prior to the start of each session the computer program recited the following instruction (from Reyes) : bout to begin. When it starts, you will see two pictures. What I want you to do is to think about which one you like more and use the mouse to click on the picture. When you select a picture, you will be able to look at it for 5 seconds, and then it will g o away. After a little while, two new pictures will appear and I want you to do the same thing, and pick which one you like more. It is important that you really pick the one you like more. You will not get in trouble for picking any of the pictures. When you are

PAGE 50

50 ready to start, click the green start button or click the other button to see a On each trial, two stimuli were presented on the computer screen. The participant was instructed to select one of the stimu li. Once the participant selected the stimulus he was allowed to view it for 5 s. If the participant did not select either stimulus within 5 s of presentation, the computer program verbally prompt ed the participant to make a selection. If the particip ant did not make a selection following the prompt, the computer program, again, prompted the participant to make a selection it should be noted that th e prompt, although programmed never occurred. Following the viewing period, the selected stimulus was re moved from the sc reen, and the next trial was initiated. This assessment was completed five times for each participant (once per day). Data a nalysis. All preference assessment data were analyzed as the percentage of opportunities each stimulus was selected Results Figure 3 1 shows assessment results for all participants. Preference assessment data are plotted on the y axis as the average percentage of opportunities each stimulus was selected across the entire assessment (horizontal bar), and the percentag e of opportunities a stimulus was selected each session (black dots). Preference assessment data are plotted on the x axis with the highest preferred stimuli on the left end of the axis and the least preferred on the right end of the axis. The ordering of the stimuli is based on graphing techniques that are typically used in the preference assessment literature (e.g., Fisher et al., 1992 ; Hanley, Iwata, & Roscoe, 2006; Roscoe, Iwata, & Kahng, 1999 ). This method of data display was chosen because it allows f or the visual analysis of assessment means, variability across sessions, and preference

PAGE 51

51 hierarchies. The top panels show assessment results for Participant 1. The top left panel show s female kindergarten assessment outcomes and the top right panel show s f emale teen assessment outcomes. Preference assessment results show clear hierarchies of preference within both the female kindergarten and female teen categories. The FK.2 stimulus was identified as the most preferred stimulus in the female kindergarten pr eference assessment, followed by the FK.3 stimulus, and the FK.1. The FTeen.3 stimulus was identified as the most preferred stimulus in the fe male teen preference assessment followed by the FTeen.1 stimulus, and the FTeen.2 stimulus. The middle panels of Figure 3 1 show assessment results for Participant 2. The middle left panel show s female teen assessmen t outcomes, and the middle right panel show s female adult assessment outcomes. Preference assessment results show clear hierarchies of preference within both the female teen and female adult categories. The FTeen.3 stimulus was identified as the most preferred stimulus in the female teen preference assessment followed by the FTeen.1 stimulus, and the FTeen.2. The F Adult.2 stimulus was identified as the m ost preferred stimulus in the female adult preference assessment, followed by the FAdult .1 stimulus, and the FAdult.3 stimulus. The bottom panels of Figure 3 1 show assessment results for Participant 3. The bottom left panel show s female 6 7 assessment out c omes, the bottom middle panel show s female 8 9 assessmen t outcomes, and the bottom right panel show s female teen assessment outcomes. Preference assessment results show a slight hierarchy of preference in the F6 7 category, and clear hierarchies of prefer ence within the female 8 9 and female t een categories. The F6 7.1 stimulus was identified as the most preferred

