<%BANNER%>

The Prediction of Risk to Recidivate among a Juvenile Offending Population

Permanent Link: http://ufdc.ufl.edu/UFE0021569/00001

Material Information

Title: The Prediction of Risk to Recidivate among a Juvenile Offending Population
Physical Description: 1 online resource (167 p.)
Language: english
Creator: Baglivio, Michael T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: juvenile, offenders, prediction, recidivism, risk
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The Florida Department of Juvenile Justice has recently implemented a new fourth-generation risk/need assessment to predict the risk to re-offend for juveniles referred to the Department. The new assessment, the Positive Achievement Change Tool, or PACT, is adapted from the validated ?Back On Track!? instrument utilized in Washington State. This study validates the PACT assessment, and examines whether the instrument is as predictive of female delinquency as it is of male delinquency utilizing official delinquency referral as the dependent measure. The study furthermore examines the predictive power of each of the twelve areas, or criminogenic domains, of the PACT to examine the strength of each domain in the prediction of juvenile delinquency. Gender differences are explored to examine whether the factors predicting female delinquency mimic those predictive of male delinquency.
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 Michael T Baglivio.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Piquero, Alexis R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021569:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021569/00001

Material Information

Title: The Prediction of Risk to Recidivate among a Juvenile Offending Population
Physical Description: 1 online resource (167 p.)
Language: english
Creator: Baglivio, Michael T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: juvenile, offenders, prediction, recidivism, risk
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The Florida Department of Juvenile Justice has recently implemented a new fourth-generation risk/need assessment to predict the risk to re-offend for juveniles referred to the Department. The new assessment, the Positive Achievement Change Tool, or PACT, is adapted from the validated ?Back On Track!? instrument utilized in Washington State. This study validates the PACT assessment, and examines whether the instrument is as predictive of female delinquency as it is of male delinquency utilizing official delinquency referral as the dependent measure. The study furthermore examines the predictive power of each of the twelve areas, or criminogenic domains, of the PACT to examine the strength of each domain in the prediction of juvenile delinquency. Gender differences are explored to examine whether the factors predicting female delinquency mimic those predictive of male delinquency.
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 Michael T Baglivio.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Piquero, Alexis R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021569:00001


This item has the following downloads:


Full Text





THE PREDICTION OF RISK TO RECIDIVATE AMONG A JUVENILE OFFENDING
POPULATION




















By

MICHAEL T. BAGLIVIO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007

































2007 Michael T. Baglivio

































To the memory of my mother, Lorraine









ACKNOWLEDGMENTS

I thank Dr. Alex Piquero for his guidance and accessibility, and the entire committee for

their patience and continued support throughout the course of this project. I thank my family for

their patience and continued faith regarding the completion of this project.










TABLE OF CONTENTS

page

A CK N O W LED G M EN T S .................................................................. ........... .............. .....

L IS T O F T A B L E S ................................................................................. 7

LIST OF FIGURES .................................. .. .... ... ...............9

A B S T R A C T ................................. ............................................................ 10

CHAPTER

1 IN T R O D U C T IO N ............... .............................. .................... .............. 11

P re -S c re e n ................... ........................................................... ................ 1 3
F u ll A ssessm ent ................................................................................ 14
A ssessm ent Scoring ...................................................... ................. .... ...... 21

2 L ITE R A TU R E R E V IE W .............................................................................. ..................... 23

Collective vs. Selective Incapacitation ......................................................... ............... 23
N ew P e n o lo g y .................................................................................................................... 2 7
E rro rs in P re d ictio n ............................................................................................................ 3 6
P reduction M ethod ................................................................38
P re d ictio n S tu d ie s ........................................................................................................4 1
W h at W ork s? ................................................................4 6
E m pirical P redictors ................................................................47
Principles of Effective Intervention.....................................51
Gender Differences in Offending ................................................59
C u rrent F o cu s ................................................................................ 64

3 METHODS .......................................69

4 D IS C U S S IO N ......................................................................................... 1 13

5 C O N C L U S IO N ................................................................................................................ 12 0

6 FU TU RE D IRECTION S ............................................................................................. ........ 122

APPENDIX

A HOW TO READ AND INTERPRET THE PACT OVERVIEW AND THE YES CASE
PL A N R E P O R T S ...............................................................124

B PACT PRE-SCREEN ASSESSMENT ................................................................. 129

C FU LL A SSE SSM EN T .............................................................................................. .. ........ 137









R E FE R E N C E S ....................................................... 153

B IO G R A PH IC A L SK E T C H ............................................................................... ............... ..... 167





















































6









LIST OF TABLES


Table page

1-1 Pre-Screen D om ains....... ....... .......... .................................. ......... 14

1-2 PACT Full A ssessm ent D om ains.................. ......................... ................. ............... 15

1-3 PA CT Scoring M atrix .............. ....... .............. ........ ................................. 22

3-1 Criminal History Item Descriptive Statistics .......................................... ............... 72

3-2 Social H history Item D escriptive Statistics .............................................. ............... 74

3-3 Community Cohort 1 Criminal History Descriptive Statistics .......................................81

3-4 Community Cohort 1 Social History Descriptive Statistics ...........................................81

3-5 Residential Cohort 1 Criminal History Descriptive Statistics .......................................82

3-6 Residential Cohort 1 Social History Descriptive Statistics ................. ....... ............83

3-7 Community Cohort 1 Covariate Descriptive Statistics.....................................................84

3-8 Community Cohort 1 Logistic Regression Results...........................................................85

3-9 Residential Cohort 1 Covariate Descriptive Statistics............... ........... ............... 85

3-10 Residential Cohort 1 Logistic Regression Results........................ .. .................86

3-11 Split Residential Cohort 1 Covariate Descriptive Statistics ...........................................88

3-12 Split Residential Cohort 1 Logistic Regression Results ......................................... 89

3-13 Community Cohort 2 Criminal History Descriptive Statistics .......................................91

3-14 Community Cohort 2 Social History Descriptive Statistics ...........................................91

3-15 Residential Cohort 2 Criminal History Descriptive Statistics .......................................92

3-16 Residential Cohort 2 Social History Descriptive Statistics......... ............. ............. 93

3-17 Community Cohort 2 Covariate Descriptive Statistics.....................................................93

3-18 Community Cohort 2 Logistic Regression Results................................. ...............94

3-19 Residential Cohort 2 Covariate Descriptive Statistics ..................................................95

3-20 Residential Cohort 2 Logistic Regression Results .......................................................96









3-21 Split Residential Cohort 2 Covariate Descriptive Statistics ...........................................96

3-22 Split Residential Cohort 2 Logistic Regression Results ......................................... 97

3-23 D escriptive Statistics By Gender .......................................................... ............... 101

3-24 Logistic Regression Results By Gender ............................................... ............... 102

3-25 Cohort 2 D escriptive Statistics By Gender .................................... ............................... 103

3-26 Cohort 2 Logistic Regression By Gender .................................... ................................. 104

3-27 Cohort 2 Community Gendered Covariate Descriptive Statistics .................................105

3-28 Cohort 2 Community Expanded Covariate Logistic Regression...................................107

3-29 Cohort 2 Community Gender-Split Covariate Descriptive Statistics ...........................107

3-30 Cohort 2 Community Gender-Split Covariate Logistic Regression .............................109

3-31 Cohort 2 Community Violent Offense Logistic Regression................................111









LIST OF FIGURES

Figure page

A PA CT Overview R eport........................................................................ ............... 124

A -2 YES Case Plan ......... .... ................ ...................... .. .. ...... ............ 127









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

THE PREDICTION OF RISK TO RECIDIVATE AMONG A JUVENILE OFFENDING
POPULATION

By

Michael T. Baglivio

December 2007

Chair: Lonn Landza-Kaduce
Cochair: Alex Piquero
Major: Criminology, Law and Society

The Florida Department of Juvenile Justice has recently implemented a new fourth-

generation risk/need assessment to predict the risk to re-offend for juveniles referred to the

Department. The new assessment, the Positive Achievement Change Tool, or PACT, is adapted

from the validated "Back On Track!" instrument utilized in Washington State. This study

validates the PACT assessment, and examines whether the instrument is as predictive of female

delinquency as it is of male delinquency utilizing official delinquency referral as the dependent

measure. The study furthermore examines the predictive power of each of the twelve areas, or

criminogenic domains, of the PACT to examine the strength of each domain in the prediction of

juvenile delinquency. Gender differences are explored to examine whether the factors predicting

female delinquency mimic those predictive of male delinquency.









CHAPTER 1
INTRODUCTION

The focus of the current study is to validate the assessment instrument used by the Florida

Department of Juvenile Justice to determine a referred youth's risk to re-offend. The validation

of an assessment or instrument used in the prediction of risk to recidivate is critical to ensure the

predictions that come from that instrument and the recommendations made based on the

predictions are indeed the best possible. The first step to being able to provide effective treatment

to juvenile offenders is the use of a validated risk/needs assessment to determine the risk level of

the youth, which determines the needed intensity of treatment, and the individualized

criminogenic needs, which must be targeted in order to reduce the likelihood of recidivism.

The Florida Department of Juvenile Justice is one of the largest juvenile justice agencies in

the United States. The Department received 150,104 delinquency referrals during FY 2005-06,

representing 94,288 individual youth. Beyond the importance of the safety and security of the

youth served, one of the most challenging tasks for a criminal justice agency of that magnitude is

to ensure the provision of appropriate, evidence-based treatment services to the youth referred in

an effort to reduce their risk of re-offending and increase public safety. A critical step to

providing empirically proven, appropriate treatment to a juvenile offending population is the

implementation of an evidence-based, validated risk/need assessment.

The Florida Department of Juvenile Justice (DJJ) has recently implemented a new

instrument designed to assess juvenile offenders with regard to risk, needs, and strengths related

to the same risk factors as outlined in the extant "what works" literature (Andrews & Bonta,

2003). The Positive Achievement Change Tool, or PACT, is heavily adapted from the validated

Washington State Juvenile Court Assessment, Back On Track! (also know as the YASITM in

some jurisdictions), which has been in use throughout the country since 1998. The Florida









Department of Juvenile Justice and Assessments.com created the PACT collaboratively in 2005.

Implementation of the PACT began in November of 2005 in Hillsborough and Pinellas counties,

and was statewide by August of 2006. Over 1,400 Juvenile Probation Officer staff members have

received the two-day PACT training including risk assessment theory, case planning, and the

techniques of motivational interviewing (Miller & Rollnick, 2002). The 1,400 staff represents all

Juvenile Probation Officers and all juvenile assessment center (JAC) screeners.

The PACT is an advancement from other fourth-generation risk/needs assessments,

including Back on Track! and the YASI on which it was based, due to its incorporation of

questions from the Massachusetts Youth Screening Instrument (MAYSI), making it an indicator

for need of further mental health assessment, and its integration with DJJ's Juvenile Justice

Information System (JJIS) database. The JJIS and PACT interface allows for automatic scoring

of the criminal history domain (official prior criminal history). Furthermore, the PACT

automates findings into the Youth Empowered Success Plan (YES), one of the first ever

supervision and treatment plans in the nation to be automated to directly apply information

gained from an evidence-based risk/needs assessment (see Appendix A: How to Read and

Interpret the PACT Overview and the YES Case Plan Reports).

The PACT identifies not only the areas (domains) in which a juvenile is most at risk, but

equally as important, those areas which he or she has strengths (protective factors), which can be

built upon in the case planning process. Risk and protective factors within the PACT include

both static and dynamic characteristics. The benefit of measuring risk and protective factors and

characteristics is that a juvenile justice professional is better able to match a youth's current

needs with appropriate programs and services.









Pre-Screen

The PACT has both a Pre-Screen and a Full-Assessment, with risk levels ranging from

low, moderate, moderate-high, and high. Youth who score moderate-high or high risk to re-

offend are administered the Full-Assessment, as are youth being placed in residential

commitment. Furthermore, youth on probation supervision are given a reassessment every ninety

days so as to measure progress on the treatment plan, and youth in residential commitment

receive an assessment within sixty days of release so as to assist in aftercare supervision

planning. The Pre-Screen is a 46-item, initial assessment instrument, which produces risk level

scores measuring a juvenile's risk of re-offending. This tool is designed as a semi-structured

interview protocol and utilizes Motivational Interviewing techniques (Miller & Rollnick, 2002).

With automated scoring, the PACT Pre-Screen identifies a total risk level for each youth

assessed, as well as a rank order of criminogenic needs and protective factors. Low-risk to re-

offend youth are generally recommended for diversion or other community-based intervention

programs and isolated from any contagion effects caused by intermingling with higher-risk youth

or separation from the protective factors present in their lives. Moderate-risk youth are generally

directed to supervised probation and other community-based interventions. Moderate-high and

high-risk youth receive the PACT Full-Assessment.

The PACT Pre-Screen collects information on four distinct domains (see Table 1-1).

Domain 1 assesses the youth's prior criminal history. This domain is pre-populated from the JJIS

database, as mentioned above, and is therefore not self-reported as are many of the other

domains. Domain 1 includes information on the age at which the youth first offended,

seriousness of prior offenses in terms of number of prior felonies and prior misdemeanors, prior

commitment placements, prior escapes, and the type of offenses (such as weapon, against person,









and sex offenses). A complete listing of Domain 1 questions, and all PACT Pre-Screen questions

can be seen in Appendix B: PACT Pre-Screen Assessment.

Table 1-1. Pre-Screen Domains
Domain # Domain Name
1 Record of Referrals
2 Social History
3 Mental Health
4 Attitude/Behavior Indicators


Domain 2 of the PACT Pre-Screen examines the youth's social history. This domain

features gender, current school status and performance, history of, and extent of current, anti-

social friends, history and current alcohol and drug use, history of abuse and mental health

problems, and history of family problems. Domain 3 assesses the youth's mental health status

including suicidal ideation, history of anger, depression, somatic complaints, thought

disturbance, and history of traumatic experiences. It is this domain that has incorporated

questions from the MAYSI and allows for the identification of the need for further mental health

assessment. Certain "hits" on Domain 3 questions trigger the need for a more in-depth

assessment of the youth's current mental status. The final Pre-Screen domain is Domain 4, which

examines the youth's extent of anti-social attitudes and beliefs, and resulting behavior. This

section covers attitudes toward law-abiding behavior, beliefs about aggression and fighting to

solve problems, self-reports of violence not included in the official criminal history domain

(Domain 1), and reports of sexual aggression.

Full Assessment

The PACT Full-Assessment is a 126-item, in-depth assessment instrument, which

produces identical risk level scores as the Pre-Screen, measuring a juvenile's risk of re-offending.

This tool is designed as a semi-structured interview protocol utilizing Motivational Interviewing









techniques as well. Every question in the Pre-Screen is included in the Full-Assessment and is

pre-populated from the Pre-Screen into the Full-Assessment to decrease the workload of the

individual completing the assessment. The Full-Assessment is required to be administered to

moderate-high and high risk to re-offend youth, and all youth being placed in residential

commitment as a means to gather the most comprehensive amount of information on each youth

related to research-validated risk and need factors as possible. The PACT Full-Assessment

instrument measures a youth's risk and protective factors in the following twelve domains:

Criminal History, School, Use of Free Time, Employment, Relationships, Family History, Living

Arrangements, Alcohol and Drugs, Mental Health, Attitudes/Behaviors, Aggression, and Skills

(see Table 1-2).

Table 1-2. PACT Full Assessment Domains
Domain # Domain Name
1 Record of Referrals
2 Gender
3A School History
3B Current School Status
4A Historic Use of Free Time
4B Current Use of Free Time
5A Employment History
5B Current Employment
6A History of Relationships
6B Current Relationships
7A Family History
7B Current Living Arrangements
8A Alcohol and Drug History
8B Current Alcohol and Drugs
9A Mental Health History
9B Current Mental Health
10 Attitudes/Behaviors
11 Aggression
12 Skills









The first domain, as in the Pre-Screen, is automated from the JJIS database so as to

decrease the errors in reporting prior official offending history. Domain 1 of the Full-Assessment

is identical to Domain 1 of the Pre-Screen and provides a comprehensive description of referral

history for the youth. Prior referrals are disaggregated by type of offense (against-person, sexual,

weapon) by felony and misdemeanor. Furthermore, the amount of prior detention confinements,

residential commitments, escapes, and pick up orders/failure-to-appear in court are included (for

a full list of Full-Assessment questions, see Appendix C: PACT Full-Assessment). The inclusion

of this domain is the empirical importance of the predictive power of past behavior on future

behavior; the notion that the best predictor of future behavior is past behavior (Sherman et al.,

1997).

The second domain distinguishes between males and females. This gender domain is

essential to capture the reality that male youth consistently offend and re-offend at higher rates

than do female youth. Males automatically receive one point towards the assessment of risk

whereas females receive zero points.

The third domain is subdivided into School History and Current School Status. The School

History section examines whether the youth has ever been a special education student, and why

(behavioral, ADHD/ADD, learning disability, mental retardation). This section also captures any

history and the amount of expulsion and out of school suspensions, age at first

expulsion/suspension, and whether the youth has been enrolled in school during the last six

months. The Current School Status section captures the youth's current enrollment status, type of

school in which the youth is enrolled, and the youth's perception of the value of education. This

section also examines the youth's conduct, involvement in school activities, number of

expulsions/suspensions, attendance, and performance, all in the most recent school term.









Information included in this domain is by self-report from the youth in combination with

corroboration from teachers from the school in which the youth was enrolled.

The next domain captures the youth's use of free time. This domain is also separated to

examine historic use of free time and the youth's current use of free time. Information for this

domain is captured by self-report from the youth. Historic use of free time is simply measured by

whether the youth has participated in structured and supervised, or unstructured, pro-social

community activity within the past five years. Current use of recreational time is captured by the

youth's current interest and involvement in structured recreational activities, which activities the

youth is currently engaged in (volunteer organizations, athletics, religious groups, etc.), and the

current interest and involvement in other positive recreational activities (such as reading or pro-

social hobbies). The notion behind this domain is social control, and that youth who are involved

in pro-social recreational and leisure activities will be less likely to commit criminal offenses

than those youth without an interest in such activities.

The fifth domain captures the youth's employment and is divided into employment history

and current employment. This domain does take into account the fact that the youth may be too

young for employment consideration. The employment history section looks at whether the

youth has ever been employed, problems while employed (fired or quit for poor performance, or

fired/quit because he/she could not get along with employer or coworkers), and history of

positive personal relationships with past employers or coworkers. The current employment

section measures whether the youth understands what is required to maintain a job, interest in

employment, current employment status, and current positive relationships with employers or

coworkers.









Relationship issues are measured by the sixth domain and are divided into history of

relationships and current relationships. History of relationships captures past positive adult non-

family relationships, who can provide support and model pro-social behavior, as well as past

antisocial friends/peers, who are hostile to or disruptive of the legal social order or engage in

criminal activities. Current relationship issues are measured by the extent of current positive

adult influences, pro-social community ties, whether current friends are pro-social or antisocial,

current romantic relationship, whether the youth currently admires anti-social peers, and the

youth's current resistance to anti-social peer influences. This domain is heavily influenced by the

notion of differential association (Akers, 1998).

Family history and current living arrangements compose the next domain. Family history

is captured by a history of out-of-home Department of Children and Families (DCF) and shelter

placements, history of running away, history of imprisonment of persons living in the household

with the youth, and whether the youth lives with a supervising adult, with peers, or is transient.

Current living arrangements examines who the youth currently lives with, annual income of the

family, imprisonment history of persons the youth is currently living with, problem history of

parents currently in the household (such as alcohol, drug, mental health, employment), problem

history of siblings in the household (identical to those of the parents), support network for the

family, level of conflict between parents and siblings, parental supervision, and the extent of

appropriate punishment and rewards. The incorporation of factors dealing with the consistency

of punishments and rewards, and the extent of parental supervision, authority, and control are

heavily influenced by Gottfredson and Hirschi's General Theory of Crime (1990).

The eighth domain captures substance abuse and is gathered from self-report from the

youth as well as corroboration with parents and teachers when possible. This domain is divided









into alcohol and drug history, and current use of alcohol and drugs. Alcohol use is measured

separately from drug use for both subgroups of questions. Past use or current use is captured as is

whether the use has disrupted, or currently disrupts, education, causes family conflict, interferes

with keeping pro-social friends, causes health problems, contributes to criminal behavior, or

leads to increased tolerance or withdrawal. History or current referral for assessment and

substance abuse classes/treatment is also captured. For the current drug use section, the types of

drugs used are identified as well.

The ninth domain measures mental health history and current mental health status. As

stated above, this domain purposively incorporated questions from the MAYSI assessment so as

to be able to flag youth for need of further mental health and suicide assessment. This inclusion

allows the individual performing the assessment to discontinue using the MAYSI altogether,

decreasing workload and increasing efficiency. The mental health history section assesses

whether the youth has ever thought about suicide, attempted suicide, engaged in self-mutilating

behavior, has a history of physical abuse, sexual abuse, neglect, a traumatic experience,

witnessing violence, mental health problems, anger or irritability, depression, somatic

complaints, or thought disturbance. The assessment takes into account any neglect or abuse self-

reported by the youth, whether substantiated or not, but excludes reports proven to be false in the

event investigations by social service agencies concluded the report to be false.

The mental health problem questions include such issues as schizophrenia, bi-polar, mood,

thought, personality, and adjustment disorders. Conduct disorder, oppositional defiant disorder,

substance abuse, and ADD/ADHD are excluded. The mental health problems are those

confirmed by a professional in the social service/healthcare field (the youth must have been

diagnosed).









The current mental health section includes whether the youth is currently prescribed

treatment, or currently prescribed medication. These questions disaggregate further by whether

the youth is actually attending the treatment, or taking the medication, or whether it is prescribed

but the youth is not attending or complying with the medication regimen. Furthermore, the

current mental health section also captures whether the individual performing the assessment

believes the mental health problems of the youth interfere with working with the youth.

The next domain assesses the attitudes and behaviors of the youth. This section captures

the youth's self-reported primary emotion when committing crimes (nervous, excited,

indifferent, confident) and the primary purpose for committing them (anger, revenge, impulse,

sexual desire, money, amusement, or peer status). The level of optimism the youth has about the

future is also examined in terms of whether the youth has high aspirations, some sense of

purpose, low aspirations, believes nothing matters, as well as his/her belief in being able to

successfully meet the conditions of the court. Several questions capture the extent to which the

youth believes he/she can control or avoid anti-social behavior, or use self-control. The youth's

empathy for victims is assessed, as is the youth's acceptance of responsibility for anti-social

behavior, respect for the property of others, authority figures, and attitude toward law abiding

behavior.

The Aggression domain is next, which measures tolerance for frustration, hostile

interpretation of the actions and intentions of others, and belief that yelling or belief that fighting

resolves disagreements. Information for this domain is generated from youth self-report. This

domain also measures reports/evidence of violence not included in the criminal history domain,

such as outbursts, deliberately inflicting physical pain, threatening with a weapon, fire starting,









and animal cruelty. Reports of problems with sexual aggression are examined, such as aggressive

sex, sex for power, young sex partners, child sex, voyeurism, and exposure.

The final domain examines skills that could potentially help the youth avoid anti-social

behavior. These skills include the ability to recognize there are consequences to one's actions,

ability to set realistic goals, ability to apply appropriate solutions and problem-solve, and

situational perception. The Skill domain also captures the youth's use of basic social skills to

deal with others, deal with difficult situations, and deal with feelings/emotions. The final few

questions measure whether the youth can monitor internal triggers, external triggers, control

impulsive behaviors, and control aggression. Information for this domain is generated from

youth self-report as well as the assessor's impression of the youth's skills.

Assessment Scoring

The PACT Pre-Screen and Full-Assessment are scored identically. The purpose of

including additional items in the Full-Assessment is to gather as much information on the youth

as possible to assist with case planning, but the overall total risk to re-offend score is identically

derived for each assessment type. The score (low, moderate, moderate-high, high) is based on

two sets of factors. The criminal history score is assessed exclusively based on the youth's prior

official criminal offending. The criminal history score ranges from zero to thirty-one. The social

history score, in contrast, is based on responses to questions from several of the other PACT

domains. The social history score ranges from zero to eighteen. The two scores, criminal and

social history, are associated within a matrix, determining the youth's overall risk to re-offend

(see Table 1-3).

The items used to score criminal history include age at first offense, number of prior

referrals (misdemeanor, felony, weapon, and person-misdemeanor, person-felony), prior

detention confinement, prior residential commitments, prior escapes from commitment, and prior









pick-up orders. Social history is scored using variables including gender, school status

(enrollment, conduct, attendance, academic performance), current anti-social peers, history of

court-ordered out of home placements, history of running away or being kicked out of home,

incarceration of a current household member, parental control, history and current alcohol or

drug use, history of violence/physical abuse, history of being a witness to violence, sexual abuse

history, history of neglect, and mental health problems. Individuals administering the

assessments are not knowledgeable about the maximum scores associated with any domain of

the PACT. The instrument, therefore, cannot be manipulated as easily as assessments without

automated scoring, or where individuals are aware of which questions are weighted more heavily

than others.

Table 1-3. PACT Scoring Matrix
Criminal History Score Social History Risk Score
Oto5 6 to 9 10 to 18
0 to 5 Low Low Moderate
6 to 8 Low Moderate Moderate-High
9 to 11 Moderate Moderate-High High
12 to 31 Moderate-High High High









CHAPTER 2
LITERATURE REVIEW

The notion of predicting individuals' risk to offend or recidivate has been a constant focus

for criminologists and policy-makers for several decades. As early as 1923, Warner examined

offender characteristics related to violation of parole (Warner, 1923; von Hirsch & Gottfredson,

1984). Prediction studies generated worldwide attention over the next few decades due to the

seminal work by Glueck and Glueck (1950; see also von Hirsch & Gottfredson, 1984). The

California Base Expectancy Score, used in the 1960's, and a similar scale adopted by the United

States Parole Commission in the 1970's, were used extensively to predict parole success and

classification decisions (von Hirsch & Gottfredson, 1984). Prediction then moved to the realm of

classifying offenders based on the likelihood of re-offending. David Greenberg (1975) first

introduced the term "selective incapacitation", and the idea of targeting particular offender

groups for imprisonment emerged, although a comprehensive strategy for implementing such a

plan was first extrapolated by Greenwood and Abrahamse (1982).

Collective vs. Selective Incapacitation

Incapacitation is a simple crime prevention strategy based on the notion that offenders

cannot commit crimes in society as long as they are incarcerated. Crime reduction therefore

occurs, as those incarcerated are prevented from committing offenses in the community. A

secondary benefit of a crime prevention strategy based on incarceration is the concept of

deterrence. Individuals in the general population are believed to be deterred from committing

crime from the fear of being sent to prison; referred to as the general deterrence effect.

Furthermore, imprisoned individuals may be deterred from committing subsequent offenses upon

release; referred to as the specific deterrent effect (Sherman, et al., 1997).









Collective incapacitation refers to the crime reduction effect resulting from physically

restraining individuals regardless of the strategy utilized (be it deterrent, incapacitative,

retributive, or rehabilitative) and without the necessity of prediction of future behavior.

Collective incapacitation was popularized by James Wilson (1975), calling for uniform sentences

imposed on all offenders of specific felonies. Removing all of these offenders from the

community for a significant portion of their criminal careers would prevent all of them who were

so inclined from re-offending while incarcerated (von Hirsch & Gottfredson, 1984).

Wilson (1975) promised a twenty percent reduction in robbery rates if a strategy of

incapacitating a larger fraction of the convicted serious robbers were implemented. The estimates

were derived from a projection technique developed by Shinnar utilizing the newly invented

lambda variable (von Hirsch & Gottfredson, 1984; Shinnar & Shinnar, 1975). The effect on the

crime rate due to collective incapacitation strategies has been found to depend critically on the

estimates of lambda, the average annual offense rate at which individual offenders commit

offenses in the community (von Hirsch & Gottfredson, 1984). When lambda is assumed to be

high, the reduction in the crime rate due to incapacitation would be substantial, but when a lower

average offense rate is assumed, the effects are much more modest.

Selective incapacitation, in contrast, implies the prevention of future criminality through

physically restraining those particular individuals who are predicted to be higher risk to commit

future criminal behavior than other groups of individuals (Greenberg, 1975). Greenwood and

Abrahamse (1982) revised the formula used in the Shinnar and Wilson studies by establishing

not one, but three lambdas (von Hirsch & Gottfredson, 1984). A separate lambda would be

estimated for low, medium, and high-risk offenders. Greenwood and Abrahamse (1982) hold that

the separate estimates can be gathered from the data by using the average self-reported annual









robbery rates for individuals falling in each of the three risk level classifications of his prediction

instrument (a seven-item, dichotomous additive index). These authors argue for the ability to

estimate the incapacitative effect for each level of offender risk, demonstrate that incarcerating

high-risk offenders is most effective, and subsequently propose the lengthening of sentences for

individuals in that high-risk classification category.

In the event accurate prediction can be achieved, and if it is coupled with focusing

incarceration only on high risk to re-offend individuals, Greenwood and Abrahamse proposed

that it would be possible to remain within the available prison resources, a statement not possible

by those adhering to a collective incapacitation mentality (Greenwood & Abrahamse, 1982, von

Hirsch & Gottfredson, 1984). In this model, the offender's risk score, not simply the seriousness

of the presenting offense, would principally determine whether that offender is incarcerated. Von

Hirsch and Gottfredson (1984) argue this policy would require imprisonment of a larger

proportion of offenders convicted of less serious crimes if their predicted level indicated they

were bad risks. This statement, while true, discounts the empirical evidence arguing against the

offense-specific specialization of offenders (discussed at greater length below).

Other researchers question the ability to estimate lambda as Greenwood and Abrahamse

have due to their use of generalizing the self-reported robbery rates of the convicted robbers in

his sample to active robbers generally, especially in light of their own admission that the

probability of being convicted then incarcerated for robbery is so low; .0258 in that study (von

Hirsch & Gottfredson, 1984). Other researchers have found incarcerated samples biased toward

having more high-rate offenders than found in the general population (Auerhahn, 1999; Canela-

Cacho et al., 1997; Chaiken & Chaiken, 1984; Shannon, 1991; Wolfgang et al., 1972; Wright &

Rossi, 1986). Any incarceration effect, be it selective or collective, would be overestimated if the









sample used for estimation is an incarcerated one. Additional problems with estimating the

benefits of incarceration strategies have been proposed as well, such as the assumptions

regarding length of criminal career, replacement/substitution effects, and the decline in offending

with age (Blumstein et al., 1978; Gottfredson & Hirschi, 1986; Spelman, 1994; Zimring &

Hawkins, 1995; 1988). Replacement and substitution effects refer to the notion that when an

individual is arrested and/or removed from a particular locale, another individual will take

his/her place and commit similar offenses. This notion is prevalent when referring to the sale of

illegal drugs where removal of a drug dealer from a location may only lead to a new individual

subsequently selling drugs at that same location.

At the time of the 1975 publication, Greenberg, and others, including researchers and

prison administrators, hypothesized the group of high-risk offenders to be targeted for selective

incapacitation would be between ten to fifteen percent of the prison population, while all other

prisoners could be safely released under varying levels of supervision with minimal threat to

society (Greenberg, 1975; see also Goldfarb & Singer, 1973; Mitford, 1973).1 Other research at

the time had found that most crimes are committed by a small percentage of the population as a

whole and that most are committed by recidivists (Belkin et al., 1972; Hughes, 1971; Shannon et

al., 1988; Shinnar & Shinnar, 1975; West & Farrington, 1977; Wolfgang et al., 1972).

Shinnar and Shinnar (1975) caution the importance of distinguishing between the

percentage of crimes committed by recidivists and the percentage of criminals who are

recidivists. They provide the example that assuming for every career criminal there are 5 one-

1 Greenberg also presented the argument that higher recidivism rates for particular offenders could demonstrate both
the need for longer sentences for public protection, or show the need for increased resources aimed at increasing
rehabilitation program effectiveness. The same data are utilized by both ends of the rehabilitative-incapacitative
spectrum with neither party realizing the high recidivism rates may be greatly due to "artifacts of administration" in
that most returns to prison are for technical violations, not new-law violations (Greenberg, 1975:556). By constantly
calling the publics' attention to high recidivism, both sides of the argument increase the prejudice against ex-
convicts.









time offenders and with an estimate of 50 crimes per career criminal, then career criminals would

compose only 16% of the criminal population, yet account for more than 90% of the crimes. This

hypothetical has been somewhat echoed by empirical research by Wolfgang who found 6.3% of

a birth cohort of Philadelphia boys (18% of the offenders) were responsible for roughly 52% of

the offenses committed by the cohort (Wolfgang et al., 1972; see also Chaiken & Chaiken, 1984;

Cohen, 1984; Mednick & Christiansen, 1977; Shannon, 1991; Shannon et al., 1988; Wilson &

Herrnstein, 1985; West & Farrington, 1977; Wright & Rossi, 1986).

New Penology

The concept of selective incapacitation came just after the now infamous "nothing works"

declaration interpreted from Martinson, questioning the possibility of rehabilitative effectiveness

in the criminal justice system (1974; see also Lipton et al., 1975; Zimring & Hawkins, 1995).2

Critics argued against the "nothing works" conclusion, faulting inadequacy of the research

methodology used to evaluate the programs and the poor implementation and delivery of the

treatment programs studied, for the inability to realize reduced recidivism or draw conclusions

from the research (Cullen & Gendreau, 1989; Cullen & Gilbert, 1982; Gendreau & Ross, 1979,

1981, 1987; Gottfredson, 1979; Halleck & Witte, 1977; Palmer, 1975, 1983; van Voorhis, 1987).

While these critics published evaluations indicating substantial evidence of effective treatments,

the "nothing works" mantra exerted tremendous power over practitioners, professionals, and the

public (Cullen & Gendreau, 1989). The Martinson report encapsulated the shifting focus evident

in the waning influence of the principle of rehabilitation for imprisonment previously held by the

criminal justice system in the United States, which had began to decline in the 1970's (Allen,

1981; Auerhahn, 1999; Flanagan, 1996; von Hirsch, 1985).


2 Several researchers have suggested the decline in support for the rehabilitative ideal may also be traced to the
rising proportion of minority individuals in U.S. prisons (Allen, 1981; see also Sklansky, 1995; Tonry, 1995).









The focus then transferred from the rehabilitative ideal of treatment of individual offenders

towards incapacitative and retributive goals emphasizing control, management, and incarceration

of particular offender groups (Auerhahn, 1999; Feeley & Simon, 1992; Shichor, 1997). The

major aim of the criminal justice system switched to the effective management of so-called

dangerous groups in the "new penology" (Feeley & Simon, 1992). The new emphasis on

efficiency, calculability, prediction, and control was based on the Weberian concept of "formal

rationality", and has resulted in the movement being compared with the fast-food industry and

the "McDonaldization" of punishment (Ritzer, 1993; Shichor, 1997; Weber, 1968).

Three distinct influences have been identified as being involved in the shift to the new

penology, including language, objectives, and techniques (Feeley & Simon, 1992). With the

intent shifted from individual treatment to managerial concerns for controlling certain offender

groups, the notions of probability and risk became increasingly relevant, replacing discourse of

clinical diagnosis and mental status. These ideas of probability and risk are instrumental in the

isolation and classification of particular high-rate offending groups as well as predicting the

likelihood of offending or recidivating. Objectives of efficient control replaced traditional goals

of rehabilitation and crime control. Finally, techniques targeting aggregate levels of offenders for

the purpose of identification, classification, and management of dangerousness groupings

supplanted traditional practices of diagnosis, intervention, and treating offenders on an

individualized basis (Cohen, 1985; Garland & Young, 1983; Feeley & Simon, 1992; Gordon,

1991; Messinger, 1969; Messinger & Berecochea, 1990; Reichman, 1986; Wilkins, 1973).

According to Feeley and Simon, a focus on risk management and incapacitation of

particular offender groups "promises to reduce the effects of crime in society not by altering

either offender or social context, but by rearranging the distribution of offenders in society"









(1992:458). The task became one of removing high-rate offenders from society, while returning

those low risk individuals who pose no significant threat to the community. Advancement in

statistical methods allowed for the expansion of discourse of actuarial and quantifying language

with "high-rate offenders" and "career criminals" defined by the distribution of criminals itself,

linking direct relationships between penal policy and strategies to the offending population.

One additional effect of the focus on control of aggregate subpopulations of offenders was

the decrease in expectations about the effectiveness of criminal sanctions (Feeley & Simon,

1992). No longer was there anticipation that a sanction, be it incarceration, probation, parole,

etc., have the potential to rehabilitate or treat the offender with the hope of reintegration into

society. Rather, the desire became that the sanction be effective at monitoring and managing

aggregate groups based on classification according to the degree of control warranted by their

classification risk levels. This "custodial continuum", as Feeley and Simon called it, had at one

extreme the prisons, which provided the highest security at the highest cost for those individuals

who present the greatest risk, and probation at the low end of the spectrum providing low-cost

surveillance for low-risk individuals.

In between the extremes are intermediate levels of supervision and surveillance techniques

as evidenced by present-day intermediate sanctions such as boot camps, intensive supervision

(with frequent drug testing) and electronic monitoring. Feeley and Simon caution that "once the

focus is on categories of offenders rather than individuals, methods naturally shift toward

mechanisms of appraising and arranging groups rather than intervening in the lives of

individuals" thus "in the end the search for causal order is at least premature" (1992:459). The

appraising of groups these scholars mention is evident in the increased number of prison

admissions due to technical parole and probation violations making prisons holding pens for









violators assessed as too dangerous to remain in society. This shifts the mission of the prison

from correctional to a management function of warehousing a particular risk level of offenders.

The differentiation of prisons by security risk level is in contrast to the differentiation by

specialized function evident in prisons in the 1970's, such as the California Rehabilitation Center

for drug users and California Medical Prison at Vacaville for the mentally ill (Feeley and Simon,

1992).

The new penology movement led to the funding of a research program in the 1970's and

1980's by the National Institute of Justice for the study of criminal careers in relation to policy

decisions impacting the criminal justice system. The 1982 Rand report by Greenwood and

Abrahamse came out of that program. This report was the first to articulate a coherent proposal

for the identification of offender groups for the purpose of selective incapacitation. The

Greenwood and Abrahamse study and the subsequent RAND studies by Chaiken and Chaiken

improved over previous prediction research in one essential aspect. These studies moved beyond

attempting to predict whether an offender would return to crime, and included also addressing

the frequency of offending (von Hirsch, 1984). High-rate offenders were to receive longer

sentences while lower-rate offenders would receive shortened sentences. This process would in

turn reduce crime and be a more efficient use of penal resources by only holding high-rate

offenders for prolonged periods (Greenwood and Abrahamse, 1982).

Shinnar and Shinnar (1975) advised that what is important is the uniform application of

penalties for any recidivist offender, regardless of the severity of the second offense

(foreshadowing current three-strikes laws further examined below). Their suggested strategy

included improved sentencing procedures, screening to recognize recidivists, and selective

increase in those sentences. These scholars did provide one caveat to their argument stating, "in









no way do we want to imply that future crime policies should be based solely on incapacitation.

Better rehabilitation would be preferable but rehabilitation must become much more successful

before it will have a significant effect" (Shinnar & Shinnar, 1975:606).

The Greenwood and Abrahamse report sparked a debate in the literature as to the potential

effectiveness of a selective incapacitation strategy (Blumstein et al., 1978; Spelman, 1994;

Zimring & Hawkins, 1995), as well as the ethical and methodological obstacles to implementing

such a plan (Blackmore & Welsh, 1983; Chaiken & Chaiken, 1984; Cohen, 1983; Gottfredson &

Gottfredson, 1985a, 1985b; Moore et al., 1984; Visher, 1986; von Hirsch, 1972; von Hirsch &

Gottfredson, 1984). Moore and colleagues (1984) argued that selective incapacitation is at its

core an assertion that individuals vary in their capacity for evil and that these differences are

inherent within the individual and not the result of any social processes, which could be altered

or remedied.3 As the individual who arguably coined the term "selective incapacitation",

Greenberg himself cautioned that "the recommendationfor deciding who is to be incarcerated

on the basis of the threat such individuals would represent if released could be implemented only

if it were possible to predict how a given prisoner would behave if released" (Greenberg,

1975:543, emphasis in the original).

The notion of selective incapacitation, coupled with the empirical research on career

criminals then provided the rationale for the habitual offender three-strikes and truth-in-



3 Feeley and Simon (1992) echo Wilson (1987) in their assertion that policy and policy-makers view a certain
segment of the population as unemployable (even if jobs were available) and incapable of being integrated into
society, labeling this group with the term "underclass". This segment of marginalized individuals is also viewed as
dangerous to the social order, distrusted for their ability to disrupt as a collective group more so than as individuals.
Feeley and Simon (1992) expand by illustrating the extent of custodial supervision enforced on young black males,
arguing that the heightened visibility of minority overrepresentation in the penal system is likely to reinforce the
notion that crime is the result of this "pathological subpopulation" upon which nothing can be done with the
exception of custodial control. They furthermore argue that rehabilitation or reintegration only makes sense if those
being rehabilitated or reintegrated share common values and norms with middle class society. The existence of
permanent marginality of a subpopulation would render rehabilitative ideals irrelevant by nature of the argument.









sentencing legislation, which has since followed (Auerhahn, 1999; Benekos & Merlo, 1995;

Clark et al., 1997; Stolzenberg & D'Alessio, 1997). Based on the selective incapacitation ideal,

the goal of this legislation is to incapacitate the relatively small number of potential career

criminals early in their criminal life-course to achieve a substantial reduction in crime. One such

policy came in 1994 when Congress enacted the Violent Crime Control and Law Enforcement

Act, also known as the Federal Crime Bill (commonly referred to as the "three strikes and you're

out" law). This law mandates life sentences for criminals convicted of three violent felonies or

drug offenses if the third conviction is a federal crime (Lewis, 1994; Shichor, 1997).

The purpose of this legislation was to increase the severity of the sanction and lengthen the

term of confinement for offender groups deemed habitual or career criminals. Some three strikes

legislation requires prosecutors to pursue the case under the statute, calling for longer sentences

for offenders with one or more prior felony convictions, and often life sentences for three time

felons. Prosecutorial discretion is limited, unless he or she can plea bargain down the charges (in

which case discretion is enhanced for the prosecution), and judicial discretion reduced due to a

requirement to impose the maximum sentence in the event of conviction. Mitigating factors are

irrelevant to sentence considerations (Stolzenberg & D'Alessio, 1997). As the laws mandate such

extreme sentences, more offenders are taking their cases to trial rather than excepting plea

agreements, which has the effect of increasing the number of individuals in county jails awaiting

trial, more court trials which increases costs, more long-term prisoners leading to a need for

more prisons, an older prison population with increasingly high medical and health care bills,

and increasing welfare expenditures for dependents of incarcerated felons for longer periods of

time (Shichor, 1997).









The other sentencing legislation, beyond the three strikes and 10-20-life laws, has been the

truth-in-sentencing movement. This legislation has been enacted by many states to secure prison

funding under the 1994 Federal Crime Act (Greenwood, 1998). These laws usually require that

all violent offenders serve at least 85% of their sentence before they are eligible for parole or

early release. This is in stark contrast to the 30% to 40% that most of these offenders had been

serving prior to the legislation (BJS, 1992; Greenwood, 1998). The percentage of sentence

served was lower pre-legislation due to the ability to accrue good-time credit and the possibility

of early release. Truth-in-sentencing doubled the length of time these offenders serve, therefore

drastically increasing the prison population.

The important empirical question then becomes whether the three strikes and other habitual

offender legislation, or the truth-in-sentencing legislation has reduced crime. Greenwood and

colleagues (1996) projected the incapacitative effect of California's three strikes laws and

estimated that, if fully implemented, they would reduce serious felonies between 22% and 34%.

Furthermore, they estimated that nearly 1/3 of that reduction would be for violent crimes, and the

remaining 2/3 for less violent felonies. As Greenwood and colleagues (1996), and Stolzenberg &

D'Alessio (1997) have hypothesized, this reduction may be an underestimation, as the study did

not take into account any deterrent effect the laws may have for individuals contemplating

commission of one of the severely sanctioned offenses.

Projected analysis aside however, empirical research directly assessing the impact of the

three strikes legislation has not been so favorable. For example, the California three-strikes law

was examined by Greenwood and colleagues (1994) who concluded, if applied in all eligible

cases, the legislation would reduce the number of serious felonies in a year by about 28%, or

329,000 crimes. They caution however that this broad application would cost an additional $5.5









billion per year in additional criminal justice funding for the additional construction and

operation of prisons (Greenwood et al., 1994). This finding can be translated to a cost of $16,000

per serious felony prevented (Sherman et al., 1997).

Stolzenberg & D'Alessio (1997) assessed the impact of the California three strikes

legislation on serious crime using an interrupted time-series design. They concluded the laws did

not reduce crime levels below that which would be expected based on the preexisting downward

trend. Possible explanations offered by the authors for the lack of reduction include the

relationship between age and crime (most offenders would age out of crime anyway by the time

they are incarcerated for their third strike) and the prevalence of juvenile crime (juveniles

account for roughly 20% of the violent crime arrests and 35% of property crime arrests, yet are

unaffected by the legislation).

The conclusions made by Stolzenberg and D'Alessio regarding the aging out of offenders

is in keeping with one of the most commonly accepted extant relationships in the field of

criminology and broader research on criminal careers; that of the age crime curve and the

declining frequency of offending with age, regardless of sex, race, country, offense, or time

period (Farrington, 1986; Gottfredson & Hirschi, 1986; 1990; Nagin, 1998; Nagin & Land, 1993,

Petersilia, 1980; Tittle & Grasmick, 1998; Wilson & Herrnstein, 1985). As far back as the early

1970's research has found the proportion of offenders who desist their criminal career, or "drop

out", after each subsequent offense is constant and around 20% to 30% (Wolfgang et al., 1972;

Auerhahn, 1999). This high drop out rate means that while it is possible to predict with some

degree of accuracy that a certain proportion of a sample will be high-rate offenders, prediction at

the individual level of analysis is plagued by the stochastic sequence at hand. For example,

assuming 10% of a sample will be high-rate offenders, predicting which particular individuals









will be career criminals, is beset by the fact that 20% to 30% of even that high risk group will

desist after each additional offense. Alfred Blumstein proposed a now classic explanation of this

phenomenon, saying:

"Any stochastic sequence of events with a non-zero probability of termination after an

event will inevitably result in a distribution of sequence lengths. In criminal-career terms... since

every statistical distribution has to have a right-hand tail, the group of 'chronic offenders' who

comprise the right-hand tail will necessarily account for disproportionately large number of

offenses. The critical question is whether the members of this group are distinguishably

different... The fundamental policy question, then, is whether the 'chronic offenders' are

identifiable in prospect, that is, during the period in which they accumulate a record...unless

such discrimination can be made, any identification of chronic offenders can only be made

retrospectively, and so is of little policy or operational value (1979; reproduced in Auerhahn,

1999:709; and Petersilia, 1980:374).

The aging out phenomenon has led to the necessity to look for other factors that are more

predictive of re-offending than is the prior criminal history of the offender. Otherwise, by the

time the individual has enough priors to be deemed high-risk, that individual is likely to be near

the end of his/her criminal career or at the very least decreasing in frequency of offending.

Gottfredson and Hirschi (1986) argue the 20-20 hindsight of criminal career research to be

misleading in that "it turns out that the particular career criminals identified in criminological

research are no longer active and their replacements cannot be identified until they too are on the

verge of 'retirement'" (pp.217). These scholars see incapacitation policies doomed to failure in

that, with the knowledge that crime declines with age and peaks around 17 years of age, then an

incapacitation strategy must intervene at an optimal age to prevent the most frequent offending









(before 17 years of age), suggesting lengthy incarceration of young juveniles in the interest of

prevention.

A further argument criminal career research brings is the notion of specialization. The

habitual offender legislation call for longer sentences for violent offenders. This contradicts life-

course criminology research, which finds little to no evidence of specialization, especially over

the long-term, for high frequency criminals (Auerhahn, 1999; Blumstein et al., 1986; Blumstein

et al., 1988; DeLisi, 2005; Farrington et al., 1988; Kempf, 1987; Piquero et al.; 2003; Simon,

1997; Wright & Rossi, 1986).4 Legislation aimed at targeting offenders who specialize in violent

crime is misguided as this vast literature shows these offenders may not even exist.

Errors in Prediction

There exist two forms of error in the prediction of offender classification based on risk to

re-offend. In the event an individual is predicted to be a low rate offender but actually commits

crimes at a high rate, the error is one of underprediction. These errors of underprediction are

called false negatives. These errors involve threats to public safety by predicting low offense

rates for individuals who will in fact end up committing offenses at a high rate. The second type

of error, while less of a public safety risk, is perhaps more ethically challenging. This type of

error involves the prediction of an individual as being a high rate offender, who would in reality

commit offenses at a low rate. These false positives lead to the incapacitation of individuals to

prevent them from committing future crimes, which would actually not be committed. False

positives are also disturbing due to the high unnecessary costs and wasted resources of




4 While the studies referenced indicate offenders having low levels of specialization, especially over the long-term,
recently Sullivan and colleagues (2006) have demonstrated empirical evidence of a "considerable amount of short-
term specialization with regard to their offending patterns" (p.221).









incarceration of individuals for prolonged periods of time that would not have been a threat to

commit future offenses.

To reduce the number of false negatives, a larger fraction of the actual recidivists must be

identified, thus requiring the researcher to lower the "cutoff" score in the prediction instrument

to make the predictions more inclusive. This will, however, increase the number of false

positives; increase the already high incidence of predicting individuals as being high rate

offenders, who would in reality, commit offenses at a low rate. Conversely, to reduce the number

of false positives, the "cutoff' score would have to be raised to be less inclusive; yet this results

in the instrument missing an even larger number of actual recidivists (von Hirsch & Gottfredson,

1984). The essential fact to gather is that reducing the occurrence of one type of error will in all

cases increase the occurrence of the other type.

The extent to which these two types of errors occur in the prediction of risk to re-offend is

substantial. Research reveals that most assessments predicting offender risk have false positive

rates greater than fifty percent (Auerhahn, 1999; 2006; Blackmore & Welsh, 1983; Cohen, 1983;

Monahan, 1981; von Hirsch & Gottfredson, 1984). High error rates in the prediction of any low

base rate phenomenon or rare outcome have been found to be quite common; the rarer the event,

the greater the tendency to overpredict (Cohen, 1983; Copas, 1983; Farrington, 1987; von

Hirsch, 1972; von Hirsch & Gottfredson, 1984). The seminal selective incapacitation strategy

outlined in the above-mentioned report by Greenwood and Abrahamse (1982) resulted in a fifty

one percent overall success rate in classifying offenders. The rate of classifying an offender as

high risk who in fact is not (false-positive rate) was forty eight percent. In four attempts to

reanalyze the Rand data used by Greenwood and Abrahamse, a false-positive rate of fifty five









percent was discovered (Auerhahn, 1999; 2006; Chaiken & Chaiken, 1982; Cohen, 1983; Visher,

1986).

The Greewood study also evidenced a false-negative rate of about sixteen percent (von

Hirsch & Gottfredson, 1984). This corresponds to one of every six offenders classified as

medium or low risks actually committing frequent offenses. These "misses" or failures to

imprison offenders assessed at lower risk levels, yet actually high risk may cause political

pressure to make the definition of high risk more inclusive (which as mentioned above would

increase the rate of false positives).

Prediction Method

The primary task of risk assessment is to predict, not to explain the behavior in question

(Gottfredson & Snyder, 2005). Prediction involves first assessing particular criterion variables at

one time to forecast a specific behavior or performance at a later time (von Hirsch &

Gottfredson, 1984). Currently, there exists an extensive history of risk prediction of offending

behavior in the criminal and juvenile justice fields (Gottfredson, 1967; Gottfredson & Tonry,

1987; Mannheim & Wilkins, 1955; Simon, 1971). Prediction has certainly progressed both in

methodology as well as accuracy and can be grouped into four "generations" (Andrews & Bonta,

2003).

Originally, risk to re-offend was "predicted" using clinical/professional judgment; the "gut

feeling" approach. A variation of this first-generation assessment is called "structured clinical

judgment" which reflects decisions made after review of specific factors, but without a validated

process to link the scores to the decisions made (Andrews et al., 2006). Meta-analyses have

illustrated the weak predictive validity of first generation assessment with an overall mean r of

.12 (Andrews et al., 2006; Grove et al., 2000). While only available for sex offender samples









thus far, structured clinical judgment assessments appear to do better than unstructured, yet not

as well as later generation assessment (Andrews et al., 2006; Hanson & Morton-Bourgon, 2004).

Prediction, as the new penology took over, evolved into utilizing statistical techniques

correlating static characteristics of the individual (such as age, race, prior criminal history, and

prior substance abuse history) with the dependent behavior of offending/re-offending but was

theoretical (Andrews et al., 2006; Shichor, 1997). These second generation assessments

continue to be used extensively. The prediction studies described extensively below by

Greenwood and Abrahamse (1982) and Auerhahn (1999) both utilized second-generation,

actuarial, static predictor instruments.

Meta-analyses have found the predictive validity of second-generation assessments much

higher than that for clinical assessments, with an r of .42 (Andrews et al., 2006). Additional past

research has shown that informal, subjective, clinical judgments are far less accurate than

actuarial/statistical methods (Andrews & Bonta, 2003; Dawes, Faust & Meehl, 1993; Hoge,

2001; Meehl & Grove, 1996). Grove and colleagues conducted a 136 study meta-analysis

comparing actuarial approaches to clinical risk prediction, finding actuarial superior in 47% of

the studies, equally as predictive in 47% and less predictive in only 6% of the studies (Grove et

al., 2000). While the Grove team analyzed only clinical psychology and psychiatry studies, three

other meta-analyses focusing more on criterion behavior and offender samples, all finding the

statistical/objective risk measures outperformed clinical judgment (Bonta et al., 1998; Hanson &

Bussiere, 1998; Mossman, 1994). These studies predicted recidivism for mentally disordered

offenders, violent behavior among psychiatric patients and sexual offending.

The third-generation of prediction examined both static factors and dynamic factors (such

as antisocial attitudes and beliefs). Meta-analysis or even accessible reports of predictive validity









could only be ascertained from one third generation instrument, the Level of Service Inventory-

Revised (LSI-R). Results show the predictive criterion validity for this instrument to have an

overall mean r of .36 (Andrews et al., 2006). This is much better than the r of .12 for first

generation, yet lower than the .42 reported for second-generation prediction instruments.

Finally, the fourth-generation of risk assessment has attempted to assess static and

dynamic risk factors as well as protective factors and follows service and supervision from intake

through case cessation (Andrews et al., 2006). It is the addition of evaluating protective factors

that may help reduce the likelihood of re-offending and the guidance with the creation of case

management plans that advances the latest generation of prediction. Risk and protective factors

are not, however, mutually exclusive categories. For example, family may be a risk factor for

one juvenile with abusive parents who engage in criminality themselves, and protective for

another youth who has a supportive, consistent family structure. The inclusion of both risk and

protective factors highlights one of the distinguishing characteristics of fourth-generation risk

assessment in the area of increased attention to the linkage between assessment and case

management (Andrews & Bonta, 2003). Individualized strengths are built upon to construct a

prosocial orientation and responsivity factors and learning styles are considered when placing for

treatment.

As in the case for third-generation instruments, the predictive validity of only one fourth-

generation instrument has been reported. Andrews and colleagues (2006) report an estimate of

.41 for the Level of Service/Case Management Inventory (LS/CMI). Andrews and colleagues

(2006) do report that the best know fourth-generation instruments are the Correctional

Assessment and Intervention System (CAIS) and the Correctional Offender Management

Profiling for Alternative Sanctions (COMPAS).









Fourth-generation assessments appear to be equally as valid in terms of predictive criterion

validity as the more popular, yet significantly more simplistic, second-generation instruments.

Andrews and his co-authors furthermore suggest, "substantial improvements in the predictive

criterion validity of risk assessments may reside in reassessments of dynamic risk factors"

(2006:16; see also Andrews & Robinson, 1984; Motiuk et al., 1994; Roynor et al., 2000). This

notion is supported by two recent dissertations showing increased predictive validity with re-

assessment as well as assessment of dynamic risk factors that can change rapidly (Brown, 2003;

Law, 2004). Andrews and colleagues (2006) suggest that re-assessment with a focus on rapidly

changing risk factors may double or even triple the variance explained by intake assessments,

touting the idea "a breakthrough" (2006:16; see also Hanson & Harris, 2000; Quinsey et al.,

1997; Zamble & Quinsey, 1997).

Prediction Studies

Most prediction studies vary in degrees of complexity, yet utilize essentially the same

method (von Hirsch & Gottfredson, 1984). The research team collects data from a sample of

convicted offenders, usually criminal history and varying degrees of social history,

demographics, education, drug abuse, family history, etc. (Gottfredson & Snyder, 2005). The

sample participants are followed for some length of time and recidivism is captured for those

who commit subsequent offending (either self-reported or official). The original data collected is

examined in how it correlates to the subsequent recidivism, using bivariate or multivariate

statistical techniques.

The factors that correlate strongly with the recidivism are constructed into a predictive

instrument that is then tested on a validation sample. The final process of testing against the

validation sample is the most essential as it shows how well the instrument actually predicts

offending, rather than how well it correlates to the offending of the population from which the









data were gathered (Gottfredson, 1967; Gottfredson & Snyder, 2005; Mannheim & Wilkins,

1955; Simon, 1971). When an item is a statistically significant predictor of the dependent

variable in question, the item improves the power of the instrument and should be used in the

construction of the scale (Gottfredson & Snyder, 2005).5

Five basic steps have been shown essential in developing and utilizing prediction

instruments (Gottfredson & Snyder, 2005). First, the criterion to be predicted must be defined.

Second, the predictors are selected. Third, the relationships between the predictors and the

criterion must be measured. Fourth, the relationships found must be validated. Lastly, the method

of prediction must be applied in situations for which it was developed.

