Juvenile Sexual Offenders: Analysis

1
Juvenile Sexual OffendersAnalysis Management
of Risk
The Field Center for Childrens Policy, Practice
Research University of Pennsylvania 05 December
2006

• Robert Prentky
• Research Department
• Justice Resource Institute
• Bridgewater, MA
• rprentky_at_jri.org

2
K.R.U.
3
When Things Go (really) Awry
4
Immanuel Kant, 250 years earlier

• From the crooked timber of humanity,
• No straight thing can be made

5
Invisible, Insidious, Incurable

• Out of a benign sense of obligation to prevent
  the worst of the worst from
  re-occurring, we may, unwittingly, unleash demons
  from which no straight thing can be made
• Twin Curses
• Protracted severe abuse neglect
• Emotional Detachment

6
Affective Risk Factors

• Emotional disconnectedness / loneliness
• Low self-esteem vulnerability to feeling
• humiliated
• Absence of empathy
• Inability to express positive feelings
• Inability to manage strong negative feelings
  

7
Emotional Detachment
8
Perry, Pollard, Blakley, Baker, Vigilante
(1995)

• The child who has been emotionally neglected
  early in life will exhibit profound attachment
  problems which are extremely insensitive to any
  replacement experiences later, including therapy.
  Examples of this include feral children,
  children in orphanages and, often, the
  remorseless, violent child, (p. 276-277).

9
Amid the extraordinary diversity and
heterogeneity of expressions of sexually coercive
behavior, there appears to be only 1 least common
denominatorthe relative inability to develop
and sustain healthy attachments with peers
10
Disordered Attachment(Levy Orlans, 1999)

• Behavior Oppositional defiant, impulsive
  destructive, hyperactive, self-destructive
• Emotions Intense anger and/or depression,
  fearful anxious
• Thoughts Negative core beliefs about ones self
  and ones relationship with others,
  attention
• learning problems
• Relationships Untrusting manipulative, unable
  to give
• or receive
  genuine affection, poor
• 
  perspective-taking
• Moral Lack of empathy compassion, little
• capacity for
  remorse, asocial or antisocial
• values.

11
Critical Outcomes of Disordered Attachment

• Impulsive (inability to regulate or control
  emotions)
• Intense emotions (anger, depression or anxiety)
• Profound lack of trust
• Incapacity for empathy perspective-taking
• Incapacity for reciprocal love
• Relationships are superficial and loveless

12
What is the Magnitude of this Problem?
13
Lyons-Ruth, 1996

• Of the 1 million substantiated cases of serious
  abuse and neglect in the United States each year,
  about 800,000 of those children have severe
  attachment disorder

14
Gallup, Moore, Schussel, 1995

• The actual number of cases of serious abuse and
  neglect may be 10-16 times higher (8 million -
  12.8 million), increasing the number of children
  with severe attachment disorder to 6,400,000 -
  10,240,000

15
The Nexus of Caregiver Instability Sexual
Abuse(Prentky, Knight, Sims-Knight, Straus,
Rokous, Cerce, 1989)

• A study of 82 adult sex offenders
• File Coded Variables
  Self-Report
• Physical Abuse longest
  time spent with caregivers
• Neglect
  total of caregivers
• Sexual Abuse total
  of changes in caregivers
• Sexual Deviation in Family total time
  with bio mother
• 
  total time with bio father
• 
  total time spent with caregivers
• 
  longest time spent in institutions
• 
  total of changes in institutions
  

16
Outcome Variables

• Severity of sexual aggression
• Severity of nonsexual aggression
• Total of sexual offenses
• Total of nonsexual, victim-Inv. offenses