PAGE 52

52 stimulus in the female 6 7 preference assessment followed by the F6 7.2 stimulu s and the F6 7.3 The F8 9 .2 stimulus was identified as the mos t preferred stimulus in the female 8 9 preference assessment, followed by the F8 9 .1 stimulu s and the F8 9 .3 stimulus. The FTeen.3 stimulus was identified as the most preferred stimulus in the female teen preference assessment followed by the FTeen.1 sti mulus, and the FTeen.3 stimulus. Although not depicted in the figure, the degree to which the results of preference assessments corresponded with the results of PPG assessments varied. In two cases (Participant 1: female teen; Participant 2: female teen) the results of the preference assessments corresponded exactly with the results of the PPG. In three cases (Participant 1: female teen; Participant 2: female teen, female adult), the most preferred stimulus in the preference assessment also produced the highest level of arousal in the PPG. In four cases (Participant 1: female teen, female adult; Participant 2: female teen; Participant 3: female teen), the least preferred stimulus in the preference assessment also produced the lowest level of arousal in th e PPG. Discussion The purpose of the current study was to evaluate if stimulus features other than age and gender (i.e., secondary stimulus characteristics) influence outcomes of preference assessments Overall, the results demonstrate that sec ondary stimu lus characteristics must influence the outcomes of preference assessments T he fact that hierarchies of preference were obtained in 7 out 7 assessments provides support for their influence given that age and gen der were held constant As noted by Reyes ( 2008), preference assessments and PPG assessments provide fundamentally different measures of sexual interest. Preference assessments

PAGE 53

53 require an individual to emit an operant response, selecting a stimulus, whereas PPG hysiological response to a presented stimulus. Given that Reyes reported t hat 4 out 5 participants preference assessment outcomes corresponded to PPG outcomes, the degree of correspondence between preference assessments and PPG in the current study could b e viewed as surprisingly low. However, it should be noted that the current study evaluated preference within age and gender categories, whereas Reyes evaluated preference across age and gender categories It is possible that the high degree of corresponden ce that was shown in the Reyes study was do to the fact that comparisons were made across more salient stimulus categories. It is important to note a limitation of the current study parallel to the limitation of Study 1 Although it appears that secondar y stimulus characteristics must influence preference assessment outcomes t he specific stimulus characteristics responsible for producing these effects are unknown. Study 3 is an exploratory evaluation of the influence of specific secondary stimulus charac teristics on preference of ID child sex offenders.

PAGE 54

54 Figure 3 1 Preference a ssessment results for assessment results are displayed in the middle pa assessment results are displayed in the bottom panels.

PAGE 55

55 CHAPTER 4 STUDY 3 As demonstrated in Studies 1 and 2, secondary stimulus characteristics must influence both PPG and preference assessment results. A limitation of Studies 1 and 2 is that the specific stimulus features responsible for producing the differentiated outcomes within age and gender categories are unknown because they were not systematically manipulated Recent technological advances may help address this limitat ion. Recent research has incorporated existing sex offender assessments (e.g., PPG) and digitally rendered avatars. The results of preliminary studies (Renaud, Rouleau, Granger, Barsetti, & Brouchard, 2002; Renaud et al., 2005; Renaud et al., 2009 Renaud et al., 2010 ) suggest that the use stimulus sets comprised of lifelike avatars are a viable alternative to stimulus sets comprised of pictures and videos of real people. The use of lifelike avatars can address several limitations of stimulus sets comprised of videos and pictures of real people. First, their use addresses an ethical concern of using videos and pictures of real people for the purpose of sex offender assessments ( Laws and Gress, 2004; Vollmer, Reyes, & Walker, 2011 ) Second, their use opens up the possibility of developing stimulus sets in which specific stimulus features (e.g., clothing type) can be modified, allowing the effects of other stimulus features to be empirically evaluated while hold ing the age and gender constant (Vollmer, Reyes, & Walker, 2011). As noted previously, there is a large body of research that has systematically evaluated the influence of physical characteristics on ratings of physical attractiveness by non offenders. T his research ha s shown that ratings of physical att ractiveness are