The steps outlined above have led some scholars to assert the seminal Greenwood and

Abrahamse study was not a prediction of risk to re-offend study at all (Auerhahn, 1999; Chaiken

& Chaiken, 1984; Clear and Barry, 1983; Visher, 1986; von Hirsch & Gottfredson, 1984).

Greenwood and Abrahamse used offender characteristics and self-reported past delinquency

from a single questionnaire to assess the correlation of the former with the later. The authors then

concluded that certain offender characteristics are "predictive" of offending behavior and

devised their seven-item scale. This process contradicts several of the steps of constructing a

prediction study in that Greenwood and Abrahamse do not have sequential time periods, only the

one-time questionnaire. They are in effect examining how current offender characteristics



5 Gottfredson and Snyder (2005) caution regarding the ethical implications of including certain significant correlates
in the scale construction which may actually simply be proxies for unmeasured items (such as race for poverty,
school failure, amount of police presence in the community, etc.). Any recidivism differences captured may be
correlated with the proxy but may not be caused by the proxy. They expand, stating that "statistically including the
race variable improves the predictions made by the risk scale... Ethically, excluding the race variable prevents the
assignment of risk scores that may inappropriately lead to greater sanctions for minority youth" (pp.iii). Gottfredson
and Snyder suggest including these variables (such as race) in the early stages of scale construction and then
removing them from the published instrument, thus removing the bias. If the race variable was never included in the
construction, other measures would be proxies for race and would keep racial bias inherent in the instrument, yet
hidden beneath the surface.









correlate with past offending behavior, and not analyzing how current characteristics predict

future behavior (which presumably was the intention).

Furthermore, Greenwood and Abrahamse did not utilize a validation sample once the

original scale was created, in order to assess "shrinkage". A replication/validation sample is

essential to determine the generalizability of the prediction instrument, as opposed to correlating

offending behavior with the original sample's idiosyncrasies (Gottfredson & Snyder, 2005; von

Hirsch & Gottfredson, 1984). "Shrinkage" refers to the portion of variation that is lost when the

prediction instrument is validated on the replication sample; usually the instrument accounts for

a smaller portion of variation on a subsequent sample than the original sample (Gottfredson &

Snyder, 2005; von Hirsch & Gottfredson, 1984).

Shrinkage is especially true with retrospective studies where the researchers are typically

not blind to outcomes in that they already are aware of which individuals in the samples have

recidivated. Researchers can then "fish for combinations of predictors that enhance the accuracy

of what is actually a 'postdiction'" which potentially "capitalizes on unique features of the

original sample -- features that will not be present in other samples" (Lanza-Kaduce et al.,

1990:28). Therefore, without the inclusion of a validation sample, the Greenwood and

Abrahamse study overstates the accuracy of their seven-factor scale. Research has consistently

shown that simple, less sophisticated statistical techniques are less vulnerable to shrinkage, such

as the Burgess basic additive approach, in comparison to complex multivariate techniques

(Lanza-Kaduce et al., 1990; Gottfredson & Gottfrdson, 1979; Gottfredson & Snyder, 2005;

Wilbanks, 1985). In addition to shrinkage considerations, the predictive accuracy of simpler,

easier to implement techniques does not differ significantly from more elaborate, sophisticated

techniques (Lanza-Kaduce et al., 1990).









Auerhahn (1999) replicated the Greenwood and Abrahamse scale on a representative

sample of California state prison inmates (n=2,188). The replication scale used identical, or

proxy measures as the Greenwood and Abrahamse scale. The seven items in the replication

include: 1) Prior conviction for any offense, 2) Free at least one year before the current

incarceration, 3) Arrest before age sixteen, 4) Served time in a juvenile detention facility, 5)

Regular use of heroin or barbiturates at any point in inmate's life, 6) Heroin or barbiturate use as

a juvenile, and 7) Employed before arrest resulting in current incarceration.

Identical to the Greenwood and Abrahamse study, the seven predictor variables were

summed and offenders grouped according to low-, medium-, and high-risk. Offenders were

classified as low-rate with scores 0-1, medium-rate with scores 2-3, and high-rate those with

scores of 4 or higher (Auerhahn, 1999). The replication scale resulted in 41% of the sample

being low-risk, 31% medium-risk, and 28% being classified as high risk to re-offend. Auerhahn

(1999) next compared the "predictive" risk to re-offend with the official prior convictions of the

sampled individuals. She notes the predictive accuracy (the percentage of offenders placed in the

right cell) of the Greenwood and Abrahamse scale was 51%, while the replication scale produced

slightly better accuracy at 60% (Auerhahn, 1999). Furthermore, 13% of the offenders who are

not high-rate were predicted as such (false positives) by the Greenwood and Abrahamse scale,

and only 5.6% were reported so by the replication. However, upon closer examination of the

Auerhahn study, the Greenwood and Abrahamse scale had 14% of the offenders not high-rate

predicted as such (not the 13% reported) and the replication scale produced 18% of all the

offenders being predicted high-rate who in actuality were not (not the 5.6% reported). This

means that only 15% of all the offenders predicted high risk in the Greenwood and Abrahamse

study (not the 29% predicted), and only 10% of the offenders predicted high risk in the









replication study (not the 28% predicted), were in actuality high risk to re-offend. The two

studies, therefore, demonstrate false positive rates around fifty percent for all the individuals

predicted as high risk to re-offend as nearly half of those predicted high risk actually recidivated.

Cronbach's alpha, a measure of internal reliability for multiple-item scales based on the

strength of correlations between the items in the scale, is reported as .38 (Auerhahn, 1999).

Alpha values range from 0 to 1, with 1 representing perfect reliability of the assessment tool. An

alpha value of .70 or higher is usually recognized as acceptable for reliability in the social

sciences (Auerhahn, 1999). Auerhahn (1999) reports all of the inter-item correlations of the scale

significant at the p=.01 level, yet cautions that some associations are small enough that their

significance is attributable mainly to the size of the sample, as it is easier to achieve significance

with larger samples. Also pointed out is the finding that many of the associations are in fact

negative, though coded for agreement, which is counter to the requirement of positive direction

association for additive scales.

The Auerhahn (1999) study intentionally elaborates on the question of internal reliability

due in part to the brevity or outright exclusion of the issue in previous research. The study

mentions prior research focusing on the reliability of using inmate self-report data (Greenwood

and Turner, 1987; Spelman, 1994; von Hirsch & Gottfredson, 1984), estimation of incapacitation

effects of selective incapacitation strategies, or the ethical issues surrounding the basis of

sentencing decisions on prediction of future offending, rather than actions the offender has

decided to take, yet little attention to the methodological issues inherent in the construction of

prediction instruments (Auerhahn, 1999; Gottfredson & Gottfredson, 1985). Auerhahn (1999)

presents the internal reliability, false positive rate, and scale alpha values for the purpose of









illustrating the shortcomings of actuarial risk assessment in terms of driving a policy of selective

incapacitation.

Even as research shows actuarial risk assessment superior to clinical judgment and "gut

feeling" approaches, there is still a roughly 50% false positive rate of predicting offenders to be

high-risk when, in actuality, their offending patterns will not play out that way. Greenberg

(1975) cautioned over thirty years ago the inaccuracy of prediction in criminal behavior suggests

the selective incarceration of individuals on a predictive basis was not a promising direction.

Nearly twenty-five years later, Auerhahn concluded that the "obstacle to realizing this seemingly

perfect solution to crime prevention lies in the prospective identification of this offender

pool...we simply cannot do it with any reliable accuracy", and that "it appears that the time has

come to open the discourse of crime and crime prevention to permit the consideration of

strategies that do not focus on incarceration in formulating our response to the crime problem"

(1999:727-728).

What Works?

The "nothing works" mantra sparked by the Martinson report pushed corrections policy

into relying on incapacitation as it's predominate strategy with regard to addressing crime, yet

also spawned a quite contradictory line of inquiry; that of the "what works?" movement. This

empirically based research implies that certain programs and interventions, when properly

targeted, reliably produce reductions in recidivism (NIC, 2004). This line of inquiry has been

fueled mainly by the results of meta-analysis, cost-benefit analysis, and clinical trials, all

providing indications of how to more effectively reduce the likelihood of re-offending (Andrews

et al., 1990a; Andrews et al., 1990b; Andrews et al., 1999; Antonwicz & Ross, 1994; Aos, 1998;

Cullen, 2002; Gendreau, 1996; McGuire, 2002; NIC, 2004; Sherman et al., 1997). The meta-

analysis technique has been especially instrumental in the identification of effective intervention









characteristics and permits the aggregation of a large body of research literature in order to

examine and compare the effect sizes for treatment versus control group comparisons. Meta-

analyses summarize a large number of studies (a rigorous "study of studies") while controlling

for important methodological issues (Sherman et al., 1997).

"What works" is now often used interchangeably with the term "evidence-based practices"

(NIC, 2004). Evidence-based practices (EBPs) have recently become a trend throughout the

social sciences and human service fields and focus primarily on outcome. The outcome with

regard to criminal justice practices is almost exclusively the reduction of risk to re-offend and

subsequent offending. Bogue and colleagues stress that "interventions within corrections are

considered effective when they reduce offender risk and subsequent recidivism and therefore

make a positive long-term contribution to public safety" (NIC, 2004:1). Research on offender

rehabilitation has progressed to the point where there exists sufficient evidence of several key

interdependent principles (Andrews & Bonta, 2003; Burrell, 2000; Carey, 2002; Corbett et al.,

1999; Currie, 1998; Latessa et al., 2002; McGuire, 2002; NIC, 2004; Sherman et al., 1997).

Meta-analyses have been used both to identify individual risk factors associated with recidivism,

as well as to determine the characteristics of the most effective delinquency treatment programs

(Andrews & Bonta, 2003; Andrews et al, 1990; Lipsey, 1989, 1992; Palmer, 1992). Research and

practitioner communities have moved progressively from "nothing works", to "what works", and

are now learning what in fact makes "what works" work (Andrews, 2006; Antonowicz & Ross,

1994).

Empirical Predictors

Meta-analyses have led to the ability to identify major and minor risk/criminogenic need

factors with respect to offending behavior (Andrews et al., 2006; Bonta et al., 1998; Gendreau et

al., 1996; Hanson & Morton Bourgon, 2004; Lipsey & Derzon, 1998). The major risk factors can









be divided into the "big four" and the "central eight" (Andrews et al., 2006; Andrews & Bonta,

2003). The big four include 1) history of antisocial behavior, 2) antisocial personality pattern, 3)

antisocial thinking, and 4) antisocial peer associations. The four additional risk factors rounding

out the central eight include 1) family problems, 2) school/work problems, 3) procriminal

leisure/recreation activities, and 4) substance abuse. Minor risk factors include such issues as

emotional distress, mental disorders, low socio-economic status, seriousness of the presenting

offense, low IQ, and physical health problems.

Research has demonstrated the minor factors to be of relatively low predictive validity, if

present at all, and that they most likely can be explained away by the contributions of the big

four, whose inclusion pushes the minor factors to insignificance (Andrews et al., 2006). For

example, the predictive power sometimes associated with mental disorders' relation to offending

is believed in actuality to be related to antisocial attitudes/thinking, antisocial personality pattern,

and substance abuse (Andrews et al., 2006; Link et al., 1992; Swanson et al., 1996).

Furthermore, the power associated with low socioeconomic status, or socially disadvantaged

neighborhoods can be attributed to the increased prevalence of the big four risk factors within the

individuals with lower socioeconomic status and the residents in those disadvantaged areas

(Stouthamer-Loeber et al., 2002).

The meta-analyses uncovering the big four and the central eight risk factors demonstrate a

significant reduction in recidivism when the factors are targeted through effective interventions

(Andrews et al., 2006; Andrews & Bonta, 2003). It is essential to understand the risk associated

with each of the eight major factors, as well as the dynamic need of the individual in terms of

what is needed to address the risk. It is a well-known criminological mantra that the best

predictor of future behavior is past behavior (Sherman et al., 1997). The first of the big four,









history of antisocial behavior, is the often-utilized criminal history item present in predictive

instruments since the inception of second-generation assessments (Andrews & Bonta, 2003).

This item is characterized by early and continuous involvement in deviant behavior in a variety

of situations and manifestations. Individuals presenting with prior criminal history (be it self-

report or official) are in need of skills to identify non-criminal alternative behaviors in risky

situations.

The second factor, antisocial personality pattern, is characterized by an impulsive,

pleasure-seeking, and risk-taking personality. This risk factor, though believed to be dynamic in

nature in this context, is similar to Gottfredson and Hirschi's (1990) concept of low self-control.

Individuals presenting with antisocial personality patterns place a high value on adventurous

thrill-seeking and immediate gratification, which increases the probability of engaging in deviant

behavior. This dynamic criminogenic need can be addressed by building problem-solving, anger

management, and coping skills (Andrews et al., 2006).

Antisocial attitudes and thinking is the third of the big four risk factors. This factor is

characterized by attitudes, beliefs, and values that are supportive of criminal behavior. These

procriminal sentiments "represent the pool of justifications and exonerating statements that the

person has available in any particular situation" (Andrews & Bonta, 2003:178). Antisocial

cognitions enable individuals to rationalize deviant behavior, thus increasing the likelihood of its

occurrence. Individuals with these beliefs and attitudes are in need of cognitive restructuring to

reduce antisocial attitudes and recognize risky thinking triggers, so as to engage in alternative

prosocial behaviors.

The fourth of the major risk factors which need to be targeted to reduce recidivism is the

extent of antisocial peer associations. Deviant peer associations have been the subject of









countless empirical investigations, and their influences have proven significant and withstood the

test of time (Akers, 1998). Association with criminal peers and the seclusion from pro-social

peers leads to procriminal modeling and adjustments to the reward and cost structure associated

with deviant behaviors, thus making the likelihood of continuation of those behaviors more

likely (Akers, 1998). This dynamic criminogenic need can be addressed by reducing the

associations with antisocial others and enhancing the exposure to prosocial peers.

Negative family situations are the fifth identified risk factor of the central eight. The key

elements of this need are nurturing and support, and monitoring and supervision. Individuals

who are reared in environments which lack support and monitoring of behavior are at increased

odds of engaging in deviant behavior (Andrews et al., 2006). The notion of behavior monitoring

is central to Gottfredson and Hirschi's explanation of deviant behavior with their General Theory

and its emphasis on the monitoring, recognition, and consistent punishment of deviance

(Gottfredson & Hirschi, 1990). Addressing this dynamic need requires the reduction of conflict

within the family environment, enhancing the three areas of effective parenting of monitoring,

recognition, and punishment, and building positive relationships within the family structure.

The sixth factor identified in the empirical research is in the domain of school and work

problems. The intent of this factor is to identify low levels of performance and satisfaction in the

areas of school and employment. Enhancing the individual's performance in school or work and

improving satisfaction or increasing rewards is an important way to address this dynamic need.

Problems in the area of recreation and leisure activities are the seventh risk factor of the

central eight. This factor is composed of low levels of involvement in structured prosocial

activities and pursuits. A low level of satisfaction with pursuing prosocial activities may also be









involved. Ways to address this dynamic need include increasing the rewards associated with

engaging in anti-criminal activities, and enhancing the involvement in prosocial pursuits.

The final risk factor domain is substance abuse. This simply points to the increased

likelihood of offending or re-offending for individuals who abuse alcohol or drugs. In order to

address this dynamic need, a reduction in the abusing behavior must occur. Drug education has

been found successful, but only when accompanied with building relapse prevention skills to

identify triggers and alternative behaviors (Sherman et al., 1997). Historically, second generation

prediction instruments have utilized an individual's history of substance abuse, but newer third

and fourth generation assessments have begun to separate prior use from current use of alcohol

or drugs.

Principles of Effective Intervention

A meta-analysis of 374 controlled experimental studies of the treatment effects of various

interventions on recidivism was conducted by the Carleton University (Andrews, et al., 1999).

This analysis helped to reveal three principles of effective intervention: 1) the risk principle, 2)

the need principle, and 3) the responsivity principle. The risk principle expresses the finding that

intensity of treatments and interventions should mimic the intensity of the risk to re-offend. More

intense services should be delivered to the highest risk youth (Andrews & Bonta, 2003; Harland,

1996, McGuire, 2002; Sherman et al., 1997). Criminological literature actually shows high

intensity services delivered to low-risk to re-offend youth are iatrogenic; having the negative

consequence of actually increasing recidivism. It is important to remember the risk to re-offend

concept is not to be confused with the seriousness of the presenting offense for which the

individual may have come in contact with the criminal justice system for (Andrews, 2006). An

individual may be lower risk to re-offend, regardless of the apparent heinousness of the

presenting offense.









Studies spanning nearly thirty years have observed the negative effect of providing intense

services to lower risk to re-offend youth (Andrews & Kiessling, 1980; Baird et al., 1979; Bonta

et al., 2000; O'Donnell et al., 1971). In these studies, low-risk to re-offend individuals had higher

recidivism percentages when they received intensive treatment compared with low-risk to re-

offend individuals who received minimal treatment, and high-risk to re-offend individuals had

lower recidivism percentages when receiving intensive treatment compared to high-risk

receiving minimal treatment, all illustrating the importance of the risk principle. Furthermore, a

greater percentage of high-risk individuals have a need for pro-social skills and thinking and will

demonstrate a greater improvement, as lower risk individuals re-offending is a much lower base-

rate phenomenon to begin with, thus making for a greater "bang-for-the-buck" as well (Andrews

& Bonta, 2003; Gendreau, 1996; Harland, 1996; McGuire, 2002; NIC, 2004; Sherman et al.,

1997).

The need principle, the second principle of effective intervention, holds that services

provided should address individualized criminogenic needs relevant to each individual.

Criminogenic needs are dynamic, changeable factors associated with re-offending behavior that,

when addressed or changed, affect the individual's risk to re-offend (NIC, 2004; Sherman et al.,

1997). These needs are in contrast to static factors, such as age at first offense, prior substance

abuse, or gender, which, while strong predictors of recidivism are not amenable to treatment and

can not be change through intervention. The strongest dynamic correlates of crime are anti-social

peer relationships, family factors, substance abuse, and antisocial attitudes, values and beliefs

toward authority, education, and employment and low self-control (Gendreau, Andrews, Cogin

& Chanteloupe, 1992; NIC, 2004). Seven of the central eight major risk factors outlined above

(all except history of antisocial behavior) are dynamic needs that can be addressed through









intervention/treatment. Programs and interventions successful in reducing these criminogenic

needs can expect corresponding reductions in recidivism, especially when the needs are

prioritized through valid assessment so that services are focused on the greatest ones (Andrews

& Bonta, 2003; Gendreau et al., 1994; Harland, 1996).

An important caveat to the need principle is the notion of appropriateness. Individuals

should only be treated, or receive interventions targeting, an actual need they possess. Research

has shown it to be iatrogenic to provide blanket approaches and non-targeted services to

individuals not presenting with the specific need addressed (Andrews, 2006; Brown & Campbell,

2005). For example, providing substance abuse services to an individual free of substance abuse

problems may lead to an actual increase in recidivism (Brown & Campbell, 2005). This is in

direct opposition to those, operating opposite to empirical findings, holding that simple education

(such as substance abuse education) will deter individuals from every following a certain path.

Andrews suggests appropriate targets including "building up low-risk alternative behavior in

high-risk situations, reducing antisocial associations, and building up levels of reward,

satisfaction, and monitoring in home, school, work, and leisure settings" (2006:598).

The third principle derived from the Carleton University databank was the responsivity

principle. There are two types of responsivity, general and specific. General responsivity asserts

that behavioral, social learning, and cognitive-behavioral strategies are the most effective when

the preferred outcome is recidivism reduction (Andrews et al., 1990c; Andrews et al., 2006; Izzo

& Ross, 1990; Lipsey, 1992). These strategies are more effective due to their use of techniques

such as role-playing, modeling, graduated reinforcement, and skill-building (Hubbard, 2002).

Lipsey (1992) evaluated over 400 studies in a meta-analysis and concluded that not only was









treatment more effective than control groups, but that type of treatment was the most important

variable.

Cognitive-behavioral treatment is at the forefront of correctional treatment. Behavior

therapy is a technique based on the principle that environment has an effect on learning and

behavior through operant and classical conditioning (Hubbard, 2002). Behavior therapy assumes

modeling to be essential in learning and that offenders have learned antisocial behaviors by

watching others and the rewards/punishments achieved, and by being rewarded themselves for

behaviors in the past. Antisocial behaviors, and attitudes conducive to them, are continued as a

result of these reward contingencies. Behavioral therapy for offenders utilizes techniques of

behavioral rehearsal, modeling, and token economies (Hubbard, 2002).

Cognitive therapy posits errors in thinking lead to problematic behavior (Yochelson &

Samenow, 1976, 1977). This faulty thinking, coupled with offenders' impulsive tendencies, lack

of problem solving skills, and failure to consider the consequences of their actions, is what drives

delinquency (Hubbard, 2002). Cognitive therapy focuses on changing these thinking processes

and patterns, and not on directly changing the behavior. Cognitive-behavioral therapy is a

combination of the aforementioned two therapies and includes both a focus on the cognitions of

offenders and changing those patterns, as well as utilizes behavioral strategies of modeling, role-

playing, behavioral rehearsal, and reward/punishment contingencies. Cognitive-behavioral

therapy teaches offenders "how to develop self-conrol, manage their anger more appropriately,

develop empathy through role playing, improve problem solving abilities, and develop their level

of moral reasoning" (Hubbard, 2002:38).

Specific responsivity is based on the assumption that not all offenders are alike and certain

personal characteristics of offenders have the possibility of interfering with treatment (Andrews









& Bonta, 2003). Specific responsivity argues for the need to match services provided to the

learning styles, personality, motivation, developmental stages, mental disorders, and

demographics such as age, gender, and ethnicity of the individual (Andrews et al., 2006; Gordon,

1970; Miller & Rollnick, 2002). Specific responsivity is based on the assumption that the

personal characteristics mentioned above may interfere with treatment and therefore need to be

addressed. The essential issue of specific responsivity is the moving from what works in

correctional treatment to which cognitive-behavioral curriculums work best for which offenders.

There exists a plethora of empirical research on general responsivity, while little more than

theoretical speculation and hypothesizing has been undertaken with regard to the notion of

specific responsivity. Van Voorhis (1987) holds that program evaluations will mask the

effectiveness of treatment if offender classification, in terms of risk level, personality

characteristics, and other responsivity characteristics, is not utilized. This lack of classification

and matching offenders to appropriate treatment may lead to the assignment of individuals to

harmful interventions that may even increase their risk of re-offending, making treatment

programs appear to be failures when they may actually just have been misdirected.

One exception to this gap in specific responsivity literature has been a dissertation

conducted by Hubbard (2002). The purpose of the study was to ascertain whether the

characteristics of gender, depression, self-esteem, history of sexual abuse, intelligence, or

personality affect treatment effectiveness. The question is whether these characteristics can affect

whether individuals are successful in treatment, are capable of understanding the treatment, can

focus while in treatment, and has the ability or willingness to alter behavior (Hubbard, 2002). Of

the characteristics listed above, only gender and history of sexual abuse were significantly

related to subsequent recidivism, with males and individuals with a history of sexual abuse more









likely to recidivate. Gender remained significant after controlling for risk to re-offend level,

quality of programming, sexual abuse, depression, and self-esteem; all possible variables for

which gender could possibly have been a proxy measure. Contrary to previous work (Fabiano et

al., 1991) intelligence was not related to the outcome measure of recidivism, and those with

lower IQ's performed equally well as those with higher IQ's.

Hubbard (2002) offered the possibility that the significance of gender may be due to

different staff quality, leadership, programming or overall implementation of the curriculum for

females. Another explanation offered was that perhaps women are more receptive to the

strategies of cognitive-behavioral treatment. While possible, this position does not explain the

fact that females, even in no-treatment control groups, always recidivate at a lower rate than

males. The final explanation offered for the gender significance was that perhaps females are

more motivated or able to take responsibility for their actions. This conclusion remains

speculative, as noted by Hubbard (2002), since no control for or assessment of motivation was

made. Overall, Hubbard (2002) found little support for specific responsivity and the possibility

of personal characteristics impacting treatment effectiveness, except for gender and history of

sexual abuse. The possibility that cognitive-behavioral treatment succeeds at reaching a wide

variety of learning styles may negate the effects of personal characteristics on treatment success

(Hubbard, 2002).

The Carleton meta-analysis uncovered the need for services provided to use behavioral,

social learning, or cognitive-behavioral strategies rather than unstructured, insight-oriented,

nondirective counseling or get tough approaches such as deterrence-oriented or "Scared Straight"

interventions (Andrews & Bonta, 2003; Sherman et al., 1997). The structuring strategies that

have been found to work effectively include differential reinforcement, motivational









interviewing, practicing of new skills, and modeling (Andrews, 2006). The meta-analysis found

that when correctional interventions adhered to the risk, need, and responsivity principles, the

mean effect size was .26 in sixty tests of treatment. This is in contrast to .18 in eighty tests when

only two of the principles are followed, and .02 in 106 tests when only one of the three are

followed (Andrews & Bonta, 2003).

Recently, Andrews has incorporated a few more essential elements of effective

programming, which he terms the "supplementary" principles of integrity and intervention

implementation (Andrews, 2006). These supplementary elements deal with management and

organizational aspects and help make for a more complete picture of what is needed for

rehabilitative effectiveness (Andrews, 2001; Andrews & Bonta, 2003; NIC, 2004). The first

additional element is the necessity of implementing and validating a structured risk/need

assessment instrument. Differential risk designations and targeting of individualized

criminogenic needs is critical to case management and treatment progression. Simply having

effective instruments is not enough, if they are never fully implemented and utilized in the case

planning process (Goggin & Gendreau, 2006). These elements illustrate the need for validated

assessment instruments and demonstrate the importance of the current study, which attempts to

examine the predictive ability, and construct validity of the PACT assessment.

Andrews (2006) also argues for the necessity of having program managers or directors and

supervisors cognizant and attentive to the relationship skills and the structuring skills of the

service delivery staff (ability of delivery staff to employ the responsivity principle).

Administrators or supervisors must carefully select, train, and supervise delivery staff according

to the risk/need/responsivity principles to ensure compliance and proper implementation of the

program. Staff competence has been found critical in enhancing the effectiveness of









programming and intervention to reduce recidivism, as evidenced by Barnoski's (2003) analysis

of the conditions necessary for "blueprint" programs to successfully work. This principle was

further evidenced by the possibility of the failure of Project Greenlight to be attributed to two

case managers with inadequate training and skills in program delivery (Brown & Campbell,

2005).

Additional research has identified other principles of effective evidence-based practice.

The fidelity principle holds that it is critical to monitor the implementation of treatment programs

to ensure the curriculums are delivered in accordance to the way in which they were intended

and designed. Many programs proven to be effective at reducing recidivism are delivered

according to manualized curriculums that are partially scripted so as to ensure all facilitators

deliver the lessons in identical fashion. The fidelity principle holds that it is beneficial to

periodically monitor the delivery of treatment groups and coach facilitators.

Research has demonstrated that even sound and empirically supported programs, when not

delivered with adherence to the protocol, will not have the extent of recidivism reduction evident

in those programs that monitor treatment staff to ensure proper delivery (Bamoski, 2004; Lipsey,

1992; Lowenkamp et al., 2006). Simply choosing an evidence-based program from available

lists, such as the University of Colorado "blueprints" programs is not enough, as even "model"

programs such as Aggression Replacement Training (ART), Multisystemic Therapy (MST), and

Functional Family Therapy (FFT) have failed when not properly implemented (Andrrews, 2006;

Bamoski, 2004). There exists extensive literature that even programs that were empirically

shown to be effective in one place and time were not so in another, due to differences in

implementation (Welsh, 2006; Welsh & Harris, 2004). The major aim of the fidelity principle is

to monitor the implementation quality and treatment fidelity to ensure programs are delivered the









way in which they were designed and intended to maximize program success and recidivism

reduction (Lipsey, 1993; McGuire, 2002; Miller & Rollnick, 2002).

Gender Differences in Offending

Gender is one of the single best predictors of delinquency, as males continue to offend at

higher rates than their female counterparts. However, in recent years, research has begun to

discover that, while rates of delinquency are declining for both males and females since 1997,

the decline has not been shared equally among the sexes, nor does it offset the substantial

increase in criminal offending between 1984 and 1997 (Snyder, 2003). Snyder and Sickmund

(1999) have found arrests of adolescent females between 1987 and 1999 to have increased 76%,

while arrests of adolescent males increased 42% (see also Chesney-Lind, 2001). Arrests for

female juveniles increased 50.3% between 1981 and 1997, while arrests for male juveniles was

observed at only a 16.5% increase, according to the Uniform Crime Reports (Chesney-Lind,

2001). A more alarming fact may be the increase in violent crime for female adolescents during

the same period of 103%, while the arrests for male juvenile violent crime only increased 27%

(Chesney-Lind, 2001). Despite the fact that male juveniles still account for the majority of

juvenile crime (as is also true of adult criminality), female adolescents are unmistakably

narrowing the gender gap. These statistics illustrate the necessity of examining the potential for

gender differences with regard to risk and protective factors to offending as well as the

predictability of gendered offender. The differences in the correlates of offending for males and

females are central to the current focus of this study.

While many factors have empirically been related to delinquency, the question at hand has

to do with whether or not the factors that predict female adolescent delinquency are different

from those that predict male juvenile criminal involvement. An additional to concern is the issue

of differential significance. Perhaps, even though identical factors may predict delinquency of









both sexes, certain factors are more important in the prediction of criminality of one sex versus

the other due to differential exposure.

Scholars have argued both sides of the coin as to whether identical prediction instruments

can be utilized for females and males. Reisig and colleagues (2006:385) state, "whether actuarial

risk tools... can effectively gauge women offenders' risk and needs continues to be debated".

Scholars more orientated toward a feminist perspective argue that existing actuarial risk

assessments ignore the gendered context of female criminality, as they were derived from male

theories of crime and delinquency. These scholars contend existing tools fail to consider the

economic disadvantages, propensity for drug-related offense involvement in the criminal justice

system, and the importance of prior victimization, many females face (Covington, 2003;

Holtfreter and Morash, 2003; Reisig et al., 2006). Proponents of identical risk assessment

utilization maintain that the "potential influence of factors highlighted by feminist scholars (e.g.

victimization, poverty, substance abuse) is captured... through offenders' attitudes, peer

associations, and the like" Reisig et al., 2006:385; see also Andrews and Bonta, 2003; Dowden

and Andrews, 1999).

The scholars oriented towards the feminist gendered context of female criminality believe

the utility of risk assessment tools is suspect since those tools were developed based on male

theories of crime and delinquency and therefore ignore the specific risks and protective factors

associated with female offending. These scholars hold the view that traditional criminological

theories and empirical research were developed by men who brought forth assumptions about

social life, society in general, and offending to predict criminality. The samples used in the

studies to test and validate their assumptions were exclusively male and therefore would not be

applicable or generalizable to female criminality, as feminists reject the notion that the same









functions and correlates explain both male and female delinquency (Chesney-Lind, 1989; Daly

& Chesney-Lind, 1988; Reisig et al., 2006; Simpson, 1989).

Reisig and colleagues argue "male-centered theories fail to appreciate a variety of elements

unique to female criminality, including the greater relative likelihood of young women fleeing

abusive homes and experiencing domestic abuse as adults" (Reisig et al., 2006:388; see also

Daly, 1992). Victimization during both childhood and as a young adult, alcohol and drug abuse,

and economic disadvantage have all been disproportionately connected to female criminality in

empirical literature (Broidy & Agnew, 1997; Covington, 2001; Reisig et al., 2006).

Daly (1992) suggests the need to differentiate women according to gendered pathways.

Daly's pathway framework examines how abuse history, substance use, economic disadvantage

and destructive intimate relationships are differentially dispersed throughout the female

population. Daly suggests disaggregating females into those who have prior histories with the

aforementioned factors from those who do not, so as to more effectively predict the risk to re-

offend associated with each group separately rather than as a whole. Daly suggests the need for

four separate pathway groupings; street women, drug-connected, harmed and harming, and

"other" (1992, 1994). Prior theory and research has not made similar assertions for the

disaggregation of male offenders.

An additional attempt to answer the fundamental question of differential involvement and

the reason for it can be extrapolated from Agnew's general strain theory. General strain theory

has been found unable to explain the higher rates of male delinquency in terms of amount of

exposure to strain, as females experience at least as much strain as males (Broidy & Agnew,

1997). Gender differences in the types of strain experienced, such as material/financial, ties to

others, family violence, and procedural justice, may foster the explanation of gender differences









in crime type; i.e. what offenses females and males most typically engage (see Broidy & Agnew,

1997).

Furthermore, gender differences in the response to emotional strain (particularly anger)

help to explain differences in delinquency. Males are more likely to respond to anger with

external aggressive behavior (inflicting harm on others), while females more likely to engage in

self-directed behaviors (self-mutilation, substance abuse, disordered eating) when experiencing

anger. Finally, general strain theory is useful for the explanation of gender differences in that

males are more likely, due to differences in coping strategies, social ties, criminal opportunities,

social control, and disposition to crime, respond to a given amount of strain with more serious

and violent behaviors (Broidy & Agnew, 1997). In tandem with the gender differences described

by the authors, Broidy and Agnew assert the importance to "note the above gender differences,

to the extent that they exist, involve differences in degree rather than kind' (1997:296, emphasis

in the original). The picture is not one in which different factors explain male and female

criminality, but that different degrees of emphasis of those factors predict delinquency.

Reisig and colleagues (2006) argue, "the research addressing the predictive accuracy of

risk assessment tools using samples of women offenders is relatively limited" (pp.387). One

study examining the creation of actuarial risk assessments separately for female and male

offenders shows the male model to be similar to the combined sample construction, but the

female model to resemble neither the male model nor the combined sample model (Funk, 1999).

By creating a female specific model, the adjusted R-squared increased from .21 for the combined

sample model to .32 for the female only model (Funk, 1999). Interestingly enough, the male-only

model achieved an R-squared of only .20, meaning the authors were able to predict female re-

offending significantly better than male re-offending.









Contrary to the view presented in the Funk study, is the meta-analysis conducted by

Simourd and Andrews (1994), who examined 60 studies and 464 correlations between

delinquency and risk factors to determine whether specific factors are more important for one

gender than for the other. These authors found no differences in risk factors across gender; the

general risk factors important for male delinquency were identical to those important for female

delinquency. Furthermore, and even more essential is that no risk factor was found to be more

important for a particular gender, however, victimization was not included as too few studies

utilized that correlate. The inability to include victimization, and even more specifically to

separate different types of abuse/neglect is a disadvantage in light of the gendered pathways

research described below. The current study will be able to examine this factor in more detail

and with substantial sample size to allow for gender comparison of the role of abuse/neglect.

Gender differences in offending are argued to have serious repercussions with respect to

utilizing scores from actuarial risk/needs assessments. For example, women charged with serious

violent offenses are most often accessories to the event and not the primary instigator, a situation

that would lead to over-classification of risk to re-offend on any actuarial instrument considering

previous engagement in a violent offense as leading to risk (Reisig et al., 2006). The over-

classification argument is rejected by supporters of actuarial tools who acknowledge a dearth of

empirical research to support the notion (Blanchette, 2002; Reisig et al., 2006). The importance

of the over-classification argument is essential to the possibility of equality in the prediction of

male and female likelihood of offending, and thus necessitates the inclusion of a focus on

gender-specific false positive calculations. Essential to the current study is the need to ascertain

whether female offenders predicted to offend yet subsequently abstain is disproportional to those

instances with males.









Current Focus

The PACT risk/needs assessment has not yet been formally validated in terms of predicting

recidivism in the state of Florida for the youth served by the Department of Juvenile Justice. An

initial pre-validation examination used proxy measures of criminal history in an attempt to assess

whether the risk levels presented closely match in percentage those attained using the

Washington State assessment upon which the PACT was based, which they in fact have. The

actual PACT assessment of official criminal history was not used, and no social history items

were included in the scoring, as the PACT includes. Additional preliminary reports and analysis

indicate that the PACT risk levels (as presented by the actual PACT assessment) do indeed

predict recidivism, with higher risk to re-offend youth recidivating at higher rates (Winokur &

Blankenship, 2007). However, these initial validation attempts have only been able to track

youth up to six months after PACT administration and have not examined whether the

assessment predicts recidivism equally well for females as it does males. Formal gender

comparisons on any level of complexity have unfortunately been lacking up to this point.

The focus of this study is on assessing the validity of the PACT assessment for predicting

recidivism, tracking juvenile offenders for over twelve months post-administration, and

providing gender comparisons for instrument effectiveness. The current study will also examine

the relationship between risk to re-offend and time to failure, hypothesizing higher risk to re-

offend youth more likely to commit a new offense in a shorter time post-assessment than lower

risk youth. The study will furthermore investigate whether identical items should be included to

predict female recidivism as used to predict male recidivism.

The PACT risk/need assessment is designed to assess a youth's overall risk to re-offend

level. Therefore, it is hypothesized youth with higher overall risk levels will be more likely to

recidivate than youth with lower overall risk levels. It is hypothesized that the assessment will be









valid for youth assessed who remain in the community, or who are assessed in residential

commitment and tracked post-release, as the risk factors for recidivism are similar for all youth.

The third hypothesis relates to the scoring of the PACT. Individuals with higher criminal history

scores (more serious official prior offending histories) will be more likely to recidivate than

individuals with a lower criminal history score. The notion of more extensive serious prior

offending being predictive of a higher likelihood of recidivism is not new to criminological

literature (Gottfredson & Hirschi, 1990; Moffitt, 1993; Sherman et al., 1997). However, the

criminal history of the PACT as it is scored has not been validated with respect to recidivism

when tracking youth twelve months post-assessment. Additionally, it is predicted youth with

higher social history scores will be more likely to recidivate. The social history score has never

been examined as it relates to recidivism for youth tracked twelve months post-assessment.

Central to the current study is the importance of gender comparison. No attempts have yet

to be made to examine the validity of the PACT assessment with regard to whether the

instrument predicts recidivism as well for female youth as it does for male youth. Additionally,

research examining the predictive accuracy of actuarial risk assessment of female offenders is

relatively sparse (Reisig et al., 2006). The overall risk score presented by the PACT instrument is

hypothesized to be a significant predictor for male recidivism and female recidivism. In other

words, it is hypothesized that there is no need for any other gender-specific scoring differences,

other than the additional one point on the social history score males automatically receive. The

PACT currently gives males one point on their social history score automatically, for being male,

and females zero points automatically.

The additional point assignment for males is in relation to the commonly accepted extant

relationship in the field of criminology between gender and offending, finding males









significantly more likely to engage in offending behavior than females (Farrington, 1986;

Gottfredson & Hirschi, 1990; Moffitt, 1993). The current study hypothesizes no additional need

for scoring alteration for females or males due to literature asserting the causes of delinquency to

be essentially the same for males and females (Gottfredson & Hirschi, 1990; Hubbard & Pratt,

2002; Lipsey, 1992; Moffitt, 1993). Hubbard and Pratt (2002) found the predictors of female

delinquency to be the same as those of male delinquency, namely a history of antisocial

behavior, antisocial attitudes, antisocial personality, and antisocial peers. These scholars did,

however, find school and family relationships and a history of physical and/or sexual abuse to be

significant predictors of female delinquency. The PACT assessment does take these factors into

account with respect to scoring, as mentioned above, making it possible to examine the

relationships observed by Hubbard and Pratt (2002).

The causes of delinquency of female offenders has historically been lacking in the extant

criminological literature (Hubbard & Pratt, 2002; Pratt et al., 1998). Criminological theory, the

"causes" of criminality, and the research on crime and deviance have been based on male

offending and then generalized to female offenders (Belknap, 1996; Chesney-Lind & Sheldon,

1992; Hubbard & Pratt, 2002). Hubbard and Pratt acknowledge the importance of discerning

whether the predictors of female delinquency are identical as those for male delinquency and

state if "the strongest predictors of female delinquency are different for females, correctional

programs may need to make adjustments to target certain gender-specific risk factors" rather

than simply apply the treatment programs found effective for male delinquency to female

offenders (2002: 2). Additional scholars have asserted the existence of genderedd pathways",

finding the importance of sexual and physical abuse in the development of female offending

(Chesney-Lind & Rodriguez, 1983; Hubbard & Pratt, 2002; Reisig et al., 2006). History of









abuse, both physical and sexual, is a predominant risk factor for both females and males, yet the

fact that females are more likely to be victims of the abuse could arguably disproportionately

increase the probability of females engaging in criminal behavior as a result of such abuse

(Hubbard & Pratt, 2002).

Additionally, females charged with serious or violent offenses are, more often than males,

accessories that did not instigate the offending, and therefore are misclassified as higher risk by

assessment instruments and become "false positives" that in fact do not re-offend (Reisig et al.,

2006). The inclusion of an assessment within the PACT of history of physical abuse, history of

sexual abuse, and school relationships will enable the current study to assess whether these

factors are more important for the prediction of female recidivism than for males. Previous

research indicates that actuarial assessments are not predictive of female offending as a whole,

but are valid for predicting the recidivism for economically motivated women (Reisig et al.,

2006). The current study will examine whether the PACT assessment is a valid predictor of

female re-offending.

Gender differences that may appear in analysis of the PACT assessment are hypothesized

to be the result of prevalence of some social history measurements utilized in the instrument

scoring. The current study hypothesizes that all significant gender differences will wash out to

insignificance once the proxy measure is taken into account (for example history of sexual abuse

being more prevalent for females; the issue is not the youth being female, but the youth having a

history of sexual abuse).

The discussion above leads to the following hypotheses:

* Overall risk to re-offend, as presented by the PACT, will significantly predict recidivism,
with higher risk youth more likely to re-offend.









* Youth assessed as higher risk to re-offend will do so regardless of whether the youth was
assessed and placed on probation supervision, or assessed during a residential commitment
placement and tracked post-release.

* The higher the criminal history score as presented by the PACT, the more likely the youth
will re-offend.

* The higher the social history score, as presented by the PACT, the more likely the youth
will re-offend.

* Overall risk to re-offend, as presented by the PACT, will be a significant predictor of male
recidivism and female recidivism.

* Males will be more likely to re-offend than females, controlling for overall risk level.

* Gender differences that may appear will be "washed out" to insignificance with the
inclusion of proxy measures (such as school relationships and history of physical or sexual
abuse).

* History of physical or sexual abuse will not be a stronger predictor of female re-offending
than of male re-offending.









CHAPTER 3
METHODS

The Florida Department of Juvenile Justice maintains a comprehensive database

called the Juvenile Justice Information System (JJIS). The database stores information on

all youth entering the system and all services provided to those youth. From the JJIS

database, the Office of Research and Planning performs monthly cumulative updates of

various data elements on youth served and placement history. Part of these monthly

extracts includes all PACT assessments administered. The monthly data extract for

February 2007 will be utilized for the current study. The extract contains all PACT

assessments from the initial November 15, 2005 implementation date up to the date of the

extract (February 4, 2007). The PACT extract contains the date each assessment was

administered, as well as the overall risk score, criminal history risk score, social history

risk score, answers to each PACT question, and gender. Furthermore the extract contains

information as to whether the assessment was a PACT Pre-Screen or a Full Assessment.

The PACT data will serve as the initial prediction of whether the youth will recidivate

and the length of time until re-offending occurs.

The PACT data will be merged with another monthly data extract that captures the

entire offense history for every youth who enters the juvenile justice system. This extract

contains demographic information including date of birth, race, ethnicity, and gender.

The extract also includes all prior referrals for each youth who has ever entered the

juvenile justice system, reliable from 1994 to the present. Currently, there are over two

million records of referrals for juvenile offenders contained in the extracts. The extract

for April 2007 will be utilized for this study, containing all referral records up to April 6,

2007.









The data is stored by referral meaning that each youth may have multiple referrals

within the database. Every referral has a unique referral identification number and every

youth has a unique identifier, called the djjid number. Every line of data is a unique

referral with information pertaining to the date of the referral, the date of the arrest, and

the date of the offense. Furthermore, information related to the most serious offense

within the referral is collected, as an arrest can be made for one event with multiple

charges. The data also include information about whether the youth was adjudicated and

the resulting disposition.

The referral extract will be used to examine recidivism. Therefore, the dependent

variable will be recidivism based on an official delinquency referral to the Department of

Juvenile Justice post-administration of a PACT assessment. The PACT data are

organized by djjid number (uniquely identifying each youth) and by PACT identification

number (uniquely identifying each PACT assessment). The two extracts will be merged

so that PACT information and referral information for each youth can be examined. All

youth included in the PACT extract must have a referral history since they are given an

initial PACT assessment at intake, after being referred to the Department of Juvenile

Justice. The referral history prior to the PACT assessment will be captured in Domain 1

of the PACT, which scores official criminal history. The referral extract is necessary,

however, to examine all referrals occurring after the date of PACT administration. This

creates a situation in which a youth is assessed using the PACT and given a score

representing that youth's risk to re-offend. In order to validate the PACT, the current

study will examine whether subsequent official delinquency referrals after PACT

administration can be predicted in relation to the youth's overall risk to re-offend.









PACT data

The February 2007 monthly PACT extract utilized for the current study contains

111,450 PACT assessments. The data includes 88,670 Pre-Screen assessments and

22,780 Full Assessments. Of the Pre-Screen assessments completed, 67,300 (75.9%)

were administered to males and 21,370 (24.1%) to females. Data for the Full

Assessments include 18,270 (80.2%) males, and 4510 (19.8%) females. Arguably, the

most important variable contained in both the Pre-Screen and Full Assessment is the

overall risk to re-offend. The 111,450 assessments reveal 50.9% of the youth to be low

risk to re-offend, 15.8% moderate, 17.9% moderate-high, and 15.5% high risk to re-

offend. The overall risk score, as mentioned above, is derived from a matrix of a criminal

history score, coded 0-31 with the higher the score the greater the previous official

referral seriousness, and a social history score, coded 0-18 with the higher the score the

more risk factors present in the youth's social environment. The mean criminal history

score for the 111,450 assessments is 8.30 (standard deviation of 5.482) and the mean

social history score 4.96 (standard deviation of 2.885). Overall risk to re-offend has been

coded from 1 to 4 representing low risk to re-offend (1), and then moderate (2),

moderate-high (3), and high risk to re-offend (4). The mean overall risk to re-offend for

the 111,450 assessments is 1.98, with a standard deviation of 1.143.

The criminal history score is derived entirely from Domain 1, which consists of

twelve questions. Two of the questions, sexual misconduct misdemeanor referrals, and

felony sex offense referrals, are not used in the criminal history scoring to count towards

overall risk to re-offend. The remaining ten questions from Domain 1 are used to

generate the criminal history score (see Table 3-1 for descriptive statistics). The answer

to all Domain 1 questions, as described above, are automated from JJIS, the Department's









"real-time" database which captures, among many other elements, the entire official

referral history for every youth entering the juvenile justice system.

The first Domain 1 question utilized in the criminal history scoring assesses the

youth on the age at the time of the offense for which the youth was first referred (again,

see Appendix B and Appendix C for exact wording of PACT questions). The item is

coded 1 to 5 with 1 being over age 16 at first referral, 2 being 16 years of age, 3 for 15

years of age, 4 for 13 to 14 years of age, and 5 for 12 years of age or under. The next two

items assess the number of misdemeanors and felonies for which the youth was referred,

respectively. The misdemeanor item is coded from 1 to 4, with 1 being none or one

referral, 2 being two referrals, 3 being three or four referrals, and 4 representing five or

more misdemeanor referrals. The felony item is also coded 1 to 4, but is more stringent

on the criteria. For this item, 1 represents no felony referrals, 2 being one referral, 3 being

two referrals, and 4 representing three or more felony referrals.

Table 3-1. Criminal History Item Descriptive Statistics
Criminal History Item Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Age at First Offense 111,450 1 5 2.2898 1.219
Adjudicated Misdemeanors 111,450 1 4 1.6912 0.93
Adjudicated Felonies 111,450 1 4 2.0778 1.01
Total Weapon Offenses 111,450 1 2 1.249 0.331
Total Against-person Misd. 111,450 1 3 1.4349 0.658
Total Against-person felonies 111,450 0 2 1.2299 0.463
Secure Detention Placements 111,450 1 4 1.8986 1.163
Commitment Placements 111,450 1 3 1.156 0.422
Total Escape Adjudications 111,450 1 3 1.0073 0.089
Total Failure to Appear PUOs 111,450 1 3 1.3957 0.698









The next three items assess the extent of official against-person and weapon referrals in the

youth's history. Weapon referrals is coded 1 to 2, with 1 being no weapon referrals, and 2

representing one or more referrals for which the most serious offense in the referral was a

firearm/weapon charge. The next two items measure against-person referrals, which are broken

down by misdemeanors and felonies. Misdemeanor against-person referrals are coded 1 to 3,

with 1 being no against-person misdemeanor referrals, 2 being one referral, and 3 representing

two or more referrals for which the most serious offense was a misdemeanor involving threats,

force, or physical harm to another person. The against-person felony item is coded 0 to 2, with 0

representing no against-person felony referrals, 1 being one or two such referrals, and 2

representing three or more against-person felony referrals.

The remaining four questions utilized in the scoring of official criminal history relate to

confinements, residential commitments, escapes, and pick up orders for failure- to-appear in

court. The confinements item measures the number of times the youth was held for at least forty-

eight hours in a detention facility. This item is coded 1 to 4, representing no detention

confinements (1), one confinement (2), two confinements (3), and three or more confinements in

a detention facility (4). The total number of residential commitment confinements is coded 1 to

3, representing no placements in a residential facility (1), one placement (2), and a history of two

or more residential placements (3). History of escapes is coded 1 to 3, representing no history of

escape (1), one attempt or actual escape (2), and two or more attempted or actual escapes (3).

The final criminal history item used in scoring gauges the number of failures-to-appear in court

that have resulted in a pick up order being issued in the youth's history. The item is coded 1 to 3,

representing no pick up orders (1), one (2), and two or more pick up orders for failure-to-appear

(3).









Table 3-2. Social History Item Descriptive Statistics
Social History Item Descriptive Statistics
N Minimum
Sex 111,450 0
Current School Enrollment 111,450 1
Recent School Conduct 111,450 1
Recent School Attendance 111,450 1
Recent Academic Performance 111,450 1
Anti-social Peers 111,450 0
Out-of-home Placements 111,450 1
History of Runing Away 111,450 1
Parental Authority/Control 111,450 1
Household Jail History 111,450 0
History of Alcohol Use 111,450 0
History of Drug Use 111,450 0
Current Alcohol Use 111,450 0
Current Drug Use 111,450 0
History of Physical Abuse 111,450 0
History of Sexual Abuse 111,450 0
History of Neglect 111,450 1
Mental Health Problems 111,450 1


Maximum
1
6
5
5
5
2
4
5
3
1
1
1
1
1
2
1
2
5


Mean
0.7678
2.5592
2.9025
2.6197
3.3858
0.6688
1.2003
1.5722
1.7241
0.9951
0.3375
0.4758
0.2203
0.3544
0.237
0.0497
1.0723
1.2738


Std. Deviation
0.422
1.23
1.183
1.533
0.946
0.552
0.593
1.11
0.68
0.07
0.473
0.499
0.414
0.478
0.608
0.217
0.259
0.847


The social history of the youth is assessed, in part, using the eighteen items illustrated in

Table 3-2. Descriptive statistics for these social history items are reported in Table 3-2. The first

item assessing social history is sex. Females are coded as 0 and males as 1. Current school

enrollment status is coded 1 to 6, representing the youth has graduated high school or has

received a GED (1), being enrolled full-time (2), enrolled part-time (3), suspended (4), dropped

out (5), and having been expelled (6). Recent school conduct is measured next, coded 1 to 5.

Conduct is broken down by recognition for good behavior, no problems with school conduct,

problems reported by teacher, problem calls to parents, and lastly, calls to police related to the

youth's conduct. The youth's recent academic attendance is coded 1 to 5, representing good

attendance with few absences (1), no unexcused absences (2), some partial-day unexcused (3),

some full-day unexcused (4), and the youth being a habitual truant (5). Academic performance









rounds out the school-related measures, and is coded 1 to 5. Honor students are coded as 1,

G.P.A above 3.0 coded 2, G.P.A. from 2.0 to 3.0 coded as 3, G.P.A. from 1.0 to 2.0 coded as 4,

and G.P.A. below 1.0 represented as 5.

The youth's current friends/companions who the youth actually spends time with is broken

down by no consistent friends or companions, pro-social friends, anti-social friends, and gang

members/associates. Within the PACT software, this item can receive multiple "checks"

meaning that the individual administering the instrument can record that the youth spends time

with any combination of the options listed above. The item has been recorded for this study to

separate youth who spend no time with anti-social friends, or have all pro-social friends, from

those who spend time with anti-social friends, and from youth who are gang members or have

friends who are gang members/associates. The new item has been coded 0 to 2, representing no

anti-social friends (0), anti-social friends (1), and the youth who are gang members or actually

spend time with gang members (3).

History of court-ordered or Department of Children and Families voluntary out-of-home

and shelter care placements exceeding thirty days is measured as part of the social history

scoring. This item is coded 1 to 4, representing no out-of-home placements exceeding thirty days

(1), one placement (2), two placements (3), and three or more out-of-home placements (4). The

youth's history of running away is coded 1 to 5, representing no history or running away or being

kicked out (1), one instance of running away (2), two to three instances (3), four to five instances

(4), and over five instances of running away (5).

The jail and imprisonment history of individuals currently living in the household with the

youth is taken into account in the social history scoring as well. This item differentiates youth

living with no one with a history from the mother having a history of jail/prison, the father









having a history, older siblings having a history, younger siblings having a history, and from

other members of the household having a history of jail or imprisonment. The individual

administering the PACT is able to "check" each of these scenarios separately, much like the

"current friends" item described above. Therefore, this item was recorded to represent the history

of jail/imprisonment of individuals currently living in the household with the youth, now coded 0

to 1. The new coding coincides with the scoring utilized by the PACT separating youth who

currently live in a household without an individual who has a history of jail or imprisonment

(coded as 0) from those living with at least one such individual (coded as 1).

The level of parental authority and control is accounted for in the social history score and

coded 1 to 3. Youth who usually obey parents and follow the rules they set are coded as 1, youth

sometimes obeying the rules coded as 2, and youth consistently disobeying rules or acting hostile

toward the parents are coded as 3. Four items measuring alcohol and drug use are measured as

well. History of alcohol use, history of drug use, current alcohol use, and current drug use are

each utilized in the social history scoring. History of use includes any use by the youth in his/her

lifetime. Current use refers to engaging in the behavior in the last six months. The two history of

use items are broken down by no past use, past use, alcohol (or drugs) disrupted education,

caused family conflict, interfered with keeping pro-social friends, caused health problems,

contributed to criminal behavior, youth needed increasing amounts to achieve the same level of

intoxication or high, and youth experienced withdrawal problems. Current use is broken down by

not currently using, currently using, alcohol (or drugs) disrupts education, causes family conflict,

interferes with keeping pro-social friends, causes health problems, contributes to criminal

behavior, youth needs increasing amounts, and youth experiences withdrawal problems. Identical

to two scenarios described above, the individual administering the PACT can "check" all









answers that item, necessitating the need to recode the items. The new items are coded 0 to 1

separating individuals with no history of use, or no current use in the case of the second two

variables, from youth with a history of use or current use. Recoding the variables in this manner

is consistent with the scoring utilized in the social history score.

Abuse and mental health issues are the last four items used in the social history scoring.

For the abuse and neglect items, suspected incidents of abuse are included, as are those disclosed

by the youth (whether or not reported or substantiated officially) but reports of abuse and neglect

investigated officially and proven to be false are excluded. Youth are assessed for a history of

being the victim of physical abuse, which is broken down by not having been a victim, having

been a victim at home, having been a victim in a foster/group home, having been victimized by a

family member, having been victimized by someone outside of the family, and having been

attacked by a weapon. Again, the individual administering the PACT can "check" all scenarios

that apply. The item has therefore been recorded 0 to 2 to separate youth not having been a victim

(coded as 0) from youth ever having been victimized by a family member or someone outside the

family (coded as 1) and from youth ever having been the victim of physical abuse at home, in a

foster/group home, or attacked with a weapon (coded as 2). Recoding the item in this manner is

consistent with the scoring used and the notion that being attacked with a weapon or having been

victimized where the youth was residing achieves some higher level of seriousness than only

having been a victim.

History of sexual abuse/rape is assessed and taken into account in the scoring of social

history. The item differentiates youth who have not been a victim of sexual abuse or rape from

youth having been abused/raped by a family member, and from youth having been abused/raped

by someone outside of the family. Youth may have been victimized both by a family member,









and by someone outside of the family. The individual administering the PACT may "check" all

that apply. The item has been recorded 0 to 1 to separate youth not having been a victim of sexual

abuse/rape from those that have, regardless of the perpetrator. Recoding the item in this fashion

is consistent with the method used in the social history scoring. History of being the victim of

neglect is assessed, coded 1 to 2. Individuals not having been the victim of neglect (coded as 1)

are separated from individuals that have been the victim of neglect (coded as 2).

Finally, the youth's history of mental health problems is assessed. Mental health problems

such as schizophrenia, bi-polar, mood, thought, personality, and adjustment disorders are

included. Conduct disorder, oppositional defiant disorder, substance abuse, and ADD/ADHD are

excluded. The problem must have been confirmed by a professional in the social

service/healthcare field to be included, as individuals administering the assessment are

predominately non-licensed, non-clinical staff (usually they are juvenile probation officers). The

item is coded 1 to 5, with 1 being no history of mental health problems, 2 being diagnosed with

mental health problems, 3 being diagnosed and only mental health medication is prescribed, 4

being diagnosed and only mental health treatment prescribed, and 5 representing the youth being

diagnosed and having mental health medication and treatment prescribed.