17
ß0.31
Caregiver Inconstancy
Sexual Aggression
ß0.43
Institutional History
General Aggression
ß0.24
Physical Abuse Neglect
Total Offenses Sexual
ß0.-0.26
ß0.32
Sexual Deviation Abuse in Family
Total Offenses Non-Sexual
plt .05, p lt.005, p lt.001
18
Prediction of Sexual Aggression(Using CI SDA)
Low High
Neither Variable Above Mean 77 23
One Variable Above Mean 48.6 51.4
Both Variables Above Mean 12.5 87.5
19
Mother Violence ß.29
Caregiver Instability
Father Violence ß.18
Emotional Abuse ß.30
Physical Abuse ß.22
Sexual Abuse ß.09
20
Health ß.77
School Problems ß.40
Caregiver Instability
Aggression ß.77
Depression ß.59
Juv. Antisocial Behavior ß.83
Ineptness ß.26
Anxiety ß.30
21
Research ProjectAssessment for Safe and
Appropriate Placement (ASAP)
Mission The Massachusetts ASAP Program evaluates
children and juveniles who engage in firesetting
and sexually abusive behaviors. The mandate
Rosenberg Law is to improve the management and
care of these children by the Department of
Social Services (DSS) and to reduce victimization
of other children.
22
ASAP Research Project (Prentky, 2006)

• Only youth with a history of sexual behavior
  problems
• Supported by DSS and NIJ
• Study Timeline
• February, 1998 (ASAP Evaluations Began)
• July, 2001 (Began Data Collection)
• June, 2003 (Phase 1 Completed N250)
• November, 2004 (Phase 2 Completed N720)
• November, 2005 (Phase 3 Completed Stopped
  Data Collection)
• Final sample N 822
• Boys 667 Girls 155

23
Adverse Early Life Experiences Short-Term
Outcome
24
Child Abuse Sexual BehaviorsLegend
SA Severity of Sexual Abuse PA
Severity of Physical Abuse SA PA Severity
of Sexual and Physical Abuse LS Number of
changes in Living Situations LS SA
Living Situation and Severity of Sexual Abuse
LS, SA, PA Living Situation, Severity of
Sexual Abuse, and Severity of Physical
Abuse
25
Child Abuse Sexual Aggression(yellow p lt
.01, .005, or .001)
SA PA SAPA LS LSSA LSSAPA
Obscene Gestures 3.50 8.93 6.11 21.17 19.28 23.36
Violation of Body Space 2.70 2.88 9.49 31.11 21.89 18.22
Sexually Aggressive Remarks 7.38 1.61 11.14 19.94 22.50 17.32
Genital Touching w/o Permission 9.98 2.43 16.18 3.25 23.04 20.72
Pulling Skirt/ Pants Down 6.77 11.53 20.28 13.50 25.65 24.82
Total of Victims 21.66 3.58 9.91 37.94 49.57 41.19
Sexualized Aggression 8.08 3.25 8.06 8.06 21.94 17.61
26
Child Abuse Sexual Aggression(yellow p lt
.01, .005, or .001)
SA PA SAPA LS LSSA LSSAPA
Obscene Gestures 21.17 19.28 23.36
Violation of Body Space 31.11 21.89 18.22
Sexually Aggressive Remarks 19.94 22.50 17.32
Genital Touching w/o Permission 16.18 23.04 20.72
Pulling Skirt/ Pants Down 11.53 20.28 13.50 25.65 24.82
Total of Victims 21.66 37.94 49.57 41.19
Sexualized Aggression 21.94 17.61
27
Child Abuse Paraphilias(yellow p lt .01,
.005, or .001)
SA PA SAPA LS LSSA LSSAPA
Peeping Agemates 25.04
Exposing Agemates 26.75 30.89 26.77
Frottage Agemates 23.92 19.18
Peeping Strangers 27.49 25.00 32.51
Exposing Strangers 14.27 14.74 18.66 23.23
Frottage Strangers
28
Take Home Message

• Sexual abuse and physical abuse, either
  alone or in combination, were not
  significantly correlated with different types of
  sexual abuse.
• However, when changes in living situation was
  added to either or both of the above very
  significant relationships were noted.