PAGE 56

56 greatly influenced by a number of physical characteristics such as waist to hip ratio ([WHR], Furnham & Reeves, 2006; Furnham, et al., 2005; Perilloux, et al. 2010; Singh, 1993; Singh & Young, 1995 ) and breast size (Schutwohl, 2006; Singh & Young, 1995 ; Zelazniewicz, 2011). N o research however, has evaluated the effects of clothing type on ratings of attractiveness and only one study has evaluated the effects of hair length on ratings of attractiveness ( Swami et al., 2008), t wo stimulus f eatures that were noted to vary within age and gender categories in Studies 1 and 2 and in a host of previous studies and clinical applications Further, these prior evaluations of attractiveness may or may not relate to preferences displayed by child sex offenders with ID. The purpose of Study 3 was to conduct an exploratory evaluation of the effects of specific stimulus features on outcomes of preference assessments evaluating sexual interest. Study 3A evaluate d the independent effects of clothing type WHR breast size, and hair length on preference These stimulus features were selected for two reasons: a) some are reported to be attractiveness variables in the research literature, and b) they are all commonly varied in an uncontrolled fashion in comme rcially available stimulus sets used in research and clinical assessment. Study 3B evaluate d relative preference of stimulus features that were shown to produce hierarchies of preference in Study 3A. Study 3A Method Participants Participants 1, 2, and 4 were included in the current study. These individuals were referred by the facilit s treatment team due to their alleged history of sexual offen s es.

PAGE 57

57 Stimuli Stimuli for the clothing type evaluation consist ed of digitally rendered avatars developed usi ng Evo lver ( www.evolver.com ) Figure 4 1 displays the stimuli that were used for clothing type and breast size evaluation s For the clothing type evaluation the only stim ulus characteristic that differed across stimul i is the type of bathing suit (Bikini, Tankini, and One Piece). Stimuli for the breast size evaluation also consist ed of pictures of digitally rendered avatars developed using Evo lver T he only stim ulus characteristic that differed across stimuli was breas t size. Evolver allowed for the adjustment of breast size along a continuum. The small breast size was based upon the extreme low end of the continuum, the medium breast size was based on the middle of the continuum, and the large breast size was based upo n the high end of the continuum. Figure 4 2 displays the stimuli that were used for the WHR and hair length evaluations. Stimuli for the WHR evaluation consist ed of pictures of digitally rendered avatars developed using Evolver a nd Adobe Photoshop 7.0. T he only stim ulus characteristic that differed across stimuli wa s WHR. Specifically, WHRs of .60, .75, and .90 were evaluated. The se WHR values were selected because they represent a range of values that have been evaluated in previous research (e.g., Furnham & Reeves, 2006; Perilloux, et al., 2010; Singh & Young, 1995). Stimuli for the hair length evaluation also consist ed of pictures of digitally rendered avatars developed using Evolver an d Adobe Photoshop 7.0. The only stim ulus characteristic that differed across stimuli is hair length. Specifically, t he short hair avatar s hair was 19 pixels shorter than the medium hair avatar s hair, and the long hair avatar s hair was 19 pixels longer than the medium avatar s hair. Across all evaluations e ach picture meas ured

PAGE 58

58 Paired Stimulus Preference Assessment and Pre session Instruction The methods described in Study 2 were used to complete study 3A. The stimuli evaluated were the only difference s between the studies. Data a nalysis. All data w ere analyzed as percentage of opportunities each stimulus was selected. Results and Discussion Figure 4 3 shows the results for the clothing type evaluation Data are displayed as average percentage of opportunities each stimulus was selected across the e ntire assessment (bar), and the percentage of opportunities a stimulus was selected in each session (black dots). Results for Participant 1 are shown in the top panel, results for Participant 2 are shown in the middle panel, and results for Participant 4 a re shown in the bottom panel. The results show that all participants had clear hierarchies of preference for specific types of clothing. The results for Participants 1 and 2 show that the Bikini stimulus was selected most often, f ollowed by the Tankini sti mulus, and then followed by the One Piece stimulus. The results for Participant 4 show that the Tankini was selected most often, followed by the Bikini stimul us, and then followed by the One Piece stimulus. Figure 4 4 shows the results for the breast size evaluation. The results show that two out of the three participants had clear hierarchies of preference for breast size. The (top panel) show no clear hierarchy of preference for breast size. The results of Participa (middle panel) show a clear hierarchy of preference with the Large Breast stimulus selected most often, followed by the Medium Breast stimulus, and the Small B reast stimulus. The results of Participant (bottom panel) also s hows a clear hierarchy of preference with the