The above section has articulated the items from the PACT that are used in deriving the

criminal history score as well as various social history items. The remaining ninety-eight items in

the Full Assessment are used to gain a comprehensive picture of the youth and his/her past and

present situation with respect to the twelve PACT domains. These remaining items, however, are

not utilized in the scoring of the overall risk to re-offend revealed by the PACT assessment. The

items used in the overall risk to re-offend scoring of the PACT assessment will be utilized in the

current validation/gender comparison study. These variables will be examined with respect to









their ability to predict official recidivism for the juvenile offending population, using Florida

Department of Juvenile Justice JJIS data as explained above.

Sample

The PACT data outlined above represent all PACT assessments completed from August of

2005 to February 2007. The PACT is administered to each youth who enters the juvenile justice

system in the state of Florida. Furthermore, each youth remaining in the custody of the

Department is re-assessed every ninety days, as well as sixty days prior to release, if placed in

residential commitment. Due to the re-assessment process, and the fact that many youth enter the

juvenile justice system on multiple occasions, the current study narrows the sample down so as

to utilize the first ever Full Assessment for each youth in the sample. In the event a Full

Assessment was never administered to the youth, the first ever Pre-Screen is utilized. The first

assessment administered to the youth was chosen so as to maximize the follow-up period

possible to track recidivism. As the PACT assessment is a new process, implemented in August

of 2005, allowing for follow-up times is a critical factor. The choice to use the first Full

Assessment was made in order to capture the most information available on each youth, due to

the Full Assessment containing 126 variables versus the forty-six included in the Pre-Screen.

The sampling procedure outlined above resulted in a final N of 48,871. There is no need to

divide the sample into subgroups for the purpose of construction and validation samples as

would be necessary in the event an instrument were being designed. The current study is

attempting to validate and examine an existing instrument and, as a result, will not benefit from

the construction and replication process nor will it suffer any shrinkage issues that would

normally accompany instrument development. Future research wishing to examine potential

improvements to the existing instrument and/or scoring protocols would be well advised to









utilize a sub-sample for construction plus multiple replication samples to ascertain shrinkage and

predictive validity.

However, the current study benefits from dividing the sample into those youth that were

assessed and subsequently remained in the community versus those youth assessed and placed in

residential commitment. The study may have kept these sub-samples together and simply

controlled for "street time", beginning the follow-up period immediately for youth remaining in

the community and waiting until release from residential commitment to begin the follow-up

period for the committed youth. The decision was instead made to separate youth into the two

categories mentioned above and analyzed independent of one another to assess the effectiveness

of the PACT for youth remaining in the community versus those youth being released from

residential placement.

Cohort 1 Descriptive Statistics

Prior to dividing the overall sample into community youth and residential youth, the

decision was made to split the sample into two cohorts. Cohort 1 will consist of youth who are

tracked for only six months, while Cohort 2 consists of youth who are followed for twelve

months. This will allow for the utilization of as many cases as possible of youth who have

received the PACT assessment thus far, as well as allow for the examination of the effectiveness

of the assessment to predict short-term and long-term success. Youth who can be tracked twelve

months, and thus placed in Cohort 2, are not also placed in Cohort 1. The youth that remain in

the community are tracked beginning right after the assessment and followed six months for

Cohort 1 and twelve months for Cohort 2. The youth who are in residential commitment are

tracked following release from placement and then followed six months post-release for Cohort 1

youth and twelve months post-release for Cohort 2 youth.









Table 3-3. Community Cohort 1 Criminal History Descriptive Statistics
Criminal History Item Descriptive Statistics


N Minimum
Age at First Offense 36,325 1
Adjudicated Misdemeanors 36,325 1
Adjudicated Felonies 36,325 1
Total Weapon Offenses 36,325 1
Total Against-person Misd. 36,325 1
Total Against-person felonies 36,038 1
Secure Detention Placements 36,325 1
Commitment Placements 36,325 1
Total Escape Adjudications 36,325 1
Total Failure to Appear PUOs 36,325 1


Maximum
5
4
4
2
3
2
4
3
3
3


Mean
2.5581
1.4332
1.7703
1.0951
1.3428
1.1984
1.5
1.0829
1.0032
1.2158


Std. Deviation
1.298
0.762
0.853
0.293
0.584
0.3988
0.931
0.315
0.058
0.533


Table 3-4. Community Cohort 1 Social History Descriptive Statistics
Social History Item Descriptive Statistics
N Minimum Maximum
Sex 36,325 0 1
Current School Enrollment 34,489 1 6
Recent School Conduct 29,982 1 5
Recent School Attendance 29,982 1 5
Recent Academic Performance 29,982 1 5
Anti-social Peers 36,325 0 2
Out-of-home Placements 36,325 1 4
History of Running Away 36,325 1 5
Parental Authority/Control 36,105 1 3
Household Jail History 36,105 0 1
History of Alcohol Use 36,325 0 1
History of Drug Use 36,325 0 1
Current Alcohol Use 32,043 0 1
Current Drug Use 32,043 0 1
History of Physical Abuse 36,325 0 2
History of Sexual Abuse 36,325 0 1
History of Neglect 36,325 1 2
Mental Health Problems 35,466 1 5


Mean
0.7335
2.4114
2.6863
2.3237
3.2571
0.6125
1.1445
1.3658
1.5628
0.2173
0.2783
0.3848
0.1088
0.2204
0.1814
0.0394
1.0506
1.2107


Std. Deviation
0.442
1.103
1.122
1.439
0.923
0.538
0.504
0.892
0.632
0.412
0.448
0.487
0.311
0.415
0.538
0.195
0.219
0.758


The resulting sub-samples of Cohort 1 community and residential youth yield a final N of

36,325 community and 3,399 residential Cohort 1 youth. The first sub-sample to be examined is

composed of youth remaining in the community with the follow-up period beginning









immediately after the PACT assessment. The community Cohort 1 sub-sample is 56.7% white,

42.7% black, and 0.6% "other". This sub-sample is 73.3% male and 26.7% female. Descriptive

statistics for the criminal history and the social history of the community Cohort 1 sub-sample

can be seen in Table 3-3 and Table 3-4, respectively.

The second Cohort 1 sub-sample is composed of youth exiting residential commitment

placement prior to beginning the six-month follow-up period. The residential sub-sample is

44.6% white, 55% black, and 0.4% "other". The residential sub-sample is 84.5% male and

15.5% female. The descriptive statistics outlining the criminal history and social history of the

residential sub-sample are reported in Table 3-5 and Table 3-6, respectively.

Table 3-5. Residential Cohort 1 Criminal History Descriptive Statistics
Criminal History Item Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Age at First Offense 3,399 1 5 1.8664 0.972
Adjudicated Misdemeanors 3,399 1 4 2.0891 1.047
Adjudicated Felonies 3,399 1 4 2.7193 1.041
Total Weapon Offenses 3,399 1 2 1.1692 0.375
Total Against-person Misd. 3,399 1 3 1.5711 0.741
Total Against-person felonies 3,232 1 2 1.3685 0.482
Secure Detention Placements 3,399 1 4 2.8653 1.154
Commitment Placements 3,399 1 3 1.4693 0.629
Total Escape Adjudications 3,399 1 3 1.0274 0.172
Total Failure to Appear PUOs 3,399 1 3 1.6973 0.837


Cohort 1 Analysis

The first step is to assess the effectiveness of the PACT to predict re-offending for the

Cohort 1 community sub-sample and the residential sub-sample, separately, before moving to

Cohort 2 comparisons. The purpose of this step is to ascertain whether the instrument is effective

at predicting recidivism for youth who remain in the community subsequent to assessment as

well as youth exiting a residential commitment placement prior to beginning the six-month

follow-up. Consistent with prior risk assessment research, re-offending is operationalized as any









subsequent delinquency referral after the assessment date (Funk, 1999:55; see also Foley, 1991;

Wiebush et al., 1995). The current study uses the date of the subsequent offense post-assessment

as it is capturing the behavior that desired, and not any indication of police or court action, so

will use the actual date the offense occurred. Recidivism, therefore, for the purposes of this

study, will be any subsequent delinquency referral to the Florida Department of Juvenile Justice.

The "success" or "failure" of a youth will determined based on whether the youth commits a new

offense for which he/she is officially referred by law enforcement to the Department and having

an offense date within six months for Cohort 1 youth and twelve months for Cohort 2 youth from

the beginning of the follow-up period.

Table 3-6. Residential Cohort 1 Social History Descriptive Statistics
Social History Item Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Sex 3,399 0 1 0.8452 0.362
Current School Enrollment 2,889 1 6 2.5147 1.158
Recent School Conduct 2,602 1 5 3.0223 1.171
Recent School Attendance 2,602 1 5 2.9181 1.596
Recent Academic Performance 2,602 1 5 3.5984 0.963
Anti-social Peers 3,399 0 2 0.7941 0.539
Out-of-home Placements 3,399 1 4 1.3113 0.718
History of Running Away 3,399 1 5 1.9412 1.343
Parental Authority/Control 3,337 1 3 2.0039 0.701
Household Jail History 3,399 0 1 0.341 0.474
History of Alcohol Use 3,399 0 1 0.474 0.499
History of Drug Use 3,399 0 1 0.6764 0.468
Current Alcohol Use 3,399 0 1 0.2304 0.421
Current Drug Use 3,399 0 1 0.5179 0.5
History of Physical Abuse 3,399 0 2 0.3033 0.676
History of Sexual Abuse 3,399 0 1 0.0691 0.254
History of Neglect 3,399 1 2 1.1068 0.309
Mental Health Problems 3,309 1 5 1.417 0.98


The dependent variable for analysis of the effectiveness of the PACT risk to re-offend

prediction for Cohort 1 community and residential youth is dichotomous (either the youth









recidivated within six months or not) and therefore the use of logistic regression is appropriate.

The analysis, beginning with Cohort 1 community youth, uses the overall level of risk to re-

offend, as measured utilizing the PACT assessment, to predict recidivism, based on a new

referral to the Department of Juvenile Justice subsequent to the assessment date. Consistent with

prior research, the analysis begins by assessing recidivism rates across risk-need categories of

overall risk to re-offend (Andrews & Bonta, 2003; Andrews et al., 2001; Bonta, LaPairie, &

Wallace-Capretta, 1997; Reisig et al., 2006). Descriptive statistics for overall risk to re-offend

can be seen in Table 3-7. Low risk to re-offend youth made up 70.7% of the sample, moderate

risk 13.2%, moderate-high risk 10%, and high risk to re-offend 6.1%. The percentage of low risk

youth who received a new referral to the Department within six months was 19.6%. The

percentage for moderate risk youth was 34.7%, 38.6% for moderate-high, and 43.4% of all high

risk youth received a new referral to the Department within six months.

Table 3-7. Community Cohort 1 Covariate Descriptive Statistics
Covariate Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 36,325 1 4 1.5158 0.904
Sex 36,325 0 1 0.7335 0.442
Race 36,325 0 1 0.4332 0.496
Referral within six months 36,325 0 1 0.2499 0.433


The analysis using logistic regression to predict six-month recidivism of youth who remain

in the community after receiving a PACT assessment reveals that overall risk to re-offend is a

statistically significant predictor, with a p-value of .000. Youth assessed as having a higher

overall risk to re-offend score are more likely to recidivate then youth assessed as lower risk. As

Table 3-8 illustrates, each increase in the independent variable, overall risk to re-offend,

produces a 1.543 increase in the dependent variable, recidivism (an over 50% increase).









Next, race was also included as a covariate, simply separating white (coded 0) and non-

white (coded as 1). The final covariate included was sex, separating males (coded 1) and females

(coded 0). Logistic regression was once again utilized to predict a delinquency referral tracked

six months post PACT assessment for youth who remain in the community, this time using

overall risk to re-offend, race, and sex as predictors. The analysis reveals all three covariates to

be significant predictors of a delinquency referral subsequent to PACT assessment, all with p-

values of .000. As above, youth assessed as higher risk are more likely to recidivate during the

six-month follow-up period, as are males and non-white youth. Results of the logistic regressions

can be seen in Table 3-8. Examination of the Exp(B) reveal overall risk to re-offend to have the

strongest relationship to recidivism, followed by sex and finally race.

Table 3-8. Community Cohort 1 Logistic Regression Results
Covariate Statistics
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.434 0.012 1226.507 1 .000* 1.543

Logistic Regression 2
Overall Risk to Re-offend 0.395 0.013 981.987 1 .000* 1.484
Sex 0.434 1 158.437 1 .000* 1.456
Race 0.434 0 223.35 1 .000* 1.453
* =significant at p=.05


Table 3-9. Residential Cohort 1 Covariate Descriptive Statistics
Covariate Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 3,399 1 4 2.9491 1.047
Sex 3,399 0 1 0.8452 0.362
Race 3,399 0 1 0.554 0.497
Referral within six months 3,399 0 1 0.3945 0.489


The Cohort 1 youth who were in residential commitment facilities at the time of the

assessment, or who entered residential subsequent to the PACT assessment were analyzed next.









These youth were analyzed identically to the Cohort 1 community youth described above

(descriptive statistics can be seen in Table 3-9). For this sub-sample, low risk youth composed

14.1% of the youth, moderate risk 15.1%, moderate high 32.5%, and high risk to re-offend youth

38.3%. These simple percentages illustrate the difference between the Cohort 1 community

youth analyzed initially from these Cohort 1 residential youth. While only 6.1% of the

community youth were assessed as high risk to re-offend, 38.3% of the residential youth are high

risk. Likewise, whereas only 14.1% of these residential youth are low risk, 70.7% of the

community youth were low risk to re-offend. These differences will help to see whether the

PACT assessment can predict equally well the recidivism of lower risk community youth as it

does higher risk residential youth.

Table 3-10. Residential Cohort 1 Logistic Regression Results
Covariate Statistics
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend -0.041 0.033 1.516 1 .218 0.96

Logistic Regression 2
Overall Risk to Re-offend -0.045 0.034 1.782 1 .182 0.956
Sex -0.015 0.097 0.024 1 .877 0.985
Race 0.072 0.071 1.03 1 .310 1.075


First, logistic regression was used to predict a delinquency referral subsequent to PACT

assessment and release from a residential commitment facility. Overall risk to re-offend is not a

significant predictor of a referral to the department when youth exiting residential facilities are

followed six months (see Table 3-10). Identically to the analysis to Cohort 1 community youth,

logistic regression was again used to predict a subsequent referral using overall risk to re-offend,

sex, and race. Similar to the community sub-sample, the residential sample is predominately

male (see Table 3-9 for descriptive statistics). However, the residential sample is majority non-

white, where the community sample was over 50% white. The results of the logistic regression









(see Table 3-10) reveal overall risk to re-offend, race, and sex are not significant predictors of a

referral to the Department for youth who exist residential facilities subsequent to a PACT

analysis.

There are several reasons why it may be expected that youth being released from a

residential facility may be different from youth who remain in the community, which could lead

to hypothesizing as to why the PACT overall risk to re-offend score is not a significant predictor

of recidivism for these youth. Prior to any theorizing as to why this is the case, the residential

sample must be further analyzed. The sample was composed of two types of scenarios. The first

was youth who were administered the PACT and subsequently placed into a residential facility

after a court hearing. The six month follow-up began after the youth was released from the

facility, but using the pre-commitment PACT assessment. The second group of youth was those

that were assessed while already in a commitment facility, and it was that assessment that was

used to predict the six months follow-up recidivism.

The second scenario, youth assessed after already being in commitment, may be an

inadequate use of the PACT assessment. The Florida Department of Juvenile Justice realizes this

inadequacy and is currently in the process of developing a PACT assessment more appropriate

for residential youth. The problem lies with the social history questions from which the overall

risk to re-offend score is derived. Remember that the overall risk score is composed of a matrix

of criminal history items and social history items. The criminal history items are static predictors

(age at first offense, number and seriousness of prior offenses, for example), so would not make

an issue for assessing youth within commitment facilities.

However, the social history items are dynamic, changeable factors and certainly could be

altered by residential placement. This alteration is not necessitated by some type of effective, or









for that matter ineffective, treatment delivered to residential youth, but more of a result of the

questions themselves. Questions related to delinquent peers, substance abuse, and school

attendance differ by necessity simply by residing in a residential facility. Youth are, arguably,

forced to be around antisocial peers, as all youth received delinquency referrals that resulted in

the commitment. Therefore, all youth would be scored higher the day they were placed in the

facility through no fault of their own with regard to choosing who they associate with (with the

realization that with committing the offense comes the possibility of residential placement if

apprehended). Furthermore, all youth are essentially forced to attend school on a daily basis,

making the score on this factor decrease for many youth who did not have stellar attendance in

the community. Also, using substances and alcohol decreases by nature of the youth being placed

in commitment. A picture is quickly painted where the results of a PACT assessment given to a

youth in residential commitment is suspect at best, as some variables used in the scoring are

artificially inflated and some deflated by the mere fact of the placement.

Table 3-11. Split Residential Cohort 1 Covariate Descriptive Statistics
Covariate Descriptive Statistics
PACT before commitment N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 2,850 1 4 2.9102 1.088
Sex 2,850 0 1 0.8435 0.363
Race 2,850 0 1 0.5411 0.498
Referral within six months 2,850 0 1 0.4119 0.492
PACT while in commitment N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 549 1 4 3.1512 0.772
Sex 549 0 1 0.8543 0.353
Race 549 0 1 0.6211 0.486
Referral within six months 549 0 1 0.3042 0.46


The scenario depicted above was tested, breaking the Cohort 1 residential sample into two

groups. The first group was composed of youth who were given the PACT assessment prior to

residential placement. This will help to test the longevity of the predictability of the PACT as









some youth are placed in residential facilities for several months and the follow-up period does

not begin until release meaning scores would have to be valid and predictive of re-offending for

months. The second group, youth assessed while in residential commitment, will show whether

the PACT assessment is appropriate for assessing the risk to re-offend after subsequent release

for these youth. The descriptive statistics for the two groups of residential youth are shown in

Table 3-11.

Table 3-12. Split Residential Cohort 1 Logistic Regression Results
Covariate Statistics
PACT before commitment
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend -0.041 0.035 1.377 1 .241 0.96

Logistic Regression 2
Overall Risk to Re-offend -0.043 0.035 1.475 1 .225 0.958
Sex -0.102 0.105 0.954 1 .329 0.903
Race 0.067 0.077 0.763 1 .383 1.07
PACT while in commitment
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.129 0.123 1.106 1 .293 1.138

Logistic Regression 2
Overall Risk to Re-offend 0.105 0.126 0.686 1 .407 1.11
Sex 0.623 0.297 4.393 1 .036* 1.864
Race 0.228 0.197 1.334 1 .248 1.256
* =significant at p=.05


Identically to previous analyses, logistic regression was used to predict a new delinquency

referral post-release from a residential facility. First, only the overall risk to re-offend level was

entered as a covariate, separately for the youth assessed prior to commitment and for youth

assessed while committed. For the first group, youth assessed prior to being placed in residential

commitment, overall risk was not significant (see Table 3-12). This is contrary to the notion

presented above that the PACT may still be predictive for youth assessed with appropriate









questions. It may, however, depict that inability to use risk assessment scores from months

previous to the beginning of the follow-up period to ascertain adequate predictions of risk to

recidivate. For youth assessed while in commitment placement, overall risk to re-offend was also

not a significant predictor of post-release recidivism (see Table 3-12). This finding was

suspected due to the inappropriateness of some PACT questions for a residential population.

As analyzed previously, logistic regression was next used to assess the predictive ability of

overall risk to re-offend, race, and sex of the two sub-samples, separately, of a subsequent

delinquency referral post-release. For the youth assessed prior to entering a facility, none of the

three covariates were significant predictors of recidivism (see Table 3-12). For the youth

assessed while in residential commitment, only sex was a significant predictor of recidivism.

Male youth are more likely to recidivate then female youth when followed post-release.

Examination of the Exp(B) for sex reveals that every increase in the B produces a 1.864 times

increase in recidivism (well over a 50% increase).

Cohort 2 Descriptive Statistics

Youth for which a one-year follow-up can be conducted are analyzed next. This will

enable the comparison of the ability of PACT to predict long-term recidivism versus short-term.

The following analyses will be conducted identically to those reported for Cohort 1 youth. After

initial descriptive statistics are reported, the residential youth will be split into two samples prior

to further analysis (separating those youth assessed prior to commitment and those youth

assessed while in a residential facility).

The Cohort 2 community sub-sample is 61.1% white, 38% black, and 0.9% "other". The

sub-sample is 69.8% male and 30.2% female. The PACT criminal history descriptive statistics

are reported in Table 3-13. Table 3-14 illustrates the descriptive statistics for the eighteen

variables composing the social history scoring for the Cohort 2 community youth sub-sample.









Table 3-13. Community Cohort 2 Criminal History Descriptive Statistics
Criminal History Item Descriptive Statistics


N Minimum
Age at First Offense 8,132 1
Adjudicated Misdemeanors 8,132 1
Adjudicated Felonies 8,132 1
Total Weapon Offenses 8,132 1
Total Against-person Misd. 8,132 1
Total Against-person felonies 8,077 1
Secure Detention Placements 8,132 1
Commitment Placements 8,132 1
Total Escape Adjudications 8,132 1
Total Failure to Appear PUOs 8,132 1


Maximum
5
4
4
2
3
2
4
3
3
3


Mean
2.4953
1.4405
1.7213
1.1024
1.3517
1.1804
1.5425
1.0637
1.0041
1.2298


Std. Deviation
1.281
0.766
0.838
0.303
0.585
0.385
0.98
0.28
0.073
0.555


Table 3-14. Community Cohort 2 Social History Descriptive Statistics
Social History Item Descriptive Statistics
N Minimum Maximum Mean
Sex 8,132 0 1 0.6981
Current School Enrollment 7,959 1 6 2.5155
Recent School Conduct 7,957 1 5 2.9773
Recent School Attendance 7,957 1 5 2.5831
Recent Academic Performance 7,955 1 5 3.2925
Anti-social Peers 8,132 0 2 0.6779
Out-of-home Placements 8,132 1 4 1.1422
History of Running Away 8,132 1 5 1.3876
Parental Authority/Control 8,102 1 3 1.5994
Household Jail History 8,102 0 1 0.2407
History of Alcohol Use 8,132 0 1 0.2907
History of Drug Use 8,132 0 1 0.3915
Current Alcohol Use 7,468 0 1 0.1197
Current Drug Use 7,468 0 1 0.2442
History of Physical Abuse 8,132 0 2 0.2167
History of Sexual Abuse 8,132 0 1 0.0454
History of Neglect 8,132 1 2 1.0594
Mental Health Problems 7,951 1 5 1.24


Std. Deviation
0.459
1.168
1.247
1.587
0.978
0.536
0.502
0.927
0.632
0.428
0.454
0.488
0.325
0.43
0.586
0.208
0.236
0.817


The second Cohort 2 sub-sample is composed of the youth who were assessed using the

PACT and had a residential commitment placement prior to the follow-up period. The Cohort 2

sub-sample is 52.2% white, 46.7% black, and 1.1% "Other". Furthermore, the sample is 84.6%









male and 14.4% female. As in the Cohort 1 residential sample, the Cohort 2 youth being placed

in residential commitment prior to the follow-up period is substantially more male (84.6% versus

the 69.8% for community youth) and less likely to be white (52.2% versus 61.1% for the

community youth). Descriptive statistics for the PACT assessed criminal history and social

history for the Cohort 2 youth who have a residential commitment placement prior to the follow-

up period can be seen in Table 3-15 and Table 3-16, respectively.

Table 3-15. Residential Cohort 2 Criminal History Descriptive Statistics
Criminal History Item Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Age at First Offense 1,015 1 5 1.802 0.896
Adjudicated Misdemeanors 1,015 1 4 2.1271 1.073
Adjudicated Felonies 1,015 1 4 2.6739 1.038
Total Weapon Offenses 1,015 1 2 1.1645 0.371
Total Against-person Misd. 1,015 1 3 1.5793 0.741
Total Against-person felonies 981 1 2 1.3191 0.466
Secure Detention Placements 1,015 1 4 2.999 1.135
Commitment Placements 1,015 1 3 1.2926 0.526
Total Escape Adjudications 1,015 1 2 1.0197 0.139
Total Failure to Appear PUOs 1,015 1 3 1.6887 0.832


Cohort 2 Analysis

The analysis of the Cohort 2 community sub-sample utilizes logistic regression to examine

the effectiveness of the overall risk to re-offend score as identified by the PACT assessment in

the prediction of a subsequent delinquency referral to the Department of Juvenile Justice for

youth who remain in the community after assessment. The Cohort 2 community sub-sample is

composed of 70.9% low risk, 13.3% moderate risk, 9.6% moderate-high risk, and 6.3% high risk

to re-offend youth. The sample is 61.1% white, and 38.9% non-white. Descriptive statistics for

the Cohort 2 community sample can be seen in Table 3-17. The percentage of low risk youth

who received a new referral to the Department within twelve months was 27.3%. The percentage

for moderate risk youth was 43.6%, and 48.3% for moderate-high risk youth, and 50.3% for high









risk youth. This means that just over half of the youth assessed as high risk to re-offend will in

fact do so within twelve months if they remain in the community post-assessment. The logistic

regression results using only overall risk to re-offend to predict a subsequent referral to the

Department for youth assessed reveal overall risk level to be a significant predictor of recidivism

at p=.000 (see Table 3-18). The Exp(B) illustrates that for every increase in overall risk to re-

offend, as assessed by the PACT, there is a 1.491 times increase in recidivism (almost a 50%

increase).

Table 3-16. Residential Cohort 2 Social History Descriptive Statistics
Social History Item Descriptive Statistics


N Minimum
Sex 1,015 0
Current School Enrollment 958 1
Recent School Conduct 957 1
Recent School Attendance 957 1
Recent Academic Performance 957 1
Anti-social Peers 1,015 0
Out-of-home Placements 1,015 1
History of Running Away 1,015 1
Parental Authority/Control 1,007 1
Household Jail History 1,007 0
History of Alcohol Use 1,015 0
History of Drug Use 1,015 0
Current Alcohol Use 894 0
Current Drug Use 894 0
History of Physical Abuse 1,015 0
History of Sexual Abuse 1,015 0
History of Neglect 1,015 1
Mental Health Problems 992 1


Maximum
1
6
5
5
5
2
4
5
3
1
1
1
1
1
2
1
2
5


Mean
0.8562
2.8758
3.3501
3.4253
3.7482
0.869
1.2916
2.0562
2.0725
0.3903
0.5251
0.7084
0.2808
0.5436
0.4
0.069
1.13
1.4849


Std. Deviation
0.351
1.422
1.188
1.603
1.019
0.536
0.681
1.372
0.679
0.488
0.5
0.455
0.45
0.498
0.766
0.254
0.337
1.094


Table 3-17. Community Cohort 2 Covariate Descriptive Statistics


Covariate


Descriptive Statistics


Overall Risk to Re-offend 8,132
Sex 8,132
Race 8,132
Referral within twelve months 8,132


Minimum Maximum Mean
1 4 1.5124
0 1 0.6981
0 1 0.3893
0 1 0.3294


Std. Deviation
0.905
0.459
0.488
0.47









Identical to the previous analyses, the next step utilized logistic regression to predict a

referral subsequent to PACT assessment using overall risk to re-offend, race (again coded white

and non-white), and sex (male and female). Consistent with the analysis of Cohort 1 community

youth who were tracked only six months post-assessment, overall risk to re-offend, race, and sex

were all significant (at p=.000) predictors of a subsequent delinquency referral (see Table 3-18).

Results reveal that youth assessed as higher risk to re-offend are in fact more likely to do so

within the one-year follow-up. Furthermore, males are more likely to receive a subsequent

delinquency referral, as are non-white youth. The second strongest relationship (race being the

strongest) to recidivism was the overall risk to re-offend as assessed by the PACT, showing for

every increase in overall risk, there is a 1.433 times increase in the dependent variable. These

results illustrate the validity of the PACT to predict short-term (six month) as well as long-term

(one year) recidivism for youth who are assessed and remain in the community post-assessment.

Table 3-18. Community Cohort 2 Logistic Regression Results
Covariate Statistics
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.4 0.025 249.763 1 .000* 1.491

Logistic Regression 2
Overall Risk to Re-offend 0.36 0.026 196.432 1 .000* 1.433
Sex 0.302 0.054 31.047 1 .000* 1.353
Race 0.428 0.049 75.848 1 .000* 1.534
* =significant at p=.05


Cohort 2 residential youth are analyzed using the same procedures utilized to examine

Cohort 1 residential youth. These youth are collective analyzed first, including both youth who

were assessed using the PACT and subsequently placed in residential commitment before one-

year follow-up, as well as youth assessed while in residential commitment and tracked following

release. Covariate descriptive statistics for these youth can be seen in Table 3-19. The Cohort 2









residential sub-sample is composed of 12.4% low risk youth (as assessed by the PACT), 17.1%

moderate risk youth, 29.2% moderate-high risk, and 41.3% high risk to re-offend youth. These

frequencies are in stark contrast to the 70.9% low risk and only 6.3% high-risk youth who

composed the Cohort 2 community sample. The Cohort 2 residential sample is also slightly less

white (52.2% compared to 61.1% for community youth) and more male (85.6% compared to

69.8% for community youth).

Table 3-19. Residential Cohort 2 Covariate Descriptive Statistics
Covariate Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 1,015 1 4 2.9931 1.04
Sex 1,015 0 1 0.8562 0.351
Race 1,015 0 1 0.4778 0.5
Referral within twelve months 1,015 0 1 0.6059 0.489


Logistic regression analysis of the full sub-sample of Cohort 2 residential youth using

overall risk to re-offend, race (white and non-white) and sex (male and female) as predictors of

referral subsequent to release reveals the covariate to be a significant predictor. The results of the

analysis reveal a severely unexpected significance (see Table 3-20). Race is significant with non-

whites more likely to recidivate than whites, as is sex with males more likely than females.

However, overall risk to re-offend is significant, but in the reverse direction as expected. Logistic

regression shows those with high risk to re-offend being significantly less likely to do so. The

same holds true when only risk to re-offend is included as a covariate predicting recidivism; the

higher the overall risk level, the less likely to recidivate for youth in residential facilities prior to

the one-year follow-up period. The strongest relationship to recidivism was with sex, followed

by race, and finally overall risk to re-offend. However, for every drop in risk to re-offend, there

was still a 0.776 times increase in the dependent variable, recidivism.









The results of the logistic regression for the Cohort 2 residential sub-sample are perplexing

and enough cause for the sample to be further disaggregated, using the same approach as that

used for the Cohort 1 residential youth. The Cohort 2 residential sample was further divided into

youth who received the PACT assessment prior to being placed in residential commitment and

then followed one year, and youth assessed at some point during a commitment placement and

then tracked post-release (see Table 3-21 for descriptive statistics). The timing of when the

PACT assessment was given (prior to, or during commitment placement) may yield some

interesting clues as to why the overall risk to re-offend is a significant predictor of recidivism,

only in the opposite direction, for Cohort 2 residential youth.

Table 3-20. Residential Cohort 2 Logistic Regression Results
Covariate Statistics
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend -0.23 0.064 13.031 1 .000* 0.794

Logistic Regression 2
Overall Risk to Re-offend -0.254 0.065 15.358 1 .000* 0.776
Sex 0.421 0.182 5.336 1 .021* 1.523
Race 0.312 0.132 5.565 1 .018* 1.366
*=significant at p=.05


Table 3-21. Split Residential Cohort 2 Covariate Descriptive Statistics
Covariate Descriptive Statistics
PACT before commitment N Minimum Maximum Mean
Overall Risk to Re-offend 984 1 4 2.9797
Sex 984 0 1 0.8598
Race 984 0 1 0.4797
Referral within twelve months 984 0 1 0.6128
PACT while in commitment N Minimum Maximum Mean
Overall Risk to Re-offend 31 1 4 3.4194
Sex 31 0 1 0.7419
Race 31 0 1 0.4194
Referral within twelve months 31 0 1 0.3871


Std. Deviation
1.043
0.347
0.5
0.487
Std. Deviation
0.848
0.445
0.502
0.495









Logistic regression was again used separately to predict recidivism for the Cohort 2

residential youth who were assessed prior to a commitment placement and the youth assessed

while in residential placement (see Table 3-22 for all Cohort 2 residential split sample analyses).

Initially, only overall risk to re-offend, as assessed by the PACT was entered in the model as a

covariate. For youth assessed prior to a commitment placement, then subsequently tracked one

year, overall risk to re-offend was again significant (p=.001), however, lower risk youth were

more likely to receive a delinquency referral within the follow-up period. This result is similar to

that found when analyzing the full Cohort 2 sub-sample. For youth assessed while in

commitment and tracked post-release, overall risk to re-offend was not significant (p=.65).

However, this result must be interpreted with caution, as there were only thirty-one youth

included in the analysis.

Table 3-22. Split Residential Cohort 2 Logistic Regression Results

Covariate Statistics
PACT before commitment
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend -0.221 0.065 11.683 1 .001* 0.802

Logistic Regression 2
Overall Risk to Re-offend -0.245 0.066 13.869 1 .000* 0.783
Sex 0.317 0.188 2.858 1 .091 1.373
Race 0.301 0.134 5.02 1 .025* 1.351
PACT while in commitment
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend -0.198 0.437 0.206 1 .65 0.82

Logistic Regression 2
Overall Risk to Re-offend 0.063 0.473 0.018 1 .893 1.066
Sex 9.516 34.54 0.076 1 .783 13580.44
Race 0.978 0.89 1.208 1 .272 2.66
* =significant at p=.05









Next, sex and race were included in the model, along with overall risk to re-offend to

predict subsequent referral to the Department of Juvenile Justice. For youth assessed by the

PACT prior to residential placement, overall risk to re-offend was significant (p=.000), again in

the opposite direction with youth assessed as lower risk more likely to receive a referral during

the one year follow-up. Race was also a significant predictor of subsequent referral (p=.025) with

non-whites more likely to be referred then whites. For youth assessed while in residential

commitment and then tracked after release, none of the three covariates (overall risk, race, or

sex) were significant predictors. Again, the logistic regression results for youth assessed while in

commitment and tracked one year must be interpreted with caution, as only thirty-one youth

were included in the analysis.

While unexpected, the result of overall risk to re-offend being significant in the opposite

direction (low risk youth more likely to recidivate) for youth assessed prior to entering

residential commitment, the finding is consistent with the plethora of research of the Principles

of Effective Intervention explored above. The first major principle, the risk principle, holds that

the duration and the intensity of treatment should mimic the risk level of the youth, with higher

risk youth receiving more intense treatment for a longer period of time. The research consistently

illustrates, for over thirty years, that low risk to re-offend youth who receive intense services

actually re-offend at higher rates than they do when receiving minimal treatment. The findings of

the current study illustrate this point, finding low risk youth who are sent to residential

commitment re-offending at higher rates than high risk youth who were provided the longer,

more intense treatment they required. This finding further illustrates the need to keep low risk

youth out of long-term residential commitment. Placing low risk to re-offend youth in long-term









residential placement actually increases their risk to re-offend and jeopardizes both the future of

these individuals and public safety.

Unfortunately, the Department of Juvenile Justice does not currently have a method to

track the type, frequency, or duration of the delinquency interventions youth receive while in

residential commitment programs. Therefore, there is no way to distinguish what type or brand

name intervention a youth received, how many times per week that intervention was delivered to

that youth, and for how many weeks it was delivered. Had the Department captured this data,

analysis could have been conducted to distinguish which treatment, or group of treatments,

delivered at what frequency and what duration was most effective at predicting a lower

likelihood of recidivism. The three covariates of treatment type, treatment frequency, and

treatment duration could have been included with risk to re-offend to investigate possible

interaction effects on the prediction of subsequent referral.

However, all that can be deduced from the current study is that the PACT is not predictive

of subsequent referral for youth who are assessed within residential commitment and tracked

post-release. This deduction holds true for both Cohort 1 residential youth tracked six months

and Cohort 2 residential youth followed for twelve months. The finding that sex and race are

sometimes significant in Cohort 1 and Cohort 2 residential analysis is not surprising and

consistent with prior empirical research that males and non-whites are more likely to re-offend

than females and whites. The finding that overall risk to re-offend predicts which youth assessed

prior to entering residential placement, again, illustrates the necessity to avoid long-term

residential placement for lower risk youth, and is in keeping with previous research described

above.









The inability of the PACT to predict recidivism for the youth placed in residential

commitment leads to the remainder of the analysis of the current study to focus on the prediction

of recidivism for youth remaining in the community after assessment. One of the original intents

of the PACT instrument was for the Department's Probation and Community Intervention branch

to more effectively determine which youth under supervision are more likely to re-offend, and to

aid in recommendations to the court in terms of the likelihood of a youth on probation

recidivating. The PACT assessment has in fact achieved this goal as the recidivism of youth who

remain in the community post-assessment has been significantly predicted in follow-ups of both

six and twelve months. The next step of the analysis is to investigate the role of gender for both

Cohort 1 and Cohort 2 youth who remain in the community.

Gender Analysis

In both the six-month and the twelve-month follow-up sub-samples, race and gender were

significant predictors with non-whites and females more likely to receive a subsequent

delinquency referral than whites and females. Previous empirical as well as theoretical research

has argued for either the existence of gender-specific prediction instruments, and even pathway-

specific prediction instruments for sub-samples of female offenders. As the PACT is utilized to

assess risk to re-offend for all youth referred to the Department, analysis of the predictive ability

for both males and females is of critical concern. This analysis will be the first known attempt to

formally validate the PACT across gender.

Logistic regression was used to examine the ability of the PACT to predict recidivism, as

measured by a subsequent referral to the Department, for both males and females who are

assessed and remain in the community. As in previous analyses, youth who are tracked six

months and remain in the community (Cohort 1) and youth tracked twelve months and remain in

the community (Cohort 2) will be assessed separately to examine the ability of the PACT to









predict short-term as well as long-term follow-up recidivism. Descriptive statistics for Cohort 1

and Cohort 2 males and females can be seen in Table 3-23.

Table 3-23. Descriptive Statistics By Gender
Covariate Descriptive Statistics
Cohort 1 Community Males N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 26,643 1 4 1.5675 0.937
Race 26,643 0 1 0.4335 0.496
Referral within six months 26,643 0 1 0.271 0.444
Cohort 1 Community Females N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 9,682 1 4 1.3735 0.789
Race 9,682 0 1 0.4323 0.495
Referral within six months 9,682 0 1 0.1918 0.394
Cohort 2 Community Males N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 5,677 1 4 1.5723 0.941
Race 5,677 0 1 0.3838 0.486
Referral within twelve months 5,677 0 1 0.3527 0.478
Cohort 2 Community Females N Minimum Maximum Mean Std. Deviation
Overall Risk to Re-offend 2,455 1 4 1.3739 0.798
Race 2,455 0 1 0.402 0.49
Referral within twelve months 2,455 0 1 0.2758 0.447


Logistic regression results reveal overall risk to re-offend and race (white and non-white)

to be significant predictors of subsequent referral, with higher risk youth and minority youth

more likely to recidivate. These results hold true for male youth who are assessed using the

PACT and remain in the community being tracked for six months and for twelve months, as well

as female youth assessed and remaining in the community both six and twelve months (see Table

3-24 for results). For both the male youth tracked six months and those tracked twelve months,

the relationship to recidivism was slightly stronger for race than overall risk to re-offend. For

both groups of female youth, overall risk to re-offend has a stronger relationship to recidivism

than race. For all youth, for every increase in overall risk to re-offend, there is between a 1.4 to

1.6 times increase in recidivism (roughly 50%). These analyses illustrate the effectiveness of the

PACT assessment in the prediction of a delinquency referral within a six-month or one-year









follow-up for all youth who remain in the community. The PACT was designed for the purpose

of doing precisely that task. The goal was to implement an instrument that could assess the

likelihood that youth that enter the juvenile justice system would recidivate if they remained in

the community.

Table 3-24. Logistic Regression Results By Gender
Covariate Statistics
Cohort 1
Logistic Regression Males B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.374 0.014 713.279 1 .000* 1.454
Race 0.405 0.028 203.312 1 .000* 1.499

Logistic Regression Females
Overall Risk to Re-offend 0.476 0.028 280.985 1 .000* 1.61
Race 0.269 0.053 25.963 1 .000* 1.309
Cohort 2
Logistic Regression Males B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.336 0.029 131.779 1 .000* 1.4
Race 0.466 0.058 64.836 1 .000* 1.593

Logistic Regression Females
Overall Risk to Re-offend 0.436 0.053 67.554 1 .000* 1.547
Race 0.328 0.093 12.398 1 .000* 1.389
* =significant at p=.05


The ability of the PACT to predict recidivism for both males and females is a critical piece

of the current study. The significant prediction of a new referral for female juveniles is essential

for the PACT to achieve the desired goal and to advance from previous prediction instruments

that were unable to predict female offending. Reisig and colleagues, when assessing the LSI-R

prediction instrument for females state, "Predictive accuracy exists if recidivism increases with

risk-need level...the relationship between risk-need and recidivism (when using the LSI-R for

females) is not statistically significant for the full sample" (Reisig et al., 2006:397). Contrary to

the results found in that analysis, the current study finds the overall risk level identified by the









PACT statistically significant in the prediction of full-sample female recidivism. These results

are extremely important with regard to literature stressing the need to have gender-specific

instruments or separate instruments for females and males. The PACT has been proven to predict

re-offending for males and for females equally well, with both a six and a twelve-month follow-

up.

While the prediction of re-offending for both males and females is an essential goal of the

current study, it is important to know whether the same aspects of the PACT are predictive for

males and females. The construction of the PACT allows for analyzing whether the criminal

history of the youth or the social history of the youth is driving the significance of the prediction.

Therefore, the next step in the analysis is to separate overall risk to re-offend into its two core

components of criminal history score and social history score to examine the predictive

significance of each separately for male and female youth. Logistic regression is used to predict a

subsequent referral to the Department for both male and female youth who are assessed and

remain in the community for twelve months (Cohort 2).

Table 3-25. Cohort 2 Descriptive Statistics By Gender
Covariate Descriptive Statistics
Cohort 2 Males N Minimum Maximum Mean Std. Deviation
Criminal History Score 5,677 0 28 6.51 4.695
Social History Score 5,677 1 16 4.59 2.425
Race 5,677 0 1 0.3838 0.486
Referral within six months 5,677 0 1 0.3527 0.478
Cohort 2 Females N Minimum Maximum Mean Std. Deviation
Criminal History Score 2,455 0 27 5.25 3.913
Social History Score 2,455 0 15 3.87 2.905
Race 2,455 0 1 0.402 0.49
Referral within twelve months 2,455 0 1 0.2758 0.447


The covariates for this analysis are criminal history, social history, and race, rather than

race and overall risk to re-offend utilized in previous analysis (see Table 3-25 for dependent









variable and covariate descriptive statistics). This allows for the determination as to whether the

extent and seriousness of previous official offending is driving the significance of the prediction,

or whether the social, educational, family, mental health, substance abuse, peer relationships, or

abuse history circumstances are driving the prediction for both males and females. Logistic

regression results show that criminal history, social history, and race are significant predictors of

subsequent referral for both males and females. The higher the criminal history and the higher

the social history, the more likely the youth will receive another referral. Both of these covariates

are significant at p=.000 for males and for females (see Table 3-26 for results). Identical analyses

were conducted on males and females who remain in the community and followed six months

(Cohort 1) post-assessment, with identical results. Furthermore, non-whites were again more

likely to receive a subsequent referral. While race has the strongest relationship to recidivism

among the predictors, for both males and females, the relationship between social history and

recidivism is stronger than that between criminal history and recidivism. For every increase in

the social history score, there is a 1.138 times increase in recidivism for males and a 1.152

increase for females.

Table 3-26. Cohort 2 Logistic Regression By Gender
Covariate Statistics
Logistic Regression Males B S.E. Wald df Sig. Exp(B)
Criminal History 0.044 0.006 50.246 1 .000* 1.045
Social History 0.129 0.012 112.92 1 .000* 1.138
Race 0.487 0.059 68.206 1 .000* 1.627
Logistic Regression Females B S.E. Wald df Sig. Exp(B)
Criminal History 0.079 0.012 42.053 1 .000* 1.083
Social History 0.141 0.017 68.163 1 .000* 1.152
Race 0.313 0.097 10.449 1 .000* 1.368
* =significant at p=.05


An additional hypothesis of this current study involving gender differences is the notion

that any significance associated with gender being a predictor of recidivism will be "washed out"









with the inclusion of additional covariates. Some researchers have argued that genderedd

pathways" exist in that the tract to delinquency for females is often different than that for males

(see Daly 1992, 1994; Daly & Chesney-Lind 1988). The current study hypothesizes it may be

possible to include variables associated with female pathways in the prediction of recidivism

which will make the seemingly significant finding that gender predicts re-offending become

insignificant in the logistic regression model.

Table 3-27. Cohort 2 Community Gendered Covariate Descriptive Statistics
Covariate Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
Sex 8,132 0 1 0.6981 0.459
History of Physical Abuse 8,132 0 1 0.1299 0.336
History of Sexual Abuse 8,132 0 1 0.0454 0.208
Drug Use 8,132 0 1 0.3915 0.488
Alcohol Use 8,132 0 1 0.2907 0.454
School Suspension/Expulsion 8,132 0 6 0.6369 1.433
Teachers/Coaches 8,132 0 5 0.3789 0.918
Positive Adult Non-family 8,132 0 4 0.4292 0.934
History of Running Away 8,132 0 5 1.3876 0.92681
Parental Supervision 8,132 0 3 0.3296 0.708
Antisocial Peers 8,132 0 2 0.6779 0.536
Referral within twelve months 8,132 0 1 0.3294 0.47


To test the above hypothesis, the Cohort 2 sub-sample of youth who are assessed using the

PACT and subsequently remain in the community to be tracked for twelve months will be used.

Gender was a significant predictor of recidivism for this sub-sample in earlier analyses. Now,

however, additional covariates will be introduced in the prediction of recidivism that have been

suggested to be issues that lead to female offending. In addition to the sex of the youth, history of

physical abuse, history of sexual abuse, history of alcohol use, history of drug use, anti-social

peer involvement, history of school suspension or expulsion, number of teachers or coaches the

youth is comfortable talking to, history of adult non-family school or job positive relationships,









history of running away or getting kicked out of the house, and level of parental supervision will

be included. All covariates are coded so that the higher the value, the greater the extent of the

issue defined (see Table 3-27 for descriptive statistics).

Logistic regression was again utilized to assess the significance of the covariates in the

prediction of recidivism within twelve months for youth who remain in the community post

PACT assessment (see Table 3-28 for results). Youth who have a history of more school

suspensions and expulsions are significantly more likely to recidivate, as are youth with more

instances of running away, inconsistent/inadequate supervision, and a higher prevalence of

antisocial peers. Interestingly, youth with more relationships with pro-social adults who are not

employers or teachers, have a greater likelihood of recidivism. History of physical abuse and

history of sexual abuse were not significant predictors in the model (though physical abuse

approached significance at p= .067), neither were history of alcohol abuse and problems related

to that use, or presence of teachers or coaches the youth feels comfortable with. Furthermore, and

congruent with hypothesis seven of the current study, sex was "washed" to insignificance in the

presence of the above-mentioned covariates. This finding runs counter with previous research by

Hubbard (2002) who found gender to remain significant, even after controlling for risk to re-

offend, quality of programming, sexual abuse, and self-esteem. The strongest relationship to the

dependent variable, recidivism, was with antisocial peers. For every increase in that independent

variable, there was a 1.34 times increase in the dependent variable. The second strongest

relationship was between parental supervision and recidivism, followed by school

suspension/expulsion and recidivism, then history of running away and recidivism, and finally

prevalence of non-family/employment/school positive adults and recidivism.









Table 3-28. Cohort 2 Community Expanded Covariate Logistic Regression

Covariate Statistics
Logistic Regression B S.E. Wald df Sig. Exp(B)
Physical Abuse 0.263 0.143 3.365 1 .067 1.3
Sexual Abuse -0.367 0.243 2.28 1 .131 0.692
Alcohol -0.066 0.131 0.257 1 .613 0.936
Drugs 0.144 0.127 1.281 1 .258 1.155
School Suspension/Expulsion 0.172 0.033 28.031 1 .000* 1.188
Teachers/Coaches -0.042 0.049 0.737 1 .391 0.959
Non-family Adults -0.112 0.056 4.058 1 .044* 0.894
Running Away 0.109 0.054 4.133 1 .042* 1.116
Parental Supervision 0.231 0.083 7.824 1 .005* 1.26
Antisocial Peers 0.293 0.098 9.03 1 .003* 1.34
Sex 0.109 0.127 0.747 1 .388 1.116
* =significant at p=.05

This finding is extremely important in that it shows that apparent sex differences may, in

actuality, be due to proportional differences in the extent to which various risk factors are present

for males and females. This supports the notion that sex differences may be differences in

degree, rather than differences in kind. While beyond the scope of the current study, future

empirical endeavors should look to see whether qualitative differences exist between the physical

and sexual abuse suffered by male and female youth and the implications of that abuse on the

youth.

The finding in the previous analysis of physical and sexual abuses not being significant

predictors of recidivism is surprising and deserves further attention. It may be the case that such

abuse is more important in the prediction of female offending as it is in male offending behavior.

For the 2,455 females and the 5,677 males in the Cohort 2 Community sub-sample, 11.5% of the

females and only 1.5% of the males reported sexual abuse, and 16.6% of the females and 11.4%

of the males reported a history of physical abuse. Due to these differences, physical abuse and

sexual abuse, along with race, history of alcohol use, history of drug use, anti-social peer









involvement, history of school suspension or expulsion, number of teachers or coaches the youth

is comfortable talking to, history of adult non-family school orjob positive relationships, history

of running away or getting kicked out of the house, and level of parental supervision, are utilized

as covariates in the prediction of recidivism separately for males and females (see Table 3-29 for

descriptive statistics). This will allow for the examination as to whether certain life events,

namely physical and sexual abuse, matter more in the prediction of re-offending for one sex than

they do for the other.


Table 3-29. Cohort 2 Community Gender-Split Covariate Descriptive Statistics
Covariate Descriptive Statistics
Cohort 2 Community Males N Minimum Maximum Mean St
Race 5,677 0 1 0.3838 0.
Physical Abuse 5,677 0 1 0.1143 0.
Sexual Abuse 5,677 0 1 0.0153 0.
Antisocial Peers 5,677 0 2 0.7013 0.
Alcohol Use 5,677 0 1 0.2869 0.
Drug Use 5,677 0 1 0.4104 0.
School Suspension/Expulsion 1,333 1 6 2.934 1.
Teachers/Coaches 1,192 1 5 1.9069 1.
Non-Family Adults 1,333 1 4 2.0645 1.
Running Away 5,677 1 5 1.291 0.
Parental Supervision 1,310 1 3 1.5351 0.
Referral within twelve months 5,677 0 1 0.3527 0.
Cohort 2 Community Females N Minimum Maximum Mean St
Race 2,455 0 1 0.402 0.
Physical Abuse 2,455 0 1 0.1658 0.
Sexual Abuse 2,455 0 1 0.1149 0.
Antisocial Peers 2,455 0 2 0.624 0.
Alcohol Use 2,455 0 1 0.2994 0.
Drug Use 2,455 0 1 0.3479 0.
School Suspension/Expulsion 442 1 6 2.8688 1.
Teachers/Coaches 414 1 5 1.9517 1.
Non-Family Adults 442 1 4 2.052 0.
Running Away 2,455 1 5 1.611 1.
Parental Supervision 434 1 3 1.5415 0.
Referral within twelve months 2,455 0 1 0.2758 0.


d. Deviation
486
318
123
541
452
492
674
145
023
795
694
478
d. Deviation
49
372
319
523
458
476
615
147
99
146
706
447









Table 3-30. Cohort 2 Community Gender-Split Covariate Logistic Regression
Covariate Statistics
Logistic Regression Males B S.E. Wald df Sig. Exp(B)
Race 0.823 0.132 38.909 1 .000* 2.278
Physical Abuse 0.324 0.179 3.292 1 .070 1.383
Sexual Abuse 0.091 0.417 0.048 1 .827 1.096
Alcohol 0.012 0.157 0.006 1 .938 1.012
Drugs 0.337 0.149 5.136 1 .023* 1.401
School Suspension/Expulsion 0.164 0.039 17.478 1 .000* 1.178
Teachers/Coaches -0.077 0.058 1.749 1 .186 0.926
Non-family Adults -0.082 0.065 1.586 1 .208 0.921
Running Away 0.073 0.072 1.047 1 .306 1.076
Parental Supervision 0.219 0.098 4.953 1 .026* 1.245
Antisocial Peers 0.278 0.114 5.929 1 .015* 1.32
Logistic Regression Females B S.E. Wald df Sig. Exp(B)
Race 0.578 0.237 5.958 1 .015* 1.783
Physical Abuse 0.379 0.257 2.177 1 .14 1.461
Sexual Abuse -0.530 0.325 2.659 1 .103 0.588
Alcohol 0.412 0.282 2.135 1 .144 1.51
Drugs -0.257 0.274 0.884 1 .347 0.773
School Suspension/Expulsion 0.097 0.065 2.227 1 .136 1.102
Teachers/Coaches 0.05 0.097 0.264 1 .607 1.051
Non-family Adults -0.286 0.116 6.057 1 .014* 0.751
Running Away 0.188 0.088 4.611 1 .032* 1.207
Parental Supervision 0.114 0.165 0.473 1 .492 1.121
Antisocial Peers 0.284 0.202 1.983 1 .159 1.329
* =significant at p=.05


Results of the analysis (reported in Table 3-30) reveal somewhat different predictors to be

significant for males and females. Both male and female re-offending is significantly predicted

by being non-white. However, all other significant covariates were different for males and

females. Starting with the independent variable with the strongest relationship to recidivism to

the weakest, greater histories of drug use and problems associated with that use (such as drugs

disrupting education, causing family conflict, causing health problems, withdrawal issues, and

tolerance increase), having antisocial peer associations, and inadequate/inconsistent parental

supervision, and a greater history of school suspensions or expulsions, is predictive of male









recidivism. The only covariates, other than race, that are predictive of female recidivism are

greater history of running away and more relationships with pro-social adults other than teachers

and employers.

Therefore, males and females have different covariates that are predictive of re-offending

behavior, even though the overall risk to re-offend and the social history score of the PACT is

predictive for both. Perhaps most surprising of the difference is that a greater presence of

antisocial peers is not predictive of female recidivism. The presence of either a history of

physical abuse or a history of sexual abuse was not significant for either the male or the female

sub-samples.

Violence Analysis

The final goal for the current study is to examine whether the PACT assessment can

predict violent re-offending6. Previous analyses in this study have illustrated the ability of the

PACT to predict recidivism, but not necessarily violent recidivism. The ability to predict which

youth will re-offend with a subsequent violent offense was never the goal of the PACT

assessment, or the Florida Department of Juvenile Justice. The PACT was designed to predict

which youth are more likely to recidivate generally, not specifically related to the level of

violence of the subsequent offense. While many studies have shown there to be a limited level of

offense specialization among juveniles (meaning a youth who has committed a violent offense is

no more likely to commit a violent offense in the future than a youth who did not commit an

initial violent offense), criminologists and the public are still interested in violent offending.




6 The current study also validated the PACT with respect to predicting recidivism for both "non-white" and "white"
youth to ensure there was no difference in the predictive ability of the instrument for different racial groups. Results
reveal the PACT to be a valid predictor for both. Further analyses illustrate the social history score to have a
stronger relationship with recidivism than the criminal history score, though both were significant at p=.000.









The current analysis will use a dependent variable similar to that used in previous analyses

above of a delinquency referral to the Department within twelve months. However, for this

analysis, only violent offenses will count for recidivism, meaning that the analysis will examine

whether the PACT can predict which youth are more likely to commit a violent offense within

twelve months post-assessment. For the purposes of this analysis, any delinquency referral

within twelve months will count as violent if the most serious charge in of the referral was

murder/manslaughter, attempted murder/manslaughter, felony sexual battery, armed robbery, or

aggravated assault and/or battery. The Cohort 2 Community sub-sample will be used in the

analysis (youth assessed who remain in the community and are tracked twelve months post-

assessment). Of the 8,132 youth in that sub-sample, 386 (4.7%) committed a violent offense

(using the criteria of this analysis), scored dichotomously as a zero or a one with a mean of

0.0475 and a standard deviation of 0.21265.

Table 3-31. Cohort 2 Community Violent Offense Logistic Regression

Covariate Statistics
Logistic Regression 1 B S.E. Wald df Sig. Exp(B)
Overall Risk to Re-offend 0.455 0.046 96.278 1 .000* 1.576
Sex 0.322 0.126 6.512 1 .011* 1.38
Race 0.817 0.109 56.599 1 .000* 2.265

Logistic Regression 2 B S.E. Wald df Sig. Exp(B)
Criminal History 0.072 0.01 53.429 1 .000* 1.075
Social History 0.103 0.02 26.732 1 .000* 1.109
Sex 0.271 0.127 4.55 1 .033* 1.311
Race 0.824 0.11 56.032 1 .000* 2.279
* =significant at p=.05


Logistic regression is used to predict violent recidivism within twelve months using the

overall risk to re-offend level from the PACT, sex, and race (white and non-white). An additional

analysis uses logistic regression to predict violent recidivism once again, but instead uses the









criminal history and social history scores composing the PACT rather than the overall risk to re-

offend level. This will determine whether it is the social history or criminal history of the youth

that are driving the prediction of violent recidivism, or both (results of both analyses can be seen

in Table 3-31).

The first analysis reveals overall risk to re-offend as evidenced by the PACT to be a

significant predictor of a violent offense subsequent to assessment for youth who remain in the

community and are tracked twelve months for follow-up recidivism. Youth who are assessed as

higher risk on by the PACT are significantly more likely to commit a violent offense during the

twelve-month follow-up than lower risk youth. The results also reveal males and non-whites to

be more likely to commit violent offenses. While race has the strongest relationship with

recidivism, for every increase in overall risk to re-offend, there is a 1.576 times increase in the

dependent variable (over 50%).