29
So Whats the Story with Sexual Abuse ???
30

• A youths history of having been a victim
  of sexual abuse / assault is generally NOT
  predictive of sexual re-offense
• (e.g., Hagan Cho, 1996 Langstrom, 2002
  Rasmussen, 1999 Worling Curwen, 2001)

31
Studies finding sex abuse history predictive of
sexual re-offense in juveniles
STUDY History Present?
Worling Curwen, 2000 no
Langstrom, Grann, Lindblad, 2000 no
Auslander, 1998 no
Sipe, Jensen, Everett, 1998 no
Kenny, Keogh, Seider, 2001 no
Kahn Chambers, 1991 no
Rasmussen, 1999 no
Hagan Cho, 1996 no
Smith Monastersky, 1986 no
32
Studies finding sex abuse history predictive of
sexual re-offense in juveniles
STUDY History Present?
Boyd, 1994 no
Ageton, 1983 no
Schram, Milloy, Rowe, 1991 no
33
Worling Langstrom, 2003

• all of the available data indicate that
  adolescents who have offended sexually and also
  acknowledge a history of child sexual abuse are
  at no greater risk of sexual assault recidivism,
  p.355

34
The Silver Bullet is Tarnished

• Recapitulation Ho of the 1980s
• cycle-of-abuse
• vampire syndrome

35
Abuse-Specific ConsiderationsMorbidity
Factors(Burton, 2000 2003 Prentky, 1999)

• Sexual abuse becomes critical under certain
  conditions or in the presence of other factors
  (Kaufman Zigler, 1987)
• 1. Age of onset
• 2. Duration
• 3. Severity
• 4. Relationship to
  offender

36
Abuse-Specific Victim Age

• Ho juveniles who victimize children are more
  likely to have been sexually abused than those
  who victimize peers
• Awad Saunders, 1991 (20 vs. 4)
• OBrien, 1991 (40 vs. 29)

37
Abuse-Specific Victim Gender

• Ho juveniles who victimize males are more
  likely to have been sexually abused than
  those who victimize females
• Becker Stein (32 male vs. 18 female)
• Worling, 1995 (75 male vs. 25 female)

38
Victim Age / Gender Confound

• Assaults against peers or adults female victims
• Assaults against children roughly 2/3 include
  male victims

39
Treatment Confound?

• Juvenile offenders with known histories of
  sexual abuse more likely to be treated? And thus
  at reduced risk?

40
Developmental Constitutional Considerations

• self-esteem,
• resilience,
• ego strength,
• capacity for
  introspection
• (Friedrich, 1998)

41
Child Sexual AbuseA Kaleidoscopic Risk Predictor

• whereas, Delinquency, reflects an invariant,
  stable pattern of behavior.
• Child Sexual Abuse, reflects idiosyncratic
  patterns of abusive experiences that interact
  with other concurrent life experiences, some
  mitigating and some aggravating, producing highly
  variable degrees of risk

42
To Further Complicate MattersNeurobiology of
Child Abuse
43
Effects of Child Abuse on the Developing Brain

• A wide range of replicated experiments have
  demonstrated that child abuse can cause permanent
  damage to the neural structure and function of
  the developing brain.

44
Abuse-induced Stress

• Persistently high levels of stress trauma
  induced hormones during childhood may permanently
  alter brain development,
• especially the hippocampus

45
Why the Hippocampus?

• It develops slowly.
• It is one of the few brain regions that continues
  to develop after birth.
• It has a higher density of cortisol receptors
  than almost any other area of the brain.

46
How does stress damage the new
brain?

• Prolonged or excessive exposure to cortisol can
  significantly change the shape of the largest
  neurons in the hippocampus, and can kill them.
• Cortisol can suppress production of new granule
  cells (small neurons), that normally continue to
  develop after birth.

47
Severe, Prolonged Abuse in Childhood

• Damage to limbic structures, associated with
  physiological effects (hyporeactivity) that may
  impair
• Socialization
  Capacity
  for identifying with others
  Internalization of values
  (i.e., conditionability)

48
De Bellis Colleagues, 1999 (Univ. Of
Pittsburgh Medical Center)

• 44 children adolescents with PTSD and 61
  matched controls
• brain volume robustly and positively correlated
  with age of onset of PTSD and negatively
  correlated with duration of abuse
• symptoms of intrusive thoughts, avoidance,
  hyperarousal or dissociation correlated
  negatively with brain volume

49
Teicher, 2002

• Human brains evolved to be molded by
  experience, and early difficulties were routine
  during our ancestral development. Is it plausible
  that the developing brain never evolved to cope
  with exposure to maltreatment and so is damaged
  in a nonadaptive manner?