PAGE 59

59 Medium Breast stimulus selected most often, followed by the Large Breast stimulus, and the Small Breast stimulus. Figure 4 5 shows the results for the WHR evaluation. Results for Participant 1 are shown in the top panel, results for Participant 2 are shown in the middle panel, and results for Participant 4 are shown in the bottom panel. The results show that none of the participants exhibited a clear hierarchy of preference for WHR. Figure 4 6 shows the result s for the hair length evaluation. Results for Participant 1 are shown in the top panel, results for Participant 2 are shown in the middle panel, and results for Participant 4 are shown in the bottom panel. The results show that none of the participants exh ibited a clear hierarchy of preference for hair length. The stimulus characteristics that were evaluated in the current study were chosen because of the attractiveness because th ey represent salient differences in the stimul i used in Studies 1 and 2 and prior studies and clinical applications, or both The results of the current study demonstrate that specific stimulus characteristics can influence the results of preference assess ments. The methods described in the current study represent a potentially important step in the process of identifying additional stimulus characteristics that can influence the results of age and gender a ssessments of sexual interest. Additionally, these results add to the literature showing that stimulus sets comprised of digitally rendered avatars are viable for the assessment of sexual inter e st or at least visual stimulus preferences. The results of the clothing type and breast size evaluations show tha t specific types of clothing and breast size influence d preference s of child sex offenders

PAGE 60

60 diagnosed with ID. However the results of the WHR and hair length evaluations evaluation show that child was not influenced by WHR and hai r length at least for these three individuals Given the relatively large literature base showing that WHR it is surprising that the WHR evaluation produced no c lear hierarchies of preference. Ho wever, the lack of effect in the WHR and hair length evaluations might be due to the limited numb er of participants. Excluding the current study, no research has evaluated the influence of clothing ratings of physical attractiveness, and on ly one previous study ( Swami et al., 2008) has evaluated the influence of hair length on males attractiveness. Furthermore, no previous research has evaluated how these variables influence preference of sex offenders either with or wit hout ID. Given the dearth of research evaluating these physical characteristics, future research should not only evaluate the influence of these physical characteristics with additional sex offender populations, but also evaluate their influence in non off ender populations. T he current study only evaluated the independent effects of these stimulus characteristics, leaving open the possibility that one of these stimulus characteristics might influence preference to a greater degree than the other. The pur pose of Study 3B was to evaluate the relative influence of clothing type an d breast size on preference, as manipulations of these physical characteristics were shown to produce hierarchies of preference when evaluated independently.

PAGE 61

61 Study 3B Method Parti cipants Because they had participated in Study 3A and showed hierarchies of preference for clothing type and breast size Participants 1, 2, and 4 were included in the current study. Stimuli Figure 4 7 shows the stimuli that were used in the current stu dy. Stimuli used in the in Study 3A were modified for the current study. Specifically, s timuli consisted of 9 avatars developed using Evolver ( www.evolver.com ) and Adobe Photoshop 7.0 representing three different type s of clothing (Bikini, Tankini, and One Piece swimming suites) and three breast sizes (Small, Medium, and Large). Each picture measure d Paired Stimulus Preference Assessment and Pre session Instruction The methods described in Stud y 2 were used to complete S tudy 3B The number of stimuli and the stimuli evaluated were the only differences. Data a nalysis. All data were analyzed as percentage of opportunities each stimulus was selected. Results and Discussion Figure s 4 8 and 4 9 sho w the results for study 3B Data are displayed as average percentage of opportunities each stimulus was selected across the entire assessment (bar), and the percentage of opportunities a stimulus was selected each session (black dots). Figu re 4 8 shows the data categorized by b reast size and Figure 4 9 shows the data categorized by clothing type. R esults for Participant 1 (top panel of Figure 4 8 ) show that clothing type was the stimulus characteristic that primarily influenced preference. Specifically, Par ticipant 1 selected avatars wearing b ikinis most often, this