The second analysis reveals that both the youth's prior criminal history, and the

seriousness of that history, as well as the social circumstances surrounding the youth's life both

significantly predict subsequent violent offending. Youth who have a greater and more serious

criminal history are more likely to commit violent offenses during the follow-up period, as are

youth with more risky social circumstances Again, males and non-white youth are also more

likely to commit violent offenses within the twelve month time frame. While race, followed by

sex have the two strongest relationships to a violent recidivism, social history, again, has a

stronger relationship with recidivism than criminal history. These two analyses validate the

PACT for youth assessed who remain in the community as an assessment capable of predicting,

on an aggregate level, which youth are more likely to commit a violent offense within twelve

months if allowed to remain in the community.









CHAPTER 4
DISCUSSION

The main purpose of the current study was to validate the Positive Achievement Change

Tool (PACT), the risk/needs assessment utilized by the Probation and Community Intervention

program area of the Florida Department of Juvenile Justice. Eight of the ten hypotheses of the

study were supported. The overall risk to re-offend level presented by the PACT is indeed a

significant predictor of recidivism for youth who are assessed and remain in the community, with

higher risk youth more likely to re-offend. The purpose of the development of the PACT was to

create an instrument that could help Juvenile Probation Officers, and the Department, classify

youth in terms of likelihood of re-offense if all youth remained in the community. Youth

assessed as lower risk to re-offend by the PACT are indeed significantly less likely to commit an

offense if allowed to remain in the community than youth assessed as higher risk. Furthermore,

the likelihood of re-offense is exactly as the PACT predicts with low risk youth less likely than

moderate risk, who are in turn less likely than moderate-high risk, which are less likely than

high-risk youth to recidivate. The analysis related to this first hypothesis confirms the PACT as a

valid risk assessment.

The second hypothesis was not supported. The hypothesis held that youth assessed as

higher risk to re-offend would indeed do so regardless of whether the youth was assessed and

remained in the community, or assessed prior to, or during, a residential commitment placement.

The youth who remained in the community re-offend as predicted by the PACT. However,

neither youth assessed prior to commitment placement, nor youth assessed while in commitment

placement, recidivated subsequent to release as predicted by the PACT. Several hypotheses for

these phenomena were presented. First, and consistent with the literature related to the Principles

of Effective Intervention (specifically the risk principle), intensive treatment provided to low risk









youth actually increases their likelihood of recidivating. Long-term residential treatment is

beneficial to high-risk youth, and they therefore improve from the PACT assessment conducted

prior to placement. However, residential placement provided to youth assessed as low risk prior

to residential placement is actually iatrogenic, making the low risk youth worse and more likely

to re-offend.

The extant literature holds that reassessments are a necessary component in order to gather

information related to progression of criminogenic needs. The Department currently has a policy

that all youth who remain in the community must be reassessed every ninety days in order to be

certain the treatment plans for youth on probation are in keeping with the risks and needs of the

youth. While beyond the scope of this study, future analyses should examine how PACT overall

risk to re-offend scores, and individual scores for the domains of the PACT rise and fall as it

relates to the treatment/services youth receive.

An alternative hypothesis as to why the PACT is not predictive of recidivism for youth

assessed prior to residential placement is related to the treatment youth receive while committed.

The current study did not control for the type of treatment youth receive while in the more than

120 residential commitment facilities in Florida operated or contracted by the Department.

Treatment variables are as of yet unavailable within the Department's JJIS database making it

impossible to know from the monthly data extracts exactly what services each youth received

while in commitment.

The finding that the PACT is not predictive for youth assessed while in commitment and

tracked subsequent to release becomes logical once the actual questions of the PACT are

examined. Several of the questions essential to the scoring of the PACT make the instrument

artificial and directly related to the circumstance of being committed. For example, "Current use









of alcohol", "Current use of drugs", "Current friends", as well as "Youth's attendance in the

most recent (school) term" are all "fixed" due to the youth being committed. There are no

opportunities to engage in the use of alcohol or drugs, all of the other youth with which the youth

interacts are arguably antisocial, as they are all in a residential commitment facility, and all youth

are required to attend school every weekday. Many of the questions themselves force all youth to

"look" a certain way and therefore many of the questions relevant to accurate assessment,

scoring, and differentiation amongst classifications, become artificial.

The Department is currently designing, along with the developers of the PACT, an

instrument that will alleviate this issue. The new instrument will attempt to measure progress

during residential placement on decreasing criminogenic needs and risk factors, and increasing

protective factors. The instrument is still in its formative stage, soon to be piloted then evaluated.

The current study does, however, raise one issue with respect to the current business rules

associated with PACT assessments. Currently all youth are assessed within sixty days of

projected release from a residential facility, using the PACT assessment. The present study

shows the assessment prior to release to not be a significant predictor of recidivism for these

youth when they are released (more than likely due to the issue of improper questions for youth

in residential, as discussed in the paragraphs above). The Department may benefit from using the

assessment conducted sixty days prior to release as a way to guide aftercare planning, but should

be sure to reassess youth within a relatively short time of actual release from commitment to

gauge reentry success more effectively.

The third and fourth hypotheses of the current study were both supported. The third

hypothesis held that the higher the criminal history score, as presented by the PACT, the more

likely the youth is to re-offend. This was indeed the case, as a greater extent of prior official









offenses, and greater seriousness of those offenses was a significant predictor of recidivism. The

identical held true for the social history score, as the higher the score, the more likely the youth is

to re-offend. The current study did not go into depth as to which components of the criminal

history score or social history score were driving the significance of those scores, except for

some analysis with respect to sex differences. Future research would benefit from more in-depth

analysis to uncover the specific components responsible for the significance.

The fifth hypothesis was also supported, and was perhaps one of the most important

findings of the current study. The hypothesis held that the overall risk to re-offend level, as

evidenced by the PACT, would be a significant predictor of both male and female recidivism.

The analysis conducted by the current study on the ability of the PACT to predict female as well

as male re-offending validate the use of the PACT in the state of Florida for all youth under the

care and custody of DJJ. The current study illustrates there exists no need, with respect to

predicting re-offending, for a "gender-specific" assessment as the PACT predicts female and

male re-offending equally well. The Pact establishes which classifications of youth will

recidivate (on an aggregate level) if they are able to remain in the community. Therefore, the

PACT is an appropriate assessment for DJJ to use to determine which youth should be

recommended for community placement and which youth are significantly more likely to be a

detriment to public safety if they are allowed to remain in the community.

The finding that the PACT significantly predicted recidivism for the full sample of female

youth assessed and subsequently remaining in the community differentiates the PACT from other

risk assessment instruments unable to predict full sample female re-offending, and only re-

offending of sub-groups of females. As mentioned in the review of the literature, previous

research indicates that actuarial assessments (specifically the LSI-R) are not predictive of female









offending as a whole, but are valid for predicting the recidivism for economically motivated

women (Reisig et al., 2006). The current study shows the PACT to be a valid instrument for the

prediction of female offending, without the necessity of having to sub-group the females, or only

being able to predict re-offend of certain of those sub-groups. This finding runs counter to much

of the gendered pathways research asserting the need to examine the ways females become

involved in offending behavior, and the rationale of why those females engage in that behavior.

The sixth hypothesis was also supported by the current study. Males are significantly more

likely to re-offend than females. This pattern held true even after controlling for overall risk to

re-offend as evidenced by the PACT and a variety of other criminal history and social history

covariates. This finding is in keeping with a plethora of criminological literature and empirical

research, such that it is a universal truth in criminology that males are more likely to recidivate

than females regardless of time period, age of the youth, country, or culture.

The seventh hypothesis was supported by the current study. The hypothesis asserted that

any sex differences that would appear would be "washed out" with the inclusion of proxy

measures. The inclusion of sexual and physical abuse history, drug and alcohol abuse, delinquent

peer associations, history of school expulsion/suspensions, number of teachers/coaches the youth

is comfortable with, positive adult non-family relationships, history of running away, and level

of parental supervision did indeed decrease the significance of simply being male or female. Sex

was reduced from being a significant predictor of recidivism, with males more likely to re-

offend, by the inclusion of the proxy measurements. Future direction for additional analyses

should include a more in-depth analysis of the more than 126 PACT measures that may be

proxies for sex to examine sex differences in re-offending predictors.









In relation to sex difference, the eighth hypothesis was supported. The hypothesis was that

a history of physical or sexual abuse would not be a stronger predictor of female re-offending

than of male re-offending. The original notion behind the hypothesis was related to the seventh

hypothesis of proxy measures being able to reduce sex (being male or female) to insignificance,

which did indeed occur. Although the basis for the prediction that physical and sexual abuse

would not be stronger predictors of female recidivism did not play out as hypothesized, the

analysis led to interesting findings. History of physical abuse and history of sexual abuse are not

stronger predictors for female re-offending than for male re-offending because neither physical

or sexual abuse history was found to significantly predict either male or female recidivism. The

seventh hypothesis (correctly) believed that inclusion of proxy measures such as physical and

sexual abuse would "wash out" the significance of being male or female, meaning that it is really

the circumstance of having been abused that led to delinquent behavior, and not being male or

female. The eighth hypothesis held that males who were abused physically or abused sexually

would be just as likely as females with those histories to re-offend. The hypothesis was

supported, but not in the way in which it was intended. The physical and sexual abuse history did

not lead to the offending behavior, as believed, because the abuse histories did not significantly

predict the behavior at all.

The final hypothesis examined whether the PACT was predictive of violent recidivism.

While the original intent of the PACT was to enable the prediction of which youth would be

more likely to re-offend if they remained in the community, and to categorize that likelihood into

four components (low, moderate, moderate-high, and high), researchers have often sought to

examine whether risk/needs assessments can predict offenses more detrimental to public safety.

Violent offense prediction is often first on the list of subcomponents of offense to attempt to









predict. Hypothesis ten was supported as the analysis showed the overall risk to re-offend score

of the PACT to be a significant predictor of violent recidivism for youth assessed who remain in

the community.

It is a significant step for the PACT to be able to significantly predict which youth, on an

aggregate level, are more likely to commit a violent offense if they remain in the community

than other youth. Furthermore, only a very narrow range of offenses was included in the

"violent" category. Only murder/manslaughter, attempted murder/manslaughter, felony sexual

battery, armed robbery, and aggravated assault and/or battery were included. This extremely

strict assortment of offenses was included so as to be certain the analysis was capturing the most

violent against person offenses possible. Future research should examine whether the PACT is

equally as successful at predicting other categories of offenses, such as property or drug offenses.

Future direction may also include examination of which of the more than 46 questions within the

Pre-Screen, and which of the more than 126 questions within the Full Assessment are driving the

prediction of violent offenses.









CHAPTER 5
CONCLUSION

The Florida Department of Juvenile Justice received over 150,000 delinquency referrals,

representing over 94,000 youth, between July 1, 2005 and June 31, 2006. The Department has

developed and implemented a risk/needs assessment to help predict which youth are more likely

to recidivate if they remain in the community, as well as to identify the criminogenic needs of all

youth referred so as to provide appropriate treatment to reduce those risk factors and increase

protective factors. The purpose of the current study was to validate that instrument, the Positive

Achievement Change Tool (PACT), for the youth referred in Florida. The current study utilized

the full population of PACT assessments, one for every youth referred making any issue

regarding the generalizabilty of the findings irrelevant.

The analysis shows that the PACT is indeed a valid instrument and that the higher the risk

to re-offend as evidenced by the PACT, the more likely recidivism will occur for youth assessed

who remain in the community. Furthermore, the PACT is valid for both males and females. The

finding that the PACT significantly predicts female offending as well as male offending is

substantial, and places the PACT above other risk/needs assessments, which have been found

incapable of predicting full-sample female recidivism. Additional analyses showing the ability of

the PACT to predict not only general recidivism, but violent offending as well are also

important. These analyses validate the use of the PACT for the prediction of risk to recidivate for

all youth referred to the Department. The PACT was able to predict both short-term (six month)

and long-term (one year) recidivism.

The PACT, for the youth in the current sample, was capable of predicting recidivism for

youth assessed prior to entering residential commitment facilities that are tracked subsequent to

release, only in the opposite direction as intended. For the youth assessed as higher risk to re-









offend, this may be mainly due to progress within specific criminogenic needs as a result of the

treatment youth receive within residential facilities. However, there currently exists no way to

discern what treatment youth receive based on the monthly data extracts from the Department's

JJIS database and contact with each of the over 120 residential facilities would have been futile

as there are often no records of the frequency or duration of particular treatments for each

individual youth. For the youth assessed as low risk to re-offend, the extant literature coupled

with the results of the current study suggest this is due to the negative implications of providing

low risk youth intensive treatment.

The PACT was also not able to predict re-offending for youth assessed while in residential

commitment facilities and tracked subsequent to release. The major limitation of this finding is

that only thirty-one youth composed the sub-sample of youth assessed while in commitment and

tracked one year. Possibilities for this inability include the small sample size lacking power to

capture statistical significance and the artificial nature of some PACT questions for youth in a

residential commitment facility. Future studies should revisit this issue when a greater number of

PACT assessments have been conducted on youth while in residential placement. In the event

the results of this study are replicated, and the PACT administered while a youth is in

commitment is still not predictive of recidivism post-release, the Department may wish to

consider ensuring an additional PACT is administered shortly after release so as to gain a more

accurate picture of the risk and protective factors for youth while in the community.









CHAPTER 6
FUTURE DIRECTIONS

Future empirical research should attempt to provide a more in-depth analysis regarding the

individual questions which compose the four domains of the PACT Pre-Screen and the twelve

domains of the Full Assessment. Each domain is composed of multiple questions and factor

analysis would be a useful tool to examine whether domains are individual constructs or multiple

factors. Furthermore, scale analysis could be used to examine the domains as they relate to the

social history score composing one half of the scoring matrix used by the PACT to determine

risk to re-offend.

A substantial amount of work remains to be done regarding the individual domains, or,

more specifically, questions, which drive the significance of male versus female offending. The

finding that the social history score significantly predicts male and female offending is

important, but it may be the case that certain individual covariates that compose that score are

more essential to female offending and others more essential to male offending. The validation

of the PACT for both males and females was the initial step, and essential if the instrument is to

be used to assess both sexes, but more extensive work should be undertaken to examine the

covariates which lead to female offending behavior and, separately, male offending behavior, so

the appropriate criminogenic needs can be targeted with treatment.

An additional stream of future research should focus on examining whether there may exist

a better alternative for scoring overall risk to re-offend based on the information gathered by the

PACT. As the PACT has been found significant in the prediction of recidivism for almost every

analysis attempted by the current study, this task may be difficult, but worthy of undertaking. As

the current study was an attempt to validate an existing instrument, there was no need to create a

construction sample and validation samples. However, ample data exist for the attempt to create









an instrument based on various PACT covariates and scoring of those covariates to then be

validated and assessed for shrinkage. Additional future directions may also wish to examine

whether more simple combinations of covariates are equally as predictive as the complex,

automated scoring of the PACT assessment.

Perhaps the most important analysis that has yet to be conducted relates to PACT

reassessments. Future work should examine the ways scores on the various domains of the

PACT fluctuate from one assessment to the next. More importantly is the examination of the

ways in which the individual domain scores fluctuate when youth are provided with specific

treatments. Considerable work should be undertaken related to how the provision of a

standardized evidence-based treatment curriculum positively, or negatively affects a domain

score, and whether a similar effect for the identical treatment is observed for both males and

females.

The current study in conjunction with the empirical work suggested above would take the

question of "What works?" to "What works for whom?" Answers to these questions will guide

the Department toward providing the best treatment possible for the risk factors and needs

presented by the youth served. The provision of proven effective, evidence-based treatment,

targeted to specific individualized criminogenic needs, as presented by a validated risk/needs

assessment as the PACT has shown itself to be in the current study, is the quintessential step to

reducing juvenile recidivism, and thereby increasing public safety, and to improving the lives of

the youth served throughout the state of Florida.










APPENDIX A
HOW TO READ AND INTERPRET THE PACT OVERVIEW AND THE YES CASE PLAN
REPORTS


PACT Overview Report


Name: Chopper Test
Created By: FL State Acct Administrator
Created Date: Aug 9 2006 1:23PM


DOB: 5/111990


DJJID: 804266
Last Modified By: FL State Acct Administrator
Last Modified Date: Aug 10 2006 1:49PM


Overall Level of Risk to Re-Offend: Low
Record of Referrals Risk Score: 3
Social History Risk Score: 4


12: Skills


-I-I-I-


Figure A-1. PACT Overview Report


Risk Factors Static and Dynamic Protective Factors
Combined
0% 25% 50% 75% 100% Domain 0% 25% 50% 75% 100%

SA: History oF Relationships
__ A: Employment History
___ 7A: Family History
Si_10: Attitulesiehaviors
-I 3A: School History
S7B: Current ULving Arrangement ___
SS g8A: Alcohol and Drug History
1: Record of ReFerrals
3B: Current School Status
4A: Historic Use oF Free Time
48: Current Use of Free Time
SB: Current Employment
SB: Current Relationships
8B: Current Alcohol and Drugs
9A: Mental Health History
9B: Current Mental Health
11: Aggression









The PACT (Positive Achievement Change Tool) is designed to accomplish four basic

objectives:

* Determine a youth's level of risk for re-offending as a way to target resources to higher-
risk youth.
* Identify the risk and protective factors linked to criminal behavior so that the rehabilitative
effort can be tailored to address the youth's unique assessment profile.
* Develop a case management approach focused on reducing risk factors and increasing
protective factors.
* Allow managers to determine if targeted factors change as a result of the court's
intervention.

The Overview Report helps Florida's juvenile justice practitioners accomplish those

objectives by showing, at-a-glance:

Overall level of risk to re-offend: In simple terms, a youth with a low overall level of

risk is much less likely to commit another crime than a youth with a high overall level of risk.

Indeed, according to the Pre-Validation Study of the PACT by the Florida Justice Research

Center conducted in September 2005, recidivism rates for higher-risk youth are 57% compared

to 17% for low-risk youth. This is key information because the research literature is clear that

"what works" in reducing recidivism is matching the levels of treatment with the risk levels of

offenders. High-risk offenders require intensive interventions to reduce recidivism, while low-

risk offenders benefit most from low intensity interventions or no intervention at all.

Risk factors are the circumstances or events in the youth's life that increase the likelihood

that the youth will start or continue criminal activities. In the Overview Report, they are

prioritized from highest to lowest; the higher the risk score, the more risk in that area. They can

be static or dynamic. Staticfactors are circumstances in a youth's life that are historic and

cannot be changed, such as a history of physical abuse. In the Overview Report, Domain 1,

Record of Referrals, and the other domains labeled with an "A" are static. All the others are

dynamic. Dynamic factors are circumstances or conditions in a youth's life that can potentially









be changed, such as the youth's friends or school performance. Dynamic factors are used to

guide the rehabilitative effort.

Protective factors are circumstances or events in the youth's life that reduce the likelihood

of the youth committing a crime, those positive things that help the youth overcome adversity.

Domains: The PACT full assessment includes 12 major domains related to juvenile

delinquency and continued criminal activity based on the research literature: (1) Criminal

History; (2) Gender; (3) School; (4) Use of Free Time; (5) Employment; (6) Relationships; (7)

Family/Current Living Arrangements; (8) Alcohol and Drugs; (9) Mental Health; (10)

Attitudes/Behaviors; (11) Aggression; and (12) Skills.

Pre-Screen vs. Full Assessment: The PACT implementation consists of a two-stage

process. The first stage is a pre-screen assessment completed for all youth placed on probation.

The pre-screen is a shortened version of the full assessment that quickly indicates whether a

youth is of low-, moderate-, moderate-high-, or high-risk to re-offend. Each of these levels is

presumed to have distinctly different recidivism rates.

The second stage, a full assessment, is required by DJJ business rules, for youth assessed

as moderate-high or high-risk on the pre-screen. The full assessment identifies a youth's risk and

protective factor profile and then forwards only those factors that are dynamic to the YES Case

Plan to guide rehabilitative efforts.

The YES Case Plan Report displays the dynamic factors that are influencing the youth's

anti-social behavior, and it provides the facts (evidence) needed to develop a Case Plan to

effectively address these prioritized criminogenic needs. In a very focused way, the PACT and

the YES Case Plan help JPOs and judges define the youth's unique behavioral problems and

create goals and actions to resolve those problems.












Juvenile Case Plans
You are currently working with Juvenile "Chopper Test" [ Return to JJs ]


Goals for Youth Empowered Success (YES) Case Plan
Based on Prioritized Criminogenic Needs

Name: Chopper Test DOB: 5/1/1990 DJJID: 804266
Created By: FL State Acct Administrator Last Modified By: FL State Acct Administrator
Created Date: Aug 9 2006 1:23PM Last Modified Date: Aug 10 2006 1:49PM
Scores from PACT Full Assessment completed 8/10/2006
Risk Protective
Lou o e Higaodhr e.
Lou Moderate Mo te- High e Doma Lo Moderate Hre- ighq
High f hrk on dumon nlame to add a gypl High




, ,,,e -r ,,-r :, :l q


View Risk Factors


View Protective Factors


Figure A-2. YES Case Plan

The responsivity principle: The Assessments.com software simplifies and organizes this

process and provides a critical reporting function for monitoring individual and systemic

progress toward the desired positive outcomes. However, it is up to the case workers to use their

professional judgment to choose which behavioral problems to work on based on responsivity

issues; and to choose those actions which the youth and the parents are most ready, willing and

able to take to help turn the youth's life around. If implemented successfully, the Case Plan will

increase the youth's protective factors and lower the risk factors, and will result in lower overall

recidivism.









The risk principle: It is important, of course, to target all this effort, not at the low-risk,

but at higher-risk youth. The PACT provides the facts (evidence) for judges to focus their

resources on higher-risk youth and assign lower-risk youth to minimum- supervision caseloads,

bringing savings and more effectiveness to the entire juvenile justice system.



Re-assessments. Either the pre-screen or the full assessments can be re-administered to

measure changes in risk and protective factors as interim outcomes for court interventions. This

comparative-over-time information can help a judge make good sentencing decisions. For

example, a youth who at intake is not attending school, but who returns to school as a result of

court-ordered sanctions and then performs well as measured by a PACT re-assessment can be

expected, on average, to be associated with reduced recidivism.











APPENDIX B
PACT PRE-SCREEN ASSESSMENT


Referrals, rather than offenses, are used to assess the persistence of re-offending by the youth. Include only
referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult
court (regardless of whether successfully completed).
1. Age at first offense: The age at the time of the offense for which the youth was referred to O Over
juvenile court for the first time on a non-traffic misdemeanor or felony that resulted in 16
diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. 0 16
O15
O 13 to
14
O 12 and
Under
Felony and misdemeanor referrals: Items 2 and 3 are mutually exclusive and should add to the total
number of referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or
referral to adult court.
2. Misdemeanor referrals: Total number of referrals for which the most serious offense was a 0 None
non-traffic misdemeanor that resulted in diversion, adjudication withheld, adjudication, or one
deferred prosecution or referral to adult court (regardless of whether successfully completed). 0 Two
O Three
or four
O Five or
more
3. Felony referrals: Total number of referrals for a felony offense that resulted in diversion, O None
adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless O One
of whether successfully completed). O Two
O Three
or
more
Against-person or weapon referrals: Items 4, 5, and 6 are mutually exclusive and should add to the total
number of referrals that involve an against-person or weapon offense, including sex offenses that resulted in
diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of
whether successfully completed).
4. Weapon referrals: Total referrals for which the most serious offense was a firearm/weapon O None
charge or a weapon enhancement finding. O One or
more
5. Against-person misdemeanor referrals: Total number of referrals for which the most O None
serious offense was an against-person misdemeanor a misdemeanor involving threats, O One
force, or physical harm to another person or sexual misconduct (assault, coercion, O Two or
harassment, intimidation, etc.). more
6. Against-person felony referrals: Number of referrals involving force or physical harm to O None
another person including sexual misconduct as defined by FDLE as violent felonies. O One or
two
O Three
or
more
Sex offense referrals: Items 7 and 8 are mutually exclusive and should add to the total number of referrals
that involve a sex offense or sexual misconduct that resulted in diversion, adjudication withheld, adjudication,
deferred prosecution or referral to adult court.











7. Sexual misconduct misdemeanor referrals: Number of referrals for which the most serious O None
offense was a sexual misconduct misdemeanor including obscene phone calls, indecent O One
exposure, obscenity, pornography, or public indecency, or misdemeanors with sexual O Two or
motivation. more
8. Felony sex offense referrals: Referrals for a felony sex offense or involving sexual O None
motivation including carnal knowledge, child molestation, communication with minor for O One
immoral purpose, incest, indecent exposure, indecent liberties, promoting pornography, rape, O Two or
sexual misconduct, or voyeurism, more
9. Confinements in secure detention where youth was held for at least 48 hours: Number O None
of times the youth was held for at least 48 hours physically confined in a detention facility. O One
O Two
O Three
or
more
10. Commitment orders where youth served at least one day confined under residential 0 None
commitment: Total number of commitment orders and modification orders for which the O One
youth served at least one day confined under residential commitment. A day served includes O Two or
credit for time served. more
11. Escapes: Total number of attempted or actual escapes that resulted in adjudication. O None
O One
O Two or
more
12. Pick Up Orders forfailure-to-appear in court or absconding supervision: Total number O None
of failures- to-appear in court or absconding supervision that resulted in a pick 0 One
up order being issued. Exclude failure-to-appear warrants for non-criminal matters. O Two or
more

Current is defined as behaviors occurring within the last six months
1. Youth's Gender O Male
O Female
2a. Youth's current school enrollment status, regardless of O Graduated, O Suspended
attendance: If the youth is in home school as a result of being GED O Dropped
expelled or dropping out, check the expelled or dropped out box, O Enrolled out
otherwise check enrolled, full-time O Expelled
O Enrolled
part-time
2b. Youth's conduct in the most recent term: Fighting or threatening O Recognition for good
students; threatening teachers/staff; overly disruptive behavior; behavior
drug/alcohol use; crimes, e.g., theft, vandalism; lying, cheating, O No problems with school
dishonesty, conduct
O Problems reported by
teachers
O Problem calls to parents
O Calls to police
2c. Youth's attendance in the most recent term: Full-day absence O Good attendance with few
means missing majority of classes. Partial-day absence means absences
attending the majority of classes and missing the minority. Habitual O No unexcused absences
truancy as defined in FS includes 15 unexcused absences in a 90-day O Some partial-day unexcused
period, absences
O Some full-day unexcused
absences
0 Habitual truant










2d. Youth's academic performance in the most recent school O Honor student (mostly As)
term: O Above 3.0 (mostly As and
Bs)
O 2.0 to 3.0 (mostly Bs and Cs,
no Fs)
O 1.0 to 2.0 (mostly Cs and Ds,
some Fs)
O Below 1.0 (some Ds and
mostly Fs)
3a. History of anti-social friends/companions: Anti-social peers are O Never had consistent
youths hostile to or disruptive of the legal social order; youths who friends or companions
violate the law and the rights of others and other delinquent youth. O Had pro-social friends
(Check all that apply.) O Had anti-social friends
O Been a gang
member/associate
3b. Current friends/companions youth actually spends time with: O No consistent friends or
(Check all that apply.) companions
O Pro-social friends
O Anti-social friends
O Gang member/associate
4. History of court-ordered or DCF voluntary out-of-home and O No out-of-home placements
shelter care placements exceeding 30 days: Exclude DJJ residential exceeding 30 days
commitments. O 1 out-of-home placement
O 2 out-of-home placements
O 3 or more out-of-home
placements
5. History of running away or getting kicked out of home: Include O No history of running
times the youth did not voluntarily return within 24 hours, and include away/being kicked out
incidents not reported by or to law enforcement O 1 instance of running
away/kicked out
O 2 to 3 instances of running
away/kicked out
O 4 to 5 instances of running
away/kicked out
O Over 5 instances of running
away/kicked out
6a. History of jaillimprisonment of persons who were ever involved in O No jail/imprisonment history
the household for at least 3 months: (Check all that apply.) in family
O Mother/female caretaker
O Father/male caretaker
O Older sibling
O Younger sibling
O Other member
6b. History of jaillimprisonment of persons who are currently O No jail/imprisonment history
involved with the household: (Check all that apply.) in family
O Mother/female caretaker
O Father/male caretaker
O Older sibling
O Younger sibling
O Other member










6c. Problem history of parents who are currently involved with the O No problem history of
household: (Check all that apply). parents in household
O Parental alcohol problem
history
O Parental drug problem
history
O Parental physical health
problem history
O Parental mental health
problem history
O Parental employment
problem history
7. Current parental authority and control: O Youth usually obeys and
follows rules
O Sometimes obeys or obeys
some rules
O Consistently disobeys, and/or
is hostile
8a. Youth's history of alcohol use: (Check all that apply.) O No past alcohol use
O Past alcohol use
O Alcohol caused family
conflict
0 Alcohol disrupted education
O Alcohol caused health
problems
O Alcohol interfered with
keeping pro-social friends
O Alcohol contributed to
criminal behavior
O Youth needed increasing
amounts of alcohol to
achieve same level of
intoxication or high
O Youth experienced
withdrawal problems
8b. Youth's history of drug use: (Check all that apply.) O No past drug use
O Past drug use
O Drugs caused family conflict
0 Drugs disrupted education
O Drugs caused health
problems
O Drugs interfered with
keeping pro-social friends
O Drugs contributed to
criminal behavior
O Youth needed increasing
amounts of drugs to achieve
same level of intoxication or
high
O Youth experienced
withdrawal problems










8c. Youth's Current alcohol use: (Check all that apply.)


8d. Youth's current drug use: (Check all that apply.)


O No current alcohol use
O Current alcohol use
O Alcohol causing family
conflict
D Alcohol disrupting education
O Alcohol causing health
problems
O Alcohol interfering with
keeping pro-social friends
O Alcohol contributing to
criminal behavior
O Youth needs increasing
amounts of alcohol to
achieve same level of
intoxication or high
O Youth experiences
withdrawal problems
O No current drug use
O Current drug use
O Drugs causing family
conflict
D Drugs disrupting education
O Drugs causing health
problems
O Drugs interfering with
keeping pro-social friends
O Drugs contributing to
criminal behavior
O Youth needs increasing
amounts of drugs to achieve
same level of intoxication or
high
O Youth experiences
withdrawal problems


For abuse and neglect, include any history that is suspected, whether or not reported or substantiated;
exclude reports of abuse or neglect proven to be false.
9a. History of violencelphysical abuse: Include suspected incidents of O Not a victim of
abuse if disclosed by youth, whether or not reported or substantiated, violence/physical abuse
but exclude reports investigated but proven to be false. (Check all that 0 Victim of violence/physical
apply.) abuse at home
O Victim of violence/physical
abuse in a foster/group
home
O Victimized by family
member
O Victimized by someone
outside the family
O Attacked with a weapon
9b History of witnessing violence: (Check all that apply) Include O Has not witnessed violence
perpetrators and victims of violence as having witnessed violence. O Has witnessed violence at
home
O Has witnessed violence in a
foster/group home
O Has witnessed violence in
the community
O Family member killed as
result of violence










9c History of sexual abuse/rape: Include suspected incidents of abuse if O Not a victim of sexual
disclosed by youth, whether or not reported or substantiated, but abuse/rape
exclude reports investigated but proven to be false. (Check all that 0 Sexually abused/raped by
apply.) family member
O Sexually abused/raped by
someone outside the family
10. History of being a victim of neglect: Include suspected incidents of O Not victim of neglect
neglect, whether or not reported or substantiated, but exclude reports O Victim of neglect
investigated but proven to be false.


11. History of mental health problems: Such as schizophrenia, bi-polar,
mood, thought, personality, and adjustment disorders. Exclude
substance abuse and special education since those issues are
considered elsewhere. Confirm by a professional in the social
service/healthcare field.


O No history of mental health
problems)
O Past history of mental health
problems) diagnosis (more
than six months ago)
O Diagnosed with mental health
problems)
O Only mental health
medication prescribed. If yes,
list

O Only mental health treatment
prescribed
O Mental health medication and
treatment prescribed


I II


1. History of suicidal ideation: Include any previous
thoughts, threats, plans and attempts even if youth
indicates they were manipulative or there was no intent.
(Check all that apply)


O Has never had serious thoughts about
suicide
O Has had serious thoughts about
suicide
O Has made a plan to commit suicide. If
yes, describe

O Has attempted to commit suicide. If
yes, describe attempts) and
date(s)
O Feels life is not worth living no hope
for future.
O Knows someone well who has
committed suicide. If yes, who, when
and how


O Engages in self-mutilating
behavior
2. History of anger or irritability: O No history of anger/irritability
O History of occasional feelings of
anger/irritability
O History of consistent feelings of
anger/irritability
O History of aggressive reactions to
feelings of anger/irritability.










3. History of depression or anxiety O No history of depression/anxiety
O History of occasional feelings of
depression/anxiety
O History of consistent feelings of
depression/anxiety
O History of impairment in every day
tasks due to depression/anxiety
4. History of Somatic Complaints: Bodily or physical O No history of somatic complaints
discomforts associated with distress, such as stomachaches
or headaches O History of one or two somatic
complaints
O History of three or four somatic
complaints
O History of 5 or more somatic
complaints
5. History of thought disturbance O No unusual thoughts or beliefs
O Presence of hallucinations (auditory or
visual)
O Presence of beliefs that the youth is
controlled by others or others control
the youth.
6. History of traumatic experience: Lifetime exposure to O No presence of traumatic event
events such as rape, abuse or observed violence, including
dreams or flashbacks O Presence of traumatic event
dreams or flashbacks
O Flashbacks to traumatic event





1. Attitude toward responsible law abiding O Abides by conventions/values
behavior: O Believes conventions/values sometimes apply to him
or her
O Does not believe conventions/values apply to him or
her
O Resents or is hostile toward responsible behavior
2. Accepts responsibility for anti-social O Accepts responsibility for anti-social behavior
behavior: O Minimizes, denies, justifies, excuses, or blames
others
O Accepts anti-social behavior as okay
O Proud of anti-social behavior
3. Belief in yelling and verbal aggression to O Believes verbal aggression is rarely appropriate
resolve a disagreement or conflict: O Believes verbal aggression is sometimes
appropriate
O Believes verbal aggression is often appropriate
4. Belief in fighting and physical aggression to O Believes physical aggression is never appropriate
resolve a disagreement or conflict: O Believes physical aggression is rarely appropriate
O Believes physical aggression is sometimes
appropriate
O Believes physical aggression is often appropriate











5. Reportslevidence of violence not included in E No reports/evidence of violence
criminal history: (Check all that apply.) O Violent outbursts, displays of temper, uncontrolled
anger indicating potential for harm
O Deliberately inflicting physical pain
O Using/threatening with a weapon
O Fire starting
O Violent destruction of property
O Animal cruelty
6. Reports of problem with sexual aggression O No reports/evidence of sexual aggression
not included in criminal history: (Check all O Aggressive sex
that apply.) O Sex for power
O Young sex partners
O Child sex
O Voyeurism
lO Exposure











APPENDIX C
FULL ASSESSMENT


Referrals, rather than offenses, are used to assess the persistence of re-offending by the youth. Include only
referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court
(regardless of whether successfully completed).


2. Age at first offense: The age at the time of the offense for which the youth was referred to
juvenile court for the first time on a non-traffic misdemeanor or felony that resulted in diversion,
adjudication withheld, adjudication, deferred prosecution or referral to adult court.


0 Over 16
016
015
0 13to 14
0 12 and
Under


Felony and misdemeanor referrals: Items 2 and 3 are mutually exclusive and should add to the total number of
referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court.
5. Misdemeanor referrals: Total number of referrals for which the most serious offense was a non- O None or
traffic misdemeanor that resulted in diversion, adjudication withheld, adjudication, deferred one
prosecution or referral to adult court (regardless of whether successfully completed). O Two
O Three
or four
O Five or
more
6. Felony referrals: Total number of referrals for a felony offense that resulted in diversion, O None
adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of O One
whether successfully completed). O Two
O Three
or more
Against-person or weapon referrals: Items 4, 5, and 6 are mutually exclusive and should add to the total number
of referrals that involve an against-person or weapon offense, including sex offenses that resulted in diversion,
adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether
successfully completed).
7. Weapon referrals: Total referrals for which the most serious offense was a firearm/weapon O None
charge or a weapon enhancement finding. O One or
more
12. Against-person misdemeanor referrals: Total number of referrals for which the most serious O None
offense was an against-person misdemeanor a misdemeanor involving threats, force, or physical 0 One
harm to another person or sexual misconduct (assault, coercion, harassment, intimidation, etc.). O Two or
more
13. Against-person felony referrals: Number of referrals involving force or physical harm to another O None
person including sexual misconduct as defined by FDLE as violent felonies. O One or
two
O Three
or more
Sex offense referrals: Items 7 and 8 are mutually exclusive and should add to the total number of referrals that
involve a sex offense or sexual misconduct that resulted in diversion, adjudication withheld, adjudication, deferred
prosecution or referral to adult court.


O None
O One
O Two or
more


14. Sexual misconduct misdemeanor referrals: Number of referrals for which the most serious
offense was a sexual misconduct misdemeanor including obscene phone calls, indecent exposure,
obscenity, pornography, or public indecency, or misdemeanors with sexual motivation.











15. Felony sex offense referrals: Referrals for a felony sex offense or involving sexual motivation O None
including carnal knowledge, child molestation, communication with minor for immoral purpose, O One
incest, indecent exposure, indecent liberties, promoting pornography, rape, sexual misconduct, or O Two or
voyeurism. more
16. Confinements in secure detention where youth was held for at least 48 hours: Number of O None
times the youth was held for at least 48 hours physically confined in a detention facility. O One
O Two
O Three
or more
17. Commitment orders where youth served at least one day confined under residential O None
commitment: Total number of commitment orders and modification orders for which the youth O One
served at least one day confined under residential commitment. A day served includes credit for O Two or
time served. more
18. Escapes: Total number of attempted or actual escapes that resulted in adjudication. O None
O One
O Two or
more
19. Pick Up Orders for failure-to-appear in court or absconding supervision: Total number of O None
failures-to-appear in court or absconding supervision that resulted in a pick up order being issued. O One
Exclude failure-to-appear warrants for non-criminal matters. O Two or
more





1. Youth is a special education student or has a formal O No special education need
diagnosis of a special education need: (Check all that 0 Learning O Mental
apply.) retardation
O Behavioral O ADHD/ADD
2. History of expulsions and out of school suspensions O No expel/suspend O 4 or 5
since the first grade: O 1 expel/suspend O 6 or 7
O 2 or 3 O More than 7
3. Age at first expulsion or suspension: O No expulsions O 14 to 15
O 5 to 9 years old years old
O 10 to 13 years old O 16 to 18
years old
4. Youth has been enrolled in a community school during the O No, graduated/GED and not
last 6 months, regardless of attendance: attending school, do not complete
Domain 3B
O No, dropped-out or expelled for more
than six months, do not complete
Domain 3B
O Yes, must complete Domain 3B


0 For Initial Assessments, current is the most recent term in last 6 months.
0 For Re-assessments and Final Assessments, current is the last 4 weeks in the most recent term.
1. Youth's current school enrollment status, regardless of O Graduated/GED O Suspended
attendance: If the youth is in home school as a result of being O Enrolled full-time O Dropped
expelled or dropping out, check expelled or dropped out; O Enrolled part-time out
otherwise check enrolled, if in home school. O Expelled
2. Type of school in which youth is enrolled: O Public academic O Private
0 Vocational academic
Name of School O Alternative O Home
O GED program school
O College
O Other










3. Youth believes there is value in getting an education: O Believes getting an education is of
value
O Somewhat believes education is of
value
O Does not believe education is of
value
4. Youth believes school provides an encouraging O Believes school is encouraging
environment for him or her: O Somewhat believes school is
encouraging
O Does not believe school is
encouraging
5. Teachers, staff, or coaches the youth likes or feels O Not close to any teachers, staff, or
comfortable talking with: coaches
0 Close to 1 0 Close to 3
0 Close to 2 0 Close to 4
or more
6. Youth's involvement in school activities during most O Involved in 2 or more activities
recent term: School leadership; social service clubs; music, O Involved in 1 activity
dance, drama, art; athletics; other extracurricular activities. O Interested but not involved in any
activities
O Not interested in school activities
7. Youth's conduct in the most recent term: Fighting or O Recognition for good behavior
threatening students; threatening teachers/staff; overly O No problems with school conduct
disruptive behavior; drug/alcohol use; crimes (e.g., theft, O Problems reported by teachers
vandalism); lying, cheating, dishonesty. O Problem calls to parents
O Calls to police
8. Number of expulsions and suspensions in the most recent 0 No expel/suspend 0 2 or 3
term: O 1 expel/suspend O More than
3
9. Youth's attendance in the most recent term: Partial-day O Good attendance; few excused
absence means attending majority of classes and missing absences
minority. Full-day absence means missing majority of classes. O No unexcused absences
Habitual truancy as defined in FS includes 15 unexcused O Some partial-day unexcused
absences in a 90 day period. absences
O Some full-day unexcused absences
O Habitual truant
10. Youth's academic performance in the most recent school O Honor student (mostly As)
term: O Above 3.0 (mostly As and Bs)
O 2.0 to 3.0 (mostly Bs and Cs, no Fs)
O 1.0 to 2.0 (mostly Cs and Ds, some
Fs)
O Below 1.0 (some Ds and mostly Fs)
11. Interviewer's assessment of likelihood the youth will stay O Very likely to stay in school and
in and graduate from high school or an equivalent graduate
vocational school: O Uncertain if youth will stay and
graduate
O Not very likely to stay and graduate












1. History of structured recreational activities 0 Involved in 2 or more structured activities
within the past 5 years: Youth has participated in O Involved in 1 structured activity
structured and supervised pro-social community O Never involved in structured activities
activities, such as religious group/church, community
group, cultural group, club, athletics, or other
community activities.
2. History of unstructured pro-social recreational O Involved in 2 or more pro-social unstructured
activities within the past 5 years: Youth has activities
engaged in activities that positively occupy the O Involved in 1 pro-social unstructured activity
youth's time, such as reading, hobbies, etc. O Never involved in pro-social unstructured
activities


0 For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.
1. Current interest and involvement in structured O Currently involved in 2 or more structured
recreational activities: Youth participates in activities
structured and supervised pro-social community O Currently involved in 1 structured activity
activities, such as religious group/church, community 0 Currently interested but not involved
group, cultural group, club, athletics, or other O Currently not interested in any structured
community activity, activities
2. Types of structured recreational activities in O None
which youth currently participates: (Check all that 0 Community/cultural group
apply.) O Hobby group or club
O Athletics
O Religious group/church
O Volunteer organization
3. Current interest and involvement in pro-social O Currently involved in 2 or more pro-social
unstructured recreational activities: Youth unstructured activities
engages in activities that positively occupy his or her 0 Currently involved in 1 pro-social unstructured
time, such as reading, hobbies, etc. activity
O Currently interested but not involved
O Not interested in any pro-social unstructured
activities





1. History of employment: O Too young for employment consideration
O Never been employed
O Has been employed
2. History of successful employment: O Never successfully employed
O Has been successfully employed
3. History of problems while employed: O Never fired or quit because of problems
O Fired or quit because of poor performance
O Fired or quit because he or she could not get
along with employer or coworkers
4. History of positive personal relationships) with O Never had any positive relationships
past employers) or adult coworker(s): O Had 1 positive relationship
O Had 2 or more positive relationships












O For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.
1. Understanding of what is required to maintain a O Lacks knowledge of what it takes to maintain a
job: job
O Has knowledge of abilities to maintain a job
O Has demonstrated ability to maintain a job
2. Current interest in employment: O Currently employed
O Not employed but highly interested in
employment
O Not employed but somewhat interested
O Not employed and not interested in employment
O Too young for employment consideration
3. Current employment status: O Not currently employed
O Employment is currently going well
O Having problems with current employment
4. Current positive personal relationships) with O Not currently employed
employers) or adult coworker(s): O Employed but no positive relationships
O At least 1 positive relationship





1. History of positive adult non-family O No positive adult relationships
relationships not connected to school or O 1 positive adult relationship
employment: Adults, who are not teachers and 0 2 positive adult relationships
not part of the youth's family, who can provide O 3 or more positive adults relationships
support and model pro-social behavior, such as
religious leader, club member, community
person, etc.
2. History of anti-social friends/companions: O Never had consistent friends or companions
Anti-social peers are youths hostile to or O Had pro-social friends
disruptive of the legal social order; youths who O Had anti-social friends
violate the law and the rights of others and O Been a gang member/associate
other delinquent youth. (Check all that apply.)


0 For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.
1. Current positive adult non-family O No positive adult relationships
relationships not connected to school or O 1 positive adult relationship
employment: Adults, who are not teachers and O 2 positive adult relationships
not part of the youth's family, who can provide O 3 or more positive adults relationships
support and model pro-social behavior, such as
religious leader, club member, community
person, etc.
2. Current pro-social community ties: Youth O No pro-social community ties
feels there are people in his or her community O Some pro-social community ties
who discourage him or her from getting into O Has strong pro-social community ties
trouble or are willing to help the youth.
3. Current friends/companions youth actually O No consistent friends or companions
spends time with: (Check all that apply.) O Pro-social friends
O Anti-social friends
O Gang member/associate











4. Currently in a "romantic," intimate, or O Not romantically involved with anyone
sexual relationship: O Romantically involved with a pro-social person
O Romantically involved with an anti-social
person/criminal
5. Currently admireslemulates anti-social O Does not admire, emulate anti-social peers
peers: O Somewhat admires, emulates anti-social peers
O Admires, emulates anti-social peers
6. Current resistance to anti-social peer O Does not associate with anti-social peers
influence: O Usually resists going along with anti-social peers
O Rarely resists goes along with anti-social peers
O Leads anti-social peers





1. History of court-ordered or DCF voluntary out-of-home O No out-of-home placements exceeding
and shelter care placements exceeding 30 days: 30 days
Exclude DJJ residential commitments. O 1 out-of-home placement
O 2 out-of-home placements
O 3 or more out-of-home placements
2. History of running away or getting kicked out of home: O No history of running away or being
Include times the youth did not voluntarily return within 24 kicked out
hours, and include incidents not reported by or to law O 1 instance of running away/kicked out
enforcement. O 2 to 3 instances of running away/kicked
out
O 4 to 5 instances of running away/kicked
out
O Over 5 instances of running away/kicked
out
3. History of petitions filed: Include all petitions regardless O No petitions filed
of whether the petition was granted. (Check all that apply.) O CINS/FINS
O Dependency
4. History of jaillimprisonment of persons who were ever O No jail/imprisonment history in family
involved in the household for at least 3 months: (Check O Mother/female caretaker
all that apply.) O Father/male caretaker
O Older sibling
O Younger sibling
O Other member
5. Youth living under any "adult supervision." Adult O No, living with peers without adult
supervision must be someone who is responsible for the supervision, do not complete Domain 7B
youth's welfare, either legally or with parental consent. O No, living alone without adult supervision,
do not complete Domain 7B
O No, transient without adult supervision,
do not complete Domain 7B
O Yes, living under adult supervision, must
complete Domain 7B

0 For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.










1. All persons with whom youth is O Living alone O Transient (street)
currently living: (Check all that 0 Biological mother O Biological father
apply.) 0 Non-biological mother O Non-biological father
O Older siblings) O Younger siblings)
O Grandparent(s) O Other relatives)
O Long-term parental partners) O Short-term parental
partners)
O Youth's romantic partner O Youth's child
O Foster/group home O Youth's friends
2. Annual combined income of youth and family: O Under $15,000
O $15,000 to $34,999
O $35,000 to $49,999
O $50,000 and over
3. Jail/imprisonment history of persons who are currently O No jail/imprisonment history in family
involved with the household: (Check all that apply.) O Mother/female caretaker
O Father/male caretaker
O Older sibling
O Younger sibling
O Other member
4. Problem history of parents who are currently involved O No problem history of parents in
with the household: (Check all that apply.) household
O Parental alcohol problem history
O Parental drug problem history
O Parental physical health problem history
O Parental mental health problem history
O Parental employment problem history



5. Problem history of siblings who are currently involved O No siblings currently in household
with the household: (Check all that apply.) O No problem history of siblings in
household
O Sibling alcohol problem history
O Sibling drug problem history
O Sibling physical health problem history
O Sibling mental health problem history
O Sibling employment problem history
6. Support network for family: Extended family and/or O No support network
family friends who can provide additional support to the O Some support network
family. 0 Strong support network
7. Family willingness to help support youth: O Consistently willing to support youth
O Inconsistently willing to support youth
O Little or no willingness to support youth
O Hostile, berating, and/or belittling of youth
8. Family provides opportunities for youth to participate O No opportunities for involvement provided
in family activities and decisions affecting the youth: O Some opportunities for involvement
provided
O Opportunities for involvement provided
9. Youth has run away or been kicked out of home: O Has not run away/kicked out of home
Include times youth did not voluntarily return within 24 O Has run away/kicked out of home
hours, and include incidents not reported by or to law O Is currently kicked out of home or is a
enforcement. runaway











10. Family members) youth feels close to or has good O Does not feel close to any family
relationship with: (Check all that apply.) member
O Feels close to mother/female caretaker
O Feels close to father/male caretaker
O Feels close to male sibling
O Feels close to female sibling
O Feels close to extended family
11. Level of conflict between parents, between youth and 0 Some conflict that is well managed
parents, among siblings: 0 Verbal intimidation, yelling, heated
arguments
0 Threats of physical abuse
O Domestic violence: physical/sexual abuse
12. Parental supervision: Parents know whom youth is with, O Consistent good supervision
when youth will return, where youth is going, and what O Sporadic supervision
youth is doing. O Inadequate supervision
13. Parental authority and control: O Youth usually obeys and follows rules
O Youth sometimes obeys or obeys some
rules
O Youth consistently disobeys and/or is
hostile
14. Consistent appropriate punishment for bad behavior: O Consistently appropriate punishment
Appropriate means clear communication, timely response, O Consistently overly severe punishment
and response proportionate to conduct. O Consistently insufficient punishment
O Inconsistent or erratic punishment
15. Consistent appropriate rewards for good behavior: O Consistently appropriate rewards
Appropriate means clear communication, timely response, O Consistently overly indulgent/overly
and response proportionate to conduct; rewards mean protective
affection, praise, etc. O Consistently insufficient rewards
O Inconsistent or erratic rewards
16. Parental characterization of youth's anti-social O Disapproves of youth's anti-social
behavior: behavior
O Minimizes, denies, justifies, excuses
behavior, or blames
others/circumstances
O Accepts youth's anti-social behavior as
okay
O Proud of youth's anti-social behavior





Disrupted functioning involves having a problem in any of these five life areas: education, family conflict, peer
relationships, crime, or health, and usually indicates treatment is warranted. Use that contributes to criminal
behavior typically precipitates the commission of a crime; there is evidence or reason to believe the youth's
1. History of alcohol use: (Check all that 0 No past alcohol use O Past alcohol use
apply.) O Alcohol caused family conflict
0 Alcohol disrupted education
O Alcohol caused health problems
O Alcohol interfered with keeping pro-social friends
O Alcohol contributed to criminal behavior
O Youth needed increasing amounts of alcohol to achieve
same level of intoxication or high
O_ Youth experienced withdrawal problems











2. Youth's current drug O No current drug use O Current drug use
use: (Check all that 0 Drugs causing family conflict
apply.) O Drugs disrupting education
O Drugs causing health problems
O Drugs interfering with keeping pro-social friends
O Drugs contributing to criminal behavior
O Youth needs increasing amounts of drugs to achieve same level of
intoxication or high
O Youth experiences withdrawal problems
2. Youth's history of drug use: (Check all O No past drug use O Past drug use
that apply.) O Drugs caused family conflict
0 Drugs disrupted education
O Drugs caused health problems
O Drugs interfered with keeping pro-social friends
O Drugs contributed to criminal behavior
O Youth needed increasing amounts of drugs to achieve
same level of intoxication or high
O Youth experienced withdrawal problems
3. Youth's history of referrals for O Never referred for drug/alcohol assessment
alcohol/drug assessment: O Diagnosed as no problem
O Referred but never assessed
O Diagnosed as abuse
O Diagnosed as dependent/addicted
4. History of attending alcohol/drug O Never attended drug/alcohol education classes
education classes for an alcohol/drug O Voluntarily attended drug/alcohol education classes
problem: O Attended classes by parent, school, or other agency
request
O Attended classes at court direction
5. History of participating in alcohol/drug O Never participated in treatment program
treatment program: 0 Participated once in treatment program
O Participated several times in treatment programs
6. Youth currently using alcohol or drugs: O No, do not complete Domain 8B
O Yes, must complete domain 8B

0 For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.

1. Youth's current alcohol use: (Check all O No current alcohol use O Current alcohol
that apply.) use
O Alcohol causing family conflict
D Alcohol disrupting education
O Alcohol causing health problems
O Alcohol interfering with keeping pro-social friends
O Alcohol contributing to criminal behavior
O Youth needs increasing amounts of alcohol to achieve
same level of intoxication or high
O Youth experiences withdrawal problems











3. Type of drugs currently O No current drug use
used: (Check all that 0 Marijuana/hashish
apply.) O Inhalants (glue/gasoline)
O Cocaine (crack/rock)
O Cocaine (coke)
O Amphetamines (Meth/uppers/speed/ecstacy)
O Barbiturates (Tuinal/Seconal/downers)
O Tranquilizers/sedatives (Valium/Libnum/Dalmane/ Ketamine)
O Hallucinogens (LSD/acid/mushrooms/GHB)
O Phencyclidine (PCP/angel dust)
O Heroin
O Other opiates (Dilaudid/Demerol/Percodan/Codeine/ Oxycontin)
4. Current alcohol/drug O Alcohol/drug treatment not warranted
treatment program O Not currently attending needed alcohol/drug treatment program
participation: O Currently attending alcohol/drug treatment program
O Successfully completed alcohol/drug treatment program





History of suicidal ideation: Include any previous O Has never had serious thoughts about suicide
thoughts, threats, plans and attempts even if youth O Has had serious thoughts about suicide
indicates they were manipulative or there was no O Has made a plan to commit suicide. If yes,
intent. (Check all that apply.) describe
O Has attempted to commit suicide. If yes,
describe attempts and dates

O Feels life is not worth living-no hope for
future
O Knows someone well who has committed
suicide, If yes, who, when and how

O Engages in self-mutilating behavior
For abuse and neglect, include suspected incidents of abuse, including those disclosed by youth, whether or
not reported or substantiated, but exclude reports of abuse/neglect investigated but proven to be false.
History of violencelphysical abuse: Include suspected O Not a victim of violence/physical abuse
incidents of abuse, whether or not substantiated, O Victim of violence/physical abuse at home
but exclude reports proven to be false. (Check all Vm of v a i
that apply) O Victim of violence/physical abuse in a
foster/group home
O Victimized or physically abused by family
member
O Victimized or physically abused by someone
outside the family
O Attacked with a weapon
History of witnessing violence: (Check all that apply.) O Has not witnessed violence
Include perpetrators and victims of violence as O Has witnessed violence at home
having witnessed violence.
O Has witnessed violence in a foster/group
home
O Has witnessed violence in the community
O Family member killed as a result of violence
History of sexual abuse/rape: Include suspected O Not a victim of sexual abuse/rape
incidents of abuse if disclosed by youth, whether or O Sexually abused/raped by family member
not reported or substantiated, but exclude reports O Sexually abused/raped by someone outside
investigated but proven to be false. (Check all that the family
apply.) _










History of being a victim of neglect: 0 Not a victim of neglect
O Victim of neglect
History of ADD/ADHD: Confirmed by a professional in O No history of ADD/ADHD
the social service/healthcare field. O Diagnosed with ADD/ADHD
O Only ADD/ADHD medication prescribed
O Only ADD/ADHD treatment prescribed
O ADD/ADHD medication and treatment
prescribed
History of mental health problems: Such as O No history of mental health problems)
schizophrenia, bi-polar, mood, thought, personality, O Past history of mental health problems)
and adjustment disorders. Exclude conduct diagnosis (more than six months ago)
disorder, oppositional defiant disorder, substance O Diagnosed with mental health problems)
abuse, and ADD/ADHD. Confirmed by a 0 Only mental health medication prescribed. If
professional in the social service/healthcare field. yes, list
O Only mental health treatment prescribed
O Mental health medication and treatment
prescribed
History of anger or irritability O No history of anger/irritability
O History of occasional feelings of
anger/irritability
O History of consistent feelings of
anger/irritability
O History of aggressive reactions to feelings of
anger/irritability



1. History of depression or anxiety O No history of depression/anxiety
O History of occasional feelings of
depression/anxiety
O History of consistent feelings of
depression/anxiety
O History of impairment in everyday tasks due to
depression/anxiety
2. History of somatic complaints: Bodily or physical O No history of somatic complaints
discomforts associated with distress, such as O History of one or two somatic complaints
stomachaches or headaches.
O History of three or four somatic complaints
O History of 5 or more somatic complaints
3. History of thought disturbance O No unusual thoughts or beliefs
O Presence of hallucinations (auditory or visual)
O Presence of beliefs that the youth is controlled
by others
4. History of traumatic experience: Lifetime exposure O No presence of traumatic event
to events such as rape, abuse or observed violence, O History of traumatic event
including dreams or flashbacks.
O History of flashbacks to traumatic event

5. Currently has health insurance: O No health insurance
O Public insurance (Medicaid, KIDCARE)
O Private insurance










6. Current mental health problem status: 0 No current mental health problemss, do not
complete Domain 9B
O Current mental health problemss, must
complete Domain 9B


0 For Initial Assessments, current means behaviors during the last 6 months.
0 For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.
1. Current suicidal ideation: O Does not have serious thoughts about suicide
O Has serious thoughts about suicide
O Has recently made a plan to commit suicide. If
yes, describe
O Has recently attempted to commit suicide. If
yes, describe attempts and dates

O Feels life is not worth living-no hope for
future
O Knows someone well who has committed
suicide. If yes, who, when and how

O Engages in self-mutilating behavior
2. Currently diagnosed with ADD/ADHD: Confirmed O No ADD/ADHD diagnosis
by a professional in the social service/healthcare field. 0 No ADD/ADHD medication currently
prescribed
Type of medication: prescribed
O Currently taking ADD/ADHD medication
O ADD/ADHD medication currently prescribed,
but not taking
3. Mental health treatment currently prescribed O No current mental health problem
excluding ADD/ADHD treatment: O No mental health treatment currently
prescribed
O Attending mental health treatment
O Treatment currently prescribed, but not
attending
4. Mental health medication currently prescribed O No current mental health problem
excluding ADD/ADHD medication: O No mental health medication currently
prescribed
Type of medication: prescribed
O Currently taking mental health medication
O Mental health medication currently prescribed,
but not taking
5. Mental health problems currently interfere in O No current mental health problem
working with the youth: O Mental health problems) do not interfere in
work with youth
O Mental health problems) interfere in work with
youth





0 For Initial Assessments, current means behaviors during the last 6 months.
0O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.