50
TeicherThis seems most unlikely

• The logical alternative is that exposure
  to early stress generates molecular and
  neurobiological effects that alter neural
  development in an adaptive way that prepares the
  adult brain to survive reproduce in a dangerous
  world.

51
Teicher (2002) concluded that

• Society reaps what it sows in the way it
  nurtures its children. Stress sculpts the brain
  to exhibit various antisocial, though adaptive,
  behaviors. stress can set off a ripple of
  hormonal changes that permanently wire a childs
  brain to cope with a malevolent world.

52
Risk Assessment / Prediction
53
Why Do We Care?

• In the Dispositional Planning Unit of Family
  Court
• By a probation officer prior to adjudication
• In court during the adjudicatory process
• By Dept Human Services (DSS, CPS) caseworkers
  responding to reports of sexually acting out kids
• By clinicians considering discharging kids from
  residential placement
• By clinicians receiving postadjudication
  referrals for treatment

54
Two common demands of clinicians that work with
sexually abusive youth

• 1. gathering, assimilating, and interpreting
• data for the purpose of evaluating a
  client
• (i.e., assessment)
• 2. using the data to make judgments about
• the likelihood of specific outcomes
• (i.e., prediction)

55
Although we may do assessments of offenders for
many reasons, generally speaking, there are 3
broad areas

• 1. Clinical (or Needs) Assessment
• 2. Risk Prediction
• 3. Risk Reduction

56
(No Transcript)
57
Focus will be on last 2 points

• A Priori Risk Assessment or Risk Prediction
• (goal predicting violent behavior).
• Risk Management or Risk Reduction
• (goal determining what aggravates (increases)
  or mitigates (reduces) an existing or preexisting
  condition of risk).

58

• Risk Prediction relies on Static factors
  (historical information about past behavior
  that, for the most part, is unchangeable or
  fixed).
• Risk Reduction relies on Dynamic factors
  (current information about existing behavior
  problems, life events, situational events or
  recent interventions that are all subject to
  change.

59
To begin withWhat ARE Risk Predictors?
60
Risk Factors Defined

• Personal characteristics or environmental
  conditions that predict the onset, continuity, or
  escalation of violence

61
What is a Dynamic Risk Factor?(Rice, 2000)

• The factor precedes is associated with outcome
• The factor changes or can be changed
• Manipulating the factor changes outcome
• True Dynamic Factor

62
Are Risk Factors Causes?

• ONLY when it has been empirically demonstrated
  that changing the risk factor produces changes in
  the onset or rate of violence

63
Assessing Risk in Children Juveniles The Top
10 Problems
64

• Developmental Flux
• Situational Flux
• Boundaries of Normative Sexual Behavior
• Developmentally-Sensitive Risk Predictors
• Low Base Rate Non-Predictors
• High Base Rate Non-Predictors
• Child Sexual Abuse
• Sexual Preference
• Hormones and Sex Drive
• Sexual Recidivism Low Base Rates?

65
The MAJOR risk-relevant difference between adults
and juveniles Age!
66
2,300 Years ago, Aristotle observed
Youth are heated by Nature as drunken men by
wine
67
400 years ago, Shakespeare quipped
I would that there were no age between 10 and
23, for theres nothing in between but getting
wenches with child, wrongdoing the ancientry,
stealing, fighting
(The Winters Tale, Act III)
68
100 years ago, the great Granville Stanley Hall
wrote of
the heightened storm and stress of adolescence
69
Under normal conditions,

• Adolescence is characterized by pervasive
  developmental maturational change
• The stability that we assume when assessing risk
  in adults can NOT be assumed with adolescents

70
Hallmarks of normal adolescent development

• RISK-TAKING
• risk-adversity increases with age the
  hedonic motive in adolescence overshadows
  rational recognition of adverse outcomes
• Dr. Steinberg (2004) increased risk taking
  in adolescence is normative, biologically-driven,
  and inevitable.
• cf., Levin Hart, 2003 Reyna, 1996 Reyna
  Farley, 2006 Rice, 1995

71

• EMOTIONAL INTENSITY LABILITY
• associated with changes in reproductive and
  stress hormones
• IMPULSIVITY
• associated with CNS immaturity - frontal lobe
  continues to mature into the mid-20s
• (Dahl Spear, 2004)