PAGE 62

62 pattern of responding was replicated across breast sizes. However, as seen in the top panel of Figure 4 9, breast size influenced preference to a degree as well. Within each clothing type categor y the Large Breast stimulus was selected most often, although there was some variation and overlap in preference for medium and small breast size. Results for Participant 2 (middle panel s of Figures 4 8 and 4 9 ) show that breast size primarily i nfluenced preference, with the Large B reast stimuli being s elected most often followed by Medium Breast stimuli, and Small B reas t stimuli. However, clothing type did influence preference. Within each breast size category the Bikini stimulus was selected most often followed by the Tankini stimulus, and the One Piece stimulus. Results for Participant 4 (lower panels of Figures 4 8 and 4 9) show that breast size was the stimulus characteristic that primarily influenced preference. Specifically, Participant 4 selecte d avatars with Large Breast and Medium Breast stimuli more often than Small Breast stimuli. The purpose of Study 3B was to evaluate the relative influence of clothing type and breast size on preference. Or, at the very least the purpose was to demonstrat e a model for such evaluations using secondary stimulus features. Overall the results demonstrate that some individual stimulus characteristics influence preference more than others. However, the effects of individual stimulus characteristics are idiosync ratic. The results for Participant 1 showed that clothing type was a critical determina nt of preference, whereas the results for Participants 2 and 4 showed that breast size was a primary determina nt of choice. These results show that these stimulus charac teristics need to be controlled in evaluations of age and gender preference and suggest the

PAGE 63

63 need for stimulus sets t hat can be easily modified to account for individual preferences (again, suggesting avatars hold promise for this approach)

PAGE 64

64 Fig ure 4 1 Avatar stimuli used in the clothing type and breast size evaluations

PAGE 65

65 Figure 4 2 Avatar stimuli used in the waist to hip ratio and hair length evaluations.

PAGE 66

66 Figure 4 3 R esults of the clothing type evaluation

PAGE 67

67 Figure 4 4 R esults o f the breast size evaluation

PAGE 68

68 Figure 4 5 R esults of the waist to hip ratio evaluation

PAGE 69

69 Figure 4 6 R esults of the hair length evaluation

PAGE 70

70 d Figure 4 7 Stimuli used in study 3B

PAGE 71

71 Figure 4 8 R esults of breast size and clothing type comparison grouped by breast size

PAGE 72

72 Figure 4 9 R esults of breast size and clothing type compariso n grouped by clothing type

PAGE 73

73 CHAPTER 5 GENERAL DISCUSSION The present series of studies examined whether secondary stimulus characteristics can influence assessments of deviant sex ual interest of ID sex offenders In general, the results show that secondary stimulus characteristics do, in fact, influence assessments of sexual interest of ID sex offenders The purpose of Study 1 was to evaluate whether arousal occur red differentially in the presence of stimuli within the same age and gender category that differed across secondary stimulus characteristics. The results showed by exclusion that in some cases secondary stimulus characteristics must influence PPG outcomes. However, the spe cific secondary stimulus characteristics responsible for producing these effects are unknown. To this end, future research c ould incorporate digitally rendered avatars, which allow for the modification of secondary stimulus characteristics, into PPG assess ments. The purpose of Study 2 was to evaluate whether preferences would be shown in the presence of stimuli within the same age and gender category that differed across secon dary stimulus characteristics. The results showed that in all cases a hierarchy o f preference was identified within age and gender categories This finding s uggest s that secondary stimulus characteristics do influence preference assessments of deviant sexual interest. Future research should compare how specific secondary stimulus chara cteristics influence the results of different types of assessment s The purpose of Study 3 was to conduct an exploratory evaluation of the effects of specific stimulus features on outcomes of preference assessm ents evaluating sexual interest, by using digi tally rendered avatars as test stimuli. Study 3A evaluated the independent effects of four secondary stimulus characteristics (breast size, clothing