1. Primary emotion when O Nervous, afraid, worried, ambivalent, uncertain, or
committing crimess: indecisive
O Hyper, excited, or stimulated
O Unconcerned or indifferent
O Confident or brags about not getting caught
2. Primary purpose for committing O Anger/ Revenge
crimes) within the last 6 O Impulse
months: 0 Sexual desire
O Money or material gain, including drugs
O Excitement, amusement, or fun
O Peer status, acceptance, or attention
3. Optimism: Youth talks about future 0 High aspirations: sense of purpose, commitment to better
in positive way with plans or life
aspirations of a better life that could 0 Normal aspirations: some sense of purpose
include employment, education, O Low aspirations: little sense of purpose or plans for better
raising a family, travel, or other pro- life
social life goals. O Believes nothing matters; he or she will be dead before long
4. Impulsive; acts before thinking: O Uses self-control; usually thinks before acting
O Some self-control; sometimes thinks before acting
O Impulsive; often acts before thinking
O Highly Impulsive; usually acts before thinking
5. Belief in control over anti-social O Believes he or she can avoid/stop anti-social behavior
behavior: O Somewhat believes anti-social behavior is controllable
O Believes his or her anti-social behavior is out of his or her
control
6. Empathy, remorse, sympathy, or O Has empathy for his or her victims)
feelings for the victims) of O Has some empathy for his or her victims)
criminal behavior: O Does not have empathy for his or her victims)
7. Respect for property of others: O Respects property of others
O Respects personal property but not publicly accessible
property: "It's not hurting anybody."
O Conditional respect for personal property: "If they are stupid
enough to leave it out, they deserve losing it."
O No respect for property: "If I want something, it should be
mine."
8. Respect for authority figures: O Respects most authority figures
O Does not respect authority figures, and may resent some
O Resents most authority figures
O Defies or is hostile toward most authority figures
9. Attitude toward responsible law O Believes pro-social rules/conventions apply to him or her
abiding behavior: O Believes some pro-social rules/conventions sometimes apply to
him or her
O Does not believe pro-social rules/conventions apply to him
or her
O Resents or is defiant toward pro-social rules/conventions
10. Accepts responsibility for anti- O Accepts responsibility for anti-social behavior
social behavior: O Minimizes, denies, justifies, excuses, or blames others
O Accepts anti-social behavior as okay
O Proud of anti-social behavior
11. Youth's belief in successfully O Believes he or she will be successful
meeting conditions of court O Unsure if he or she will be successful
supervision: O Does not believe he or she will be successful















Items 1. through 4.:
O For Initial Assessments, rate items based on behavior during the last 6 months.
0 For Re-assessments and Final Assessments, rate items based on behavior during the last 4 weeks.
1. Tolerance for frustration: O Rarely gets upset over small things or has temper
tantrums
O Sometimes gets upset over small things or has
temper tantrums
O Often gets upset over small things or has temper
tantrums
2. Hostile interpretation of actions and O Primarily positive view of intentions of others
intentions of others in a common non- O Primarily negative view of intentions of others
confrontational setting: O Primarily hostile view of intentions of others
3. Belief in yelling and verbal aggression to O Believes verbal aggression is rarely appropriate
resolve a disagreement or conflict: O Believes verbal aggression is sometimes appropriate
O Believes verbal aggression is often appropriate
4. Belief in fighting and physical aggression to 0 Believes physical aggression is never appropriate
resolve a disagreement or conflict: O Believes physical aggression is rarely appropriate
O Believes physical aggression is sometimes
appropriate
O Believes physical aggression is often appropriate
Items 5. and 6.:
0 For Initial Assessments, include the entire history of report.
0 For Re-assessments and Final Assessments, include reports within the last 4 weeks.
5. Reportslevidence of violence not included O No reports/evidence of violence
in criminal history: (Check all that apply.) O Violent outbursts, displays of temper, uncontrolled
anger indicating potential for harm
O Deliberately inflicting physical pain
O Using/threatening with a weapon
O Fire starting
O Violent destruction of property
O Animal cruelty
6. Reports of problem with sexual aggression O No reports/evidence of sexual aggression
not included in criminal history: (Check all O Aggressive sex
that apply.) O Sex for power
O Young sex partners
O Child sex
O Voyeurism
O Exposure





O For Initial Assessments, current means behaviors during the last 6 months.
O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.
O Use a general pattern of current behaviors and not a single incident.











1. Consequential thinking: O Does not understand there are consequences to
actions
O Understands there are consequences to actions
O Identifies consequences of actions
O Acts to obtain desired consequences-good
consequential thinking
2. Goal setting: O Does not set goals
O Sets unrealistic goals
O Sets somewhat realistic goals
O Sets realistic goals
3. Problem-solving: O Cannot identify problem behaviors
O Identifies problem behaviors
O Thinks of solutions for problem behaviors
O Applies appropriate solutions to problem behaviors
4. Situational perception: Ability to analyze O Cannot analyze the situation for use of a pro-social
the situation, choose the best pro-social skill, and skill
select the best time and place to use the pro- O Can analyze but not choose the best pro-social
social skill. skill
O Can choose the best skill but cannot select the
best time and place
O Can select the best time and place to use the best
pro-social skill
5. Dealing with others: Basic social skills O Lacks basic social skills in dealing with others
include listening, starting a conversation, having O Has basic social skills, lacks advanced skills in
a conversation, asking a question, saying thank dealing with others
you, introducing yourself, introducing other O Sometimes uses advanced social skills in dealing
people, and giving a compliment. Advanced with others
social skills include asking for help, joining in, O Often uses advanced social skills in dealing with
giving instructions, following instructions, others
apologizing, and convincing others.
6. Dealing with difficult situations: Incl. making a 0 Lacks skills in dealing with difficult situations
complaint, answering a complaint, dealing with O Rarely uses skills in dealing with difficult situations
embarrassment, dealing with being left out, standing O Sometimes uses skills in dealing with difficult
up for a friend, responding to frustration, responding situations
to failure, dealing with contradictory messages, O Often uses skills in dealing with difficult situations
dealing with accusation, getting ready for a difficult
conversation, and dealing with group pressure.
7. Dealing with feelingslemotions: Includes O Lacks skills in dealing with feelings/emotions
knowing his or her feelings, expressing feelings, O Rarely uses skills in dealing with feelings/emotions
understanding the feelings of others, dealing with 0 Sometimes uses skills in dealing with
someone else's anger, expressing affection, feelings/emotions
dealing with fear, and rewarding oneself O Often uses skills in dealing with feelings/emotions
8. Monitoring of internal triggers, distorted O Cannot identify internal triggers
thoughts, that can lead to trouble: O Identifies internal triggers
O Actively monitors/controls internal triggers
9. Monitoring of external triggers, events or 0 Cannot identify external triggers
situations, that can lead to trouble: O Identifies external triggers
O Actively monitors/controls external triggers
10. Control of impulsive behaviors that get O Never had a problem with impulsive behavior
youth into trouble: Reframing, replacing anti- O Does not know techniques to control impulsive
social thoughts with pro-social thoughts, behavior
diversion, relaxation, problem solving, O Knows techniques to control impulsive behavior
negotiation, relapse prevention. O Uses techniques to control impulsive behavior











11. Control of aggression: Includes asking 0 Never had a problem with aggression
permission, sharing thoughts, helping others, 0 Lacks alternatives to aggression
negotiating, using self control, standing up for 0 Rarely uses alternatives to aggression
one's rights, responding to teasing, avoiding 0 Sometimes uses alternatives to aggression
trouble with others, and keeping out of fights. 0 Often uses alternatives to aggression









REFERENCES


Akers, Ronald L. (1998). Social Learning and Social Structure: A General Theory of Crime and
Deviance. Boston: Northeastern University Press.

Allen, Francis A. (1981). The Decline of the Rehabilitative Ideal: Penal Policy and Social
Purpose. New Haven, CN: Yale University Press.

Andrews, Don A. (2001). Principles of effective correctional programs. In Laurence L. Motiuk
and Ralph C. Serin (eds.), Compendium 2000 on Effective Correctional Programming.
Ottawa, Ont., Canada: Correctional Service Canada.

Andrews, Don A. (2006). Enhancing adherence to risk-need-responsivity: Making quality a
matter of policy. Criminology andPublic Policy, 5:595-602.

Andrews, Don A., & Bonta, James. (2003). The Psychology of Criminal Conduct, 3rd edition.
Cincinnati, OH: Anderson Publishing Co.

Andrews, Don A., Bonta, James, & Hoge, Robert D. (1990a). Classification for effective
rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17:19-52.

Andrews, Don A, Bonta, James, & Wormith, J. Stephen. (2006). The recent past and near future
of risk and/or need assessment. Crime & Delinquency, 52:7-27.

Andrews, Don A., Dowden, Craig, & Gendreau, Paul. (1999). Clinically Relevant and
Psychologically Informed Approaches to Reduced Re-offending: A Meta-analytic Study
of Human Service, Risk, Need, Responsivity and Other Concerns in Justice Contexts.
Carleton University, Ottawa, Canada.

Andrews, Don A., & Kiessling, Jerry J. (1980). Program structure and effective correctional
practices: A summary of the CaVIC research. In R.R. Ross & P. Gendreau (eds.),
Effective Correctional Treatment. Toronto, Canada: Butterworth.

Andrews, Don A., Zinger, Ivan, Hoge, Robert D., Bonta, James, Gendreau, Paul, and Cullen,
Francis T. (1990b). Does correctional treatment work? A clinically-relevant and
psychologically informed meta-analysis. Criminology, 28:369-404.

Andrews, Don A., Zinger, Ivan, Bonta, James, Gendreau, Paul, and Cullen, Francis T. (1990c).
Does correctional treatment work? A psychologically informed meta-analysis.
Criminology, 28:369-404.

Antonowicz, D.H., & Ross, R.R. (1994). Essential components of successful rehabilitation
programs for offenders. International Journal of Offender and Comparative Criminology,
38:97-104.

Aos, Steve. (1998). Watching the Bottom Line: Cost-Effective Interventions for Reducing Crime
in Washington. Olympia, WA: Washington State Institute for Public Policy.










Auerhahn, Kathleen. (1999). Selective incapacitation and the problem of prediction.
Criminology, 37(4):703-735.

Auerhahn, Kathleen. (2006). Conceptual and methodological issues in the prediction of
dangerous behavior. Criminology & Public Policy, 5(4):771-778.

Baird, S.C., Heinz, R.C., & Bemus, B.J. (1979). Project Report #14: A Two-year Follow-up.
Madison, WI: Department of Health and Social Services, Case Classification/ Staff
Deployment Project, Bureau of Community Corrections.

Bamoski, Robert. (2004). Outcome Evaluation of Washington State's Research-Based Programs
for Juvenile Offenders. Olympia, WA: Washington State Institute for Public Policy.

Belkin, Jacob, Blumstein, Alfred, and Glass, William. (1972). Recidivism as a feedback process:
An analytical model and empirical validation. Journal of Criminal Justice, 1:7.

Belknap, J. (1996). The Invisible Woman: Gender, Crime, and Justice. Cincinnati: Wadsworth.

Benekos, Peter J. & Merlo, Alida V. (1995). Three strikes and you're out! The political
sentencing game. Federal Probation, 59:3-9.

Blackmore, John & Welsh, Jane. (1983). Selective incapacitation: Sentencing according to risk.
Crime & Delinquency, 29:504-529.

Blumstein, Alfred, Cohen, Jacqueline, and Nagin, Daniel, (eds.), (1978). Deterrence and
Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates.
Washington, D.C.: National Academy of Sciences.

Blumstein, Alfred, Cohen, Jacqueline, Roth, Jeffrey A., and Visher, Christy. (1986). Criminal
Careers and "Career Criminals", Vol.1. Washington, D.C.: National Academy Press.

Blumstein, Alfred, Cohen, Jacqueline, Das, Somnath, and Moitra, Soumyo D. (1988).
Specialization and seriousness during adult criminal careers. Journal of Quantitative
Criminology, 4:303-345.

Broidy, Lisa, & Agnew, Robert. (1997). Gender and crime: A General Strain Theory perspective.
Journal ofResearch in Crime andDelinquency, 34:275-306.

Bonta, James. (1995). The responsivity principle and offender rehabilitation. Forum on
Corrections Research, 7:34-37.

Bonta, James, Law, Moira, & Hanson, Karl. (1998). The prediction of criminal and violent
recidivism among mentally disordered offenders: A meta-analysis. Psychological
Bulletin, 123:123-142.









Bonta, James, Wallace-Capretta, Suzanne, & Rooney, Jennifer. (1998). Restorative Justice: An
Evaluation of the Restorative Resolutions Project (User Report 1998-05). Ottawa,
Ontario: Solicitor General Canada.

Bonta, James, Wallace-Capretta, Suzanne, & Rooney, Jennifer. (2000). A quasi-experimental
evaluation of an intensive rehabilitation supervision program. Criminal Justice and
Behavior, 27:312-329.

Brown, Shelley L. (2003). The dynamic prediction of criminal recidivism: A three-wave
prospective study 1995-2002. Ph.D. dissertation, Queen's University, Kingston, Ontario,
Canada.

Brown, Brenner, & Campbell, Robin. (2005). S. iumiing the Path from Prison to Home: A
Roundtable Discussion on the Lessons of Project Greenlight. New York: Vera Institute of
Justice. Available at http://www.vera.org.

Bureau of Justice Statistics. (1992). National Corrections Reporting Program 1988. Washington,
D.C.: U.S. Department of Justice.

Burrell, William D. (2000). Reinventing probation: Organizational culture and change.
Community Corrections Report, 7:49-64.

Canela-Cacho, Jose A., Blumstein, Alfred, and Cohen, Jacqueline. (1997). Relationship between
the offending frequency (A) of imprisoned and free offenders. Criminology, 35:133-175.

Carey, Mark. (2002). Social learning, social capital and correctional theories: Seeking an
integrated model. Paper presented at International Community Corrections Association
conference, November, 2002.

Chaiken, Jan M., & Chaiken, Marcia R. (1982). Varieties of Criminal Behavior. Santa Monica,
CA: Rand.

Chaiken, Marcia R., & Chaiken, Jan M. (1984). Offender types and public policy. Crime&
Delinquency, 30:195-226.

Chesney-Lind, M. (1989). Girls' crime and a woman's place: Toward a feminist model of female
delinquency. Crime & Delinquency, 35:5-29.

Chesney-Lind, M. (2001). What about the girls? Delinquency programming as if gender
mattered. Corrections Today, 63:38.

Chesney-Lind, M., & Rodriguez, N. (1983). Women under lock and key. Prison Journal, 63:47-
65.

Chesney-Lind, M., & Sheldon, R.G. (1992). Girls, Delinquency and Juvenile Justice. Pacific
Grove, CA: Brooks Cole.










Clark, John, Austin, James, and Henry, D. Alan (1997). Three Strikes and You're Out: A Review
of State Legislation. Washington, D.C.: National Institute of Justice. Available at
http://www.ncjrs.org.

Clear, Todd R., and Barry, Donald M. (1983). Some conceptual issues in incapacitating
offenders. Crime and Delinquency, 29:529-545.

Cohen, Jacqueline. (1983). Incapacitation as a strategy for crime control: Possibilities and
pitfalls. In Michael Tonry & Norval Morris (eds.), Crime and Justice: An Annual Review
ofResearch, vol.2. Chicago: University of Chicago Press.

Cohen, Jacqueline. (1984). Incapacitation. University ofIllinois Law Review, 2:253-290.

Cohen, Stanley. (1985). Visions of Social Control: Crime, Punishment and Classification.
Oxford: Polity Press.

Copas, John. (1983). Some statistical questions in the prediction of dangerous offending. In
Hinton, John, W. (ed.), Dangerousness: Problems of Assessment and Prediction. London:
Allen & Unwin.

Corbett, Ronald P., Beto, Dan R., Coen, Brian, Dilulio, John J., Faulkner, J. Richard, Fitzgerald,
Bernard L., Gregg, Irwin, Helber, Norman, Hinzman, Gerald R., Malvestuto, Robert,
Paparozzi, Mario, Perry, John, Pozzi, Rocco, & Rhine, Edward E. (1999). "Broken
Windows" Probation: The Next Step in Fighting Crime. Center for Civic Innovation,
Manhattan Institute, New York.

Covington, S.S. (2001). Creating gender-responsive programs: The next step for women's
services. Corrections Today, 61:85-87.

Covington, S.S. (2003). A woman's journey home: Challenges for female offenders. In J. Travis
and M. Waul (Eds.), Prisoners Once Removed: The Impact ofIncarceration and Reentry
on Children, Families, and Communities. (pp.67-103). Washington, D.C.: The Urban
Institute.

Cullen, Francis T. (2002). Rehabilitation and treatment programs. In James Q. Wilson and Joan
Petersilia (eds.), Crime andPublic Policy, 2nd Edition. San Francisco, CA: ICS Press.

Cullen, Francis T., & Gendreau, Paul. (1989). The effectiveness of correctional rehabilitation:
Reconsidering the "nothing works" debate. In L. Goodstein and D. MacKenzie (eds.),
American Prisons: Issues in Research and Policy. New York: Plenum.

Cullen, Francis T., & Gilbert, Karen E. (1982). Reaffirming Rehabilitation. Cincinnati, OH:
Anderson.

Currie, Elliot. (1998). Crime and Punishment in America. New York, NY: Metropolitan Books.










Daly, Kathleen. (1992). Women's pathways to felony court: Feminist theories of lawbreaking
and problems of representation. S.nlihel n Review ofLaw and Women's Studies, 2:11-52.

Daly, Kathleen. (1994). Gender, Crime, and Punishment. New Haven, CT: Yale University
Press.

Daly, Kathleen., & Chesney-Lind, M. (1988). Feminism and criminology. Justice Quarterly,
5:497-538.

Dawes, Robyn M., Faust, David, & Meehl, Paul E. (1993). Statistical prediction versus clinical
prediction: Improving what works. In G. Karen & C. Lewis (eds.), A Handbookfor Data
Analysis in the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.

DeLisi, Matt. (2005). Career Criminals in Society. Thousand Oaks, CA: Sage Publications.

Dershowitz, Alan M. (1973). Preventive confinement: A suggested framework for Constitutional
analysis. Texas Law Review, 51:1277-1324.

Dowden, C., & Andrews, Don A. (1999). What works for female offenders: A meta-analytic
review. Crime& Delinquency, 45:438-452.

Fabiano, Elizabeth A., Porporino, Frank J., & Robinson, David. (1991). Canada's cognitive skills
program corrects offenders' faulty thinking. Corrections Today, 53:102-108.

Farrington, David P. (1986).Age and crime. In M. Tonry & N. Morris (eds.), Crime and
Justice: An Annual Review ofResearch, vol. 7:189-250). Chicago, IL:
University of Chicago Press.

Farrington, David P. (1987). Predicting individual crime rates. In Don Gottfredson and Michael
Tonry, (eds.), Prediction and Classification: Criminal Decision Making. Chicago:
University of Chicago Press.

Farrington, David P., Snyder, Howard N., and Finnegan, Terrence A. (1988). Specialization in
juvenile court careers. Criminology, 26:461-485.

Feeley, Malcolm M. & Simon, Jonathan. (1992). The new penology: Notes on the Emerging
strategy of corrections and its implications. Criminology, 30:449-474.

Flanagan, Timothy J. (1996). Reform or punish: Americans' views of the correctional system. In
Timothy J. Flanagan and Dennis R. Longmire (eds.), Americans View Crime and Justice:
A National Public Opinion Survey. Thousand Oaks, CA: Sage.

Foley, M. (1991). Risk assessment: The development of an instrument to predict "failure on
probation" for juveniles. Unpublished doctoral dissertation, Fordham University, New
York.










Funk, Stephanie J. (1999). Risk assessment for juveniles on probation: A focus on gender.
Criminal Justice and Behavior, 26:44-68.

Garland, David & Young, Peter. (eds.). (1983). The Power to Punish: Contemporary Penalty
and Social Analysis. London: Heinemann.

Gendreau, Paul. (1996). The principles of effective interventions with offenders. In Alan T.
Harland (ed.), Choosing Correctional Options That Work: Defining the Demand and
Evaluating the Supply. Thousand Oaks, CA: Sage.

Gendreau, Paul, Andrews, Don A., Goggin, Claire, & Chanteloupe, F. (1992). The development
of clinical and policy guidelines for the prediction of criminal behavior in criminal justice
settings. Unpublished manuscript available from the Department of Psychology,
University of New Brunswick, St. John, New Brunswick.

Gendreau, Paul, Cullen, Francis T., Bonta, James. (1994). The Effects of Prison Sentences on
Recidivism. Ottawa, Ontario: Solicitor General Canada.

Gendreau, Paul, Little, T., & Goggin, C. (1996). A meta-analysis of the predictors of adult
offender recidivism: What works! Criminology, 34:575-607.

Gendreau, Paul, & Ross, Robert R. (1979). Effective correctional treatment: Bibliotherapy for
cynics. Crime & Delinquency, 25:463-489.

Gendreau, Paul, & Ross, Robert R. (1981). Correctional potency: Treatment and deterrence on
trial. In R. Roesch & R.R. Corrado (eds.), Evaluation and Criminal Justice Policy.
Beverly Hills, CA: Sage.

Gendreau, Paul, & Ross, Robert R. (1987). Revivication of rehabilitation: Evidence from the
1980s. Justice Quarterly, 4:349-408.

Glueck, Sheldon, & Glueck, Eleanor. (1950). Unraveling Juvenile Delinquency. Cambridge,
MA: Harvard University Press.

Goggin, Claire, & Gendreau, Paul. (2006). The implementation and maintenance of quality
services in offender rehabilitation programs. In Clive R. Hollin and Emma J. Palmer
(eds.), Offending Behavior Programs: Development, Application, and Controversies.
Chichester, U.K.: Wiley. In Press.

Goldfarb, Ronald, & Singer, Linda. (1973). After Conviction. New York: Simon and Schuster.

Gordon, Diana, R. (1991). The Power to Punish: Contemporary Penality and Social Analysis.
London: Heinemann.

Gordon, T. (1970). Parent Effectiveness Training. New York, NY: Wyden.










Gottfredson, Don M. (1967). Assessment and prediction methods in crime and delinquency. In
Task Force Report: Juvenile Delinquency and Youth Crime. Washington, D.C.: U.S.
Department of Justice. 171-187.

Gottfredson, Don M., & Snyder, Howard N. (2005). The AI tulwtiiL of Risk Classification:
Changing Data into Valid Instruments for Juvenile Courts. Office of Juvenile Justice and
Delinquency Prevention, National Center for Juvenile Justice, U.S. Department of
Justice.

Gottfredson, Don M., & Tonry, Michael. (1987). Prediction and classification: Criminal justice
decision making. In Crime and Justice: A Review ofResearch, vol.9. Chicago: University
of Chicago Press.

Gottfredson, Michael R. (1979). Parole guidelines and the reduction of sentencing disparity: A
preliminary study. Journal ofResearch in Crime & Delinquency, 16(1):39-54.

Gottfredson, Michael R., & Hirschi, Travis. (1986). The true value of lambda would appear to be
zero: An essay on career criminals, criminal careers, selective incapacitation, cohort
studies, and related topics. Criminology, 24:213-234.

Gottfredson, Michael R., & Hirschi, Travis. (1990). A General Theory of Crime. Stanford, CA:
Stanford University Press.

Gottfredson, Stephen D., & Gottfredson, Don M. (1979). Screening for risk: A comparison of
methods. Washington D.C.: U.S. Department of Justice.

Gottfredson, Stephen D., & Gottfredson, Don M. (1985a). Selective incapacitation? Annals of the
American Academy of Political and Social Sciences, 478:135-149.

Gottfredson, Stephen D., & Gottfredson, Don M. (1985b). Screening for risk among Parolees:
Policy, practice, and method. In David P. Farrington and Roger Tarling (eds.), Prediction
in Criminology. Albany, NY: SUNY Press.

Greenberg, David. (1975). The incapacitative effect of imprisonment: Some estimates. Law &
Society Review, 9:541-580.

Greenwood, Peter W. (1998). Investing in prisons or prevention: The state policy makers'
dilemma (Special Issue: NCCD 90th Anniversary). Crime and Delinquency, 44:136-142.

Greenwood, Peter W., & Abrahamse, Allan. (1982). Selective Incapacitation. Santa Monica, CA:
Rand.

Greenwood, Peter W., & Turner, Susan. (1987). Selective Incapacitation Revisited: Why the
High-Rate Offenders are Hard to Predict. Santa Monica, CA: Rand.









Grove, William M., Zald, David H., Lebow, Boyd S., Snitz, Beth E., & Nelson, Chad. (2000).
Clinical versus mechanical prediction: A meta-analysis. Psychology Assessment, 12:19-
30.

Halleck, Seymour L., & Witte, Ann D. (1977). Is rehabilitation dead? Crime & Delinquency,
23(4):372-383.

Hanson, R. Karl, & Bussiere, Monique T. (1998). Predicting relapse: A meta-analysis of sexual
offender recidivism studies. Journal of Consulting and Clinical Psychology, 66:348-362.

Hanson, R. Karl, & Harris, A.J.R. (2000). Where should we intervene? Dynamic predictors of
sexual offense recidivism. CriminalJustice and Behavior, 27:6-35.

Hanson, R. Karl, & Morton-Bourgon, K. (2004). Predictors of Sexual Recidivism: An Updated
Meta-Analysis (User report 2004-02). Public Safety and Emergency Preparedness,
Ottawa, Canada.

Harland, Alan T. (1996). Choosing Correctional Options that Work: Defining the Demand and
Evaluating the Supply. Thousand Oaks, CA: Sage.

Hoge, Robert D. (2001). The Juvenile Offender: Theory, Research, andApplication. Norwell,
MA: Kluwer.

Holtfreter, Kristy, & Morash, Merry. (2003). The needs of women offenders: Implications for
correctional programming. Women and Criminal Justice, 14:137-160.

Hubbard, Dana J. (2002). Cognitive-behavioral treatment: An analysis of gender and other
responsivity characteristics and their effects on success in offender rehabilitation. Ph.D
dissertation, University of Cincinnati, Cincinnati, Ohio.

Hubbard, Dana J., & Pratt, Travis C. (2002). A meta-analysis of the predictors of delinquency
among girls. Journal of Offender Rehabilitation, 34:1-13.

Hughes, John. (1971). Crime: Its causes, control, and effect on society. In 1971 Report of the
New York State Joint Legislative Committee. State of New York Legislation Document
Vol. 26.

Izzo, Rhena L. & Ross, Robert R. (1990). Meta-analysis of rehabilitation programs or juvenile
delinquents: A brief report. Criminal Justice and Behavior, 17:134-142.

Kempf-Leonard, Kimbery, (1987). Specialization and the criminal career. Criminology, 25:399-
420.

Lanza-Kaduce, Lonn, Gartin, P., Dundes, L., Blount, W.R., Boulinois-Manning, S., Kuhns III,
J.B., Holten, N.G., Mahan, S., Taylor, D.W., Terry III, W.C., Frank, H., and Caouette, B.
(1990). Prison utilization study: Risk assessment techniques and Florida's inmates. Vol.









I: Application to Males and Females. Division of Economic and Demographic Research
of the Joint Legislative Management Committee, Florida Legislature, Tallahassee, FL.

Latessa, Edward, Cullen, Francis, & Gendreau, Paul. (2002). Beyond correctional quakery:
Professionalism and the possibility of professional treatment. Federal Probation, 66:43-
49.

Law, M. (2004). A longitudinal follow-up of federally sentenced women in the community:
Assessing the predictive validity of the dynamic characteristics of the Community
Intervention Scale. Ph.D. dissertation, Carleton University, Ottawa, Canada.

Lewis, Neal A. (1994). President Foresees Safer U.S.: Others See Crime Bill as Symbolic. The
New York Times, August 27:6.

Link, Bruce G., Andrews, Howard, & Cullen, Francis T. (1992). The violent and illegal behavior
of mental patients reconsidered. American Sociological Review, 57:275-292.

Lipsey, Mark W. (1989). The efficacy of intervention for juvenile delinquency: Results from 400
studies. Paper presented at the 41st annual meeting of the American Society of
Criminology, Reno, NV.

Lipsey, Mark W. (1992). Juvenile delinquency treatment: A meta-analytic inquiry into the
variability of effects. In T.D. Cook et al. (eds.), Meta-Analysisfor Exploration: A
Casebook. New York, NY: Russell Sage Foundation.

Lipsey, Mark W., & Derzon, James H. (1998). Predictors of violent or serious delinquency in
adolescence and early childhood: A synthesis of longitudinal research. In R. Loeber & D.
Farrington (eds.), Serious and Violent Juvenile Offenders: Risk Factors and Successful
Interventions. Thousand Oaks, CA: Sage.

Lipsey, Mark W., & Wilson, David B. (1993). The efficacy of psychological, educational, and
behavioral treatment. American Psychologist, 48:1181-1209.

Lipton, Douglas, Martinson, Robert, and Wilks, Judith. (1975). The Effectiveness of Correctional
Treatment: A Survey of Treatment Evaluation Studies. New York: Praeger.

Lowenkamp, Christoper T., Latessa, Edward J., & Smith, Paula. (2006). Does correctional
program quality really matter? The impact of adhering to the principles of effective
intervention. Criminology and Public Policy, 5:575-594.

Mannheim, Hermann, & Wilkins, Leslie T. (1955). Prediction Methods in Relation to Borstal
Training. London: Her Majesty's Stationery Office.

Martinson, Robert. (1974). What works? Questions and answers about prison reform. The Public
Interest, 35:22-54.









McGuire, James. (2002). Evidence-based programming today. Paper presented at the
International Community Corrections Association conference, Boston, MA, November,
2002.

Mednick, Sarnoff, & Christiansen, Karl O. (1977). BiosocialBases of Criminal Behavior. New
York: Gardner.

Meehl, Paul E., & Grove, William M. (1996). Comparative efficiency of informal (subjective,
impressionistic) and formal (mechanical, algorithmic) prediction procedures: The
clinical-statistical controversy. Psychology, Public Policy, and Law, 2:293-323.

Messinger, Sheldon. (1969). Strategies of control. Ph.D. dissertation, Department of Sociology,
University of California at Los Angeles.

Messinger, Sheldon, & Berecochea, John. (1990). Don't Stay Too Long but do Come Back Soon.
Proceedings, Conference on Growth and Its Influence on Correctional Policy, Center for
the Study of Law and Society, University of California at Berkley.

Miller, William R., & Rollnick, Steven. (2002). Motivational Interviewing: Preparing People for
Change. New York, NY: Guilford Press.

Mitford, Jessica. (1973). Kind and Usual Punishment. New York: Alfred A. Knopf.

Moffitt, T.E. (1993). Adolescent-limited and life-course persistent antisocial behavior: A
developmental taxonomy. Psychological Review, 100(4), 674-701.

Monahan, John. (1981). Predicting Violent Behavior: An Assessment of Clinical Techniques.
Beverly Hills, CA: Sage.

Moore, Mark H., Estrich, Susan R., McGillis, Daniel, and Spelman, William. (1984). Dangerous
Offenders: The Elusive Target of Justice. Cambridge, MA: Harvard University Press.

Mossman, Douglas. (1994). Assessing predictions of violence: Being accurate about accuracy.
Journal of Consulting and Clinical Psychology, 62:783-792.

Nagin, Danniel S. (1998). Deterrence and incapacitation. In Michael Tonry (ed.), The Oxford
Handbook of Crime and Punishment. New York: Oxford University Press.

Nagin, Danniel S. & Land, Kenneth C. (1993). Age, criminal career, and population
heterogeneity: Specification and estimation of a nonparametric mixed poisson
model. Criminology, 31:327.

National Institute of Corrections. (2004). Implementing Evidence-Based Practice in Community
Corrections: The Principles of Effective Intervention. Washington, D.C.: National
Institute of Corrections.









O'Donnell, C.R., Lydgate, T., & Fo, W.S.O (1971). The buddy system: Review and follow-up.
Child Behavior Therapy, 1:161-169.

Palmer, Ted. (1975). Martinson revisited. Journal ofResearch in Crime andDelinquency,
12:133-152.

Palmer, Ted. (1983). The effectiveness issue today: An overview. Federal Probation, 46:3-10.

Palmer, Ted. (1992). The Re-Emergence of Correctional Intervention. Newbury Park, CA: Sage.

Petersilia, Joan. (1980). Criminal career research: A review of recent evidence. In Norval Morris
and Michael Tonry (eds.), Crime and Justice: A Review ofResearch, vol.2. Chicago:
University of Chicago Press.

Piquero, Alex R., Farrington, David P., and Blumstein, Alfred. (2003). The criminal career
paradigm. In Michael Tonry (ed.), Crime and Justice: A Review ofResearch, vol.30.
Chicago: University of Chicago Press.

Pratt, Travis C., Maahs, J., and Stehr. (1998). The symbolic ownership of the corrections
'problem:' A framework for understanding the development of corrections policy in the
United States. Prison Journal, 78:451-464.

Quinsey, Vernon L., Coleman, Garth, Jones, Brian, & Altrows, Irwin F. (1997). Proximal
antecedents of eloping and re-offending among supervised mentally disordered offenders.
Journal ofInterpersonal Violence, 12:794-813.

Reichman, Nancy. (1986). Managing crime risks: Toward an insurance-based model of social
control. Research in Law, Deviance and Social Control, 8:151-172

Reisig, Michael D., Holtfreter, Kristy, and Morash, Merry. (2006). Assessing recidivism risk
across female pathways to crime. Justice Quarterly, 23:384-405.

Ritzer, George. (1993). The McDonaldization of Society. Newbury Park, CA: Pine Forge.

Shannon, Lyle W. (1991). Changing Patterns of Delinquency and Crime: A Longitudinal Study
in Racine. Boulder, CO: Westview Press.

Shannon, Lyle W., McKim, Judith L, Curry, James P., and Haffner, Lawrence J. (1988).
Criminal Career Continuity: Its Social Context. New York: Human Science Press.

Sherman, Lawrence W., Gottfredson, Denise C., MacKenzie, Doris L., Eck, John, Reuter, Peter,
& Bushway, Shawn D. (1997). Preventing Crime: What Works, What Doesn't, What's
Promising. Washington, DC: National Institute of Justice.

Shichor, David. (1997). Three strikes as a public policy: The convergence of the new penology
and the McDonaldization of punishment. Crime andDelinquency, 43:470-493.










Shinar, Shlomo & Shinnar, Reuel. (1975). The effects of the criminal justice system on the
control of crime: A quantitative approach. Law and Society Review, 9(4):581-610.

Simon, Frances H. (1971). Prediction Methods in Criminology. London, Her Majesty's
Stationery Office.

Simon, Leonore M.J. (1997). Do criminal offenders specialize in crime types? Applied &
Preventive Psychology, 6:35-53.

Simourd, L., & Andrews, Don A. (1994). Correlates of delinquency: A look at gender
differences. Forum on Corrections Research, 6:26-31.

Simpson, S. (1989). Feminist theory, crime, and justice. Criminology, 27:605-631.

Slansky, David A. (1995). Cocaine, race, and equal protection. Stanford Law Review, 47:1283-
1322.

Snyder, H.N. (2003). Juvenile Arrests 2000. Washington, D.C.: Office of Juvenile Justice and
Delinquency Prevention. Retrieved from
http://ojjdp.ncjrs.org/ojstatbb/asp/JAR Display.asp?ID=qa2300031502 (September 19,
2003).

Snyder, H.N., & Sickmund, M. (1999). Juvenile Offenders and Victims: 1999 National Report.
Washington, D.C. Office of Juvenile Justice and Delinquency Prevention: 115.

Spelman, William. (1994). Criminal Incapacitation. New York: Plenum.

Stolzenberg, Lisa, & D'Alessio, Stewart J. (1997). "Three strikes and you're out": The impact of
California's new mandatory sentencing law on serious crime rates. Crime and
Delinquency, 34:457-470.

Stouthamer-Loeber, Magda, Loeber, Rolf, Wei, Evelyn, Farrington, David P., & Wikstrom, Per-
Olof, H. (2002). Risk and promotive effects in the explanation of persistent serious
delinquency in boys. Journal of Consulting and Clinical Psychology, 70:111-124.

Sullivan, Christopher J., McGlogin, Jean M., Pratt, Travis C., and Piquero, Alex R. (2006).
Rethinking the "norm" of offender generality: Investigating specialization in the short-
term. Criminology, 44:199-234.

Swanson, Jeffrey W., Borum, Randy, Swartz, Marvin S., & Monahan, John. (1996). Psychotic
symptoms and the risk of violent behavior in the community. Criminal Behaviour and
Mental Health, 6:309-329.

Tittle, Charles R., Grasmick, Harold G. (1998). Criminal behavior and age: A test of three
provocative hypotheses. The Journal of Criminal Law & Criminology, 88(1):309-342.










Tonry, Michael. (1995). Malign Neglect: Race, Crime and Punishment in America. New York:
Oxford University Press.

Van Voorhis, Patricia. (1987). Correctional effectiveness: The high cost of ignoring success.
Federal Probation, 51:56-62.

Visher, Christy A. (1986). The Rand inmate survey: A reanalysis. In Alfred Blumstein,
Jacqueline Cohen, Jeffrey A. Roth, and Christy A. Visher (eds.), Criminal Careers and
"Career Criminals", vol.2. Washington, D.C.: National Academy Press.

von Hirsch, Andrew. (1972). Prediction of criminal conduct and preventive confinement of
convicted persons. Buffalo Law Review, 21:717-758.

von Hirsch, Andrew. (1985). Past or Future Crimes: Deservedness and Dangerousness in the
Sentencing of Criminals. New Brunswick, NJ: Rutgers University Press.

von Hirsch, Andrew, & Gottfredson, Don M. (1984). Selective incapacitation: Some queries on
research design and equity. Review ofLaw and Social Change, 12:11-51.

Warner, Sam B. (1923). Factors determining parole from the Massachusetts reformatory. Journal
of Criminal Law & Criminology, 14:172-207.

Weber, Max. (1968). Economy and Society. Totowa, NJ: Bedminster.

Welsh, Wayne N. (2006). The need for a comprehensive approach to program planning,
development, and evaluation. Criminology & Public Policy, 5:603-614.

Welsh, Wayne N., & Harris, Philip W. (2004). Criminal Justice Policy and Planning, 2nd edition.
Cincinnati, OH: LexisNexis, Anderson Publishing Co.

West, Donald J., and Farrington, David P. (1997). The Delinquent Way ofLife. London:
Heinenmann.

Wiebush, R., Baird, C., Krisberg, B., & Onek, D. (1995). Risk assessment and classification for
serious, violent, and chronic juvenile offenders. In J. Howell (Ed.), Guide for
Implementing the Comprehensive Strategy for Serious, Violent, and Chronic Juvenile
Offenders. Washington, D.C.: U.S. Department of Justice, Office of Juvenile Justice and
Delinquency Prevention.

Wilbanks, W.L. (1985). Predicting failure on parole. In David P. Farrington and Roger Tarling
(eds.), Prediction in Criminology. Albany, NY: SUNY Press.

Wilkins, Leslie T. (1973). Crime and criminal justice at the turn of the century. Annals of the
American Academy of Political and Social Science, 408:13-29.









Wilson, James Q. (1975). Thinking About Crime. New York: Basic Books.


Wilson, James Q., and Herrnstein, Richard. (1985). Crime andhuman nature. New York:
Simon and Schuster.

Winokur, Kristin P. & Blankenship, Julia L. (2007). Evidence-based risk assessment in the
juvenile justice system: Florida's Positive Achievement Change Tool (PACT).
PowerPoint presentation, The Justice Research Center.

Wright, James D. & Rossi, Peter H. (1986). Armed and Considered Dangerous: A Survey of
Felons and Their Firearms. New York: Aldine de Gruyter.

Wolfgang, Marvin, Figlio, Robert M., and Sellin, Thorsten. (1972). Delinquency in a Birth
Cohort. Chicago: Chicago University Press.

Zamble, Edward, & Quinsey, Vernon L. (1997). The Criminal Recidivism Process. New York:
Cambridge University Press.

Zimring, Franklin E., & Hawkins, Gordon. (1988). The new mathematics of imprisonment. Crime
& Delinquency, 34:425-436.

Zimring, Franklin E., & Hawkins, Gordon. (1995). Incapacitation: Penal confinement and the
restraint of crime. New York: Oxford University Press.









BIOGRAPHICAL SKETCH

Michael Baglivio is a doctoral candidate in the Department of Criminology, Law, &

Society at the University of Florida. He has a Bachelor of Science in psychology, a Master of

Health Sciences in rehabilitation counseling, and a Master of Arts in sociology, all from the

University of Florida. Michael is currently a co-Associate Project Director on a field

demonstration grant sponsored by the Office of Juvenile Justice and Delinquency Prevention

involving the effectiveness of various treatments of juvenile offending youth with maltreatment

histories. His research interests include criminological theory, risk assessment, and life-course

criminology.





PAGE 1

1 THE PREDICTION OF RISK TO RECI DIVATE AMONG A JUVENILE OFFENDING POPULATION By MICHAEL T. BAGLIVIO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Michael T. Baglivio

PAGE 3

3 To the memory of my mother, Lorraine

PAGE 4

4 ACKNOWLEDGMENTS I thank Dr. Alex Piquero for his guidance a nd accessibility, and the entire committee for their patience and continued suppor t throughout the course of this pr oject. I thank my family for their patience and continued faith rega rding the completion of this project.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION..................................................................................................................11 Pre-Screen..................................................................................................................... ..........13 Full Assessment................................................................................................................ ......14 Assessment Scoring............................................................................................................. ...21 2 LITERATURE REVIEW.......................................................................................................23 Collective vs. Selec tive Incapacitation...................................................................................23 New Penology................................................................................................................... ......27 Errors in Prediction........................................................................................................... ......36 Prediction Method.............................................................................................................. ....38 Prediction Studies............................................................................................................. ......41 What Works?.................................................................................................................... ......46 Empirical Pr edictors........................................................................................................47 Principles of Effective Intervention.................................................................................51 Gender Differences in Offending....................................................................................59 Current Focus.................................................................................................................. ........64 3 METHODS........................................................................................................................ .....69 4 DISCUSSION..................................................................................................................... ..113 5 CONCLUSION.....................................................................................................................120 6 FUTURE DIRECTIONS......................................................................................................122 APPENDIX A HOW TO READ AND INTERPRET THE PA CT OVERVIEW AND THE YES CASE PLAN REPORTS.................................................................................................................124 B PACT PRE-SCREEN ASSESSMENT.................................................................................129 C FULL ASSESSMENT..........................................................................................................137

PAGE 6

6 REFERENCES..................................................................................................................... .......153 BIOGRAPHICAL SKETCH.......................................................................................................167

PAGE 7

7 LIST OF TABLES Table page 1-1 Pre-Screen Domains......................................................................................................... ..14 1-2 PACT Full Assessment Domains.......................................................................................15 1-3 PACT Scoring Matrix........................................................................................................22 3-1 Criminal History Item Descriptive Statistics.....................................................................72 3-2 Social History Item Descriptive Statistics.........................................................................74 3-3 Community Cohort 1 Criminal History Descriptive Statistics..........................................81 3-4 Community Cohort 1 Social Hi story Descriptive Statistics..............................................81 3-5 Residential Cohort 1 Criminal History Descriptive Statistics...........................................82 3-6 Residential Cohort 1 Social History Descriptive Statistics................................................83 3-7 Community Cohort 1 Covari ate Descriptive Statistics......................................................84 3-8 Community Cohort 1 Logi stic Regression Results............................................................85 3-9 Residential Cohort 1 Covari ate Descriptive Statistics.......................................................85 3-10 Residential Cohort 1 L ogistic Regression Results.............................................................86 3-11 Split Residential Cohort 1 Cova riate Descriptive Statistics..............................................88 3-12 Split Residential Cohort 1 Logistic Regression Results....................................................89 3-13 Community Cohort 2 Criminal History Descriptive Statistics..........................................91 3-14 Community Cohort 2 Social Hi story Descriptive Statistics..............................................91 3-15 Residential Cohort 2 Criminal History Descriptive Statistics...........................................92 3-16 Residential Cohort 2 Social History Descriptive Statistics................................................93 3-17 Community Cohort 2 Covari ate Descriptive Statistics......................................................93 3-18 Community Cohort 2 Logi stic Regression Results............................................................94 3-19 Residential Cohort 2 Covari ate Descriptive Statistics.......................................................95 3-20 Residential Cohort 2 L ogistic Regression Results.............................................................96

PAGE 8

8 3-21 Split Residential Cohort 2 Cova riate Descriptive Statistics..............................................96 3-22 Split Residential Cohort 2 Logistic Regression Results....................................................97 3-23 Descriptive Statistics By Gender.....................................................................................101 3-24 Logistic Regression Results By Gender..........................................................................102 3-25 Cohort 2 Descriptive Statistics By Gender......................................................................103 3-26 Cohort 2 Logistic Regression By Gender........................................................................104 3-27 Cohort 2 Community Gendered C ovariate Descriptive Statistics...................................105 3-28 Cohort 2 Community Expanded C ovariate Logistic Regression.....................................107 3-29 Cohort 2 Community Gender-Split C ovariate Descriptive Statistics..............................107 3-30 Cohort 2 Community Gender-Split Covariate Logistic Regression................................109 3-31 Cohort 2 Community Violent Offense Logistic Regression............................................111

PAGE 9

9 LIST OF FIGURES Figure page A-1 PACT Overview Report...................................................................................................124 A-2 YES Case Plan.............................................................................................................. ...127

PAGE 10

10 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE PREDICTION OF RISK TO RECI DIVATE AMONG A JUVENILE OFFENDING POPULATION By Michael T. Baglivio December 2007 Chair: Lonn Landza-Kaduce Cochair: Alex Piquero Major: Criminology, Law and Society The Florida Department of Juvenile Jus tice has recently implemented a new fourthgeneration risk/need assessment to predict the risk to re-offend for juveniles referred to the Department. The new assessment, the Positive Achievement Change Tool, or PACT, is adapted from the validated Back On Track! instrume nt utilized in Washi ngton State. This study validates the PACT assessment, and examines whet her the instrument is as predictive of female delinquency as it is of male de linquency utilizing official deli nquency referral as the dependent measure. The study furthermore examines the predic tive power of each of the twelve areas, or criminogenic domains, of the PACT to examine the strength of each domai n in the prediction of juvenile delinquency. Gender differences are explor ed to examine whether the factors predicting female delinquency mimic those pr edictive of male delinquency.

PAGE 11

11 CHAPTER 1 INTRODUCTION The focus of the current study is to validate the assessment instrument used by the Florida Department of Juvenile Justice to determine a referred youths risk to re-offend. The validation of an assessment or instrument used in the predicti on of risk to recidivate is critical to ensure the predictions that come from that instrument and the recommendations made based on the predictions are indeed the best pos sible. The first step to being ab le to provide effective treatment to juvenile offenders is the use of a validated ri sk/needs assessment to determine the risk level of the youth, which determines the needed inte nsity of treatment, and the individualized criminogenic needs, which must be targeted in order to reduce the likelihood of recidivism. The Florida Department of Juvenile Justice is one of the largest juvenile justice agencies in the United States. The Department receive d 150,104 delinquency referrals during FY 2005-06, representing 94,288 individual youth Beyond the importance of the safety and security of the youth served, one of the most challenging tasks for a criminal justice agency of that magnitude is to ensure the provision of appropriate, evidence-b ased treatment services to the youth referred in an effort to reduce their risk of re-offending and increase public safet y. A critical step to providing empirically proven, appr opriate treatment to a juven ile offending population is the implementation of an evidence-based, validated risk/need assessment. The Florida Department of Juvenile Jus tice (DJJ) has recently implemented a new instrument designed to assess juvenile offenders w ith regard to risk, needs, and strengths related to the same risk factors as outlined in the ex tant what works litera ture (Andrews & Bonta, 2003). The Positive Achievement Change Tool, or PACT, is heavily adapted from the validated Washington State Juvenile Court Assessment, Back On Track! (also know as the YASITM in some jurisdictions), which has been in use thro ughout the country since 1998. The Florida

PAGE 12

12 Department of Juvenile Justice and Assessments .com created the PACT collaboratively in 2005. Implementation of the PACT began in November of 2005 in Hillsborough and Pinellas counties, and was statewide by August of 2006. Over 1,400 Juve nile Probation Officer staff members have received the two-day PACT tr aining including risk assessmen t theory, case planning, and the techniques of motivational inte rviewing (Miller & Rollnick, 2002) The 1,400 staff represents all Juvenile Probation Officers and all juve nile assessment cente r (JAC) screeners. The PACT is an advancement from other fourth-generation risk/needs assessments, including Back on Track! and the YASI on whic h it was based, due to its incorporation of questions from the Massachusetts Youth Screeni ng Instrument (MAYSI), making it an indicator for need of further mental health assessment, and its integration with DJJs Juvenile Justice Information System (JJIS) database. The JJIS an d PACT interface allows for automatic scoring of the criminal history domain (official prio r criminal history). Furthermore, the PACT automates findings into the Youth Empowered Success Plan (YES), one of the first ever supervision and treatment plans in the nation to be automated to directly apply information gained from an evidence-based risk/needs a ssessment (see Appendix A: How to Read and Interpret the PACT Overview a nd the YES Case Plan Reports). The PACT identifies not only the areas (domains) in which a juvenile is most at risk, but equally as important, those areas which he or she has strengths (p rotective factors), which can be built upon in the case planning process. Risk and protective factors with in the PACT include both static and dynamic characteris tics. The benefit of measuring risk and protective factors and characteristics is that a juvenile justice professional is better able to match a youths current needs with appropriate programs and services.

PAGE 13

13 Pre-Screen The PACT has both a Pre-Screen and a Full-As sessment, with risk levels ranging from low, moderate, moderate-high, and high. Youth w ho score moderate-high or high risk to reoffend are administered the Full-Assessment, as are youth being placed in residential commitment. Furthermore, youth on probation superv ision are given a rea ssessment every ninety days so as to measure progress on the treatm ent plan, and youth in residential commitment receive an assessment within sixty days of rele ase so as to assist in aftercare supervision planning. The Pre-Screen is a 46item, initial assessment instrument, which produces risk level scores measuring a juvenile's risk of re-offendi ng. This tool is designed as a semi-structured interview protocol and utili zes Motivational Interviewing t echniques (Miller & Rollnick, 2002). With automated scoring, the PACT Pre-Screen identifies a total risk level for each youth assessed, as well as a rank order of criminogenic needs and protective factors. Low-risk to reoffend youth are generally recommended for dive rsion or other community-based intervention programs and isolated from any contagion effect s caused by intermingli ng with higher-risk youth or separation from the protective factors present in their lives. Moderate-risk youth are generally directed to supervised probation and other comm unity-based intervention s. Moderate-high and high-risk youth receive th e PACT Full-Assessment. The PACT Pre-Screen collects information on four distinct domain s (see Table 1-1). Domain 1 assesses the youths prior criminal hi story. This domain is pre-populated from the JJIS database, as mentioned above, and is therefor e not self-reported as are many of the other domains. Domain 1 includes information on the age at which the youth first offended, seriousness of prior offenses in terms of number of prior felonies and prior misdemeanors, prior commitment placements, prior escapes, and the type of offenses (such as weapon, against person,

PAGE 14

14 and sex offenses). A complete listing of Domain 1 questions, and all PACT Pre-Screen questions can be seen in Appendix B: PACT Pre-Screen Assessment. Table 1-1. Pre-Screen Domains Domain # Domain Name 1 Record of Referrals 2 Social History 3 Mental Health 4 Attitude/Behavior Indicators Domain 2 of the PACT Pre-Screen examin es the youths social history. This domain features gender, current school st atus and performance, history of, and extent of current, antisocial friends, history and curre nt alcohol and drug use, histor y of abuse and mental health problems, and history of family problems. Doma in 3 assesses the youths mental health status including suicidal ideation, history of a nger, depression, somatic complaints, thought disturbance, and history of traumatic experiences. It is th is domain that has incorporated questions from the MAYSI and allows for the identif ication of the need for further mental health assessment. Certain hits on Domain 3 questio ns trigger the need for a more in-depth assessment of the youths current me ntal status. The final Pre-Scre en domain is Domain 4, which examines the youths extent of an ti-social attitudes and beliefs, and resulting behavior. This section covers attitudes toward law-abiding be havior, beliefs about a ggression and fighting to solve problems, self-reports of violence not included in the offi cial criminal history domain (Domain 1), and reports of sexual aggression. Full Assessment The PACT Full-Assessment is a 126-item, in -depth assessment instrument, which produces identical risk level scores as the Pre-Sc reen, measuring a juvenile's risk of re-offending. This tool is designed as a semi-structured inte rview protocol utilizing Motivational Interviewing

PAGE 15

15 techniques as well. Every question in the Pre-Sc reen is included in the Full-Assessment and is pre-populated from the Pre-Screen into the Fu ll-Assessment to decrease the workload of the individual completing the assessment. The Full-A ssessment is required to be administered to moderate-high and high risk to re-offend youth, and all youth being pl aced in residential commitment as a means to gather the most comprehensive amount of information on each youth related to research-validated risk and need factors as possible. The PACT Full-Assessment instrument measures a youth's risk and protecti ve factors in the follo wing twelve domains: Criminal History, School, Use of Free Time, Em ployment, Relationships, Family History, Living Arrangements, Alcohol and Drugs Mental Health, Attitudes/Behaviors, Aggression, and Skills (see Table 1-2). Table 1-2. PACT Full Assessment Domains Domain # Domain Name 1 Record of Referrals 2 Gender 3A School History 3B Current School Status 4A Historic Use of Free Time 4B Current Use of Free Time 5A Employment History 5B Current Employment 6A History of Relationships 6B Current Relationships 7A Family History 7B Current Living Arrangements 8A Alcohol and Drug History 8B Current Alcohol and Drugs 9A Mental Health History 9B Current Mental Health 10 Attitudes/Behaviors 11 Aggression 12 Skills

PAGE 16

16 The first domain, as in the Pre-Screen, is automated from the JJIS database so as to decrease the errors in reporting prior official offending history. Domain 1 of the Full-Assessment is identical to Domain 1 of the Pre-Screen a nd provides a comprehensive description of referral history for the youth. Prior referrals are disaggregated by type of offense (against-person, sexual, weapon) by felony and misdemeanor. Furthermore, the amount of prior detention confinements, residential commitments, escapes, and pick up orde rs/failure-to-appear in court are included (for a full list of Full-Assessment questions, see A ppendix C: PACT Full-Assessment). The inclusion of this domain is the empirical importance of the predictive power of past behavior on future behavior; the notion that the best predictor of future behavior is past behavior (Sherman et al., 1997). The second domain distinguishes between ma les and females. This gender domain is essential to capture the reality that male youth consistently of fend and re-offend at higher rates than do female youth. Males automatically receive one point towards the assessment of risk whereas females receive zero points. The third domain is subdivided into School History and Current School Status. The School History section examines whether the youth has ev er been a special educ ation student, and why (behavioral, ADHD/ADD, learning di sability, mental retardation). Th is section also captures any history and the amount of expulsion and out of school suspensions, age at first expulsion/suspension, and whether the youth has been enrolled in school during the last six months. The Current School Status section captures the youths current enro llment status, type of school in which the youth is enrolled, and the yo uths perception of the value of education. This section also examines the youths conduct, i nvolvement in school activities, number of expulsions/suspensions, attendance, and perfor mance, all in the most recent school term.

PAGE 17

17 Information included in this domain is by self-report from the youth in combination with corroboration from teachers from the school in which the youth was enrolled. The next domain captures the youths use of fr ee time. This domain is also separated to examine historic use of free time and the youths current use of free time. Information for this domain is captured by self-report from the youth. Historic use of free time is simply measured by whether the youth has participated in structured and supervise d, or unstructured, pro-social community activity within the past five years. Current use of recreationa l time is captured by the youths current interest a nd involvement in structured recreati onal activities, which activities the youth is currently engaged in (vol unteer organizations, athletics, religious groups, etc.), and the current interest and involvement in other positiv e recreational activities (s uch as reading or prosocial hobbies). The notion behind this domain is so cial control, and that youth who are involved in pro-social recreational and le isure activities will be less likel y to commit criminal offenses than those youth without an in terest in such activities. The fifth domain captures the youths employment and is divided into employment history and current employment. This domain does take in to account the fact that the youth may be too young for employment consideration. The employme nt history section looks at whether the youth has ever been employed, problems while empl oyed (fired or quit for poor performance, or fired/quit because he/she could not get along w ith employer or coworkers), and history of positive personal relationships with past employers or coworkers. The current employment section measures whether the youth understands what is required to maintain a job, interest in employment, current employment status, and curr ent positive relationships with employers or coworkers.