72
Juveniles are not fully cooked

• Research on
• Physical Development
• Buchanan, Eccles, Becker, 1992
• Psychosocial Development
• Cauffman Steinberg, 2000 Steinberg
  Cauffman,
• 1996
• Brain Development
• Baird et al., 1999 Giedd et al., 1999 Sowell
  et al., 1999

73
Metaphor for Assessing Risk with Adolescents
74
Relationship of Instability to Age
Inverse under normal conditions i.e.,
instability tends to decrease with age
75
Developmentally-SensitiveRisk Predictors
76
Selected Effect Sizes of Risk Factors

• Early Onset Juvenile
  Onset
• (6-11) (12-14)
• Aggression .21 .19
• Low IQ .12 .11
• Broken Home .09 .10
• Substance Abuse .30 .06
• Low Family SES .24 .10
• Being Male .26 .19
• Antisocial Parents .23 .16
• Weak Social Ties .15 .39
• Antisocial Peers .04 .37
• Gang Memberships - .31
• Large ES r gt .30
• Moderate ES r .20-.29
• Small ES r lt .20

77

• Low Incidence Non-Predictors

78
Condition 1 for Risk Factors to Predict

• Present among a relatively large proportion of
  those who re-offend sexually
• Violation?
• deviant sexual arousal ?

79
Deviant Sexual Arousal A Low Incidence Risk
Factor ?
Non-Reoffenders
Re-offenders
Wrong 80 of the time!
80
High Incidence Non-Predictors
81
Condition 2 for Risk Factors to
Predict

• Present among a higher proportion of those who
  re-offend than those who dont re-offend
• Violations?
• dysfunction chaos in the home poor social
  skills, substance abuse, denial of problems,
  child abuse?

82
Substance Abuse A High Incidence Risk Factor?
Non-Reoffenders
Re-offenders
Wrong 60 of the time!
83
Substance Abuse(Bonta et al., 1998)
General Violent
Substance Abuse (any) .11 .08
Drug Abuse .09 NR
Alcohol Abuse .06 NR
(NR Not Reported)
84
Low Incidence of Re-offense
85
JUVENILE RECIDIVISM
Study Follow-up Time Sexual Recidivism Nonsexual Recidivism
Doshay, 43 (N 108) 6 yrs., adult incarcerations 7 40
Smith Monastersky 86 (N 112) M28 m., Charged offenses 14 35
Becker, 90 (N 52) 1 yr., referrals/ self-report 10
Kahn Chambers, 91 (N 221) M20 m., convictions 7.5 45
Schram et al., 91 (N 197) M 5 yrs., Arrests 12 63
Lab, et al., 93 (N 155) 1-3 yrs. 2 treated 4 comparison 22 treated 13 comp.
Milloy, 94 (N 59) 3 yrs. reconvictions 0 19 violent 37 nonviolent
86
Study Follow-up Sexual Nonsexual
Wiebush, 96, (3 sites, N 517 adjudicated JSOs) 25-35 mo. Arrests 6 10 Total 40-45
Hagan Cho 96 (N 100) 2 5 yrs., Convictions 9 46
Miner, 02 (N 86) M51 m. Arrests 8 55
Rasmussen 99 (N 170) M5 yrs. 14 54
Rubinstein et al., 93 very assaultive sample M8 yrs. 37 89
Worling Curwen 00 (N 58) M6 yrs. 5 treated 18 control 35 treated 54 control
Langstrom Grann 00 (N 46) M61 m. 20 65
Hagan Gust-Brey, 00, Rapists, (N 50) 10 yrs., convictions 16 68
87
Study Follow-up Sexual Nonsexual
Hecker, et al., 02 (N 54) 10 yrs., Convictions 11 37
Prentky, et al., 00 (N 75) 1 yr., court records, parent/self-rpt. 4 11
Hagan, et al, 01 (N 100) 8 yrs., Convictions CSA 20 Rapists 16 Non-sex 10 66
Nisbet, et al., 04 (N 292) M7 yrs., Conviction 25 Before Age 18 9 As Adults 61
Waite, et al., 05 (N 253) 9 yrs., Arrests 4 60
Righthand, et al., 04 (N 160) M26.88 mo. Charges 2.5 14 (victim) 23 (no)
88
Overview Incidence