PAGE 74

74 type, WHR, hair length). The results of Study 3A showed that breast size and clothing type influence d pref erence, whereas hair length and WHR d id not appear to influence preference. Although the results of Study 3A showed that preference was influenced by breast size and clothing type, it is possible that one of these stimulus characteristics might have influe nce d preference to a greater degree than the other. The purpose of Study 3B was to evaluate the relative preference of breast size and clothing type. The results of Study 3B show that some individual stimulus characteristics influence d preference more than others; however, the effects of individual stimulus characteristics we re idiosyncratic. Some potential limitations of Study 3 should be noted. First the lack of effect in the WHR and hair length evaluations might be due to the limited number of particip ants. Future research should replicate the WHR and hair length evaluations with additional participants. Second, a limited number of stimulus chara cteristics were evaluated in Study 3A The attractiveness literature suggests some other secondary stimulus c haracteristics that could influence assessments of sexual interest (e.g., face symmetry, skin tone) Future research should evaluate the effects of additional secondary stimulus characteristics. Third, the base stimulus used in Studies 3A and 3B was a fema le adult. It is possible that the effects of some secondary stimulus characteristics are only evident when comparisons are made within age and gender categories. The degree to which the age and gender of the presented stimulus and secondary stimulus charac teristics interact to influence preference remains unknown. Future research is needed to evaluate the extent to which secondary stimulus characteristics influence other age and gender categories and comparisons across age and gender categories.

PAGE 75

75 Finally, so me of the specific stimulus features evaluated in the current study were not applicable to other age and gender categories (e.g., breast size), suggesting that the generality of the findings of the study are limited. However, it should be noted that adult female stimuli are commonly included in assessments of deviant sexual interest as a test stimulus, and serve a critical role as a comparison in assessments of deviant sexual interest. The use of digitally rendered avatars potentially represents a signifi cant development in the assessment of deviant sexual interests; however, their development could also prove to be significant contribution to the treatment of child sex offenders. The use of digitally rendered avatars allows for the development of an indiv idualized profile of sexual interest (Renaud et al., 2009; Renaud et al., 2010). For instance, Renaud et al. (2010) demonstrated that digitally rendered avatars based upon verbally reported preferences are possible to develop; however, the clinical utility of such stimuli has yet to be evaluated and needs to be evaluated in future research. Furthermore, these stimuli could be incorporated into Individualized treatment plans by a llowing for a more tailored treatment by basing the stimuli used in treatment on individualized preferences. Future research could evaluate the efficacy of these stimuli in a variety of treatments such as teaching individuals to discriminate key physical characteristics of pubescent and pre pubescent individuals, or in other more trad itional behavioral therapies (e.g., covert sensitization, biofeedback). Although the current series of studies evaluated the extent to which secondary stimulus characteristics influence assessments of sexual interest with ID child sex offenders, it is poss ible that secondary stimulus characteristics could influence

PAGE 76

76 assessments of sexual interest of non ID child sex offenders. Future research should evaluate the degree to which secondary stimulus characteristics influence assessments of sexual interest of no n ID sex offenders.

PAGE 77

77 LIST OF REFERENCES Fisher, W., Piazza, C. C., Bowman, L. G., Hagopian, L. P., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying reinforcers for persons with severe and profound disa bilities. Journal of Applied Behavior Analysis, 25 491 498. Furnham, A., Petrides, K. V. & Constantinides, A. (2005). The effects of body mass index and waist to hip ratio on ratings of female attracti veness, fecundity, and health. Personality and Indivi dual Differences, 38, 1823 1834. Furnham, A., & Reeves, E. (2006). The relative influence of facial neoteny and waist to hip ratio on judgments of female attractiveness and fecundity. Psychology, Health & Medicine, 11 129 141. Hanley, G. P., Iwata, B. A., & Roscoe, E. M. (2006). Some determinants of changes in preference over time. Journal of Applied Behavior Analysis, 39 189 202 Hanson, R. K. & Morton Bourgon, K. (2004). Predictors of sexual recidivism: An updated meta analysis. (Cat. No.: PS3 1/20 04 2E PDF). Public Works and Government Services Canada (ISBN: 0 662 36397 3) Kahng, S., & Iwata, B. A. (1999). Correspondence between outcomes of brief and extended functional analyses. Journal of Applied Behavior Analysis, 32, 149 159. Lalumiere, M. L., & Harris, G. T. (1998). Common questions regarding the use of phallometric testing with sexual offenders. Sexual Abuse: Journal of Research and Treatment, 10(3), 227 237. Laws, D.R., & Gress, C.L. Z. (2004). Seeing things differently: The viewing time alt ernative to penile plethysmography. Legal and Criminological Psychology (2004), 9 183 196. Laws, D. R., & Osborn, C. A. (1983). How to Build and Operate a Behavioral Laboratory to Evaluate and Treat Sexual Deviance. In J. Greer, G. & I. Stuart, R. (Eds.) The Sexual Aggressor: Current Perspectives on Treatment (pp. 293 335). New York, NY: Van Nostrand Reinhold Company Inc. Laws, D.R. (2009). Penile Plethysmography: Strengths, Limitations, Innovations. In D. Thornton & D. R. Laws (Eds.), Cognative Approac hes to the Assessment of Sexual Interest in Sexual Offenders (pp. 7 29). New York, NY: John Wiley & Sons Ltd. Murphy, W. D., Coleman, E. M., & Haynes, M. R. (1983). Treatment and Evaluation Issues with the Mentally Retarded Sex Offender. In J. Greer, G. & I. Stuart, R. (Eds.), The Sexual Aggressor: Current Perspectives on Treatment. (pp. 22 41). New York: Van Nostrand Reinhold Company Inc.