PAGE 18

18 Relationship issues are measured by the sixt h domain and are divided into history of relationships and curren t relationships. History of relations hips captures past positive adult nonfamily relationships, who can provide support and model pro-social behavior, as well as past antisocial friends/peers, who are hos tile to or disruptive of the le gal social order or engage in criminal activities. Current relationship issues are measured by the extent of current positive adult influences, pro-social community ties, whethe r current friends are pro-social or antisocial, current romantic relationship, whether the youth currently admires anti-social peers, and the youths current resistance to anti-social peer infl uences. This domain is heavily influenced by the notion of differential association (Akers, 1998). Family history and current living arrangement s compose the next domain. Family history is captured by a history of out -of-home Department of Children and Families (DCF) and shelter placements, history of running away, history of imprisonment of persons living in the household with the youth, and whether the youth lives with a supervising adult, with peers, or is transient. Current living arrangements examines who the y outh currently lives with annual income of the family, imprisonment history of persons the youth is currently living wit h, problem history of parents currently in the household (such as alco hol, drug, mental health, employment), problem history of siblings in the house hold (identical to those of the parents), support network for the family, level of conflict between parents and sibl ings, parental supervision, and the extent of appropriate punishment and rewards. The incorpor ation of factors deali ng with the consistency of punishments and rewards, and the extent of parental supervision, au thority, and control are heavily influenced by Gottfredson and Hirs chis General Theory of Crime (1990). The eighth domain captures substance abuse a nd is gathered from self-report from the youth as well as corroboration wi th parents and teachers when possible. This domain is divided

PAGE 19

19 into alcohol and drug history, a nd current use of alcohol and drugs. Alcohol use is measured separately from drug use for both subgroups of questions Past use or current use is captured as is whether the use has disrupted, or currently disrupts, education, cau ses family conflict, interferes with keeping pro-social friends, causes health pr oblems, contributes to criminal behavior, or leads to increased tolerance or withdrawal. History or current referral for assessment and substance abuse classes/treatment is also captured. For the current drug use section, the types of drugs used are identified as well. The ninth domain measures mental health hi story and current mental health status. As stated above, this domain purposively incorporated questions from the MAYSI assessment so as to be able to flag youth for need of further ment al health and suicide a ssessment. This inclusion allows the individual performing the assessment to discontinue using the MAYSI altogether, decreasing workload and increasing efficiency. The mental health history section assesses whether the youth has ever thought about suicide, attempted suicide, enga ged in self-mutilating behavior, has a history of physical abuse, se xual abuse, neglect, a traumatic experience, witnessing violence, mental health problems, anger or irritability, depression, somatic complaints, or thought disturbance. The assessmen t takes into account any neglect or abuse selfreported by the youth, whether substantiated or not, but excludes reports proven to be false in the event investigations by social service agen cies concluded the report to be false. The mental health problem questions include such issues as schiz ophrenia, bi-polar, mood, thought, personality, and adjustment disorders. Conduct disorder, oppositional defiant disorder, substance abuse, and ADD/ADHD are excluded. The mental health problems are those confirmed by a professional in the social service/h ealthcare field (the youth must have been diagnosed).

PAGE 20

20 The current mental health section includes whether the youth is currently prescribed treatment, or currently prescr ibed medication. These questions disaggregate further by whether the youth is actually attendi ng the treatment, or taking the medicat ion, or whether it is prescribed but the youth is not attending or complying wi th the medication regimen. Furthermore, the current mental health section also captures wh ether the individual performing the assessment believes the mental health problems of the youth interfere with working with the youth. The next domain assesses the attitudes and behaviors of the youth. This section captures the youths self-reported primary emotion wh en committing crimes (nervous, excited, indifferent, confident) and the primary purpose for committing them (anger, revenge, impulse, sexual desire, money, amusement, or peer status ). The level of optimis m the youth has about the future is also examined in terms of whethe r the youth has high aspira tions, some sense of purpose, low aspirations, believes nothing matters as well as his/her belief in being able to successfully meet the conditions of the court. Several questions capture the extent to which the youth believes he/she can control or avoid anti-social behavior, or use self-control. The youths empathy for victims is assessed, as is the yout hs acceptance of responsibility for anti-social behavior, respect for the property of others, authority figures, and attitude toward law abiding behavior. The Aggression domain is next, which meas ures tolerance for frustration, hostile interpretation of the actions and intentions of othe rs, and belief that yellin g or belief that fighting resolves disagreements. Information for this domai n is generated from youth self-report. This domain also measures reports/evi dence of violence not included in the criminal history domain, such as outbursts, deliberately inflicting physical pain, threatening with a weapon, fire starting,

PAGE 21

21 and animal cruelty. Reports of problems with sexua l aggression are examined, such as aggressive sex, sex for power, young sex partners, ch ild sex, voyeurism, and exposure. The final domain examines skills that coul d potentially help the youth avoid anti-social behavior. These skills include th e ability to recognize there are consequences to ones actions, ability to set realistic goals, ability to appl y appropriate solutions and problem-solve, and situational perception. The Skill domain also capture s the youths use of basic social skills to deal with others, deal with difficult situations and deal with feelings/emotions. The final few questions measure whether the youth can monitor internal triggers, exte rnal triggers, control impulsive behaviors, and contro l aggression. Information for this domain is generated from youth self-report as well as the assess ors impression of the youths skills. Assessment Scoring The PACT Pre-Screen and Full-Assessment are scored identically. The purpose of including additional items in the Full-Assessment is to gather as much information on the youth as possible to assist with case planning, but the ov erall total risk to re-o ffend score is identically derived for each assessment type. The score (l ow, moderate, moderate-high, high) is based on two sets of factors. The crimin al history score is assessed excl usively based on the youths prior official criminal offending. The criminal history sc ore ranges from zero to thirty-one. The social history score, in contrast, is based on responses to questions from seve ral of the other PACT domains. The social history sc ore ranges from zero to eighteen. The two scores, criminal and social history, are associated w ithin a matrix, determining the youths overall risk to re-offend (see Table 1-3). The items used to score criminal history incl ude age at first offense, number of prior referrals (misdemeanor, felony, weapon, and person-misdemeanor, person-felony), prior detention confinement, prior residential commitm ents, prior escapes from commitment, and prior

PAGE 22

22 pick-up orders. Social history is scored us ing variables including gender, school status (enrollment, conduct, attendance, academic perfor mance), current anti-soc ial peers, history of court-ordered out of home placeme nts, history of running away or being kicked out of home, incarceration of a current househol d member, parental control, hi story and current alcohol or drug use, history of violence/physi cal abuse, history of being a w itness to violence, sexual abuse history, history of neglect, and mental health problems. Individuals administering the assessments are not knowledgeable about the maxi mum scores associated with any domain of the PACT. The instrument, therefore, cannot be manipulated as easily as assessments without automated scoring, or where individuals are aware of which questions are weighted more heavily than others. Table 1-3. PACT Scoring Matrix Criminal History Score So cial History Risk Score 0 to 5 6 to 9 10 to 18 0 to 5 Low Low Moderate 6 to 8 Low Moderate Moderate-High 9 to 11 Moderate Moderate-High High 12 to 31 Moderate-High High High

PAGE 23

23 CHAPTER 2 LITERATURE REVIEW The notion of predicting individual s risk to offend or recidiva te has been a constant focus for criminologists and policy-makers for severa l decades. As early as 1923, Warner examined offender characteristics related to violation of parole (Warner, 1923; von Hirsch & Gottfredson, 1984). Prediction studies generated worldwide atte ntion over the next few decades due to the seminal work by Glueck and Glueck (1950; see also von Hirsch & Gottfredson, 1984). The California Base Expectancy Score, used in the 1960s, and a similar scale adopted by the United States Parole Commission in the 1970s, were used extensively to predict parole success and classification decisions (von Hirsch & Gottfreds on, 1984). Prediction then moved to the realm of classifying offenders based on the likelihood of re-offending. David Greenberg (1975) first introduced the term selective incapacitation, a nd the idea of targeting particular offender groups for imprisonment emerged, although a comp rehensive strategy for implementing such a plan was first extrapolated by Greenwood and Abrahamse (1982). Collective vs. Selective Incapacitation Incapacitation is a simple crime prevention strategy based on the notion that offenders cannot commit crimes in society as long as th ey are incarcerated. Crime reduction therefore occurs, as those incarcerated are prevented from committing offenses in the community. A secondary benefit of a crime prevention strate gy based on incarcerati on is the concept of deterrence. Individuals in the general population are believed to be deterred from committing crime from the fear of being sent to prison; referred to as the ge neral deterrence effect. Furthermore, imprisoned individuals may be de terred from committing subsequent offenses upon release; referred to as th e specific deterrent effect (Sherman, et al., 1997).

PAGE 24

24 Collective incapacitation refers to the crime reduction effect resulting from physically restraining individuals regardle ss of the strategy utilized (b e it deterrent, incapacitative, retributive, or rehabilitative) and without the necessity of prediction of future behavior. Collective incapacitation was popularized by Jame s Wilson (1975), calling for uniform sentences imposed on all offenders of specific felonies Removing all of these offenders from the community for a significant portion of their crimin al careers would prevent all of them who were so inclined from re-offending while in carcerated (von Hirsch & Gottfredson, 1984). Wilson (1975) promised a twenty percent re duction in robbery rates if a strategy of incapacitating a larger fr action of the convicted serious robbe rs were implemented. The estimates were derived from a projection technique deve loped by Shinnar utilizing the newly invented lambda variable (von Hirsch & Gottfredson, 198 4; Shinnar & Shinnar, 1975). The effect on the crime rate due to collective incapacitation stra tegies has been found to depend critically on the estimates of lambda, the average annual offens e rate at which indivi dual offenders commit offenses in the community (von Hirsch & Gottf redson, 1984). When lambda is assumed to be high, the reduction in the crime rate due to incapa citation would be substantial, but when a lower average offense rate is assumed, the effects are much more modest. Selective incapacitation, in contrast, implies the prevention of future criminality through physically restraining thos e particular individuals who are pred icted to be higher risk to commit future criminal behavior than other groups of individuals (Greenbe rg, 1975). Greenwood and Abrahamse (1982) revised the formula used in the Shinnar and Wilson studies by establishing not one, but three lambdas (von Hirsch & Go ttfredson, 1984). A separate lambda would be estimated for low, medium, and high-risk offe nders. Greenwood and Abrahamse (1982) hold that the separate estimates can be gathered from th e data by using the aver age self-reported annual

PAGE 25

25 robbery rates for individuals falling in each of the three risk level classifications of his prediction instrument (a seven-item, dichotomous additive index). These authors ar gue for the ability to estimate the incapacitative effect for each level of offender risk, demonstrate that incarcerating high-risk offenders is most effective, and subs equently propose the lengthening of sentences for individuals in that high-risk classification category. In the event accurate prediction can be achie ved, and if it is coupled with focusing incarceration only on high risk to re-offend individuals, Gree nwood and Abrahamse proposed that it would be possible to rema in within the available prison re sources, a statement not possible by those adhering to a collec tive incapacitation mentality (Greenwood & Abrahamse, 1982, von Hirsch & Gottfredson, 1984). In this model, the offenders risk sc ore, not simply the seriousness of the presenting offense, would principally dete rmine whether that offender is incarcerated. Von Hirsch and Gottfredson (1984) ar gue this policy would require imprisonment of a larger proportion of offenders convicted of less serious crimes if their predicted level indicated they were bad risks. This statement, while true, di scounts the empirical evidence arguing against the offense-specific specialization of offenders (discussed at greater length below). Other researchers question the ability to estimate lambda as Greenwood and Abrahamse have due to their use of genera lizing the self-reported robbery ra tes of the convicted robbers in his sample to active robbers generally, especia lly in light of their own admission that the probability of being convicted then incarcerated for robbery is so low; .0258 in that study (von Hirsch & Gottfredson, 1984). Other researchers ha ve found incarcerated samples biased toward having more high-rate offenders than found in the general population (Auerhahn, 1999; CanelaCacho et al., 1997; Chaiken & Chaiken, 1984; Shannon, 1991; Wolfgang et al., 1972; Wright & Rossi, 1986). Any incarceration effect, be it selectiv e or collective, would be overestimated if the

PAGE 26

26 sample used for estimation is an incarcerat ed one. Additional problem s with estimating the benefits of incarceration strate gies have been proposed as well, such as the assumptions regarding length of criminal career, replacement/s ubstitution effects, and the decline in offending with age (Blumstein et al., 1978; Gottfreds on & Hirschi, 1986; Spelman, 1994; Zimring & Hawkins, 1995; 1988). Replacement and substitution effects refer to the notion that when an individual is arrested and/or removed from a particular locale, anot her individual will take his/her place and commit similar offenses. This no tion is prevalent when referring to the sale of illegal drugs where removal of a drug dealer fr om a location may only lead to a new individual subsequently selling drugs at that same location. At the time of the 1975 publication, Greenbe rg, and others, incl uding researchers and prison administrators, hypothesized the group of high-risk offenders to be targeted for selective incapacitation would be between te n to fifteen percent of the prison population, while all other prisoners could be safely releas ed under varying levels of supe rvision with minimal threat to society (Greenberg, 1975; see also Go ldfarb & Singer, 1973; Mitford, 1973).1 Other research at the time had found that most crimes are committe d by a small percentage of the population as a whole and that most are committed by recidivist s (Belkin et al., 1972; Hughes, 1971; Shannon et al., 1988; Shinnar & Shinnar, 1975; West & Farrington, 1977; Wolfga ng et al., 1972). Shinnar and Shinnar (1975) caution the im portance of distinguishing between the percentage of crimes committed by recidivists and the percentage of criminals who are recidivists. They provide the example that a ssuming for every career criminal there are 5 one1 Greenberg also presented the argument that higher recidivism rates for particular offenders could demonstrate both the need for longer sentences for public protection, or sh ow the need for increased re sources aimed at increasing rehabilitation program effectiveness. The same data are utilized by both ends of the rehabilitative-incapacitative spectrum with neither party realizing the high recidivism ra tes may be greatly due to artifacts of administration in that most returns to prison are for technical violations, not new-law violations (Greenberg, 1975:556). By constantly calling the publics attention to high recidivism, both si des of the argument increas e the prejudice against exconvicts.

PAGE 27

27 time offenders and with an estimate of 50 crimes per career criminal, then career criminals would compose only 16% of the criminal population, yet account for more than 90% of the crimes. This hypothetical has been somewhat echoed by empiri cal research by Wolfga ng who found 6.3% of a birth cohort of Philadelphia boys (18% of the offenders) were responsible for roughly 52% of the offenses committed by the cohort (Wolfgang et al., 1972; see also Chaiken & Chaiken, 1984; Cohen, 1984; Mednick & Christiansen, 1977; Shannon, 1991; Shannon et al., 1988; Wilson & Herrnstein, 1985; West & Farringt on, 1977; Wright & Rossi, 1986). New Penology The concept of selective incap acitation came just after the no w infamous nothing works declaration interpreted from Martinson, questioning the possibility of reha bilitative effectiveness in the criminal justice system (1974; see also Lipton et al., 1975; Zimring & Hawkins, 1995).2 Critics argued against the not hing works conclusion, faulting inadequacy of the research methodology used to evaluate the programs and the poor implementation and delivery of the treatment programs studied, for the inability to re alize reduced recidivism or draw conclusions from the research (Cullen & Gendreau, 1989; Cullen & Gilbert, 1982 ; Gendreau & Ross, 1979, 1981, 1987; Gottfredson, 1979; Halleck & Witte, 1977; Palmer, 1975, 1983; van Voorhis, 1987). While these critics published evaluations indicati ng substantial evidence of effective treatments, the nothing works mantra exerted tremendous power over practitioners, professionals, and the public (Cullen & Gendreau, 1989). Th e Martinson report en capsulated the shifting focus evident in the waning influence of the principle of reha bilitation for imprisonment previously held by the criminal justice system in the United States, which had began to decl ine in the 1970s (Allen, 1981; Auerhahn, 1999; Flanagan, 1996; von Hirsch, 1985). 2 Several researchers have suggested the decline in suppor t for the rehabilitative ideal may also be traced to the rising proportion of minority individuals in U.S. prisons (Allen, 1981; see also Sklansky, 1995; Tonry, 1995).

PAGE 28

28 The focus then transferred from the rehabilita tive ideal of treatment of individual offenders towards incapacitative and retributive goals empha sizing control, management, and incarceration of particular offender groups (Auerhahn, 1999; Feeley & Simon, 1992; Shichor, 1997). The major aim of the criminal just ice system switched to the effective management of so-called dangerous groups in the new penology (Feeley & Simon, 1992). The new emphasis on efficiency, calculability, predic tion, and control was based on the Weberian concept of formal rationality, and has resulted in the movement being compared with the fast-food industry and the McDonaldization of punishment (R itzer, 1993; Shichor, 1997; Weber, 1968). Three distinct influences have been identified as being involved in the shift to the new penology, including language, objec tives, and techniques (Feeley & Simon, 1992). With the intent shifted from individual treatment to mana gerial concerns for c ontrolling certain offender groups, the notions of probability and risk became increasingly relevant, replacing discourse of clinical diagnosis and mental status. These ideas of probability and risk are instrumental in the isolation and classification of particular high-rate offending gr oups as well as predicting the likelihood of offending or recidivating. Objectives of efficient control re placed traditional goals of rehabilitation and crime contro l. Finally, techniques targeting ag gregate levels of offenders for the purpose of identification, classification, and management of dangerousness groupings supplanted traditional practices of diagnosis intervention, and treating offenders on an individualized basis (Cohe n, 1985; Garland & Young, 1983; F eeley & Simon, 1992; Gordon, 1991; Messinger, 1969; Messinge r & Berecochea, 1990; Reichman, 1986; Wilkins, 1973). According to Feeley and Simon, a focus on risk management and incapacitation of particular offender groups promises to reduce the effects of crime in society not by altering either offender or social context, but by rearra nging the distribution of offenders in society

PAGE 29

29 (1992:458). The task became one of removing high -rate offenders from so ciety, while returning those low risk individuals who pose no significa nt threat to the community. Advancement in statistical methods allowed for the expansion of discourse of actuarial and quantifying language with high-rate offenders and career criminals defined by the distribution of criminals itself, linking direct relationships be tween penal policy and strategi es to the offending population. One additional effect of the focus on control of aggregate subpopulations of offenders was the decrease in expectations about the effectiv eness of criminal sanc tions (Feeley & Simon, 1992). No longer was there anticipation that a sanction, be it incarcera tion, probation, parole, etc., have the potential to rehabi litate or treat the offender with the hope of reintegration into society. Rather, the desire became that the sa nction be effective at monitoring and managing aggregate groups based on classification accordi ng to the degree of control warranted by their classification risk levels. This custodial continuu m, as Feeley and Simon called it, had at one extreme the prisons, which provided the highest secu rity at the highest cost for those individuals who present the greatest risk, and probation at the low end of the spectrum providing low-cost surveillance for low-ri sk individuals. In between the extremes are in termediate levels of supervis ion and surveillance techniques as evidenced by present-day intermediate sancti ons such as boot camps, intensive supervision (with frequent drug testing) and electronic mon itoring. Feeley and Simon caution that once the focus is on categories of offenders rather than individuals, methods naturally shift toward mechanisms of appraising and arranging groups rather than interven ing in the lives of individuals thus in the end th e search for causal order is at least premature (1992:459). The appraising of groups these scholars mention is evident in the increased number of prison admissions due to technical parole and probati on violations making pr isons holding pens for

PAGE 30

30 violators assessed as too dangerous to remain in society. This shifts the mission of the prison from correctional to a management function of wa rehousing a particular risk level of offenders. The differentiation of prisons by s ecurity risk level is in cont rast to the differentiation by specialized function evident in prisons in the 1970 s, such as the California Rehabilitation Center for drug users and California Medical Prison at Vacaville for the mental ly ill (Feeley and Simon, 1992). The new penology movement led to the funding of a research program in the 1970s and 1980s by the National Institute of Justice for the study of criminal careers in relation to policy decisions impacting the criminal justi ce system. The 1982 Rand report by Greenwood and Abrahamse came out of that program This report was the first to articulate a coherent proposal for the identification of offender groups for the purpose of selec tive incapacitation. The Greenwood and Abrahamse study and the subseque nt RAND studies by Chaiken and Chaiken improved over previous prediction research in one essential as pect. These studies moved beyond attempting to predict whether an offender would return to crime, and included also addressing the frequency of offending (von Hirsch, 1984). Hi gh-rate offenders were to receive longer sentences while lower-rate offende rs would receive shortened sent ences. This process would in turn reduce crime and be a more efficient us e of penal resources by only holding high-rate offenders for prolonged periods (G reenwood and Abrahamse, 1982). Shinnar and Shinnar (1975) advi sed that what is important is the uniform application of penalties for any recidivist offender, regard less of the severity of the second offense (foreshadowing current three-strikes laws furt her examined below). Their suggested strategy included improved sentencing procedures, screen ing to recognize recidi vists, and selective increase in those sentences. These scholars did provide one caveat to their argument stating, in

PAGE 31

31 no way do we want to imply that future crime po licies should be based solely on incapacitation. Better rehabilitation would be pr eferable but rehabilitation must become much more successful before it will have a significant effect (Shinnar & Shinnar, 1975:606). The Greenwood and Abrahamse report sparked a deba te in the literature as to the potential effectiveness of a selective incapacitation st rategy (Blumstein et al., 1978; Spelman, 1994; Zimring & Hawkins, 1995), as well as the ethical and methodological obst acles to implementing such a plan (Blackmore & Welsh, 1983; Chai ken & Chaiken, 1984; Cohen, 1983; Gottfredson & Gottfredson, 1985a, 1985b; Moore et al., 1984; Vi sher, 1986; von Hirsch, 1972; von Hirsch & Gottfredson, 1984). Moore and colle agues (1984) argued that selec tive incapacitation is at its core an assertion that individuals vary in thei r capacity for evil and th at these differences are inherent within the individual and not the result of any social pr ocesses, which could be altered or remedied.3 As the individual who arguably coined the term selective incapacitation, Greenberg himself cautioned that the recommendation for deciding who is to be incarcerated on the basis of the threat such individuals would represent if rele ased could be implemented only if it were possible to predict how a given prisoner would behave if released (Greenberg, 1975:543, emphasis in the original). The notion of selective incapacitation, coupl ed with the empirical research on career criminals then provided the rati onale for the habitual offender three-strikes and truth-in3 Feeley and Simon (1992) echo Wilson (1987) in thei r assertion that policy and po licy-makers view a certain segment of the population as unemployable (even if jobs were available) and incapable of being integrated into society, labeling this group with the term underclass. This segment of marginalized indi viduals is also viewed as dangerous to the social order, distrusted for their ability to disrupt as a collective group more so than as individuals. Feeley and Simon (1992) expand by illustrating the extent of custodial supervision enforced on young black males, arguing that the heightened visibility of minority overrepresentation in the penal system is likely to reinforce the notion that crime is the result of this pathological subpopulation upon which nothing can be done with the exception of custodial control. They furthermore argue that rehabilitation or reintegration only makes sense if those being rehabilitated or reintegrated share common values and norms with middle class society. The existence of permanent marginality of a subpopulation would render re habilitative ideals irrelevant by nature of the argument.

PAGE 32

32 sentencing legislation, which has since fo llowed (Auerhahn, 1999; Benekos & Merlo, 1995; Clark et al., 1997; Stolzenberg & DAlessio, 1997). Based on the se lective incapacitation ideal, the goal of this legislation is to incapacitate the relatively small number of potential career criminals early in their criminal life-course to achieve a substantial reduc tion in crime. One such policy came in 1994 when Congress enacted the Violent Crime Control and Law Enforcement Act, also known as the Federal Crime Bill (commonly referred to as the three strikes and youre out law). This law mandates life sentences for crim inals convicted of three violent felonies or drug offenses if the third conviction is a federal crime (Lewis, 1994; Shichor, 1997). The purpose of this legislation was to increase the severity of the sanction and lengthen the term of confinement for offender groups deemed hab itual or career criminals. Some three strikes legislation requires prosecutors to pursue the case under the statut e, calling for longer sentences for offenders with one or more prior felony c onvictions, and often life sentences for three time felons. Prosecutorial discretion is limited, unless he or she can plea bargain down the charges (in which case discretion is enhanced for the prosecu tion), and judicial disc retion reduced due to a requirement to impose the maximum sentence in the event of conviction. Mitigating factors are irrelevant to sentence consider ations (Stolzenberg & DAlessio, 1997). As the laws mandate such extreme sentences, more offenders are taking th eir cases to trial rather than excepting plea agreements, which has the effect of increasing th e number of individuals in county jails awaiting trial, more court trials which increases costs, more long-term prisoners leading to a need for more prisons, an older prison population with in creasingly high medical and health care bills, and increasing welfare expenditure s for dependents of incarcerated felons for longer periods of time (Shichor, 1997).

PAGE 33

33 The other sentencing legislati on, beyond the three strikes and 10-20-life laws, has been the truth-in-sentencing movement. This legislation has been enacted by many states to secure prison funding under the 1994 Federal Crime Act (Greenw ood, 1998). These laws usually require that all violent offenders serve at leas t 85% of their sentence before th ey are eligible for parole or early release. This is in stark contrast to the 30% to 40% that most of these offenders had been serving prior to the legisl ation (BJS, 1992; Greenwood, 1998). The percentage of sentence served was lower pre-legislation due to the abil ity to accrue good-time cr edit and the possibility of early release. Truth-in-sente ncing doubled the length of time th ese offenders serve, therefore drastically increasing the prison population. The important empirical question then becomes whether the three strike s and other habitual offender legislation, or the truth-in-sentenci ng legislation has reduced crime. Greenwood and colleagues (1996) projected the incapacitative effect of Califor nias three strikes laws and estimated that, if fully implemented, they woul d reduce serious felonies between 22% and 34%. Furthermore, they estimated that nearly 1/3 of that reduction would be fo r violent crimes, and the remaining 2/3 for less violent felonies. As Gr eenwood and colleagues (1996), and Stolzenberg & DAlessio (1997) have hypothesize d, this reduction may be an underestimation, as the study did not take into account any deterrent effect the laws may have for individuals contemplating commission of one of the severely sanctioned offenses. Projected analysis aside however empirical research directly assessing the impact of the three strikes legislation has not been so favorable. For example, the California three-strikes law was examined by Greenwood and colleagues (1994) who concluded, if app lied in all eligible cases, the legislation would reduc e the number of serious felonies in a year by about 28%, or 329,000 crimes. They caution however that this br oad application would cost an additional $5.5

PAGE 34

34 billion per year in additional criminal justice funding for the additional construction and operation of prisons (Greenwood et al., 1994). This finding can be translated to a cost of $16,000 per serious felony prevented (Sherman et al., 1997). Stolzenberg & DAlessio (1997) assessed the impact of th e California three strikes legislation on serious crime usi ng an interrupted time-series desi gn. They concluded the laws did not reduce crime levels below that which would be expected based on the preexisting downward trend. Possible explanations offered by the au thors for the lack of reduction include the relationship between age and crime (most offender s would age out of crime anyway by the time they are incarcerated for thei r third strike) and the prevalen ce of juvenile crime (juveniles account for roughly 20% of the violent crime arrest s and 35% of property crime arrests, yet are unaffected by the legislation). The conclusions made by Stolzenberg and DAl essio regarding the aging out of offenders is in keeping with one of the most commonly accepted extant relationships in the field of criminology and broader research on criminal car eers; that of the age crime curve and the declining frequency of offending with age, rega rdless of sex, race, country, offense, or time period (Farrington, 1986; Gottfredson & Hirs chi, 1986; 1990; Nagin, 1998; Nagin & Land, 1993, Petersilia, 1980; Tittle & Gras mick, 1998; Wilson & Herrnstein, 1985). As far back as the early 1970s research has found the proportion of offenders who desist their crimin al career, or drop out, after each subsequent offense is consta nt and around 20% to 30% (Wolfgang et al., 1972; Auerhahn, 1999). This high drop out rate means that while it is possible to predict with some degree of accuracy that a certain proportion of a sample will be high-rate offenders, prediction at the individual level of analysis is plagued by the stochastic sequence at hand. For example, assuming 10% of a sample will be high-rate o ffenders, predicting which particular individuals

PAGE 35

35 will be career criminals, is beset by the fact th at 20% to 30% of even that high risk group will desist after each additional offense. Alfred Blum stein proposed a now classic explanation of this phenomenon, saying: Any stochastic sequence of ev ents with a non-zero probabil ity of termination after an event will inevitably result in a distribution of sequence lengths. In crim inal-career termssince every statistical distribution has to have a ri ght-hand tail, the group of chronic offenders who comprise the right-hand tail will necessarily account for disproportionately large number of offenses. The critical questi on is whether the members of this group are distinguishably differentThe fundamental policy question, th en, is whether the chronic offenders are identifiable in prospect, that is, during the period in which they accumulate a recordunless such discrimination can be made, any identifi cation of chronic offenders can only be made retrospectively, and so is of little policy or operational value (1979; reproduced in Auerhahn, 1999:709; and Petersilia, 1980:374). The aging out phenomenon has led to the necessity to look for other factors that are more predictive of re-offending than is the prior crim inal history of the offe nder. Otherwise, by the time the individual has enough priors to be deemed hi gh-risk, that individual is likely to be near the end of his/her criminal career or at the very least decreasing in frequency of offending. Gottfredson and Hirschi (1986) argue the 20-20 hinds ight of criminal career research to be misleading in that it turns out that the partic ular career criminals iden tified in criminological research are no longer active and their replacements cannot be iden tified until they too are on the verge of retirement (pp.217). These scholars se e incapacitation policies doomed to failure in that, with the knowledge that crim e declines with age and peaks around 17 years of age, then an incapacitation strategy must intervene at an opti mal age to prevent the most frequent offending

PAGE 36

36 (before 17 years of age), suggesting lengthy incar ceration of young juveniles in the interest of prevention. A further argument criminal career research brings is the notion of specialization. The habitual offender legislations call for longer sent ences for violent offenders. This contradicts lifecourse criminology research, which finds little to no evidence of specializ ation, especially over the long-term, for high frequency criminals (A uerhahn, 1999; Blumstein et al., 1986; Blumstein et al., 1988; DeLisi, 2005; Farri ngton et al., 1988; Kempf, 1987; Piquero et al.; 2003; Simon, 1997; Wright & Rossi, 1986).4 Legislation aimed at targeting o ffenders who specialize in violent crime is misguided as this vast literature shows these offenders may not even exist. Errors in Prediction There exist two forms of error in the predicti on of offender classification based on risk to re-offend. In the event an individu al is predicted to be a low ra te offender but actually commits crimes at a high rate, the error is one of unde rprediction. These errors of underprediction are called false negatives. These erro rs involve threats to public sa fety by predicting low offense rates for individuals who will in fact end up committing offenses at a high rate. The second type of error, while less of a public safety risk, is perhaps more ethically ch allenging. This type of error involves the prediction of an individual as being a high rate offender, who would in reality commit offenses at a low rate. These false positives lead to the incapacitation of individuals to prevent them from committing future crimes, which would actually not be committed. False positives are also disturbing due to the high unnecessary costs and wasted resources of 4 While the studies referenced indicate offenders having low levels of specialization, especially over the long-term, recently Sullivan and colleagues (2006) have demonstrated empirical evidence of a considerable amount of shortterm specialization with regard to their offending patterns (p.221).

PAGE 37

37 incarceration of individuals for prolonged periods of time that would not have been a threat to commit future offenses. To reduce the number of false negatives, a larg er fraction of the actua l recidivists must be identified, thus requiring the res earcher to lower the cutoff scor e in the prediction instrument to make the predictions more inclusive. This will, however, increase the number of false positives; increase the already high incidence of predicting individuals as being high rate offenders, who would in reality, commit offenses at a low rate. Conversely, to reduce the number of false positives, the cutoff score would have to be raised to be less inclusive; yet this results in the instrument missing an even larger number of actual recidivi sts (von Hirsch & Gottfredson, 1984). The essential fact to gather is that reducing the occurrence of one type of error will in all cases increase the occurr ence of the other type. The extent to which these two types of errors oc cur in the prediction of risk to re-offend is substantial. Research reveals th at most assessments predicting offender risk have false positive rates greater than fifty percent (Auerha hn, 1999; 2006; Blackmore & Welsh, 1983; Cohen, 1983; Monahan, 1981; von Hirsch & Gottfredson, 1984). High error rates in the pr ediction of any low base rate phenomenon or rare outcome have been found to be quite common; the rarer the event, the greater the tendency to overpredict (Cohen, 1983; Copas, 1983; Farrington, 1987; von Hirsch, 1972; von Hirsch & Gottfredson, 1984). Th e seminal selective in capacitation strategy outlined in the above-mentioned report by Greenwood and Abrahamse (1982) resulted in a fifty one percent overall success rate in classifying offenders. The rate of classifying an offender as high risk who in fact is not (false-positive rate) was forty eight percent. In four attempts to reanalyze the Rand data used by Greenwood and Abra hamse, a false-positive rate of fifty five

PAGE 38

38 percent was discovered (Auerh ahn, 1999; 2006; Chaiken & Chaike n, 1982; Cohen, 1983; Visher, 1986). The Greewood study also evidenced a false-negati ve rate of about si xteen percent (von Hirsch & Gottfredson, 1984). This corresponds to one of every six offenders classified as medium or low risks actually committing frequent offenses. These misses or failures to imprison offenders assessed at lower risk leve ls, yet actually high risk may cause political pressure to make the definition of high risk mo re inclusive (which as mentioned above would increase the rate of false positives). Prediction Method The primary task of risk assessment is to pred ict, not to explain the behavior in question (Gottfredson & Snyder, 2005). Pred iction involves first assessing pa rticular criteri on variables at one time to forecast a specific behavior or performance at a later time (von Hirsch & Gottfredson, 1984). Currently, there exists an extensive history of risk prediction of offending behavior in the criminal and juvenile just ice fields (Gottfredson, 1967; Gottfredson & Tonry, 1987; Mannheim & Wilkins, 1955; Simon, 1971). Prediction has certainly progressed both in methodology as well as accuracy and can be groupe d into four generations (Andrews & Bonta, 2003). Originally, risk to re-offend was predicted using clinical/professiona l judgment; the gut feeling approach. A variation of this first-ge neration assessment is called structured clinical judgment which reflects decisions made after revi ew of specific factors, but without a validated process to link the scores to the decisions ma de (Andrews et al., 2006). Meta-analyses have illustrated the weak predictive validity of firs t generation assessment with an overall mean r of .12 (Andrews et al., 2006; Grove et al., 2000). While only available for sex offender samples

PAGE 39

39 thus far, structured c linical judgment assessments appear to do better than unstructured, yet not as well as later generation assessment (Andrews et al., 2006 ; Hanson & Morton-Bourgon, 2004). Prediction, as the new penology took over, evolve d into utilizing statistical techniques correlating static characteristics of the individual (such as age, race, prior criminal history, and prior substance abuse history) with the depe ndent behavior of offe nding/re-offending but was atheoretical (Andrews et al., 2006; Shichor, 1997). These second generation assessments continue to be used extensively. The predic tion studies described extensively below by Greenwood and Abrahamse (1982) and Auerhahn (1999) both utilized second-generation, actuarial, static predictor instruments. Meta-analyses have found the predictive vali dity of second-genera tion assessments much higher than that for clinic al assessments, with an r of .42 (Andrews et al., 2006). Additional past research has shown that informal, subjective, clinical judgments are far less accurate than actuarial/statistical methods (Andrews & Bonta, 2003; Dawes, Faust & Meehl, 1993; Hoge, 2001; Meehl & Grove, 1996). Grove and coll eagues conducted a 136 study meta-analysis comparing actuarial approaches to clinical risk predic tion, finding actuarial superior in 47% of the studies, equally as predictive in 47% and less predictive in only 6% of the studies (Grove et al., 2000). While the Grove team analyzed only c linical psychology and psyc hiatry studies, three other meta-analyses focusing more on criterion behavior and offender sa mples, all finding the statistical/objective risk measures outperformed clinical judgment (Bont a et al., 1998; Hanson & Bussiere, 1998; Mossman, 1994). These studies pred icted recidivism for mentally disordered offenders, violent behavior among psychi atric patients and sexual offending. The third-generation of prediction examined both static factors and dynamic factors (such as antisocial attitudes and beliefs). Meta-analysi s or even accessible reports of predictive validity

PAGE 40

40 could only be ascertained from one third genera tion instrument, the Level of Service InventoryRevised (LSI-R). Results show the predictive cr iterion validity for this instrument to have an overall mean r of .36 (Andrews et al., 2006). This is much better than the r of .12 for first generation, yet lower than th e .42 reported for second-genera tion prediction instruments. Finally, the fourth-generati on of risk assessment has attempted to assess static and dynamic risk factors as well as pr otective factors and follows servic e and supervision from intake through case cessation (Andrews et al., 2006). It is the addition of evaluati ng protective factors that may help reduce the likelihood of re-offendi ng and the guidance with the creation of case management plans that advances the latest generation of prediction. Risk a nd protective factors are not, however, mutually exclusive categories. Fo r example, family may be a risk factor for one juvenile with abusive parents who engage in criminality themselves, and protective for another youth who has a supportive, consistent family structure. The inclusion of both risk and protective factors highlights one of the distinguishing characteristi cs of fourth-generation risk assessment in the area of increased attention to the linkage between assessment and case management (Andrews & Bonta, 2003). Individua lized strengths are bu ilt upon to construct a prosocial orientation and responsiv ity factors and learning styles are considered when placing for treatment. As in the case for third-generation instrument s, the predictive valid ity of only one fourthgeneration instrument has been reported. Andr ews and colleagues (2006) report an estimate of .41 for the Level of Service/Case Management Inventory (LS/CMI). Andrews and colleagues (2006) do report that the best know fourth-generation inst ruments are the Correctional Assessment and Intervention System (CAIS) and the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS).

PAGE 41

41 Fourth-generation assessments appear to be eq ually as valid in terms of predictive criterion validity as the more popular, yet significantly more simplistic, second-generation instruments. Andrews and his co-authors furthermore suggest substantial improvements in the predictive criterion validity of risk assessments may resi de in reassessments of dynamic risk factors (2006:16; see also Andrews & Robinson, 1984; Mo tiuk et al., 1994; Roynor et al., 2000). This notion is supported by two recent dissertations showing increased predictive validity with reassessment as well as assessment of dynamic risk factors that can change rapidly (Brown, 2003; Law, 2004). Andrews and colleagues (2006) suggest that re-assessment with a focus on rapidly changing risk factors may double or even triple the variance explained by intake assessments, touting the idea a breakthrough (2006:16; see also Hanson & Ha rris, 2000; Quinsey et al., 1997; Zamble & Quinsey, 1997). Prediction Studies Most prediction studies vary in degrees of complexity, yet utilize essentially the same method (von Hirsch & Gottfredson, 1984). The resear ch team collects data from a sample of convicted offenders, usually criminal histor y and varying degrees of social history, demographics, education, drug abuse, family history, etc. (Gottfredson & Snyder, 2005). The sample participants are followed for some length of time and recidivism is captured for those who commit subsequent offending (eith er self-reported or official). The original data collected is examined in how it correlates to the subsequent recidivism, using bivariate or multivariate statistical techniques. The factors that correlate strongly with the recidivism are construc ted into a predictive instrument that is then tested on a validation sample. The final proceess of testing against the validation sample is the most essential as it s hows how well the instrument actually predicts offending, rather than how well it correlates to the offending of the population from which the

PAGE 42

42 data were gathered (Gottfredson, 1967; Gottf redson & Snyder, 2005; Mannheim & Wilkins, 1955; Simon, 1971). When an item is a statistica lly significant predic tor of the dependent variable in question, the item im proves the power of the instrume nt and should be used in the construction of the scale (G ottfredson & Snyder, 2005).5 Five basic steps have been shown essen tial in developing and utilizing prediction instruments (Gottfredson & Snyder, 2005). First, the criterion to be predicted must be defined. Second, the predictors are select ed. Third, the relationships be tween the predictors and the criterion must be measured. Fourth, the relations hips found must be validated. Lastly, the method of prediction must be applied in si tuations for which it was developed. The steps outlined above have led some sc holars to assert the seminal Greenwood and Abrahamse study was not a prediction of risk to re-offend study at all (Auerhahn, 1999; Chaiken & Chaiken, 1984; Clear and Ba rry, 1983; Visher, 1986; von Hirsch & Gottfredson, 1984). Greenwood and Abrahamse used offender characte ristics and self-report ed past delinquency from a single questionnaire to assess the correlation of the former w ith the later. The authors then concluded that certain offende r characteristics are predictiv e of offending behavior and devised their seven-item scale. This process contradicts several of the steps of constructing a prediction study in that Greenwood and Abrahams e do not have sequentia l time periods, only the one-time questionnaire. They are in effect ex amining how current offender characteristics 5 Gottfredson and Snyder (2005) caution regarding the ethical implications of including certain significant correlates in the scale construction which may act ually simply be proxies for unmeasur ed items (such as race for poverty, school failure, amount of police presence in the community, etc.). Any recidivism differences captured may be correlated with the proxy but may not be caused by the proxy. They expand, stating that statistically including the race variable improves the predictions ma de by the risk scale Ethically, excl uding the race variable prevents the assignment of risk scores that may inappropriately lead to greater sanctions for minority youth (pp.iii). Gottfredson and Snyder suggest including these vari ables (such as race) in the early st ages of scale construction and then removing them from the published instrument, thus removing the bias. If the race variable was never included in the construction, other measures would be proxies for race and would keep racial bias inherent in the instrument, yet hidden beneath the surface.

PAGE 43

43 correlate with past offending behavior, and not analyzing how current characteristics predict future behavior (which presumably was the intention). Furthermore, Greenwood and Abrahamse did not utilize a validation sample once the original scale was created, in or der to assess shrinkage. A replication/validation sample is essential to determine the genera lizability of the pred iction instrument, as opposed to correlating offending behavior with the original samples idiosyncrasies (Gottfre dson & Snyder, 2005; von Hirsch & Gottfredson, 1984). Shrinka ge refers to the portion of va riation that is lost when the prediction instrument is validated on the replication sample; us ually the instrument accounts for a smaller portion of variation on a subsequent sample than the original sample (Gottfredson & Snyder, 2005; von Hirsch & Gottfredson, 1984). Shrinkage is especially true with retrospect ive studies where the re searchers are typically not blind to outcomes in that they already are aware of which individuals in the samples have recidivated. Researchers can then fish for combin ations of predictors that enhance the accuracy of what is actually a postdic tion which potentially capitali zes on unique features of the original sample -features that will not be pr esent in other samples (Lanza-Kaduce et al., 1990:28). Therefore, without the inclusion of a validation sample, the Greenwood and Abrahamse study overstates the accuracy of their se ven-factor scale. Research has consistently shown that simple, less sophisticat ed statistical techniques are less vulnerable to shrinkage, such as the Burgess basic additive approach, in comparison to complex multivariate techniques (Lanza-Kaduce et al., 1990; Gottfredson & Go ttfrdson, 1979; Gottfredson & Snyder, 2005; Wilbanks, 1985). In addition to shrinkage consider ations, the predictive accuracy of simpler, easier to implement techniques doe s not differ significantly from mo re elaborate, sophisticated techniques (Lanza-Kaduce et al., 1990).

PAGE 44

44 Auerhahn (1999) replicated the Greenwood and Abrahamse scale on a representative sample of California state pris on inmates (n=2,188). The replica tion scale used identical, or proxy measures as the Greenwood and Abrahamse scale. The seven items in the replication include: 1) Prior conviction for any offense, 2) Free at least one year before the current incarceration, 3) Arrest before ag e sixteen, 4) Served time in a juvenile detention facility, 5) Regular use of heroin or barbiturat es at any point in inmates life, 6) Heroin or barbiturate use as a juvenile, and 7) Employed before arre st resulting in current incarceration. Identical to the Greenwood and Abrahamse study, the seven predictor variables were summed and offenders grouped according to low, medium-, and high-risk. Offenders were classified as low-rate with sc ores 0-1, medium-rate with scor es 2-3, and high-rate those with scores of 4 or higher (Auerhahn, 1999). The repl ication scale resulted in 41% of the sample being low-risk, 31% medium-risk, and 28% being cl assified as high risk to re-offend. Auerhahn (1999) next compared the predict ive risk to re-offend with the o fficial prior convictions of the sampled individuals. She notes the predictive accuracy (the percenta ge of offenders placed in the right cell) of the Greenwood and Abrahamse scal e was 51%, while the rep lication scale produced slightly better accuracy at 60% (Auerhahn, 1999). Furthermore, 13% of the offenders who are not high-rate were predicted as such (false positives) by the Greenwood and Abrahamse scale, and only 5.6% were reported so by the replication. However, upon closer examination of the Auerhahn study, the Greenwood and Abrahamse scal e had 14% of the offenders not high-rate predicted as such (not the 13% reported) and the replication scale produced 18% of all the offenders being predicted high-rate who in actuality were not (not the 5.6% reported). This means that only 15% of all the offenders pred icted high risk in the Greenwood and Abrahamse study (not the 29% predicted), and only 10% of the offenders predicted high risk in the

PAGE 45

45 replication study (not the 28% pr edicted), were in actuality hi gh risk to re-offend. The two studies, therefore, demonstrate false positive ra tes around fifty percent for all the individuals predicted as high risk to re-offend as nearly half of those predic ted high risk actually recidivated. Cronbachs alpha, a measure of internal reliability for multiple-item scales based on the strength of correlations between the items in the scale, is reported as .38 (Auerhahn, 1999). Alpha values range from 0 to 1, with 1 representi ng perfect reliability of the assessment tool. An alpha value of .70 or higher is usually recognize d as acceptable for reliab ility in the social sciences (Auerhahn, 1999). Auerhahn (1999) reports all of the inter-item co rrelations of the scale significant at the p=.01 level, ye t cautions that some associati ons are small enough that their significance is attributable mainly to the size of th e sample, as it is easier to achieve significance with larger samples. Also pointed out is the fi nding that many of the associations are in fact negative, though coded for agreement, which is c ounter to the requirement of positive direction association for additive scales. The Auerhahn (1999) study intentionally elaborat es on the question of internal reliability due in part to the brevity or outright exclusion of the issue in previous research. The study mentions prior research focusing on the reliabil ity of using inmate self-report data (Greenwood and Turner, 1987; Spelman, 1994; von Hirsch & Go ttfredson, 1984), estimation of incapacitation effects of selective incapacitation strategies, or the ethical issues su rrounding the basis of sentencing decisions on prediction of future o ffending, rather than actions the offender has decided to take, yet little atte ntion to the methodological issues inherent in the construction of prediction instruments (Auerhahn, 1999; Gottf redson & Gottfredson, 1985). Auerhahn (1999) presents the internal reliability, false positive ra te, and scale alpha values for the purpose of

PAGE 46

46 illustrating the shortcomings of actuarial risk assessment in terms of driving a policy of selective incapacitation. Even as research shows actuarial risk assessm ent superior to clini cal judgment and gut feeling approaches, there is st ill a roughly 50% false positive rate of predicting offenders to be high-risk when, in actuality, their offending pa tterns will not play out that way. Greenberg (1975) cautioned over thirty years ago the ina ccuracy of prediction in cr iminal behavior suggests the selective incarceratio n of individuals on a predictive ba sis was not a promising direction. Nearly twenty-five years later, Auerhahn concluded that the obstacle to realizing this seemingly perfect solution to crime prevention lies in the prospective identification of this offender poolwe simply cannot do it with a ny reliable accuracy, and that it appears th at the time has come to open the discourse of crime and crim e prevention to permit the consideration of strategies that do not focus on incarceration in formulating our response to the crime problem (1999:727-728). What Works? The nothing works mantra sparked by the Martinson report pushed corrections policy into relying on incapacitation as its predominat e strategy with regard to addressing crime, yet also spawned a quite contradictor y line of inquiry; that of the what works? movement. This empirically based research implies that cert ain programs and interventions, when properly targeted, reliably produce reductions in recidivism (NIC, 2004). Th is line of inquiry has been fueled mainly by the results of meta-analysis, cost-benefit analysis, and clinical trials, all providing indications of how to more effectively reduce the likelihood of re-offending (Andrews et al., 1990a; Andrews et al., 1990b; Andrews et al., 1999; Ant onwicz & Ross, 1994; Aos, 1998; Cullen, 2002; Gendreau, 1996; McGuire, 2002; NI C, 2004; Sherman et al., 1997). The metaanalysis technique has been especially instrument al in the identification of effective intervention

PAGE 47

47 characteristics and permits the aggregation of a large body of research li terature in order to examine and compare the effect sizes for trea tment versus control group comparisons. Metaanalyses summarize a large number of studies (a rigorous study of studies ) while controlling for important methodological i ssues (Sherman et al., 1997). What works is now often used interchangeab ly with the term evid ence-based practices (NIC, 2004). Evidence-based practices (EBPs) ha ve recently become a trend throughout the social sciences and human service fields and focus primarily on outcome. The outcome with regard to criminal justice practices is almost exclusively the reduction of risk to re-offend and subsequent offending. Bogue and co lleagues stress that interven tions within corrections are considered effective when they reduce offender risk and subsequent recidivism and therefore make a positive long-term contribution to public safety (NIC, 2004:1). Research on offender rehabilitation has progressed to the point where there exists sufficient evidence of several key interdependent principles (Andrews & Bonta, 2003; Burrell, 2000; Carey, 2002; Corbett et al., 1999; Currie, 1998; Latessa et al., 2002; McGu ire, 2002; NIC, 2004; Sherman et al., 1997). Meta-analyses have been used both to identify i ndividual risk factors asso ciated with recidivism, as well as to determine the characteristics of the most effective delinquency treatment programs (Andrews & Bonta, 2003; Andrews et al, 1990; Lipsey, 1989, 1992; Palmer, 1992). Research and practitioner communities have moved progressively from nothing works, to what works, and are now learning what in fact makes what works work (Andrews, 2006; Antonowicz & Ross, 1994). Empirical Predictors Meta-analyses have led to the ability to id entify major and minor risk/criminogenic need factors with respect to offending behavior (Andr ews et al., 2006; Bonta et al., 1998; Gendreau et al., 1996; Hanson & Morton Bourgon, 2004; Lipsey & Derzon, 1998). The major risk factors can

PAGE 48

48 be divided into the big four and the central eight (Andrews et al., 2006; Andrews & Bonta, 2003). The big four include 1) history of antisocial behavior, 2) antisocial personality pattern, 3) antisocial thinking, and 4) antisoci al peer associations. The four additional risk factors rounding out the central eight include 1) family probl ems, 2) school/work problems, 3) procriminal leisure/recreation activities, and 4) substance abuse. Minor risk factors include such issues as emotional distress, mental disorders, low soci o-economic status, seriousness of the presenting offense, low IQ, and physical health problems. Research has demonstrated the minor factors to be of relatively low predictive validity, if present at all, and that they most likely can be explained away by the contributions of the big four, whose inclusion pushes the minor factor s to insignificance (Andr ews et al., 2006). For example, the predictive power sometimes associated with mental disorder s relation to offending is believed in actuality to be rela ted to antisocial attitudes/thinki ng, antisocial personality pattern, and substance abuse (Andrews et al., 2006; Link et al., 1992; Swanson et al., 1996). Furthermore, the power associated with low so cioeconomic status, or socially disadvantaged neighborhoods can be attributed to the increased prevalence of the bi g four risk factors within the individuals with lower socioeconomic status an d the residents in those disadvantaged areas (Stouthamer-Loeber et al., 2002). The meta-analyses uncovering the big four and the central eight risk factors demonstrate a significant reduction in recidivism when the fact ors are targeted through effective interventions (Andrews et al., 2006; Andrews & Bonta, 2003). It is essential to understand the risk associated with each of the eight major factors, as well as the dynamic need of the individual in terms of what is needed to address the risk. It is a well-known criminological mantra that the best predictor of future behavior is past behavior (Sherman et al ., 1997). The first of the big four,

PAGE 49

49 history of antisocial behavior, is the often-utilized criminal hist ory item present in predictive instruments since the inception of second-ge neration assessments (Andrews & Bonta, 2003). This item is characterized by ear ly and continuous involvement in deviant behavior in a variety of situations and manifestations. Individuals pr esenting with prior criminal history (be it selfreport or official) are in need of skills to id entify non-criminal alterna tive behaviors in risky situations. The second factor, antisocial personality pattern, is ch aracterized by an impulsive, pleasure-seeking, and risk-taking personality. This risk factor, though believed to be dynamic in nature in this context, is sim ilar to Gottfredson and Hirschis ( 1990) concept of low self-control. Individuals presenting with anti social personality patterns place a high value on adventurous thrill-seeking and immediat e gratification, which increases the probability of engaging in deviant behavior. This dynamic criminogenic need can be addressed by building problem-solving, anger management, and coping sk ills (Andrews et al., 2006). Antisocial attitudes and thinking is the third of the big four risk factors. This factor is characterized by attitudes, belief s, and values that are supportive of criminal behavior. These procriminal sentiments represent the pool of ju stifications and exonerati ng statements that the person has available in any particular situ ation (Andrews & Bonta, 2003:178). Antisocial cognitions enable individuals to rationalize devian t behavior, thus increasing the likelihood of its occurrence. Individuals with these beliefs and att itudes are in need of cognitive restructuring to reduce antisocial attitudes and recognize risky thi nking triggers, so as to engage in alternative prosocial behaviors. The fourth of the major risk factors which need to be targeted to re duce recidivism is the extent of antisocial peer associations. Devian t peer associations have been the subject of

PAGE 50

50 countless empirical investigations and their influences have pr oven significant and withstood the test of time (Akers, 1998). Associ ation with criminal peers and the seclusion from pro-social peers leads to procriminal modeling and adjustment s to the reward and cost structure associated with deviant behaviors, thus making the likelihood of continua tion of those behaviors more likely (Akers, 1998). This dynamic criminogenic need can be addressed by reducing the associations with antisocial others and enhancing the expos ure to prosocial peers. Negative family situations are the fifth identified risk factor of the central eight. The key elements of this need are nurturing and suppor t, and monitoring and supervision. Individuals who are reared in environments which lack suppo rt and monitoring of behavior are at increased odds of engaging in deviant behavior (Andrews et al., 2006). The notion of behavior monitoring is central to Gottfredson and Hirs chis explanation of deviant beha vior with their General Theory and its emphasis on the mon itoring, recognition, and consiste nt punishment of deviance (Gottfredson & Hirschi, 1990). A ddressing this dynamic need re quires the reduction of conflict within the family environment, enhancing the three areas of e ffective parenting of monitoring, recognition, and punishment, and building positive relationships within the family structure. The sixth factor identified in the empirical research is in the domain of school and work problems. The intent of this factor is to identif y low levels of performance and satisfaction in the areas of school and employment. Enhancing the individuals perf ormance in school or work and improving satisfaction or increasing rewards is an important way to address this dynamic need. Problems in the area of recreation and leisure ac tivities are the sevent h risk factor of the central eight. This factor is composed of low levels of involvement in structured prosocial activities and pursuits. A low level of satisfacti on with pursuing prosocial activities may also be

PAGE 51

51 involved. Ways to address this dynamic need in clude increasing the rewards associated with engaging in anti-criminal activities, and enha ncing the involvement in prosocial pursuits. The final risk factor domain is substance a buse. This simply points to the increased likelihood of offending or re-offending for individua ls who abuse alcohol or drugs. In order to address this dynamic need, a reduction in the abus ing behavior must occur. Drug education has been found successful, but only when accompanied with building relapse prevention skills to identify triggers and alternativ e behaviors (Sherman et al., 1997). Historically, second generation prediction instruments have utilized an individual s history of substance abuse, but newer third and fourth generation assessments have begun to se parate prior use from current use of alcohol or drugs. Principles of Effective Intervention A meta-analysis of 374 controlled experimental studies of the treatment effects of various interventions on recidivism was conducted by th e Carleton University (Andrews, et al., 1999). This analysis helped to reveal three principles of effective intervention: 1) the risk principle, 2) the need principle, and 3) the responsivity princi ple. The risk principle expresses the finding that intensity of treatments and interventions should mi mic the intensity of the risk to re-offend. More intense services should be deliv ered to the highest risk youth (Andrews & Bonta, 2003; Harland, 1996, McGuire, 2002; Sherman et al., 1997). Crim inological literature actually shows high intensity services delivered to low-risk to re-offend youth are iatroge nic; having the negative consequence of actually increasing recidivism. It is important to remember the risk to re-offend concept is not to be confused with the seri ousness of the presenting offense for which the individual may have come in contact with the criminal justice system for (Andrews, 2006). An individual may be lower risk to re-offend, re gardless of the apparent heinousness of the presenting offense.