• Most studies, sexual recidivism rates 5-14
• Higher rates in some older, more aggressive
  offenders
• Higher rates in residential samples
• Higher rates of nonsexual offending
• True incidence is unknown

89
DSS / NIJ Follow-up Study

• Supported by DSS and NIJ
• Study Timeline
• November, 2005 (Phase 3 Completed stop data
  collection)
• Final sample N 822
• Boys 667 Girls 155

90
Sources of Follow-Up Data

• Department of Social Service Records (DSS)
• Phone Conversations w/Social Workers (SW)
• Criminal History Systems Board (CHSB)
• Sex Offender Registry Board (SORB)

91
Followup Time

• Minimum 6 Months
• Maximum 7 years

Age Range (ASAP Age)

• 1998 3 years old to 20 years old
• 2005 10 years old to 27 years

92
Group Assignments

• Group 1 DSS Sexual Reoffend
• Youth who sexually reoffend and the reoffense was
  only reported in the DSS record and/or DSS SW
• Group 2 CJ Sexual Reoffend
• Youth whose sexual reoffense was known to the
  criminal justice system and resulted in a
  criminal sanction
• Group 3 CJ Non-Sexual Reoffend
• Youth who reoffended non-sexually and was known
  to the criminal justice system and resulted in a
  criminal sanction
• Group 4 No Reoffense
• Youth who had no sexual reoffense and no
  non-sexual reoffense, per CJ DSS records

93
Survival Analysis
N Reoffend Failure Rate N / () Censored Log-Ranka Log-2 log (LR)a
Boys 626 19 116 (27.6) 510 (81) 6.32 6.92
Girls 147 9 13 (10.1) 134 (91)
Pre-Adolescent Boys 331 24 79 (32.6) 252 (76) 7.43 6.79
Adolescent Boys 220 13 29 (22.0) 191 (87)
a test of equality, df 1 p lt .001
94
JSOAP Static1, Dynamic2, Total2 Scores by
Groups (boys only)
DSS Sexual Reoffend CJ Sexual Reoffend CJ Non-Sexual Reoffend Non-Reoffend F
Scale 1 2 M s 17.23a 3.57 17.76a 4.63 15.36b 3.80 13.05c 4.38 36.91
Scale 3 4 M s 17.17a 5.01 15.27ab 5.35 13.53b 5.14 9.30c 5.49 25.20
Total Score M s 36a 6.56 33ab 10.71 31b 6.62 24c 8.09 29.04
1 df 3, 648 2 df 3, 205 abc reflects SNK
group difference (plt.05)
95
ROC curve for Pre-Adolescent Boys (J-SOAP-II
Total Score)
96
ROC Pre-adolescent Boys(optimum probability
levels for classification)
Parameter Concordant C / AUC P level sensitivity specificity
Total Score 81.2 .824 .28 84.8 73.8
.30 75.8 76.2
.32 75.8 76.2
97
ROC curve for Adolescent Boys (J-SOAP-II Total
Score)
98
ROC Adolescent Boys(optimum probability levels
for classification)
Parameter Concordant C / AUC P level sensitivity specificity
Total Score 79.4 .803 .22 78.6 70.4
.24 78.6 77.8
.26 71.4 77.8
99
ROC AUC Comparison of Adolescent
Pre-adolescent Boys
Parameter Pre-Adolescents Adolescents
Scale 1 .775 .830
Scale 2 .553 .671
Scale 3 .771 .643
Scale 4 .751 .816
TOTAL .824 .803
100
Empirical Risk Factors for Juvenile Sex Offenders

• Prior sexual offenses
• Impulsive antisocial behavior, delinquency,
  Conduct Disorder, ADHD, truancy, fighting
• Marked deficits in social skills
• Severe abuse / maltreatment
• Sexual preoccupation / paraphilias
• Deviant sexual arousal

101
Weak and/or Inconsistent Risk Predictors

• Victim Sex
• Substance Abuse
• Psychiatric Problems
• Denial of Problems
• Poor Social Interpersonal Skills
• Dysfunction Chaos in the Home
• Domestic Violence
• Recent Failures or Disappointments
• Failure to Comply with Probation