PAGE 78

78 Perilloux, H. K., Webster, G. D., & Gaulin, S. J. (2010). Signals of genetic quality and material investment capacity: The dynamic effects of fluctuating asymmetry and waist to Social Psychology and Personality Science, 1, 34 42. Renaud, P., Rouleau, J. L., Granger, L., Barsetti, D. P., & Bouchard, S. (2002). Measurin g sexual preferences in virtual reality: A pilot study. Cyberpsychology and Behavior, 5 ,1 9. Renaud, P., Rouleau, J. L., Proulx, J., Trottier, D., Goyette, M., Bradford, J. P., Fedoroff, P., Dufreasne, M., Dassylva, B., Cote, G., & Bouchard, S. (2010). Vir tual characters designed for forensic assessment and rehabilitation of sex offenders: standardized and made to measure. Journal of Virtual Reality and Broadcasting, 7. Renaud, P., Proulx, P., Rouleau, J.L., Bouchard, S., Madrigrano, G., Bradford, J., & Fed oroff, P. (2005). The recording of observational behaviors In virtual immersion: A new Research and Clinical Tool to Address the Problem of Sexual Preferences with Paraphiliacs. Annual Review of Cyber Therapy and Telemedicine, 3 85 92. Renaud, P., Charti er, S., Rouleau, J.L., Proulx, J., Decarie, J., Trottier, D., Bradford, J. P., Fedoroff, P., & Bouchard, S. (2009). Gaze behavior nonlinear dynamics assessed in virtual immersion as a diagnostic index of sexual deviancy: preliminary results. Journal of Vir tual Reality and Broadcasting, 6, 1 11. Reyes, J. R., Vollmer, T. R., Sloman, K. N., Hall, A., Reed, R., Jansen, G., et al. (2006). Assessment of deviant arousal in adult male sex offenders with developmental disabilities. Journal of Applied Behavior Anal ysis, 39 173 188. Reyes, J.R. (2008). Assessment of sex offenders with developmental disabilities. (Doctoral Dissertation). Retrieved from ProQuest Dissertations and Thesis. (Accession Order No. ATT 3334500). Reyes, J. R., Vollmer, T. R., & Hall, A. (2011 a). Replications and extensions in arousal assessment for sex offenders with developmental disabilities. Journal of Applied Behavior Analysis, 44 369 373. Reyes, J. R., Vollmer, T. R., & Hall, A. (2011b). The influence of presession factors in the assessm ent of deviant arousal. Journal of Applied Behavior Analysis, 44 707 717. Roscoe, E. M., Iwata, B.A., & Kahng, S. (1999). Relative versus absolute reinforcement effects: Implications for preference assessments. Journal of Applied Behavior Analysis, 32, 47 9 493. Schutzwohl, A. (2006). Judging female figure: A new methodological approach to male attractiveness judgments of female waist to hip ratio. Biological Psychology, 71, 223 229.