PAGE 52

52 Studies spanning nearly thirty years have obser ved the negative effect of providing intense services to lower risk to re-offend youth (Andr ews & Kiessling, 1980; Baird et al., 1979; Bonta et al., 2000; ODonnell et al., 1971). In these studies, low-risk to re-offend individuals had higher recidivism percentages when they received intens ive treatment compared with low-risk to reoffend individuals who received minimal treatment and high-risk to re-offend individuals had lower recidivism percentages when receiving intensive treatment compared to high-risk receiving minimal treatment, all illustrating the im portance of the risk principle. Furthermore, a greater percentage of high -risk individuals have a need for prosocial skills and thinking and will demonstrate a greater improvement, as lower risk individuals re-offending is a much lower baserate phenomenon to begin with, thus making for a greater bang-for-the-buck as well (Andrews & Bonta, 2003; Gendreau, 1996; Harland, 1996; McGuire, 2002; NIC, 2004; Sherman et al., 1997). The need principle, the sec ond principle of effective inte rvention, holds that services provided should address individu alized criminogenic needs relevant to each individual. Criminogenic needs are dynamic, changeable factor s associated with re-o ffending behavior that, when addressed or changed, affect the individua ls risk to re-offend (N IC, 2004; Sherman et al., 1997). These needs are in contrast to static factors, such as age at first offense, prior substance abuse, or gender, which, while strong predictors of recidivism are not amenable to treatment and can not be change through intervention. The stronge st dynamic correlates of crime are anti-social peer relationships, family factors, substance abus e, and antisocial attitu des, values and beliefs toward authority, education, and employment a nd low self-control (Ge ndreau, Andrews, Cogin & Chanteloupe, 1992; NIC, 2004). Seven of the centr al eight major risk factors outlined above (all except history of antisocial behavior) are dynamic needs that can be addressed through

PAGE 53

53 intervention/treatment. Programs and interventi ons successful in reduc ing these criminogenic needs can expect corresponding re ductions in recidivism, esp ecially when the needs are prioritized through valid assessment so that se rvices are focused on th e greatest ones (Andrews & Bonta, 2003; Gendreau et al., 1994; Harland, 1996). An important caveat to the need principle is the notion of appropr iateness. Individuals should only be treated, or receive interventions targeting, an actual need they possess. Research has shown it to be iatrogenic to provide bl anket approaches and nontargeted services to individuals not presenting with the specific need addressed (A ndrews, 2006; Brown & Campbell, 2005). For example, providing substance abuse servic es to an individual free of substance abuse problems may lead to an actual increase in r ecidivism (Brown & Campbell, 2005). This is in direct opposition to those, operati ng opposite to empirical findings, holding that simple education (such as substance abuse education) will deter in dividuals from every following a certain path. Andrews suggests appropriate targets including building up low-risk alternative behavior in high-risk situations, reducing antisocial associations, and building up levels of reward, satisfaction, and monitoring in home, school work, and leisure settings (2006:598). The third principle derived from the Carlet on University databank was the responsivity principle. There are two types of responsivity, ge neral and specific. General responsivity asserts that behavioral, social learning, and cognitive-behavioral strategies are the most effective when the preferred outcome is recidivism reduction (A ndrews et al., 1990c; Andrews et al., 2006; Izzo & Ross, 1990; Lipsey, 1992). These strategies are more effective due to their use of techniques such as role-playing, modeling, graduated rein forcement, and skill-building (Hubbard, 2002). Lipsey (1992) evaluated over 400 studies in a me ta-analysis and concluded that not only was

PAGE 54

54 treatment more effective than control groups, but that type of treatment was the most important variable. Cognitive-behavioral treatment is at the forefront of correctional treatment. Behavior therapy is a technique based on the principle that environment has an effect on learning and behavior through operant and classical conditio ning (Hubbard, 2002). Behavior therapy assumes modeling to be essential in le arning and that offenders have learned antisocial behaviors by watching others and the rewards/punishments ach ieved, and by being rewarded themselves for behaviors in the past. Antisocial behaviors, and attitudes conducive to them, are continued as a result of these reward continge ncies. Behavioral therapy for o ffenders utilizes techniques of behavioral rehearsal, modeling, and token economies (Hubbard, 2002). Cognitive therapy posits errors in thinking l ead to problematic behavior (Yochelson & Samenow, 1976, 1977). This faulty thinking, coupled with offenders impulsive tendencies, lack of problem solving skills, and failure to consider the consequences of their actions, is what drives delinquency (Hubbard, 2002). Cogni tive therapy focuses on changing these thinking processes and patterns, and not on directly changing the behavior. Cognitive-behavioral therapy is a combination of the aforementioned two therapies and includes both a focus on the cognitions of offenders and changing those patter ns, as well as utilizes behavior al strategies of modeling, roleplaying, behavioral rehearsal, and reward/puni shment contingencies. Cognitive-behavioral therapy teaches offenders how to develop self -conrol, manage their anger more appropriately, develop empathy through role playing, improve pr oblem solving abilities, and develop their level of moral reasoning (Hubbard, 2002:38). Specific responsivity is based on the assumption that not all offenders are alike and certain personal characteristics of offenders have the pos sibility of interfering with treatment (Andrews

PAGE 55

55 & Bonta, 2003). Specific responsivity argues for the need to match services provided to the learning styles, personality, motivation, deve lopmental stages, mental disorders, and demographics such as age, gender, and ethnicity of the individual (Andr ews et al., 2006; Gordon, 1970; Miller & Rollnick, 2002). Specific respons ivity is based on the assumption that the personal characteristics mentioned above may interf ere with treatment and therefore need to be addressed. The essential issue of specific responsivity is the moving from what works in correctional treatment to which cognitive-behavior al curriculums work best for which offenders. There exists a plethora of empirical research on general responsivity, while little more than theoretical speculation and hypothe sizing has been undertaken with regard to the notion of specific responsivity. Van Voorhis (1987) holds that program evaluations will mask the effectiveness of treatment if offender classifi cation, in terms of risk level, personality characteristics, and other responsiv ity characteristics, is not utili zed. This lack of classification and matching offenders to appropriate treatment may lead to the assignment of individuals to harmful interventions that may even increase their risk of re-o ffending, making treatment programs appear to be failures when they may actually just have been misdirected. One exception to this gap in specific responsivity literat ure has been a dissertation conducted by Hubbard (2002). The purpose of the study was to ascertain whether the characteristics of gender, depression, self-esteem, history of sexual abuse, intelligence, or personality affect treatm ent effectiveness. The question is whet her these characteri stics can affect whether individuals are successful in treatment, are capable of understanding the treatment, can focus while in treatment, and has the ability or willingness to alter be havior (Hubbard, 2002). Of the characteristics listed above only gender and history of se xual abuse were significantly related to subsequent recidivism, with males and individuals with a histor y of sexual abuse more

PAGE 56

56 likely to recidivate. Gender rema ined significant after controlli ng for risk to re-offend level, quality of programming, sexual abuse, depression, and self-esteem; all possible variables for which gender could possibly have been a proxy meas ure. Contrary to previous work (Fabiano et al., 1991) intelligence was not related to the ou tcome measure of recidi vism, and those with lower IQs performed equally we ll as those with higher IQs. Hubbard (2002) offered the possibility that the significance of gender may be due to different staff quality, leadershi p, programming or overall implement ation of the curriculum for females. Another explanation offered was th at perhaps women are more receptive to the strategies of cognitive-behaviora l treatment. While possible, th is position does not explain the fact that females, even in no-treatment control groups, always recidivate at a lower rate than males. The final explanation offered for the ge nder significance was that perhaps females are more motivated or able to take responsibil ity for their actions. This conclusion remains speculative, as noted by Hubbard (2002), since no control for or assessment of motivation was made. Overall, Hubbard (2002) f ound little support for specific re sponsivity and the possibility of personal characteristics impacting treatment effectiveness, except for gender and history of sexual abuse. The possibility that cognitive-behavioral treatment succeeds at reaching a wide variety of learning styles may ne gate the effects of personal char acteristics on treatment success (Hubbard, 2002). The Carleton meta-analysis uncovered the need for services provide d to use behavioral, social learning, or cognitive-behavioral strategi es rather than unstru ctured, insight-oriented, nondirective counseling or get tough approaches such as deterrenceoriented or Scared Straight interventions (Andrews & Bonta, 2003; Sherman et al., 1997). The structuring strategies that have been found to work effectively include differential reinfor cement, motivational

PAGE 57

57 interviewing, practicing of new skills, and m odeling (Andrews, 2006). The meta-analysis found that when correctional interventions adhered to the risk, need, and respon sivity principles, the mean effect size was .26 in sixty test s of treatment. This is in contrast to .18 in eighty tests when only two of the principles are followed, and .02 in 106 tests when only one of the three are followed (Andrews & Bonta, 2003). Recently, Andrews has incorporated a few more essential elements of effective programming, which he terms the supplementary principles of integrity and intervention implementation (Andrews, 2006). These supplementa ry elements deal with management and organizational aspects and help make for a more complete picture of what is needed for rehabilitative effectiveness (Andrews, 2001; Andrews & Bonta, 2003; NIC, 2004). The first additional element is the necessity of implem enting and validating a structured risk/need assessment instrument. Differential risk de signations and targeting of individualized criminogenic needs is critical to case management and treatme nt progression. Simply having effective instruments is not enough, if they are ne ver fully implemented and utilized in the case planning process (Goggin & Gendreau, 2006). These el ements illustrate the need for validated assessment instruments and demonstrate the import ance of the current study, which attempts to examine the predictive ability, and construct validity of the PACT assessment. Andrews (2006) also argues for the necessity of having program managers or directors and supervisors cognizant and a ttentive to the relationship skills and the structuring skills of the service delivery staff (ability of delivery staff to employ the responsivity principle). Administrators or supervisors must carefully se lect, train, and supervise delivery staff according to the risk/need/responsiv ity principles to ensure complian ce and proper implementation of the program. Staff competence has been found criti cal in enhancing the effectiveness of

PAGE 58

58 programming and intervention to reduce recidivism as evidenced by Barnos kis (2003) analysis of the conditions necessary for blueprint progr ams to successfully work. This principle was further evidenced by the possibility of the failure of Project Green light to be attributed to two case managers with inadequate training and skills in program delivery (Brown & Campbell, 2005). Additional research has identified other princi ples of effective evid ence-based practice. The fidelity principle holds that it is critical to monitor the implementation of treatment programs to ensure the curriculums are delivered in acco rdance to the way in which they were intended and designed. Many programs proven to be eff ective at reducing recidivism are delivered according to manualized curriculums that are partially scripted so as to ensure all facilitators deliver the lessons in identical fashion. The fideli ty principle holds that it is beneficial to periodically monitor the de livery of treatment groups and coach facilitators. Research has demonstrated that even sound and empirically supported programs, when not delivered with adherence to the protocol, will not have the extent of recidivism reduction evident in those programs that monitor treatment staff to ensure proper delivery (Barnoski, 2004; Lipsey, 1992; Lowenkamp et al., 2006). Simply choosing an evidence-based program from available lists, such as the University of Colorado bluep rints programs is not enough, as even model programs such as Aggression Replacement Traini ng (ART), Multisystemic Therapy (MST), and Functional Family Therapy (FFT) have failed when not properly implemented (Andrrews, 2006; Barnoski, 2004). There exists ex tensive literature that even programs that were empirically shown to be effective in one place and time were not so in another, due to differences in implementation (Welsh, 2006; Welsh & Harris, 2004). The major aim of the fidelity principle is to monitor the implementation quality and treatment fidelity to ensure programs are delivered the

PAGE 59

59 way in which they were designed and intende d to maximize program success and recidivism reduction (Lipsey, 1993; McGuire, 2002; Miller & Rollnick, 2002). Gender Differences in Offending Gender is one of the single best predictors of delinquency, as males continue to offend at higher rates than their female counterparts. Howe ver, in recent years, research has begun to discover that, while rates of de linquency are declining for bo th males and females since 1997, the decline has not been shared equally among the sexes, nor does it o ffset the substantial increase in criminal offending between 1984 and 1997 (Snyder, 2003). Snyder and Sickmund (1999) have found arrests of adolescent female s between 1987 and 1999 to have increased 76%, while arrests of adolescent males increased 42% (see also Chesney-Lind, 2001). Arrests for female juveniles increased 50.3% between 1981 and 1997, while arrests for male juveniles was observed at only a 16.5% increase, according to the Uniform Crime Reports (Chesney-Lind, 2001). A more alarming fact may be the increase in violent crime for female adolescents during the same period of 103%, while the arrests for ma le juvenile violent cr ime only increased 27% (Chesney-Lind, 2001). Despite the fact that male juveniles still account for the majority of juvenile crime (as is also true of adult cr iminality), female adolescents are unmistakably narrowing the gender gap. These statistics illustrate the necessity of examining the potential for gender differences with regard to risk and protective factors to offending as well as the predictability of gendered offender. The differen ces in the correlates of offending for males and females are central to the curren t focus of this study. While many factors have empirically been re lated to delinquency, the question at hand has to do with whether or not the factors that pred ict female adolescent delinquency are different from those that predict male juvenile criminal involvement. An additional to concern is the issue of differential significance. Perh aps, even though identical factors may predict delinquency of

PAGE 60

60 both sexes, certain factors are more important in the prediction of criminality of one sex versus the other due to di fferential exposure. Scholars have argued both sides of the coin as to whether identical prediction instruments can be utilized for females and males. Reisig a nd colleagues (2006:385) stat e, whether actuarial risk toolscan effectively gauge women offenders risk and needs continues to be debated. Scholars more orientated toward a feminist pe rspective argue that existing actuarial risk assessments ignore the gendered context of female criminality, as they were derived from male theories of crime and delinquency. These scholar s contend existing tools fail to consider the economic disadvantages, propensity for drug-related offense involvement in the criminal justice system, and the importance of prior victimization, many females face (Covington, 2003; Holtfreter and Morash, 2003; Reis ig et al., 2006). Proponents of identical risk assessment utilization maintain that the potential influence of factors high lighted by feminist scholars (e.g. victimization, poverty, substance abuse) is captured through offenders attitudes, peer associations, and the like Reisig et al., 2006: 385; see also Andrews and Bonta, 2003; Dowden and Andrews, 1999). The scholars oriented towards the feminist gendered context of female criminality believe the utility of risk assessment tools is suspect since those tool s were developed based on male theories of crime and delinquency and therefore ignore the specific risks and protective factors associated with female offending. These scholars hold the view that traditional criminological theories and empirical research were devel oped by men who brought forth assumptions about social life, society in general, and offending to predict criminality. The samples used in the studies to test and validate thei r assumptions were exclusively ma le and therefore would not be applicable or generalizable to female criminality, as feminists reject the notion that the same

PAGE 61

61 functions and correlates explain both male and female delinque ncy (Chesney-Lind, 1989; Daly & Chesney-Lind, 1988; Reisig et al., 2006; Simpson, 1989). Reisig and colleagues argue male-centered theori es fail to appreciate a variety of elements unique to female criminality, including the gr eater relative likeli hood of young women fleeing abusive homes and experiencing domestic abuse as adults (Reisig et al., 2006:388; see also Daly, 1992). Victimization during both childhood a nd as a young adult, alcohol and drug abuse, and economic disadvantage have a ll been disproportionately connect ed to female criminality in empirical literature (Broidy & Agnew, 1997; Covington, 2001; Re isig et al., 2006). Daly (1992) suggests the need to different iate women according to gendered pathways. Dalys pathway framework examines how abuse history, substance use, economic disadvantage and destructive intimat e relationships are differentially dispersed throughout the female population. Daly suggests disaggrega ting females into those who have prior histories with the aforementioned factors from those who do not, so as to more effectively predict the risk to reoffend associated with each group separately rather than as a whole. Daly suggests the need for four separate pathway groupi ngs; street women, drug-connect ed, harmed and harming, and other (1992, 1994). Prior theory and research has not made similar assertions for the disaggregation of male offenders. An additional attempt to answer the fundament al question of differential involvement and the reason for it can be extrapolat ed from Agnews gene ral strain theory. General strain theory has been found unable to explain the higher rates of male delinque ncy in terms of amount of exposure to strain, as females experience at leas t as much strain as males (Broidy & Agnew, 1997). Gender differences in the types of strain e xperienced, such as mate rial/financial, ties to others, family violence, and procedural justice, may foster the explanation of gender differences

PAGE 62

62 in crime type; i.e. what offenses females and ma les most typically engage (see Broidy & Agnew, 1997). Furthermore, gender differences in the respons e to emotional strain (particularly anger) help to explain differences in delinquency. Ma les are more likely to respond to anger with external aggressive behavior (inf licting harm on others), while fema les more likely to engage in self-directed behaviors (self-mutilation, substance abuse, disordered eating) when experiencing anger. Finally, general strain theo ry is useful for the explanati on of gender differences in that males are more likely, due to differences in copi ng strategies, social ties criminal opportunities, social control, and disposition to crime, respond to a given amount of strain with more serious and violent behaviors (Broidy & Agnew, 1997). In tandem with the gender differences described by the authors, Broidy and Agnew assert the impo rtance to note the above gender differences, to the extent that they exist, involve differences in degree rather than kind (1997:296, emphasis in the original). The picture is not one in wh ich different factors e xplain male and female criminality, but that different degrees of emphasis of those factors predict delinquency. Reisig and colleagues (2006) ar gue, the research addressing the predictive accuracy of risk assessment tools using samples of women offenders is relatively limited (pp.387). One study examining the creation of actuarial risk assessments separately for female and male offenders shows the male model to be similar to the combined sample construction, but the female model to resemble neither the male model nor the combined sample model (Funk, 1999). By creating a female specific model, the adjust ed R-squared increased from .21 for the combined sample model to .32 for the female only mode l (Funk, 1999). Interestingly enough, the male-only model achieved an R-squared of only .20, meaning the authors were able to predict female reoffending significantly better than male re-offending.

PAGE 63

63 Contrary to the view presented in the Funk study, is the meta-analysis conducted by Simourd and Andrews (1994), who examined 60 studies and 464 correlations between delinquency and risk factors to determine whethe r specific factors are more important for one gender than for the other. Thes e authors found no differences in risk factors across gender; the general risk factors important fo r male delinquency were identica l to those important for female delinquency. Furthermore, and even more essentia l is that no risk factor was found to be more important for a particular gender, however, vi ctimization was not incl uded as too few studies utilized that correlate. The inab ility to include victimization, and even more specifically to separate different types of abus e/neglect is a disadvantage in light of the gendered pathways research described below. The current study will be able to examine this factor in more detail and with substantial sample size to allow for ge nder comparison of the role of abuse/neglect. Gender differences in offending are argued to ha ve serious repercussions with respect to utilizing scores from actuarial risk/needs assess ments. For example, women charged with serious violent offenses are most often accessories to the event and not the primary instigator, a situation that would lead to over-classifi cation of risk to re-offend on any actuarial instrument considering previous engagement in a violent offense as leading to risk (Reisig et al., 2006). The overclassification argument is rejected by supporters of actuarial t ools who acknowledge a dearth of empirical research to support th e notion (Blanchette, 2002; Reisig et al., 2006). The importance of the over-classification argument is essential to the possibility of equality in the prediction of male and female likelihood of offending, and th us necessitates the in clusion of a focus on gender-specific false positive calculations. Essentia l to the current study is the need to ascertain whether female offenders predicted to offend yet subsequently abstain is disproportional to those instances with males.

PAGE 64

64 Current Focus The PACT risk/needs assessment has not yet been formally validated in terms of predicting recidivism in the state of Florida for the youth se rved by the Department of Juvenile Justice. An initial pre-validation examination used proxy measures of criminal history in an attempt to assess whether the risk levels presented closely ma tch in percentage thos e attained using the Washington State assessment upon which the PACT was based, which they in fact have. The actual PACT assessment of official criminal hist ory was not used, and no social history items were included in the scoring, as the PACT incl udes. Additional preliminary reports and analysis indicate that the PACT risk levels (as pres ented by the actual PACT assessment) do indeed predict recidivism, with higher risk to re-offe nd youth recidivating at hi gher rates (Winokur & Blankenship, 2007). However, these initial valida tion attempts have only been able to track youth up to six months after PACT administ ration and have not examined whether the assessment predicts recidivism equally well for females as it does males. Formal gender comparisons on any level of complexity have unf ortunately been lacking up to this point. The focus of this study is on assessing the va lidity of the PACT assessment for predicting recidivism, tracking juvenile offenders for over twelve months post-administration, and providing gender comparisons for instrument effectiveness. The current study will also examine the relationship between risk to re-offend and tim e to failure, hypothesizing higher risk to reoffend youth more likely to commit a new offense in a shorter time post-assessment than lower risk youth. The study will furthermore investigate whether identical items should be included to predict female recidivism as us ed to predict male recidivism. The PACT risk/need assessment is designed to assess a youths overa ll risk to re-offend level. Therefore, it is hypothesized youth with hi gher overall risk levels will be more likely to recidivate than youth with lower overall risk levels. It is hypothesi zed that the assessment will be

PAGE 65

65 valid for youth assessed who remain in the community, or who are assessed in residential commitment and tracked post-release, as the risk factors for recidivism are similar for all youth. The third hypothesis relates to th e scoring of the PACT. Individuals with higher criminal history scores (more serious official pr ior offending histories) will be more likely to recidivate than individuals with a lower crimin al history score. The notion of more extensive serious prior offending being predictive of a higher likelihood of recidivism is not new to criminological literature (Gottfredson & Hirschi, 1990; Moff itt, 1993; Sherman et al., 1997). However, the criminal history of the PACT as it is scored ha s not been validated with respect to recidivism when tracking youth twelve months post-assessme nt. Additionally, it is predicted youth with higher social history scor es will be more likely to recidivate. The social history score has never been examined as it relates to recidivism fo r youth tracked twelve months post-assessment. Central to the current study is the importance of gender comparison. No attempts have yet to be made to examine the validity of the PACT assessment with re gard to whether the instrument predicts recidivism as well for female youth as it does for male youth. Additionally, research examining the predictive accuracy of actuarial risk asse ssment of female offenders is relatively sparse (Reisig et al., 2006). The overall risk score pres ented by the PACT instrument is hypothesized to be a significant pr edictor for male recidivism and female recidivism. In other words, it is hypothesized that there is no need for any other gender-specific scoring differences, other than the additional one point on the social history score males automatically receive. The PACT currently gives males one point on their soci al history score automa tically, for being male, and females zero points automatically. The additional point assignment for males is in relation to the commonly accepted extant relationship in the field of criminology between gender and offending, finding males

PAGE 66

66 significantly more likely to e ngage in offending behavior th an females (Farrington, 1986; Gottfredson & Hirschi, 1990; Mo ffitt, 1993). The current study hypothesizes no additional need for scoring alteration for females or males due to literature asserting the ca uses of delinquency to be essentially the same for males and females (Gottfredson & Hirschi, 1990; Hubbard & Pratt, 2002; Lipsey, 1992; Moffitt, 1993). Hubbard and Pra tt (2002) found the predictors of female delinquency to be the same as those of male delinquency, namely a history of antisocial behavior, antisocial attitudes, an tisocial personality, and antiso cial peers. These scholars did, however, find school and family relationships and a history of physical and/or sexual abuse to be significant predictors of female delinquency. The PACT assessment does take these factors into account with respect to scoring, as mentione d above, making it possible to examine the relationships observed by H ubbard and Pratt (2002). The causes of delinquency of female offenders ha s historically been lacking in the extant criminological literature (Hubbard & Pratt, 2002; Pratt et al., 1 998). Criminological theory, the causes of criminality, and the research on crime and deviance have been based on male offending and then generalized to female of fenders (Belknap, 1996; Chesney-Lind & Sheldon, 1992; Hubbard & Pratt, 2002). Hubbard and Prat t acknowledge the importance of discerning whether the predictors of female delinquency ar e identical as those for male delinquency and state if the strongest predicto rs of female delinquency are different for females, correctional programs may need to make adjustments to target certain gender-specific risk factors rather than simply apply the treatment programs found effective for male delinquency to female offenders (2002: 2). Additional scholars have as serted the existence of gendered pathways, finding the importance of sexual and physical abus e in the development of female offending (Chesney-Lind & Rodriguez, 1983; Hubbard & Pratt, 2002; Reisig et al., 2006). History of

PAGE 67

67 abuse, both physical and sexual, is a predominant risk factor for both females and males, yet the fact that females are more likely to be victim s of the abuse could ar guably disproportionately increase the probability of females engaging in criminal behavior as a result of such abuse (Hubbard & Pratt, 2002). Additionally, females charged with serious or vi olent offenses are, more often than males, accessories that did not instigate the offending, and therefore are misclassif ied as higher risk by assessment instruments and become false positives that in fact do not re-offend (Reisig et al., 2006). The inclusion of an assessment within the PA CT of history of physic al abuse, history of sexual abuse, and school relationships will enab le the current study to assess whether these factors are more important for the prediction of female recidivism than for males. Previous research indicates that actuarial assessments are not predictive of female offending as a whole, but are valid for predicting the recidivism for economically mo tivated women (Reisig et al., 2006). The current study will examine whether the PACT assessment is a valid predictor of female re-offending. Gender differences that may appear in anal ysis of the PACT assessment are hypothesized to be the result of prevalence of some social history measurements utilized in the instrument scoring. The current study hypothesizes that all si gnificant gender differences will wash out to insignificance once the proxy measure is taken into account (for ex ample history of sexual abuse being more prevalent for females; the issue is not the youth being female, but the youth having a history of sexual abuse). The discussion above leads to the following hypotheses: Overall risk to re-offend, as presented by th e PACT, will significantly predict recidivism, with higher risk youth mo re likely to re-offend.

PAGE 68

68 Youth assessed as higher risk to re-offend will do so regardless of whether the youth was assessed and placed on probation supervision, or assessed during a residential commitment placement and tracked post-release. The higher the criminal history score as pr esented by the PACT, the more likely the youth will re-offend. The higher the social history score, as presented by the PACT, the more likely the youth will re-offend. Overall risk to re-offend, as presented by the PACT, will be a significant predictor of male recidivism and female recidivism. Males will be more likely to re-offend than females, controlling for overall risk level. Gender differences that may appear will be washed out to insignificance with the inclusion of proxy measures (such as school re lationships and history of physical or sexual abuse). History of physical or sexual abuse will not be a stronger predictor of female re-offending than of male re-offending.

PAGE 69

69 CHAPTER 3 METHODS The Florida Department of Juvenile Justice maintains a comprehensive database called the Juvenile Justice Information System (JJIS). The database stores information on all youth entering the system and all servic es provided to those youth. From the JJIS database, the Office of Research and Planni ng performs monthly cumulative updates of various data elements on youth served and pl acement history. Part of these monthly extracts includes all PACT assessments ad ministered. The monthly data extract for February 2007 will be utilized for the current study. The extract contains all PACT assessments from the initial November 15, 2005 implementation date up to the date of the extract (February 4, 2007). The PACT extrac t contains the date each assessment was administered, as well as the overall risk score, criminal history risk score, social history risk score, answers to each PACT question, and gender. Furthermore the extract contains information as to whether the assessment wa s a PACT Pre-Screen or a Full Assessment. The PACT data will serve as the initial pred iction of whether the youth will recidivate and the length of time until re-offending occurs. The PACT data will be merged with anothe r monthly data extract that captures the entire offense history for every youth who enters the juvenile justice system. This extract contains demographic information including da te of birth, race, ethnicity, and gender. The extract also includes a ll prior referrals for each yout h who has ever entered the juvenile justice system, reliable from 1994 to the present. Currently, there are over two million records of referrals for juvenile offenders contained in the extracts. The extract for April 2007 will be utilized for this study, containing all referral records up to April 6, 2007.

PAGE 70

70 The data is stored by referral meaning that each youth may have multiple referrals within the database. Every referral has a uni que referral identificat ion number and every youth has a unique identifier, ca lled the djjid number. Every line of data is a unique referral with information pertaining to the date of the referral, the date of the arrest, and the date of the offense. Furthermore, info rmation related to the most serious offense within the referral is collected, as an arre st can be made for one event with multiple charges. The data also include informati on about whether the yout h was adjudicated and the resulting disposition. The referral extract will be used to exam ine recidivism. Therefore, the dependent variable will be recidivism based on an offici al delinquency referral to the Department of Juvenile Justice post-administration of a PACT assessment. The PACT data are organized by djjid number (uniquely identify ing each youth) and by PACT identification number (uniquely identifying each PACT asse ssment). The two extracts will be merged so that PACT information and referral information for each youth can be examined. All youth included in the PACT extract must have a referral history since they are given an initial PACT assessment at intake, after bei ng referred to the Department of Juvenile Justice. The referral history prior to the PA CT assessment will be captured in Domain 1 of the PACT, which scores official criminal history. The referral extract is necessary, however, to examine all referrals occurring af ter the date of PACT administration. This creates a situation in which a youth is a ssessed using the PACT and given a score representing that youths risk to re-offend. In order to validate the PACT, the current study will examine whether subsequent o fficial delinquency re ferrals after PACT administration can be predicted in relati on to the youths overa ll risk to re-offend.

PAGE 71

71 PACT data The February 2007 monthly PACT extract utilized for the current study contains 111,450 PACT assessments. The data include s 88,670 Pre-Screen assessments and 22,780 Full Assessments. Of the Pre-Screen assessments completed, 67,300 (75.9%) were administered to males and 21,370 (24.1%) to females. Data for the Full Assessments include 18,270 (80.2%) males, and 4510 (19.8%) females. Arguably, the most important variable cont ained in both the Pre-Screen and Full Assessment is the overall risk to re-offend. The 111,450 assessmen ts reveal 50.9% of the youth to be low risk to re-offend, 15.8% moderate, 17.9% m oderate-high, and 15.5% high risk to reoffend. The overall risk score, as mentioned above, is derived from a matrix of a criminal history score, coded 0-31 with the higher th e score the greater the previous official referral seriousness, and a social history scor e, coded 0-18 with th e higher the score the more risk factors present in the youths social environment. The mean criminal history score for the 111,450 assessments is 8.30 (sta ndard deviation of 5.482) and the mean social history score 4.96 (standa rd deviation of 2.885). Overall risk to re-offend has been coded from 1 to 4 representing low risk to re-offend (1), and then moderate (2), moderate-high (3), and high risk to re-offend (4). The mean overall risk to re-offend for the 111,450 assessments is 1.98, with a standard deviation of 1.143. The criminal history score is derived enti rely from Domain 1, which consists of twelve questions. Two of the questions, sexual misconduct misdemeanor referrals, and felony sex offense referrals, are not used in the criminal history sc oring to count towards overall risk to re-offend. The remaining ten questions from Domain 1 are used to generate the criminal history score (see Tabl e 3-1 for descriptive statistics). The answer to all Domain 1 questions, as described above, are automated from JJIS, the Departments

PAGE 72

72 real-time database which captures, among many other elements, the entire official referral history for every youth ente ring the juvenile justice system. The first Domain 1 question utilized in th e criminal history scoring assesses the youth on the age at the time of the offense fo r which the youth was first referred (again, see Appendix B and Appendix C for exact wo rding of PACT questions). The item is coded 1 to 5 with 1 being over age 16 at firs t referral, 2 being 16 years of age, 3 for 15 years of age, 4 for 13 to 14 years of age, a nd 5 for 12 years of age or under. The next two items assess the number of misdemeanors and fe lonies for which the youth was referred, respectively. The misdemeanor item is coded from 1 to 4, with 1 being none or one referral, 2 being two referrals, 3 being three or four referrals, and 4 representing five or more misdemeanor referrals. The felony item is also coded 1 to 4, but is more stringent on the criteria. For this item, 1 represents no felony referrals, 2 being one referral, 3 being two referrals, and 4 representing th ree or more felony referrals. Table 3-1. Criminal History Item Descriptive Statistics Criminal History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Age at First Offense 111,450 1 5 2.2898 1.219 Adjudicated Misdemeanors 111,450 1 4 1.6912 0.93 Adjudicated Felonies 111,450 1 4 2.0778 1.01 Total Weapon Offenses 111,450 1 2 1.249 0.331 Total Against-person Misd. 111,450 1 3 1.4349 0.658 Total Against-person felonies 111,450 0 2 1.2299 0.463 Secure Detention Placements 111,450 1 4 1.8986 1.163 Commitment Placements 111,450 1 3 1.156 0.422 Total Escape Adjudications 111,450 1 3 1.0073 0.089 Total Failure to Appear PUOs 111,450 1 3 1.3957 0.698

PAGE 73

73 The next three items assess the extent of offi cial against-person and weapon referrals in the youths history. Weapon referrals is coded 1 to 2, with 1 being no weapon referrals, and 2 representing one or more referrals for which th e most serious offense in the referral was a firearm/weapon charge. The next two items meas ure against-person referr als, which are broken down by misdemeanors and felonies. Misdemeanor against-person referral s are coded 1 to 3, with 1 being no against-person misdemeanor refe rrals, 2 being one referral, and 3 representing two or more referrals for which the most serious offense was a misdemeanor involving threats, force, or physical harm to another person. The ag ainst-person felony item is coded 0 to 2, with 0 representing no against-person felony referrals, 1 being one or two such referrals, and 2 representing three or more ag ainst-person felony referrals. The remaining four questions utilized in the sc oring of official crimin al history relate to confinements, residential commitments, escapes, and pick up orders for failureto-appear in court. The confinements item measures the number of times the youth was held for at least fortyeight hours in a detention fac ility. This item is coded 1 to 4, representing no detention confinements (1), one confinement (2), two confinements (3), and three or more confinements in a detention facility (4). The to tal number of residential commitment confinements is coded 1 to 3, representing no placements in a residential facil ity (1), one placement (2), and a history of two or more residential placements (3). History of es capes is coded 1 to 3, representing no history of escape (1), one attempt or actual escape (2), a nd two or more attempted or actual escapes (3). The final criminal history item used in scoring gauges the number of failures-to-appear in court that have resulted in a pick up order being issued in the youths history. The item is coded 1 to 3, representing no pick up orders (1), one (2), and two or more pick up orde rs for failure-to-appear (3).

PAGE 74

74 Table 3-2. Social History Item Descriptive Statistics Social History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 111,450 0 1 0.7678 0.422 Current School Enrollment 111,450 1 6 2.5592 1.23 Recent School Conduct 111,450 1 5 2.9025 1.183 Recent School Attendance 111,450 1 5 2.6197 1.533 Recent Academic Performance 111,450 1 5 3.3858 0.946 Anti-social Peers 111,450 0 2 0.6688 0.552 Out-of-home Placements 111,450 1 4 1.2003 0.593 History of Runing Away 111,450 1 5 1.5722 1.11 Parental Authority/Control 111,450 1 3 1.7241 0.68 Household Jail History 111,450 0 1 0.9951 0.07 History of Alcohol Use 111,450 0 1 0.3375 0.473 History of Drug Use 111,450 0 1 0.4758 0.499 Current Alcohol Use 111,450 0 1 0.2203 0.414 Current Drug Use 111,450 0 1 0.3544 0.478 History of Physical Abuse 111,450 0 2 0.237 0.608 History of Sexual Abuse 111,450 0 1 0.0497 0.217 History of Neglect 111,450 1 2 1.0723 0.259 Mental Health Problems 111,450 1 5 1.2738 0.847 The social history of the youth is assessed, in part, using the eighteen items illustrated in Table 3-2. Descriptive statistics for these social hi story items are reported in Table 3-2. The first item assessing social history is sex. Females are coded as 0 and males as 1. Current school enrollment status is coded 1 to 6, represen ting the youth has graduated high school or has received a GED (1), being enrolled full-time (2), enrolled part-time (3), suspended (4), dropped out (5), and having been expelled (6). Recent school conduct is measured next, coded 1 to 5. Conduct is broken down by recognition for good be havior, no problems with school conduct, problems reported by teacher, problem calls to parent s, and lastly, calls to police related to the youths conduct. The youths recent academic at tendance is coded 1 to 5, representing good attendance with few absences (1), no unexcused absences (2), so me partial-day unexcused (3), some full-day unexcused (4), a nd the youth being a hab itual truant (5). Academic performance

PAGE 75

75 rounds out the school-related measures, and is coded 1 to 5. Honor st udents are coded as 1, G.P.A above 3.0 coded 2, G.P.A. from 2.0 to 3.0 c oded as 3, G.P.A. from 1.0 to 2.0 coded as 4, and G.P.A. below 1.0 represented as 5. The youths current friends/companions who th e youth actually spends time with is broken down by no consistent friends or companions, pr o-social friends, anti-s ocial friends, and gang members/associates. Within the PACT software, this item can receive multiple checks meaning that the individual administering the in strument can record that the youth spends time with any combination of the options listed above The item has been recoded for this study to separate youth who spend no time w ith anti-social friends, or have all pro-social friends, from those who spend time with anti-social friends, and from youth who are gang members or have friends who are gang members/associates. The ne w item has been coded 0 to 2, representing no anti-social friends (0), anti-so cial friends (1), and the youth who are gang members or actually spend time with gang members (3). History of court-ordered or Department of Children and Families voluntary out-of-home and shelter care placements exceeding thirty days is measured as part of the social history scoring. This item is coded 1 to 4, representing no out-of-home placements exceeding thirty days (1), one placement (2), two placements (3), a nd three or more out-of-home placements (4). The youths history of running away is coded 1 to 5, representing no hi story or running away or being kicked out (1), one instance of running away (2), tw o to three instances (3), four to five instances (4), and over five instances of running away (5). The jail and imprisonment hist ory of individuals currently liv ing in the household with the youth is taken into account in the social histor y scoring as well. This item differentiates youth living with no one with a history from the mo ther having a history of jail/prison, the father

PAGE 76

76 having a history, older siblings having a hist ory, younger siblings havi ng a history, and from other members of the household having a histor y of jail or imprisonment. The individual administering the PACT is able to check each of these scenarios separately, much like the current friends item described a bove. Therefore, this item was recoded to represent the history of jail/imprisonment of individuals currently li ving in the household with the youth, now coded 0 to 1. The new coding coincides with the scori ng utilized by the PACT separating youth who currently live in a house hold without an individual who has a history of jail or imprisonment (coded as 0) from those living with at least one such individual (coded as 1). The level of parental authority and control is accounted for in the social history score and coded 1 to 3. Youth who usually obey parents an d follow the rules they set are coded as 1, youth sometimes obeying the rules coded as 2, and youth c onsistently disobeying rules or acting hostile toward the parents are coded as 3. Four items measuring alcohol and drug use are measured as well. History of alcohol use, hi story of drug use, current alcohol use, and current drug use are each utilized in the social history scoring. Histor y of use includes any use by the youth in his/her lifetime. Current use refers to e ngaging in the behavior in the last six months. The two history of use items are broken down by no past use, past use, alcohol (or drugs) disrupted education, caused family conflict, interfered with keepi ng pro-social friends, caused health problems, contributed to criminal behavior youth needed increas ing amounts to achieve the same level of intoxication or high, and youth experienced withdr awal problems. Current use is broken down by not currently using, currently using, alcohol (or drugs) disrupts education, causes family conflict, interferes with keeping pro-social friends, cau ses health problems, contributes to criminal behavior, youth needs increasing amounts, and youth experiences w ithdrawal problems. Identical to two scenarios described above, the indivi dual administering the PACT can check all

PAGE 77

77 answers that item, necessitating th e need to recode the items. The new items are coded 0 to 1 separating individuals with no hist ory of use, or no current use in the case of the second two variables, from youth with a histor y of use or current use. Recodi ng the variables in this manner is consistent with the scoring util ized in the social history score. Abuse and mental health issues are the last f our items used in the social history scoring. For the abuse and neglect items, suspected incident s of abuse are included, as are those disclosed by the youth (whether or not reported or substantia ted officially) but reports of abuse and neglect investigated officially and prove n to be false are excluded. Yout h are assessed for a history of being the victim of physical abuse, which is broken down by not having been a victim, having been a victim at home, having been a victim in a foster/group home, having been victimized by a family member, having been victimized by some one outside of the family, and having been attacked by a weapon. Again, the individual admini stering the PACT can check all scenarios that apply. The item has therefore been recoded 0 to 2 to separate youth not having been a victim (coded as 0) from youth ever having been victimi zed by a family member or someone outside the family (coded as 1) and from youth ever having b een the victim of physical abuse at home, in a foster/group home, or attacked with a weapon (c oded as 2). Recoding the item in this manner is consistent with the scoring used and the notion that being attack ed with a weapon or having been victimized where the youth was residing achieve s some higher level of seriousness than only having been a victim. History of sexual abuse/rape is assessed and taken into acco unt in the scoring of social history. The item differentiates youth who have not been a victim of sexual abuse or rape from youth having been abused/raped by a family me mber, and from youth having been abused/raped by someone outside of the family. Youth may ha ve been victimized both by a family member,

PAGE 78

78 and by someone outside of the family. The indi vidual administering the PACT may check all that apply. The item has been recoded 0 to 1 to separate youth not having been a victim of sexual abuse/rape from those that have, regardless of th e perpetrator. Recoding th e item in this fashion is consistent with the method used in the social history scoring. History of being the victim of neglect is assessed, coded 1 to 2. Individuals not having been the vi ctim of neglect (coded as 1) are separated from individuals that have b een the victim of neglect (coded as 2). Finally, the youths history of me ntal health problems is asse ssed. Mental health problems such as schizophrenia, bi-polar, mood, thought personality, and adjustment disorders are included. Conduct disorder, opposi tional defiant disorder, substance abuse, and ADD/ADHD are excluded. The problem must have been c onfirmed by a professional in the social service/healthcare field to be included, as individuals administering the assessment are predominately non-licensed, non-clini cal staff (usually they are j uvenile probation officers). The item is coded 1 to 5, with 1 being no history of mental health problems, 2 being diagnosed with mental health problems, 3 being diagnosed and on ly mental health medication is prescribed, 4 being diagnosed and only mental health treatmen t prescribed, and 5 repr esenting the youth being diagnosed and having mental health me dication and treatment prescribed. The above section has articulated the items fr om the PACT that are used in deriving the criminal history score as well as various social history items. The remaining ninety-eight items in the Full Assessment are used to gain a comprehe nsive picture of the youth and his/her past and present situation with respect to the twelve PACT domains. Thes e remaining items, however, are not utilized in the scoring of the overall risk to re-offend reveal ed by the PACT assessment. The items used in the overall risk to re-offend scor ing of the PACT assessment will be utilized in the current validation/gender comparison study. These va riables will be examined with respect to

PAGE 79

79 their ability to predict official recidivism for the juvenile offending population, using Florida Department of Juvenile Justic e JJIS data as explained above. Sample The PACT data outlined above represent all PACT assessments completed from August of 2005 to February 2007. The PACT is administered to each youth who enters the juvenile justice system in the state of Florida. Furthermor e, each youth remaining in the custody of the Department is re-assessed every ninety days, as we ll as sixty days prior to release, if placed in residential commitment. Due to the re-assessment pr ocess, and the fact that many youth enter the juvenile justice system on multiple occasions, the current study narrows the sample down so as to utilize the first ever Full Assessment for each youth in the sample. In the event a Full Assessment was never administered to the youth, th e first ever Pre-Screen is utilized. The first assessment administered to the youth was chosen so as to maximize the follow-up period possible to track recidivism. As the PACT asse ssment is a new process, implemented in August of 2005, allowing for follow-up times is a critical factor. The choice to use the first Full Assessment was made in order to capture the mo st information available on each youth, due to the Full Assessment containing 126 variables versus the forty-six included in the Pre-Screen. The sampling procedure outlined above resulted in a final N of 48,871. There is no need to divide the sample into subgroups for the purpos e of construction and validation samples as would be necessary in the ev ent an instrument were bei ng designed. The current study is attempting to validate and examine an existing instrument and, as a result, will not benefit from the construction and replication process nor will it suffer any shrinkage issues that would normally accompany instrument development. Future research wishing to examine potential improvements to the existing instrument and/or scoring protocols woul d be well advised to

PAGE 80

80 utilize a sub-sample for construction plus multiple replication samples to ascertain shrinkage and predictive validity. However, the current study benefits from divi ding the sample into those youth that were assessed and subsequently remained in the comm unity versus those youth assessed and placed in residential commitment. The study may have ke pt these sub-samples together and simply controlled for street time, beginning the follo w-up period immediately for youth remaining in the community and waiting until release from residential commitment to begin the follow-up period for the committed youth. The decision was in stead made to separate youth into the two categories mentioned above and analyzed independe nt of one another to assess the effectiveness of the PACT for youth remaining in the commun ity versus those youth being released from residential placement. Cohort 1 Descriptive Statistics Prior to dividing the overall sample into community youth and residential youth, the decision was made to split the sample into two cohorts. Cohort 1 will consist of youth who are tracked for only six months, while Cohort 2 cons ists of youth who are followed for twelve months. This will allow for the utilization of as many cases as possible of youth who have received the PACT assessment thus far, as well as allow for the examination of the effectiveness of the assessment to predict short-term and l ong-term success. Youth who can be tracked twelve months, and thus placed in C ohort 2, are not also placed in C ohort 1. The youth that remain in the community are tracked begi nning right after the assessmen t and followed six months for Cohort 1 and twelve months for Cohort 2. The youth who are in residential commitment are tracked following release from placement and then followed six months post-release for Cohort 1 youth and twelve months post-release for Cohort 2 youth.

PAGE 81

81 Table 3-3. Community Cohort 1 Crimin al History Descriptive Statistics Criminal History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Age at First Offense 36,325 1 5 2.5581 1.298 Adjudicated Misdemeanors 36,325 1 4 1.4332 0.762 Adjudicated Felonies 36,325 1 4 1.7703 0.853 Total Weapon Offenses 36,325 1 2 1.0951 0.293 Total Against-person Misd. 36,325 1 3 1.3428 0.584 Total Against-person felonies 36,038 1 2 1.1984 0.3988 Secure Detention Placements 36,325 1 4 1.5 0.931 Commitment Placements 36,325 1 3 1.0829 0.315 Total Escape Adjudications 36,325 1 3 1.0032 0.058 Total Failure to Appear PUOs 36,325 1 3 1.2158 0.533 Table 3-4. Community Cohort 1 Social History Descriptive Statistics Social History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 36,325 0 1 0.7335 0.442 Current School Enrollment 34,489 1 6 2.4114 1.103 Recent School Conduct 29,982 1 5 2.6863 1.122 Recent School Attendance 29,982 1 5 2.3237 1.439 Recent Academic Performance 29,982 1 5 3.2571 0.923 Anti-social Peers 36,325 0 2 0.6125 0.538 Out-of-home Placements 36,325 1 4 1.1445 0.504 History of Running Away 36,325 1 5 1.3658 0.892 Parental Authority/Control 36,105 1 3 1.5628 0.632 Household Jail History 36,105 0 1 0.2173 0.412 History of Alcohol Use 36,325 0 1 0.2783 0.448 History of Drug Use 36,325 0 1 0.3848 0.487 Current Alcohol Use 32,043 0 1 0.1088 0.311 Current Drug Use 32,043 0 1 0.2204 0.415 History of Physical Abuse 36,325 0 2 0.1814 0.538 History of Sexual Abuse 36,325 0 1 0.0394 0.195 History of Neglect 36,325 1 2 1.0506 0.219 Mental Health Problems 35,466 1 5 1.2107 0.758 The resulting sub-samples of Cohort 1 commun ity and residential youth yield a final N of 36,325 community and 3,399 residential Cohort 1 youth. The first sub-sample to be examined is composed of youth remaining in the commun ity with the follow-up period beginning

PAGE 82

82 immediately after the PACT assessment. The community Cohort 1 sub-sample is 56.7% white, 42.7% black, and 0.6% other. This sub-sample is 73.3% male and 26.7% female. Descriptive statistics for the criminal history and the soci al history of the community Cohort 1 sub-sample can be seen in Table 3-3 and Table 3-4, respectively. The second Cohort 1 sub-sample is composed of youth exiting residential commitment placement prior to beginning the six-month follo w-up period. The residential sub-sample is 44.6% white, 55% black, and 0.4% other. The residential sub-sample is 84.5% male and 15.5% female. The descriptive statis tics outlining the criminal hist ory and social history of the residential sub-sample are reported in Table 3-5 and Table 3-6, respectively. Table 3-5. Residential Cohort 1 Crimin al History Descriptive Statistics Criminal History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Age at First Offense 3,399 1 5 1.8664 0.972 Adjudicated Misdemeanors 3,399 1 4 2.0891 1.047 Adjudicated Felonies 3,399 1 4 2.7193 1.041 Total Weapon Offenses 3,399 1 2 1.1692 0.375 Total Against-person Misd. 3,399 1 3 1.5711 0.741 Total Against-person felonies 3,232 1 2 1.3685 0.482 Secure Detention Placements 3,399 1 4 2.8653 1.154 Commitment Placements 3,399 1 3 1.4693 0.629 Total Escape Adjudications 3,399 1 3 1.0274 0.172 Total Failure to Appear PUOs 3,399 1 3 1.6973 0.837 Cohort 1 Analysis The first step is to assess the effectivene ss of the PACT to predict re-offending for the Cohort 1 community sub-sample and the residen tial sub-sample, separately, before moving to Cohort 2 comparisons. The purpose of this step is to ascertain whether the in strument is effective at predicting recidivism for youth who remain in the community subsequent to assessment as well as youth exiting a residential commitment placement prior to beginning the six-month follow-up. Consistent with prior risk assessment research, re-offending is operationalized as any

PAGE 83

83 subsequent delinquency referral after the asse ssment date (Funk, 1999:55; see also Foley, 1991; Wiebush et al., 1995). The current study uses the da te of the subsequent offense post-assessment as it is capturing the beha vior that desired, and not any indi cation of police or court action, so will use the actual date the offense occurred. Re cidivism, therefore, for the purposes of this study, will be any subsequent delinqu ency referral to the Florida Department of Juvenile Justice. The success or failure of a youth will determined based on whether the youth commits a new offense for which he/she is officially referred by law enforcement to the Department and having an offense date within six months for Cohort 1 youth and twelve months for Cohort 2 youth from the beginning of the follow-up period. Table 3-6. Residential Cohort 1 Soci al History Descriptive Statistics Social History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 3,399 0 1 0.8452 0.362 Current School Enrollment 2,889 1 6 2.5147 1.158 Recent School Conduct 2,602 1 5 3.0223 1.171 Recent School Attendance 2,602 1 5 2.9181 1.596 Recent Academic Performance 2,602 1 5 3.5984 0.963 Anti-social Peers 3,399 0 2 0.7941 0.539 Out-of-home Placements 3,399 1 4 1.3113 0.718 History of Running Away 3,399 1 5 1.9412 1.343 Parental Authority/Control 3,337 1 3 2.0039 0.701 Household Jail History 3,399 0 1 0.341 0.474 History of Alcohol Use 3,399 0 1 0.474 0.499 History of Drug Use 3,399 0 1 0.6764 0.468 Current Alcohol Use 3,399 0 1 0.2304 0.421 Current Drug Use 3,399 0 1 0.5179 0.5 History of Physical Abuse 3,399 0 2 0.3033 0.676 History of Sexual Abuse 3,399 0 1 0.0691 0.254 History of Neglect 3,399 1 2 1.1068 0.309 Mental Health Problems 3,309 1 5 1.417 0.98 The dependent variable for analysis of the e ffectiveness of the PACT risk to re-offend prediction for Cohort 1 community and resident ial youth is dichotomous (either the youth

PAGE 84

84 recidivated within six months or not) and therefore the use of l ogistic regression is appropriate. The analysis, beginning with C ohort 1 community youth, uses the overall level of risk to reoffend, as measured utilizing the PACT assessm ent, to predict recidivism, based on a new referral to the Department of Juve nile Justice subsequent to the assessment date. Consistent with prior research, the analysis begi ns by assessing recidivism rates across risk-need categories of overall risk to re-offend (Andrews & Bonta, 2003; Andrews et al., 2001; Bonta, LaPairie, & Wallace-Capretta, 1997; Reisig et al., 2006). Descri ptive statistics for overall risk to re-offend can be seen in Table 3-7. Low risk to re-o ffend youth made up 70.7% of the sample, moderate risk 13.2%, moderate-high risk 10% and high risk to re-offend 6.1% The percentage of low risk youth who received a new referral to the Depa rtment within six months was 19.6%. The percentage for moderate risk youth was 34.7%, 38.6% for moderate-high, and 43.4% of all high risk youth received a new re ferral to the Department within six months. Table 3-7. Community Cohort 1 C ovariate Descriptive Statistics Covariate Descriptive Statistics N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 36,325 1 4 1.5158 0.904 Sex 36,325 0 1 0.7335 0.442 Race 36,325 0 1 0.4332 0.496 Referral within six months 36,325 0 1 0.2499 0.433 The analysis using logistic regression to pred ict six-month recidivism of youth who remain in the community after receiving a PACT assessmen t reveals that overall risk to re-offend is a statistically significant predicto r, with a p-value of .000. Youth assessed as having a higher overall risk to re-offend score are more likely to recidivate then youth asse ssed as lower risk. As Table 3-8 illustrates, each increase in the i ndependent variable, overall risk to re-offend, produces a 1.543 increase in the dependent va riable, recidivism (an over 50% increase).

PAGE 85

85 Next, race was also included as a covariate, simply separating white (coded 0) and nonwhite (coded as 1). The final covariate include d was sex, separating male s (coded 1) and females (coded 0). Logistic regression was once again utili zed to predict a delinque ncy referral tracked six months post PACT assessment for youth who remain in the community, this time using overall risk to re-offend, race, and sex as predicto rs. The analysis reveals all three covariates to be significant predictors of a de linquency referral subsequent to PACT assessment, all with pvalues of .000. As above, youth assessed as higher risk are more likely to recidivate during the six-month follow-up period, as are males and non-white youth. Results of the logistic regressions can be seen in Table 3-8. Examination of the Exp( B) reveal overall risk to re-offend to have the strongest relationship to recidivism followed by sex and finally race. Table 3-8. Community Cohort 1 Logistic Regression Results Covariate Statistics Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.434 0.012 1226.507 1 .000* 1.543 Logistic Regression 2 Overall Risk to Re-offend 0.395 0.013 981.987 1 .000* 1.484 Sex 0.434 1 158.437 1 .000* 1.456 Race 0.434 0 223.35 1 .000* 1.453 =significant at p=.05 Table 3-9. Residential Cohort 1 Covariate Descriptive Statistics Covariate Descriptive Statistics N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 3,399 1 4 2.9491 1.047 Sex 3,399 0 1 0.8452 0.362 Race 3,399 0 1 0.554 0.497 Referral within six months 3,399 0 1 0.3945 0.489 The Cohort 1 youth who were in residential co mmitment facilities at the time of the assessment, or who entered residential subsequent to the PACT assessment were analyzed next.

PAGE 86

86 These youth were analyzed identically to the Cohort 1 community youth described above (descriptive statistics can be seen in Table 39). For this sub-sample, low risk youth composed 14.1% of the youth, moderate risk 15.1%, moderate high 32.5%, and high risk to re-offend youth 38.3%. These simple percentages illustrate the difference be tween the Cohort 1 community youth analyzed initially from these Cohort 1 residential youth. While only 6.1% of the community youth were assessed as high risk to re-offend, 38.3% of th e residential youth are high risk. Likewise, whereas only 14.1% of these residential yout h are low risk, 70.7% of the community youth were low risk to re-offend. Thes e differences will help to see whether the PACT assessment can predict equally well the reci divism of lower risk community youth as it does higher risk residential youth. Table 3-10. Residential Cohort 1 Logistic Regression Results Covariate Statistics Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend -0.041 0.033 1.516 1 .218 0.96 Logistic Regression 2 Overall Risk to Re-offend -0.045 0.034 1.782 1 .182 0.956 Sex -0.015 0.097 0.024 1 .877 0.985 Race 0.072 0.071 1.03 1 .310 1.075 First, logistic regression was used to predict a delinquency referral subsequent to PACT assessment and release from a residential commitment facility. Overall risk to re-offend is not a significant predictor of a referral to the department when youth ex iting residentia l facilities are followed six months (see Table 3-10). Identical ly to the analysis to Cohort 1 community youth, logistic regression was again used to predict a subsequent referra l using overall risk to re-offend, sex, and race. Similar to the community sub-samp le, the residential sample is predominately male (see Table 3-9 for descriptive statistics). However, the residential sample is majority nonwhite, where the community sample was over 50% white. The results of the logistic regression

PAGE 87

87 (see Table 3-10) reveal overall risk to re-offend, r ace, and sex are not significant predictors of a referral to the Department for youth who exist residential fac ilities subsequent to a PACT analysis. There are several reasons why it may be expected that you th being released from a residential facility may be diffe rent from youth who remain in the community, which could lead to hypothesizing as to why the PACT overall risk to re-offend score is not a significant predictor of recidivism for these youth. Prio r to any theorizing as to why th is is the case, the residential sample must be further analyzed. The sample was composed of two types of scenarios. The first was youth who were administered the PACT and s ubsequently placed into a residential facility after a court hearing. The six month follow-up began after the youth was released from the facility, but using the pre-commitment PACT assessment. The second group of youth was those that were assessed while already in a commitment facility, and it was that assessment that was used to predict the six months follow-up recidivism. The second scenario, youth a ssessed after already being in commitment, may be an inadequate use of the PACT assessment. The Florid a Department of Juvenile Justice realizes this inadequacy and is currently in the process of developing a PACT assessment more appropriate for residential youth. The problem lies with the social history questions from which the overall risk to re-offend score is derive d. Remember that the overall risk score is composed of a matrix of criminal history items and soci al history items. The criminal hi story items are static predictors (age at first offense, number a nd seriousness of prior offenses, for example), so would not make an issue for assessing youth w ithin commitment facilities. However, the social history items are dynamic, changeable factors a nd certainly could be altered by residential placement. This alteration is not necessitated by some type of effective, or

PAGE 88

88 for that matter ineffective, treatment delivered to residential youth, but more of a result of the questions themselves. Questions related to delinquent peers, substance abuse, and school attendance differ by necessity simply by residing in a residential facility. Youth are, arguably, forced to be around antisocial p eers, as all youth received delinque ncy referrals that resulted in the commitment. Therefore, all youth would be sc ored higher the day they were placed in the facility through no fault of their own with regard to choosing who they associat e with (with the realization that with committing the offense come s the possibility of residential placement if apprehended). Furthermore, all youth are essentia lly forced to attend sc hool on a daily basis, making the score on this factor decrease for ma ny youth who did not have stellar attendance in the community. Also, using substances and alcoho l decreases by nature of the youth being placed in commitment. A picture is quickly painted wher e the results of a PACT assessment given to a youth in residential commitment is suspect at be st, as some variables used in the scoring are artificially inflated and some deflat ed by the mere fact of the placement. Table 3-11. Split Residential Cohort 1 Covariate Descriptive Statistics Covariate Descriptive Statistics PACT before commitment N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 2,850 1 4 2.9102 1.088 Sex 2,850 0 1 0.8435 0.363 Race 2,850 0 1 0.5411 0.498 Referral within six months 2,850 0 1 0.4119 0.492 PACT while in commitment N Mini mum Maximum Mean Std. Deviation Overall Risk to Re-offend 549 1 4 3.1512 0.772 Sex 549 0 1 0.8543 0.353 Race 549 0 1 0.6211 0.486 Referral within six months 549 0 1 0.3042 0.46 The scenario depicted above was tested, breaki ng the Cohort 1 residential sample into two groups. The first group was composed of youth w ho were given the PACT assessment prior to residential placement. This will help to test the longevity of the predictability of the PACT as

PAGE 89

89 some youth are placed in residential facilities for several months and the follow-up period does not begin until release meaning scores would have to be valid and predictive of re-offending for months. The second group, youth assessed while in residential commitment, will show whether the PACT assessment is appropriate for assessing the risk to re-offend after subsequent release for these youth. The descriptive statistics for th e two groups of resident ial youth are shown in Table 3-11. Table 3-12. Split Residential Cohor t 1 Logistic Regression Results Covariate Statistics PACT before commitment Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend -0.041 0.035 1.377 1 .241 0.96 Logistic Regression 2 Overall Risk to Re-offend -0.043 0.035 1.475 1 .225 0.958 Sex -0.102 0.105 0.954 1 .329 0.903 Race 0.067 0.077 0.763 1 .383 1.07 PACT while in commitment Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.129 0.123 1.106 1 .293 1.138 Logistic Regression 2 Overall Risk to Re-offend 0.105 0.126 0.686 1 .407 1.11 Sex 0.623 0.297 4.393 1 .036* 1.864 Race 0.228 0.197 1.334 1 .248 1.256 =significant at p=.05 Identically to previous analyses, logistic re gression was used to pr edict a new delinquency referral post-release from a reside ntial facility. First, only the overall risk to re-offend level was entered as a covariate, separately for the yout h assessed prior to commitment and for youth assessed while committed. For the first group, youth a ssessed prior to being placed in residential commitment, overall risk was not significant (see Table 3-12). This is contrary to the notion presented above that the PACT may still be predictive for youth assessed with appropriate

PAGE 90

90 questions. It may, however, depict that inability to use risk assessment scores from months previous to the beginning of the follow-up period to ascertain adequate predictions of risk to recidivate. For youth assessed while in commitment placement, overall risk to re-offend was also not a significant predictor of post-release recidivism (see Table 3-12). This finding was suspected due to the inappropria teness of some PACT questions for a residential population. As analyzed previously, logistic regression was next used to a ssess the predictive ability of overall risk to re-offend, race, and sex of the two sub-samples, separately, of a subsequent delinquency referral post-release. For the youth asse ssed prior to entering a facility, none of the three covariates were significant predictors of recidivism (see Table 3-12). For the youth assessed while in residential commitment, only sex was a significant predictor of recidivism. Male youth are more likely to recidivate then female youth when followed post-release. Examination of the Exp(B) for sex reveals that every increase in the B produces a 1.864 times increase in recidivism (w ell over a 50% increase). Cohort 2 Descriptive Statistics Youth for which a one-year follow-up can be conducted are analyzed next. This will enable the comparison of the ability of PACT to predict long-term recidivism versus short-term. The following analyses will be conducted identica lly to those reported for Cohort 1 youth. After initial descriptive statistics are reported, the resi dential youth will be split into two samples prior to further analysis (separati ng those youth assessed prior to commitment and those youth assessed while in a residential facility). The Cohort 2 community sub-sample is 61.1% white, 38% black, and 0.9% other. The sub-sample is 69.8% male and 30.2% female. The PA CT criminal history de scriptive statistics are reported in Table 3-13. Table 3-14 illustrate s the descriptive statistics for the eighteen variables composing the social history scori ng for the Cohort 2 community youth sub-sample.