PAGE 79

79 Singh, D. (1993). Adaptive significance of female physical attractiveness : Role of waist to hip ratio. Journal of Personality and Social Psychology, 65, 293 307. Singh, D., & Young, R. K. (1995). Body weight, waist to hip ratio, breasts, and hips: Role in judgments of female attractiveness and desirability for relationships. Et hology and Sociobiology, 16 483 507. Swami, V., Furnham, A., & Joshi, K. (2008). The influence of skin tone, hair length, and hair colour on ratings of women's physical attractiveness, health and fertility. Scandinavian Journal of Psychology, 49, 429 437 Vollmer, T.R., Reyes, J., & Walker, S.F. (2011). Behavioral assessment and Intervention for sex offenders with intellectual and developmental disabilities. In J. Luiselli (Ed), The Handbook of High Risk Challenging Behaviors: Assessment and Interventi on. Paul H. Brookes. Zelazniewicz, A.M., & Pawlowski, B. (2011). Female breast size attractiveness for men as a function of sociosexual orientation (restricted vs. unrestricted). Archives of Sexual Behavior, 40, 1129 1135.

PAGE 80

80 BIOGRAPHICAL SKETCH Prior t o attending the University of Florida (UF) Stephen attended the Behavior Analysis program at the Universit y of North Texas (UNT), where he received a Bachelor of Science Degree in May of 2005 and a Master of Science Degree in May of 2009. A s a part of the master s degree requirements at UNT he c ompleted an internship at the Seguin Unit under the direction of Dr. Timothy Vollmer. Following this intern ship experience, he applied to the doctoral program in Behavior Analysis at UF, and was admitted in the fall o f 2008. While at UNT his research focused on the assessment and treatment of severe behavior disorders. Working as a graduate research assistant, under the di rection of Dr. Richard Smith, he served as the primary student researcher on projects investigatin g the immediate and subsequent effects of function based treatments and parametric evaluations of pre session attention on problem behavior reinforc ed by attention. He also participated in research under the direction of other faculty at UNT, including Dr. Rosales, and Dr. Jesus Rosales Ruiz. Since being accep ted to UF in the fall of 2008, he has served as the graduate research assistant and Co PI for the Seguin Behavioral Services contract, under the direct ion of Dr. Timothy Vollmer. His research focused on assessment and treatment of intellectually disabled (ID) sex offenders. In addition to his dissertation research, he developed protocols evaluating the effects of unpredictable supervision as a treatment for high risk behavior in sex offenders with ID and the ability of sex offenders with ID to differentially suppress penile erections. teaching experience was in the fall of 2003 as a Teaching Assistant (TA) for an introductory beh avior an alysis course at UNT. He continued as a TA for this

PAGE 81

81 course until the spring of 2006 when he was offered the position of Teaching Fel low (TF) for the same course. He continued to serve as a TF for this cours e until the fall of 2007 when he was offered TA po sitions for a graduate course, Introduction to Behavior Analysis, and an advanced undergraduate course, Behavior Analysis Capstone Course, taught by Dr. Richard Smith. During his time at UF he has taught Applied Behavior Analysis and Introduction to Psycho logy, served as a TA for a large online section of Introduction to Psychology, and tutored university athletes in Introduction to Psychology. Additionally, he was recently nominated for the Graduate Student Teaching Award at UF. Stephen has also worked in a variety of clinical settings. During his time at UNT he gained clinical experience in traumatic brain injury, OBM, and early autism intervention. However, the majority of his clinical experience focused on the assessment and treatment of severe behavior disorders. Working under the di rection of Dr. Richard Smith, he volunteered at the Behavior Analysis Resource Center (BARC) at the Denton State School as an undergraduate student and continued to work there until he graduated from UNT. His responsibilitie s at BARC included conducting functional assessments of severe behavior disorders, conducting preference assessments, developing behavior intervention plans, and serving as a member of both the local review committee and client interdisciplinary teams He continued to work in a similar capacity at t he Seguin Unit. Currently he serves as a consultant for the Davenport Independent School District and Gilchrist County School District where he provides behavior analytic services.