PAGE 91

91 Table 3-13. Community Cohort 2 Crimin al History Descriptive Statistics Criminal History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Age at First Offense 8,132 1 5 2.4953 1.281 Adjudicated Misdemeanors 8,132 1 4 1.4405 0.766 Adjudicated Felonies 8,132 1 4 1.7213 0.838 Total Weapon Offenses 8,132 1 2 1.1024 0.303 Total Against-person Misd. 8,132 1 3 1.3517 0.585 Total Against-person felonies 8,077 1 2 1.1804 0.385 Secure Detention Placements 8,132 1 4 1.5425 0.98 Commitment Placements 8,132 1 3 1.0637 0.28 Total Escape Adjudications 8,132 1 3 1.0041 0.073 Total Failure to Appear PUOs 8,132 1 3 1.2298 0.555 Table 3-14. Community Cohort 2 Social History Descriptive Statistics Social History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 8,132 0 1 0.6981 0.459 Current School Enrollment 7,959 1 6 2.5155 1.168 Recent School Conduct 7,957 1 5 2.9773 1.247 Recent School Attendance 7,957 1 5 2.5831 1.587 Recent Academic Performance 7,955 1 5 3.2925 0.978 Anti-social Peers 8,132 0 2 0.6779 0.536 Out-of-home Placements 8,132 1 4 1.1422 0.502 History of Running Away 8,132 1 5 1.3876 0.927 Parental Authority/Control 8,102 1 3 1.5994 0.632 Household Jail History 8,102 0 1 0.2407 0.428 History of Alcohol Use 8,132 0 1 0.2907 0.454 History of Drug Use 8,132 0 1 0.3915 0.488 Current Alcohol Use 7,468 0 1 0.1197 0.325 Current Drug Use 7,468 0 1 0.2442 0.43 History of Physical Abuse 8,132 0 2 0.2167 0.586 History of Sexual Abuse 8,132 0 1 0.0454 0.208 History of Neglect 8,132 1 2 1.0594 0.236 Mental Health Problems 7,951 1 5 1.24 0.817 The second Cohort 2 sub-sample is composed of the youth who were assessed using the PACT and had a residential commitment placem ent prior to the follo w-up period. The Cohort 2 sub-sample is 52.2% white, 46.7% black, and 1.1% Other. Furthermore, the sample is 84.6%

PAGE 92

92 male and 14.4% female. As in the Cohort 1 resi dential sample, the Cohor t 2 youth being placed in residential commitment prior to the follow-up period is substant ially more male (84.6% versus the 69.8% for community youth) and less likely to be white (52.2% versus 61.1% for the community youth). Descriptive st atistics for the PACT assessed criminal history and social history for the Cohort 2 youth who have a reside ntial commitment placement prior to the followup period can be seen in Table 3-15 and Table 3-16, respectively. Table 3-15. Residential Cohort 2 Crimin al History Descriptive Statistics Criminal History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Age at First Offense 1,015 1 5 1.802 0.896 Adjudicated Misdemeanors 1,015 1 4 2.1271 1.073 Adjudicated Felonies 1,015 1 4 2.6739 1.038 Total Weapon Offenses 1,015 1 2 1.1645 0.371 Total Against-person Misd. 1,015 1 3 1.5793 0.741 Total Against-person felonies 981 1 2 1.3191 0.466 Secure Detention Placements 1,015 1 4 2.999 1.135 Commitment Placements 1,015 1 3 1.2926 0.526 Total Escape Adjudications 1,015 1 2 1.0197 0.139 Total Failure to Appear PUOs 1,015 1 3 1.6887 0.832 Cohort 2 Analysis The analysis of the Cohort 2 community sub-sa mple utilizes logistic regression to examine the effectiveness of the overall risk to re-offe nd score as identified by the PACT assessment in the prediction of a subsequent delinquency refe rral to the Department of Juvenile Justice for youth who remain in the community after assess ment. The Cohort 2 community sub-sample is composed of 70.9% low risk, 13.3% moderate ris k, 9.6% moderate-high risk, and 6.3% high risk to re-offend youth. The sample is 61.1% white, and 38.9% non-white. Descriptive statistics for the Cohort 2 community sample can be seen in Table 3-17. The percentage of low risk youth who received a new referral to the Department w ithin twelve months was 27.3%. The percentage for moderate risk youth was 43.6%, and 48.3% for moderate-high risk youth, and 50.3% for high

PAGE 93

93 risk youth. This means that just over half of the youth assessed as high risk to re-offend will in fact do so within twelve months if they remain in the community post-assessment. The logistic regression results using only overall risk to re-o ffend to predict a subsequent referral to the Department for youth assessed reveal overall risk le vel to be a significant pr edictor of recidivism at p=.000 (see Table 3-18). The Exp(B) illustrates th at for every increase in overall risk to reoffend, as assessed by the PACT, there is a 1.491 times increase in recidivism (almost a 50% increase). Table 3-16. Residential Cohort 2 Soci al History Descriptive Statistics Social History Item Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 1,015 0 1 0.8562 0.351 Current School Enrollment 958 1 6 2.8758 1.422 Recent School Conduct 957 1 5 3.3501 1.188 Recent School Attendance 957 1 5 3.4253 1.603 Recent Academic Performance 957 1 5 3.7482 1.019 Anti-social Peers 1,015 0 2 0.869 0.536 Out-of-home Placements 1,015 1 4 1.2916 0.681 History of Running Away 1,015 1 5 2.0562 1.372 Parental Authority/Control 1,007 1 3 2.0725 0.679 Household Jail History 1,007 0 1 0.3903 0.488 History of Alcohol Use 1,015 0 1 0.5251 0.5 History of Drug Use 1,015 0 1 0.7084 0.455 Current Alcohol Use 894 0 1 0.2808 0.45 Current Drug Use 894 0 1 0.5436 0.498 History of Physical Abuse 1,015 0 2 0.4 0.766 History of Sexual Abuse 1,015 0 1 0.069 0.254 History of Neglect 1,015 1 2 1.13 0.337 Mental Health Problems 992 1 5 1.4849 1.094 Table 3-17. Community Cohort 2 C ovariate Descriptive Statistics Covariate Descriptive Statistics N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 8,132 1 4 1.5124 0.905 Sex 8,132 0 1 0.6981 0.459 Race 8,132 0 1 0.3893 0.488 Referral within twelve months 8,132 0 1 0.3294 0.47

PAGE 94

94 Identical to the previous analyses, the next step utilized logistic regression to predict a referral subsequent to PACT assessment using ove rall risk to re-offend, race (again coded white and non-white), and sex (male and female). Consis tent with the analysis of Cohort 1 community youth who were tracked only six months post-assessm ent, overall risk to re-offend, race, and sex were all significant (at p=.000) predictors of a subsequent de linquency referral (see Table 3-18). Results reveal that youth assessed as higher risk to re-offend are in fact more likely to do so within the one-year follow-up. Furthermore, males are more lik ely to receive a subsequent delinquency referral, as are non-white youth. The second strongest relatio nship (race being the strongest) to recidivism was th e overall risk to re-offend as a ssessed by the PACT, showing for every increase in overall risk, there is a 1.433 ti mes increase in the dependent variable. These results illustrate the validity of the PACT to predict short-term (six month) as well as long-term (one year) recidivism for youth who are assessed and remain in the community post-assessment. Table 3-18. Community Cohort 2 Logistic Regression Results Covariate Statistics Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.4 0.025 249.763 1 .000* 1.491 Logistic Regression 2 Overall Risk to Re-offend 0.36 0.026 196.432 1 .000* 1.433 Sex 0.302 0.054 31.047 1 .000* 1.353 Race 0.428 0.049 75.848 1 .000* 1.534 =significant at p=.05 Cohort 2 residential youth are analyzed using the same procedures utilized to examine Cohort 1 residential youth. Thes e youth are collective analyzed first, including both youth who were assessed using the PACT and subsequently placed in residential commitment before oneyear follow-up, as well as youth assessed while in residential commitment and tracked following release. Covariate descriptive statistics for th ese youth can be seen in Table 3-19. The Cohort 2

PAGE 95

95 residential sub-sample is composed of 12.4% low risk youth (as assessed by the PACT), 17.1% moderate risk youth, 29.2% moderate-high ris k, and 41.3% high risk to re-offend youth. These frequencies are in stark contra st to the 70.9% low risk and only 6.3% high-risk youth who composed the Cohort 2 community sample. The Cohor t 2 residential sample is also slightly less white (52.2% compared to 61.1% for community youth) and more male (85.6% compared to 69.8% for community youth). Table 3-19. Residential Cohort 2 Covariate Descriptive Statistics Covariate Descriptive Statistics N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 1,015 1 4 2.9931 1.04 Sex 1,015 0 1 0.8562 0.351 Race 1,015 0 1 0.4778 0.5 Referral within twelve months 1,015 0 1 0.6059 0.489 Logistic regression analysis of the full s ub-sample of Cohort 2 residential youth using overall risk to re-offend, race (wh ite and non-white) and sex (male and female) as predictors of referral subsequent to release rev eals the covariate to be a significant predictor. The results of the analysis reveal a severely unexpected significance (see Table 3-20). Race is significant with nonwhites more likely to recidivate than whites, as is sex with males more likely than females. However, overall risk to re-offend is significant, but in the reverse direction as expected. Logistic regression shows those with high risk to re-offend being significantly less likely to do so. The same holds true when only risk to re-offend is included as a covariate pr edicting recidivism; the higher the overall risk level, the le ss likely to recidivate for youth in residential facilities prior to the one-year follow-up period. The strongest relati onship to recidivism was with sex, followed by race, and finally overall risk to re-offend. However, for every drop in risk to re-offend, there was still a 0.776 times increase in the dependent variable, recidivism.

PAGE 96

96 The results of the logistic regression for the Cohort 2 residential sub-sample are perplexing and enough cause for the sample to be further di saggregated, using the same approach as that used for the Cohort 1 residential youth. The Cohort 2 residential sa mple was further divided into youth who received the PACT assessment prior to being placed in residential commitment and then followed one year, and youth assessed at so me point during a commitment placement and then tracked post-release (see Table 3-21 for descriptive statistics). The timing of when the PACT assessment was given (prior to, or dur ing commitment placement) may yield some interesting clues as to why the overall risk to re-offend is a significant pr edictor of recidivism, only in the opposite direction, for Cohort 2 residential youth. Table 3-20. Residential Cohort 2 Logistic Regression Results Covariate Statistics Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend -0.23 0.064 13.031 1 .000* 0.794 Logistic Regression 2 Overall Risk to Re-offend -0.254 0.065 15.358 1 .000* 0.776 Sex 0.421 0.182 5.336 1 .021* 1.523 Race 0.312 0.132 5.565 1 .018* 1.366 *=significant at p=.05 Table 3-21. Split Residential Cohort 2 Covariate Descriptive Statistics Covariate Descriptive Statistics PACT before commitment N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 984 1 4 2.9797 1.043 Sex 984 0 1 0.8598 0.347 Race 984 0 1 0.4797 0.5 Referral within twelve months 984 0 1 0.6128 0.487 PACT while in commitment N Mini mum Maximum Mean Std. Deviation Overall Risk to Re-offend 31 1 4 3.4194 0.848 Sex 31 0 1 0.7419 0.445 Race 31 0 1 0.4194 0.502 Referral within twelve months 31 0 1 0.3871 0.495

PAGE 97

97 Logistic regression was again used separate ly to predict recidivism for the Cohort 2 residential youth who were assessed prior to a commitment placement and the youth assessed while in residential placement (see Table 3-22 for all Cohort 2 residential split sample analyses). Initially, only overall risk to re-offend, as asse ssed by the PACT was entered in the model as a covariate. For youth assessed prior to a commitment placement, then subsequently tracked one year, overall risk to re-offend was again signif icant (p=.001), however, lower risk youth were more likely to receive a delinquency referral within the follow-up period. This result is similar to that found when analyzing the full Cohort 2 sub-sample. For youth assessed while in commitment and tracked post-release, overall ri sk to re-offend was not significant (p=.65). However, this result must be interpreted with caution, as there were only thirty-one youth included in the analysis. Table 3-22. Split Residential Cohor t 2 Logistic Regression Results Covariate Statistics PACT before commitment Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend -0.221 0.065 11.683 1 .001* 0.802 Logistic Regression 2 Overall Risk to Re-offend -0.245 0.066 13.869 1 .000* 0.783 Sex 0.317 0.188 2.858 1 .091 1.373 Race 0.301 0.134 5.02 1 .025* 1.351 PACT while in commitment Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend -0.198 0.437 0.206 1 .65 0.82 Logistic Regression 2 Overall Risk to Re-offend 0.063 0.473 0.018 1 .893 1.066 Sex 9.516 34.54 0.076 1 .783 13580.44 Race 0.978 0.89 1.208 1 .272 2.66 =significant at p=.05

PAGE 98

98 Next, sex and race were included in the mode l, along with overall risk to re-offend to predict subsequent referral to the Department of Juvenile Jus tice. For youth assessed by the PACT prior to residential placement, overall risk to re-offend was significa nt (p=.000), again in the opposite direction with youth assessed as lower risk more likely to receive a referral during the one year follow-up. Race was also a significant predictor of subsequent referral (p=.025) with non-whites more likely to be referred then wh ites. For youth assessed while in residential commitment and then tracked after release, none of the three covariates (overall risk, race, or sex) were significant predictors. Ag ain, the logistic regression resu lts for youth assessed while in commitment and tracked one year must be inte rpreted with caution, as only thirty-one youth were included in the analysis. While unexpected, the result of overall risk to re-offend being significant in the opposite direction (low risk youth more likely to recidivate) for yout h assessed prior to entering residential commitment, the finding is consistent with the plethora of research of the Principles of Effective Intervention explored above. The first major principl e, the risk principle, holds that the duration and the intensity of treatment should mimic the risk level of the youth, with higher risk youth receiving more intense treatment for a l onger period of time. The research consistently illustrates, for over thirty years, that low risk to re-offend youth who receive intense services actually re-offend at higher rates than they do when receiving minimal treatment. The findings of the current study illustrate this point, findi ng low risk youth who are sent to residential commitment re-offending at higher rates than high risk youth who were provided the longer, more intense treatment they required. This finding further illustrates the need to keep low risk youth out of long-term residential commitment. Placing low risk to re-o ffend youth in long-term

PAGE 99

99 residential placement actually increases their risk to re-offend and jeopardizes both the future of these individuals and public safety. Unfortunately, the Department of Juvenile Justice does not currently have a method to track the type, frequency, or duration of the delinquency interventions youth receive while in residential commitment programs. Therefore, ther e is no way to distinguish what type or brand name intervention a youth received, how many times per week that intervention was delivered to that youth, and for how many w eeks it was delivered. Had the De partment captured this data, analysis could have been conducted to disti nguish which treatment, or group of treatments, delivered at what frequency and what duration was most ef fective at predicting a lower likelihood of recidivism. The three covariates of treatment type, treatment frequency, and treatment duration could have been included w ith risk to re-offend to investigate possible interaction effects on the predic tion of subsequent referral. However, all that can be deduced from the curre nt study is that the PACT is not predictive of subsequent referral for youth who are asse ssed within residential commitment and tracked post-release. This deduction holds true for bot h Cohort 1 residential yout h tracked six months and Cohort 2 residential youth followed for twel ve months. The finding that sex and race are sometimes significant in Cohort 1 and Cohort 2 reside ntial analysis is not surprising and consistent with prior empirical research that males and non-whites are more likely to re-offend than females and whites. The finding that overall risk to re-offend pred icts which youth assessed prior to entering residential placement, again, illustrates the necessity to avoid long-term residential placement for lower risk youth, and is in keeping with previous research described above.

PAGE 100

100 The inability of the PACT to predict reci divism for the youth placed in residential commitment leads to the remainder of the analys is of the current study to focus on the prediction of recidivism for youth remaining in the community after assessment. One of the original intents of the PACT instrument was for the Department s Probation and Community Intervention branch to more effectively determine which youth under s upervision are more likely to re-offend, and to aid in recommendations to the court in terms of the likelihood of a youth on probation recidivating. The PACT assessment has in fact ac hieved this goal as the recidivism of youth who remain in the community post-assessment has been significantly predicte d in follow-ups of both six and twelve months. The next step of the analys is is to investigate th e role of gender for both Cohort 1 and Cohort 2 youth who remain in the community. Gender Analysis In both the six-month and the twelve-month follow-up sub-samples, race and gender were significant predictors with non-whites and fema les more likely to receive a subsequent delinquency referral than whites and females. Previ ous empirical as well as theoretical research has argued for either the existence of gender-sp ecific prediction instrume nts, and even pathwayspecific prediction instruments for sub-samples of female offenders. As the PACT is utilized to assess risk to re-offend for all yout h referred to the Department, anal ysis of the predictive ability for both males and females is of critical concern. This analysis will be the first known attempt to formally validate the PACT across gender. Logistic regression was used to examine the ability of the PACT to predict recidivism, as measured by a subsequent referral to the De partment, for both males and females who are assessed and remain in the community. As in previous analyses, youth who are tracked six months and remain in the community (Cohort 1) and youth tracked twelve months and remain in the community (Cohort 2) will be assessed separa tely to examine the ability of the PACT to

PAGE 101

101 predict short-term as well as long-term followup recidivism. Descriptive statistics for Cohort 1 and Cohort 2 males and females can be seen in Table 3-23. Table 3-23. Descriptive Statistics By Gender Covariate Descriptive Statistics Cohort 1 Community Males N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 26,643 1 4 1.5675 0.937 Race 26,643 0 1 0.4335 0.496 Referral within six months 26,643 0 1 0.271 0.444 Cohort 1 Community Females N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 9,682 1 4 1.3735 0.789 Race 9,682 0 1 0.4323 0.495 Referral within six months 9,682 0 1 0.1918 0.394 Cohort 2 Community Males N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 5,677 1 4 1.5723 0.941 Race 5,677 0 1 0.3838 0.486 Referral within twelve months 5,677 0 1 0.3527 0.478 Cohort 2 Community Females N Minimum Maximum Mean Std. Deviation Overall Risk to Re-offend 2,455 1 4 1.3739 0.798 Race 2,455 0 1 0.402 0.49 Referral within twelve months 2,455 0 1 0.2758 0.447 Logistic regression results reve al overall risk to re-offend and race (white and non-white) to be significant predictors of subsequent referral, with hi gher risk youth and minority youth more likely to recidivate. These results hold true for male youth who are assessed using the PACT and remain in the community being tracked for six months and for twelve months, as well as female youth assessed and remaining in the community both six and twelve months (see Table 3-24 for results). For both the male youth tracked six months and those tracked twelve months, the relationship to recidivism was slightly str onger for race than overall risk to re-offend. For both groups of female youth, overa ll risk to re-offend has a stronge r relationship to recidivism than race. For all youth, for every increase in ove rall risk to re-offend, th ere is between a 1.4 to 1.6 times increase in recidivism (roughly 50%). Thes e analyses illustrate th e effectiveness of the PACT assessment in the prediction of a delinque ncy referral within a six-month or one-year

PAGE 102

102 follow-up for all youth who remain in the community. The PACT was designed for the purpose of doing precisely that task. The goal was to im plement an instrument that could assess the likelihood that youth that en ter the juvenile justice system would recidivate if they remained in the community. Table 3-24. Logistic Regr ession Results By Gender Covariate Statistics Cohort 1 Logistic Regression Males B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.374 0.014 713.279 1 .000* 1.454 Race 0.405 0.028 203.312 1 .000* 1.499 Logistic Regression Females Overall Risk to Re-offend 0.476 0.028 280.985 1 .000* 1.61 Race 0.269 0.053 25.963 1 .000* 1.309 Cohort 2 Logistic Regression Males B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.336 0.029 131.779 1 .000* 1.4 Race 0.466 0.058 64.836 1 .000* 1.593 Logistic Regression Females Overall Risk to Re-offend 0.436 0.053 67.554 1 .000* 1.547 Race 0.328 0.093 12.398 1 .000* 1.389 =significant at p=.05 The ability of the PACT to predict recidivism for both males and females is a critical piece of the current study. The significant prediction of a new referral for female juveniles is essential for the PACT to achieve the desired goal and to advance from previous prediction instruments that were unable to predict female offending. Re isig and colleagues, wh en assessing the LSI-R prediction instrument for females state, Predictive accuracy exists if recidivism increases with risk-need levelthe relationship between risk-need a nd recidivism (when using the LSI-R for females) is not statistically significant for the full sample (Reisig et al ., 2006:397). Contrary to the results found in that analysis the current study finds the over all risk level identified by the

PAGE 103

103 PACT statistically significant in the prediction of full-sample female recidivism. These results are extremely important with regard to literat ure stressing the need to have gender-specific instruments or separate instruments for females and males. The PACT has been proven to predict re-offending for males and for females equally well, with both a six and a twelve-month followup. While the prediction of re-offending for both male s and females is an essential goal of the current study, it is important to know whether th e same aspects of the PACT are predictive for males and females. The construction of the PACT allows for analyzing whether the criminal history of the youth or the social history of the youth is driving th e significance of the prediction. Therefore, the next step in the analysis is to se parate overall risk to re-offend into its two core components of criminal history score and soci al history score to examine the predictive significance of each separately for male and female youth. Logistic regression is used to predict a subsequent referral to the Department for bot h male and female youth who are assessed and remain in the community fo r twelve months (Cohort 2). Table 3-25. Cohort 2 Descriptive Statistics By Gender Covariate Descriptive Statistics Cohort 2 Males N Minimum Maximum Mean Std. Deviation Criminal History Score 5,677 0 28 6.51 4.695 Social History Score 5,677 1 16 4.59 2.425 Race 5,677 0 1 0.3838 0.486 Referral within six months 5,677 0 1 0.3527 0.478 Cohort 2 Females N Minimum Maximum Mean Std. Deviation Criminal History Score 2,455 0 27 5.25 3.913 Social History Score 2,455 0 15 3.87 2.905 Race 2,455 0 1 0.402 0.49 Referral within twelve months 2,455 0 1 0.2758 0.447 The covariates for this analysis are criminal history, social history, and race, rather than race and overall risk to re-offend utilized in previous analysis (see Table 3-25 for dependent

PAGE 104

104 variable and covariate descriptive statistics). This allows for the determination as to whether the extent and seriousness of previous official offending is driving th e significance of the prediction, or whether the social, educational, family, mental health, substance abuse, peer relationships, or abuse history circumstances are driving the pr ediction for both males and females. Logistic regression results show that criminal history, soci al history, and race are significant predictors of subsequent referral for both males and females. The higher the criminal history and the higher the social history, the more likely the youth will receive another refe rral. Both of these covariates are significant at p=.000 for males and for females (see Table 3-26 for results). Identical analyses were conducted on males and females who remain in the community and followed six months (Cohort 1) post-assessment, with identical resu lts. Furthermore, non-whites were again more likely to receive a subsequent referral. While ra ce has the strongest rela tionship to recidivism among the predictors, for both males and females, the relationship between social history and recidivism is stronger than that between crimin al history and recidivism For every increase in the social history score, there is a 1.138 tim es increase in recidivism for males and a 1.152 increase for females. Table 3-26. Cohort 2 Logist ic Regression By Gender Covariate Statistics Logistic Regression Males B S.E. Wald df Sig. Exp(B) Criminal History 0.044 0.006 50.246 1 .000* 1.045 Social History 0.129 0.012 112.92 1 .000* 1.138 Race 0.487 0.059 68.206 1 .000* 1.627 Logistic Regression Females B S.E. Wald df Sig. Exp(B) Criminal History 0.079 0.012 42.053 1 .000* 1.083 Social History 0.141 0.017 68.163 1 .000* 1.152 Race 0.313 0.097 10.449 1 .000* 1.368 =significant at p=.05 An additional hypothesis of this current study involving gender differences is the notion that any significance associated with gender being a predictor of recidivism will be washed out

PAGE 105

105 with the inclusion of additional covariates. Some researchers have argued that gendered pathways exist in that the tract to delinquency fo r females is often different than that for males (see Daly 1992, 1994; Daly & Chesney-Lind 1988) The current study hypothesizes it may be possible to include variables associated with fe male pathways in the prediction of recidivism which will make the seemingly significant findi ng that gender predicts re-offending become insignificant in the logist ic regression model. Table 3-27. Cohort 2 Community Gendere d Covariate Descriptive Statistics Covariate Descriptive Statistics N Minimum Maximum Mean Std. Deviation Sex 8,132 0 1 0.6981 0.459 History of Physical Abuse 8,132 0 1 0.1299 0.336 History of Sexual Abuse 8,132 0 1 0.0454 0.208 Drug Use 8,132 0 1 0.3915 0.488 Alcohol Use 8,132 0 1 0.2907 0.454 School Suspension/Expulsion 8,132 0 6 0.6369 1.433 Teachers/Coaches 8,132 0 5 0.3789 0.918 Positive Adult Non-family 8,132 0 4 0.4292 0.934 History of Running Away 8,132 0 5 1.3876 0.92681 Parental Supervision 8,132 0 3 0.3296 0.708 Antisocial Peers 8,132 0 2 0.6779 0.536 Referral within twelve months 8,132 0 1 0.3294 0.47 To test the above hypothesis, the Cohort 2 subsample of youth who are assessed using the PACT and subsequently remain in the community to be tracked for twelve months will be used. Gender was a significant predictor of recidivism for this sub-sample in earlier analyses. Now, however, additional covariates will be introduced in the prediction of recidivism that have been suggested to be issues th at lead to female offending. In addition to the sex of the youth, history of physical abuse, history of sexual abuse, history of alcohol use, history of drug use, anti-social peer involvement, history of school suspension or expulsion, number of teachers or coaches the youth is comfortable talking to, history of adu lt non-family school or job positive relationships,

PAGE 106

106 history of running away or getting kicked out of the house, and level of parental supervision will be included. All covariates are coded so that th e higher the value, the gr eater the extent of the issue defined (see Table 3-27 fo r descriptive statistics). Logistic regression was again utilized to asse ss the significance of the covariates in the prediction of recidivism within twelve mont hs for youth who remain in the community post PACT assessment (see Table 3-28 for results). Youth who have a history of more school suspensions and expulsions are si gnificantly more likely to recidi vate, as are youth with more instances of running away, inconsistent/inadequa te supervision, and a higher prevalence of antisocial peers. Interestingly, yout h with more relationships with pro-social adults who are not employers or teachers, have a greater likelihood of recidivism. History of physical abuse and history of sexual abuse were not significant predictors in the mode l (though physical abuse approached significance at p= .067) neither were history of alc ohol abuse and problems related to that use, or presence of teachers or coaches the youth feels comfortable with. Furthermore, and congruent with hypothesis seven of the current st udy, sex was washed to insignificance in the presence of the above-mentioned covariates. This finding runs counter with previous research by Hubbard (2002) who found gender to remain signi ficant, even after controlling for risk to reoffend, quality of programming, sexual abuse, and se lf-esteem. The stronges t relationship to the dependent variable, recidivism, was with antisocial peers. For ever y increase in that independent variable, there was a 1.34 times increase in the dependent variable. The second strongest relationship was between pa rental supervision and recidivism, followed by school suspension/expulsion and recidivism, then histor y of running away and recidivism, and finally prevalence of non-family/employment/sc hool positive adults and recidivism.

PAGE 107

107 Table 3-28. Cohort 2 Community Expande d Covariate Logistic Regression Covariate Statistics Logistic Regression B S.E. Wald df Sig. Exp(B) Physical Abuse 0.263 0.143 3.365 1 .067 1.3 Sexual Abuse -0.367 0.243 2.28 1 .131 0.692 Alcohol -0.066 0.131 0.257 1 .613 0.936 Drugs 0.144 0.127 1.281 1 .258 1.155 School Suspension/Expulsion 0.172 0.033 28.031 1 .000* 1.188 Teachers/Coaches -0.042 0.049 0.737 1 .391 0.959 Non-family Adults -0.112 0.056 4.058 1 .044* 0.894 Running Away 0.109 0.054 4.133 1 .042* 1.116 Parental Supervision 0.231 0.083 7.824 1 .005* 1.26 Antisocial Peers 0.293 0.098 9.03 1 .003* 1.34 Sex 0.109 0.127 0.747 1 .388 1.116 =significant at p=.05 This finding is extremely important in that it shows that apparent sex differences may, in actuality, be due to proportional differences in the extent to which various risk factors are present for males and females. This supports the notion that sex differences may be differences in degree, rather than differences in kind. While beyond the scope of the current study, future empirical endeavors should look to see whether qu alitative differences exist between the physical and sexual abuse suffered by male and female yout h and the implications of that abuse on the youth. The finding in the previous analysis of phys ical and sexual abuses not being significant predictors of recidivism is surp rising and deserves further attenti on. It may be the case that such abuse is more important in the prediction of fema le offending as it is in male offending behavior. For the 2,455 females and the 5,677 males in the Cohort 2 Community sub-sample, 11.5% of the females and only 1.5% of the males reported sexu al abuse, and 16.6% of the females and 11.4% of the males reported a history of physical abus e. Due to these differences, physical abuse and sexual abuse, along with race, history of alcohol use, history of drug use, anti-social peer

PAGE 108

108 involvement, history of school suspension or expu lsion, number of teachers or coaches the youth is comfortable talking to, history of adult non-fa mily school or job positiv e relationships, history of running away or getting kicked out of the house and level of parental supervision, are utilized as covariates in the prediction of recidivism se parately for males and females (see Table 3-29 for descriptive statistics). This will allow for the examination as to whether certain life events, namely physical and sexual abuse, matter more in the prediction of re-offending for one sex than they do for the other. Table 3-29. Cohort 2 Community Gender-Spl it Covariate Descriptive Statistics Covariate Descriptive Statistics Cohort 2 Community Males N Minimum Maximum Mean Std. Deviation Race 5,677 0 1 0.3838 0.486 Physical Abuse 5,677 0 1 0.1143 0.318 Sexual Abuse 5,677 0 1 0.0153 0.123 Antisocial Peers 5,677 0 2 0.7013 0.541 Alcohol Use 5,677 0 1 0.2869 0.452 Drug Use 5,677 0 1 0.4104 0.492 School Suspension/Expulsion 1,333 1 6 2.934 1.674 Teachers/Coaches 1,192 1 5 1.9069 1.145 Non-Family Adults 1,333 1 4 2.0645 1.023 Running Away 5,677 1 5 1.291 0.795 Parental Supervision 1,310 1 3 1.5351 0.694 Referral within twelve months 5,677 0 1 0.3527 0.478 Cohort 2 Community Females N Minimum Maximum Mean Std. Deviation Race 2,455 0 1 0.402 0.49 Physical Abuse 2,455 0 1 0.1658 0.372 Sexual Abuse 2,455 0 1 0.1149 0.319 Antisocial Peers 2,455 0 2 0.624 0.523 Alcohol Use 2,455 0 1 0.2994 0.458 Drug Use 2,455 0 1 0.3479 0.476 School Suspension/Expulsion 442 1 6 2.8688 1.615 Teachers/Coaches 414 1 5 1.9517 1.147 Non-Family Adults 442 1 4 2.052 0.99 Running Away 2,455 1 5 1.611 1.146 Parental Supervision 434 1 3 1.5415 0.706 Referral within twelve months 2,455 0 1 0.2758 0.447

PAGE 109

109 Table 3-30. Cohort 2 Community Gender-S plit Covariate Logistic Regression Covariate Statistics Logistic Regression Males B S.E. Wald df Sig. Exp(B) Race 0.823 0.132 38.909 1 .000* 2.278 Physical Abuse 0.324 0.179 3.292 1 .070 1.383 Sexual Abuse 0.091 0.417 0.048 1 .827 1.096 Alcohol 0.012 0.157 0.006 1 .938 1.012 Drugs 0.337 0.149 5.136 1 .023* 1.401 School Suspension/Expulsion 0.164 0.039 17.478 1 .000* 1.178 Teachers/Coaches -0.077 0.058 1.749 1 .186 0.926 Non-family Adults -0.082 0.065 1.586 1 .208 0.921 Running Away 0.073 0.072 1.047 1 .306 1.076 Parental Supervision 0.219 0.098 4.953 1 .026* 1.245 Antisocial Peers 0.278 0.114 5.929 1 .015* 1.32 Logistic Regression Females B S.E. Wald df Sig. Exp(B) Race 0.578 0.237 5.958 1 .015* 1.783 Physical Abuse 0.379 0.257 2.177 1 .14 1.461 Sexual Abuse -0.530 0.325 2.659 1 .103 0.588 Alcohol 0.412 0.282 2.135 1 .144 1.51 Drugs -0.257 0.274 0.884 1 .347 0.773 School Suspension/Expulsion 0.097 0.065 2.227 1 .136 1.102 Teachers/Coaches 0.05 0.097 0.264 1 .607 1.051 Non-family Adults -0.286 0.116 6.057 1 .014* 0.751 Running Away 0.188 0.088 4.611 1 .032* 1.207 Parental Supervision 0.114 0.165 0.473 1 .492 1.121 Antisocial Peers 0.284 0.202 1.983 1 .159 1.329 =significant at p=.05 Results of the analysis (reported in Table 3-30) reveal somewhat different predictors to be significant for males and females. Both male a nd female re-offending is significantly predicted by being non-white. However, all other significant covariates were different for males and females. Starting with the indepe ndent variable with the strongest relationship to recidivism to the weakest, greater histories of drug use and problems associated with that use (such as drugs disrupting education, causing family conflict, cau sing health problems, w ithdrawal issues, and tolerance increase), having antisocial peer associ ations, and inadequate/in consistent parental supervision, and a greater histor y of school suspensions or expul sions, is predictive of male

PAGE 110

110 recidivism. The only covariates, other than race, that are predictive of female recidivism are greater history of running away and more relationshi ps with pro-social adu lts other than teachers and employers. Therefore, males and females have different covariates that are predictive of re-offending behavior, even though the overall risk to re-offend and the social history score of the PACT is predictive for both. Perhaps most surprising of the difference is that a greater presence of antisocial peers is not predictive of female r ecidivism. The presence of either a history of physical abuse or a history of sexual abuse was no t significant for either the male or the female sub-samples. Violence Analysis The final goal for the current study is to examine whether the PACT assessment can predict violent re-offending6. Previous analyses in this study ha ve illustrated the ability of the PACT to predict recidivism, but not necessarily violent recidivism. The ability to predict which youth will re-offend with a subsequent violent offense was never the goal of the PACT assessment, or the Florida Department of Juveni le Justice. The PACT was designed to predict which youth are more likely to recidivate genera lly, not specifically re lated to the level of violence of the subsequent offense. While many studies have shown there to be a limited level of offense specialization among juveniles (meaning a youth who has committed a violent offense is no more likely to commit a violent offense in the future than a youth who did not commit an initial violent offense), criminologists and the pub lic are still interested in violent offending. 6 The current study also validated the PACT with respect to predicting recidi vism for both nonwhite and white youth to ensure there was no difference in the predictive ab ility of the instrument for different racial groups. Results reveal the PACT to be a valid predictor for both. Further analyses illustrate the social history score to have a stronger relationship with recidivism than the criminal history score, though both were significant at p=.000.

PAGE 111

111 The current analysis will use a dependent variable similar to that used in previous analyses above of a delinquency referral to the Department within twelve months. However, for this analysis, only violent offenses will count for reci divism, meaning that the analysis will examine whether the PACT can predict which youth are more likely to commit a violent offense within twelve months post-assessment. For the purposes of this analysis, any delinquency referral within twelve months will count as violent if th e most serious charge in of the referral was murder/manslaughter, attempted murder/manslaught er, felony sexual battery, armed robbery, or aggravated assault and/or battery. The Cohort 2 Community sub-sample will be used in the analysis (youth assessed who remain in the community and are tracked twelve months postassessment). Of the 8,132 youth in that sub-sample, 386 (4.7%) committed a violent offense (using the criteria of this analys is), scored dichotomously as a zero or a one with a mean of 0.0475 and a standard deviation of 0.21265. Table 3-31. Cohort 2 Community Viol ent Offense Logistic Regression Covariate Statistics Logistic Regression 1 B S.E. Wald df Sig. Exp(B) Overall Risk to Re-offend 0.455 0.046 96.278 1 .000* 1.576 Sex 0.322 0.126 6.512 1 .011* 1.38 Race 0.817 0.109 56.599 1 .000* 2.265 Logistic Regression 2 B S.E. Wald df Sig. Exp(B) Criminal History 0.072 0.01 53.429 1 .000* 1.075 Social History 0.103 0.02 26.732 1 .000* 1.109 Sex 0.271 0.127 4.55 1 .033* 1.311 Race 0.824 0.11 56.032 1 .000* 2.279 =significant at p=.05 Logistic regression is used to predict violent recidivism w ithin twelve months using the overall risk to re-offend level from the PACT, se x, and race (white and non-white). An additional analysis uses logistic regressi on to predict violent recidivism once again, but instead uses the

PAGE 112

112 criminal history and social history scores composi ng the PACT rather than the overall risk to reoffend level. This will determine whether it is the social history or criminal history of the youth that are driving the prediction of violent recidivism, or both (results of both analyses can be seen in Table 3-31). The first analysis reveals ove rall risk to re-offend as ev idenced by the PACT to be a significant predictor of a violent offense subseque nt to assessment for youth who remain in the community and are tracked twelve months for follow-up recidivism. Yout h who are assessed as higher risk on by the PACT are significantly more likely to commit a violent offense during the twelve-month follow-up than lower risk youth. The results also reveal males and non-whites to be more likely to commit violent offenses. Wh ile race has the strongest relationship with recidivism, for every increase in overall risk to re-offend, ther e is a 1.576 times increase in the dependent variable (over 50%). The second analysis reveals that both the youths prior criminal history, and the seriousness of that history, as well as the soci al circumstances surroundi ng the youths life both significantly predict subsequent violent offendi ng. Youth who have a greater and more serious criminal history are more likely to commit violen t offenses during the follow-up period, as are youth with more risky social circumstances Again, males and non-white youth are also more likely to commit violent offenses within the tw elve month time frame. While race, followed by sex have the two strongest relationships to a violent recidivism, soci al history, again, has a stronger relationship with recidivism than crim inal history. These two analyses validate the PACT for youth assessed who remain in the comm unity as an assessment capable of predicting, on an aggregate level, which youth are more likel y to commit a violent o ffense within twelve months if allowed to remain in the community.

PAGE 113

113 CHAPTER 4 DISCUSSION The main purpose of the current study was to validate the Positive Achievement Change Tool (PACT), the risk/needs assessment utili zed by the Probation and Community Intervention program area of the Florida Department of Juven ile Justice. Eight of the ten hypotheses of the study were supported. The overall risk to re-offe nd level presented by the PACT is indeed a significant predictor of recidivism for youth who are assessed and remain in the community, with higher risk youth more likely to re-offend. The purpose of the development of the PACT was to create an instrument that coul d help Juvenile Probation Officer s, and the Department, classify youth in terms of likelihood of re-offense if all youth remained in the community. Youth assessed as lower risk to re-offend by the PACT ar e indeed significantly less likely to commit an offense if allowed to remain in the community than youth assessed as higher risk. Furthermore, the likelihood of re-offense is exactly as the PACT predicts with low risk youth less likely than moderate risk, who are in turn less likely than moderate-high risk, which are less likely than high-risk youth to recidivate. The analysis related to this first hypothesis confirms the PACT as a valid risk assessment. The second hypothesis was not supported. Th e hypothesis held that youth assessed as higher risk to re-offend would indeed do so re gardless of whether the youth was assessed and remained in the community, or assessed prior t o, or during, a residentia l commitment placement. The youth who remained in the community re-o ffend as predicted by the PACT. However, neither youth assessed prior to commitment placem ent, nor youth assessed while in commitment placement, recidivated subsequent to release as predicted by the PACT. Several hypotheses for these phenomena were presented. First, and consistent with the literature related to the Principles of Effective Intervention (specifically the risk pr inciple), intensive treatment provided to low risk

PAGE 114

114 youth actually increases their lik elihood of recidivating. Long-ter m residential treatment is beneficial to high-risk youth, and they theref ore improve from the PACT assessment conducted prior to placement. However, re sidential placement provided to youth assessed as low risk prior to residential placement is actu ally iatrogenic, making the low ri sk youth worse and more likely to re-offend. The extant literature holds that reassessments are a necessary component in order to gather information related to progression of criminogeni c needs. The Department currently has a policy that all youth who remain in the community must be reassessed every ninety days in order to be certain the treatment plans for youth on probation ar e in keeping with the risks and needs of the youth. While beyond the scope of this study, future analyses should examine how PACT overall risk to re-offend scores, and indi vidual scores for the domains of the PACT rise and fall as it relates to the treatment/s ervices youth receive. An alternative hypothesis as to why the PACT is not predictive of recidivism for youth assessed prior to residential placement is rela ted to the treatment yout h receive while committed. The current study did not control fo r the type of treatment youth re ceive while in the more than 120 residential commitment faciliti es in Florida operated or c ontracted by the Department. Treatment variables are as of yet unavailable wi thin the Departments JJIS database making it impossible to know from the monthly data extrac ts exactly what services each youth received while in commitment. The finding that the PACT is not predictive for youth assessed while in commitment and tracked subsequent to release becomes logica l once the actual questi ons of the PACT are examined. Several of the questions essential to the scoring of the PACT make the instrument artificial and directly related to the circumst ance of being committed. For example, Current use

PAGE 115

115 of alcohol, Current use of drugs, Current friends, as well as Youths attendance in the most recent (school) term are all fixed due to the youth being committed. There are no opportunities to engage in the use of alcohol or drugs, all of the other you th with which the youth interacts are arguably antisocial, as they are all in a residential co mmitment facility, and all youth are required to attend school every weekday. Many of the questions themselves force all youth to look a certain way and therefore many of the questions relevant to accurate assessment, scoring, and differentiation amongst clas sifications, become artificial. The Department is currently designing, al ong with the developers of the PACT, an instrument that will alleviate this issue. Th e new instrument will attempt to measure progress during residential placement on decreasing crimi nogenic needs and risk f actors, and increasing protective factors. The instrument is still in its formative stage, soon to be piloted then evaluated. The current study does, however, raise one issue with respect to the current business rules associated with PACT assessments. Currently all youth are assessed w ithin sixty days of projected release from a residential facilit y, using the PACT assessm ent. The present study shows the assessment prior to release to not be a significant predictor of recidivism for these youth when they are released (more than likely du e to the issue of improper questions for youth in residential, as discussed in the paragraphs above). The Department may benefit from using the assessment conducted sixty days prior to release as a way to guide aftercare planning, but should be sure to reassess youth within a relatively s hort time of actual release from commitment to gauge reentry success more effectively. The third and fourth hypothese s of the current study were both supported. The third hypothesis held that the higher the criminal hist ory score, as presented by the PACT, the more likely the youth is to re-offend. This was indeed th e case, as a greater extent of prior official

PAGE 116

116 offenses, and greater seriousness of those offenses was a significant predictor of recidivism. The identical held true for the social history score, as the higher the score, th e more likely the youth is to re-offend. The current study did not go into de pth as to which components of the criminal history score or social history score were driving the significan ce of those scores, except for some analysis with respect to sex differences. Fu ture research would bene fit from more in-depth analysis to uncover the specific com ponents responsible for the significance. The fifth hypothesis was also supported, and was perhaps one of the most important findings of the current study. The hypothesis held th at the overall risk to re-offend level, as evidenced by the PACT, would be a significant pr edictor of both male and female recidivism. The analysis conducted by the current study on the ab ility of the PACT to predict female as well as male re-offending validate the use of the PACT in the state of Florida for all youth under the care and custody of DJJ. The current study illust rates there exists no need, with respect to predicting re-offending, for a ge nder-specific assessment as the PACT predicts female and male re-offending equally well. The Pact esta blishes which classifications of youth will recidivate (on an aggregate level) if they are able to remain in the community. Therefore, the PACT is an appropriate assessment for DJJ to use to determine which youth should be recommended for community placement and which youth are significantly more likely to be a detriment to public safety if they ar e allowed to remain in the community. The finding that the PACT significantly predicte d recidivism for the full sample of female youth assessed and subsequently re maining in the community differe ntiates the PACT from other risk assessment instruments unable to predic t full sample female re-offending, and only reoffending of sub-groups of females. As mentione d in the review of th e literature, previous research indicates that actuarial assessments (sp ecifically the LSI-R) are no t predictive of female

PAGE 117

117 offending as a whole, but are valid for predic ting the recidivism for economically motivated women (Reisig et al., 2006). The current study shows the PACT to be a valid instrument for the prediction of female offending, with out the necessity of having to sub-group the females, or only being able to predict re-offend of certain of thos e sub-groups. This finding runs counter to much of the gendered pathways research asserting th e need to examine the ways females become involved in offending behavior, and the rationale of why those fema les engage in that behavior. The sixth hypothesis was also supported by th e current study. Males ar e significantly more likely to re-offend than females. This pattern held true even after contro lling for overall risk to re-offend as evidenced by the PACT and a variety of other criminal hist ory and social history covariates. This finding is in keeping with a plethora of crimi nological literature and empirical research, such that it is a univers al truth in criminology that males are more likely to recidivate than females regardless of time period, age of the youth, country, or culture. The seventh hypothesis was supported by the cu rrent study. The hypothesis asserted that any sex differences that would appear would be washed out with the inclusion of proxy measures. The inclusion of sexual and physical a buse history, drug and alcohol abuse, delinquent peer associations, history of school expulsion/sus pensions, number of teachers/coaches the youth is comfortable with, positive adu lt non-family relationships, history of running away, and level of parental supervision did ind eed decrease the significance of simp ly being male or female. Sex was reduced from being a significant predictor of recidivism, with males more likely to reoffend, by the inclusion of the proxy measurements Future direction for additional analyses should include a more in-depth analysis of th e more than 126 PACT measures that may be proxies for sex to examine sex diffe rences in re-offending predictors.

PAGE 118

118 In relation to sex difference, the eighth hypothesis was supported. The hypothesis was that a history of physical or sexual abuse would not be a stronger predictor of female re-offending than of male re-offending. The original notion behind the hypothesis was related to the seventh hypothesis of proxy measures being ab le to reduce sex (being male or female) to insignificance, which did indeed occur. Although the basis for the prediction that physical and sexual abuse would not be stronger predictors of female r ecidivism did not play out as hypothesized, the analysis led to interesting findi ngs. History of physical abuse and history of sexual abuse are not stronger predictors for female re-offending than for male re-offending because neither physical or sexual abuse history was found to significantly predict either ma le or female recidivism. The seventh hypothesis (correctly) be lieved that inclusion of proxy measures such as physical and sexual abuse would wash out the si gnificance of being male or female, meaning that it is really the circumstance of having been abused that led to delinquent behavior, and not being male or female. The eighth hypothesis held that males w ho were abused physically or abused sexually would be just as likely as females with those histories to re -offend. The hypothesis was supported, but not in the way in which it was in tended. The physical and se xual abuse history did not lead to the offending behavior as believed, because the abuse histories did not significantly predict the behavior at all. The final hypothesis examined whether the PA CT was predictive of violent recidivism. While the original intent of the PACT was to enable the prediction of which youth would be more likely to re-offend if they remained in th e community, and to categor ize that likelihood into four components (low, moderate, moderate-high, and high), researchers have often sought to examine whether risk/needs assessments can predic t offenses more detrimental to public safety. Violent offense prediction is often first on the li st of subcomponents of offense to attempt to

PAGE 119

119 predict. Hypothesis ten was supporte d as the analysis showed the ove rall risk to re-offend score of the PACT to be a significant predictor of violent recidivism for youth assessed who remain in the community. It is a significant step for the PACT to be able to significantly pr edict which youth, on an aggregate level, are more likely to commit a violent offense if they remain in the community than other youth. Furthermore, only a very na rrow range of offenses was included in the violent category. Only murder/manslaughter, attempted murder/manslaughter, felony sexual battery, armed robbery, and aggrav ated assault and/or battery we re included. This extremely strict assortment of offenses was included so as to be certain the analysis was capturing the most violent against person offenses possible. Future research should examine whether the PACT is equally as successful at predicti ng other categories of offenses, such as property or drug offenses. Future direction may also include examination of which of the mo re than 46 questions within the Pre-Screen, and which of the more than 126 ques tions within the Full Assessment are driving the prediction of violent offenses.

PAGE 120

120 CHAPTER 5 CONCLUSION The Florida Department of Juvenile Justic e received over 150,000 delinquency referrals, representing over 94,000 youth, between July 1, 2005 and June 31, 2006. The Department has developed and implemented a risk/needs assessment to help predict which youth are more likely to recidivate if they remain in the community, as well as to identify the criminogenic needs of all youth referred so as to provide appropriate treat ment to reduce those ri sk factors and increase protective factors. The purpose of the current study was to validate that instrument, the Positive Achievement Change Tool (PACT), for the youth referred in Florida. The current study utilized the full population of PACT assessments, one for every youth referred making any issue regarding the generalizabilty of the findings irrelevant. The analysis shows that the PACT is indeed a valid instrument and that the higher the risk to re-offend as evidenced by the PACT, the more likely recidivism will occur for youth assessed who remain in the community. Furthermore, the PACT is valid for both males and females. The finding that the PACT significan tly predicts female offending as well as male offending is substantial, and places the PACT above other risk/needs assessments, which have been found incapable of predicting full-sample female recidi vism. Additional analyses showing the ability of the PACT to predict not only general recidi vism, but violent offending as well are also important. These analyses validate th e use of the PACT for the predic tion of risk to recidivate for all youth referred to the Departme nt. The PACT was able to predic t both short-term (six month) and long-term (one year) recidivism. The PACT, for the youth in the current sample, was capable of predicting recidivism for youth assessed prior to entering residential commitm ent facilities that are tracked subsequent to release, only in the opposite direction as intende d. For the youth assessed as higher risk to re-

PAGE 121

121 offend, this may be mainly due to progress within specific criminogenic needs as a result of the treatment youth receive within resi dential facilities. However, ther e currently exists no way to discern what treatment youth receive based on the monthly data extracts from the Departments JJIS database and contact with each of the over 1 20 residential facilities would have been futile as there are often no records of the frequency or duration of particul ar treatments for each individual youth. For the youth a ssessed as low risk to re-offend, the extant literature coupled with the results of the current study suggest this is due to the negative implications of providing low risk youth intensive treatment. The PACT was also not able to predict re-o ffending for youth assessed while in residential commitment facilities and tracked subsequent to release. The ma jor limitation of this finding is that only thirty-one youth composed the sub-sa mple of youth assessed while in commitment and tracked one year. Possibilities for this inability include the small sample size lacking power to capture statistical significance and the artificial nature of some PACT questions for youth in a residential commitment facility. Future studies should revisit this issue when a greater number of PACT assessments have been conducted on youth wh ile in residential pla cement. In the event the results of this study are replicated, and the PACT admi nistered while a youth is in commitment is still not predictive of recidivi sm post-release, the Department may wish to consider ensuring an additional PACT is administered shortly after release so as to gain a more accurate picture of the risk and protectiv e factors for youth while in the community.

PAGE 122

122 CHAPTER 6 FUTURE DIRECTIONS Future empirical research shoul d attempt to provide a more in -depth analysis regarding the individual questions which compose the four dom ains of the PACT Pre-Screen and the twelve domains of the Full Assessment. Each domain is composed of multiple questions and factor analysis would be a useful tool to examine whether domains are individual constructs or multiple factors. Furthermore, scale analysis could be used to examine the domains as they relate to the social history score composing one half of the scoring matrix used by the PACT to determine risk to re-offend. A substantial amount of work remains to be done regarding the i ndividual domains, or, more specifically, questions, which drive the sign ificance of male versus female offending. The finding that the social histor y score significantly predicts male and female offending is important, but it may be the case that certain indi vidual covariates that compose that score are more essential to female offending and others mo re essential to male offending. The validation of the PACT for both males and females was the initi al step, and essential if the instrument is to be used to assess both sexes, but more extens ive work should be undertaken to examine the covariates which lead to female offending beha vior and, separately, male offending behavior, so the appropriate criminogenic needs ca n be targeted with treatment. An additional stream of future research s hould focus on examining whether there may exist a better alternative for scoring overall risk to re-offend based on the information gathered by the PACT. As the PACT has been found significant in the prediction of recidivism for almost every analysis attempted by the current study, this task may be diff icult, but worthy of undertaking. As the current study was an attempt to validate an ex isting instrument, there was no need to create a construction sample and validation samples. However, ample data exist for the attempt to create

PAGE 123

123 an instrument based on various PACT covariates and scoring of those c ovariates to then be validated and assessed for shrinkage. Additional future directions may also wish to examine whether more simple combinations of covariat es are equally as predictive as the complex, automated scoring of the PACT assessment. Perhaps the most important analysis that has yet to be conducted relates to PACT reassessments. Future work should examine the ways scores on the various domains of the PACT fluctuate from one assessment to the next More importantly is the examination of the ways in which the individual domain scores fluc tuate when youth are provided with specific treatments. Considerable work should be unde rtaken related to how the provision of a standardized evidence-based treatment curriculum positively, or negatively affects a domain score, and whether a similar effect for the id entical treatment is observed for both males and females. The current study in conjunction with the empi rical work suggested above would take the question of What works? to What works for whom? Answers to these questions will guide the Department toward providing the best treatment possible for the risk factors and needs presented by the youth served. The provision of proven effective, evidence-based treatment, targeted to specific individuali zed criminogenic needs, as presented by a validated risk/needs assessment as the PACT has shown itself to be in the current study, is the quintessential step to reducing juvenile recidivism, a nd thereby increasing public safety, and to improving the lives of the youth served throughout the state of Florida.

PAGE 124

124 APPENDIX A HOW TO READ AND INTERPRET THE PACT OVERVIEW AND THE YES CASE PLAN REPORTS Figure A-1. PACT Overview Report

PAGE 125

125 The PACT ( Positive Achievement Change Tool ) is designed to accomplish four basic objectives: Determine a youths level of risk for re-offending as a way to target resources to higherrisk youth. Identify the risk and protective factors linked to criminal behavi or so that the rehabilitative effort can be tailored to address th e youths unique assessment profile. Develop a case management approach focu sed on reducing risk factors and increasing protective factors. Allow managers to determine if targeted factors change as a result of the courts intervention. The Overview Report helps Floridas juvenile justice practitioners accomplish those objectives by showing, at-a-glance: Overall level of risk to re-offend: In simple terms, a youth with a low overall level of risk is much less likely to commit another crime than a youth with a high overall level of risk. Indeed, according to the Pre-Va lidation Study of the PACT by the Florida Justice Research Center conducted in September 2005, recidivism rates for higher-risk youth are 57% compared to 17% for low-risk youth. This is key informati on because the research l iterature is clear that what works in reducing recidivism is matching th e levels of treatment with the risk levels of offenders. High-risk offenders require intensiv e interventions to reduce recidivism, while lowrisk offenders benefit most from low intens ity interventions or no intervention at all. Risk factors are the circumstances or events in the youth s life that increase the likelihood that the youth will start or continue criminal activities. In the Overview Report, they are prioritized from highest to lowest; the higher the risk score, the more risk in that area. They can be static or dynamic. Static factors are circumstances in a youths life that are historic and cannot be changed, such as a hi story of physical abuse. In the Overview Report, Domain 1, Record of Referrals, and the other domains labele d with an A are static. All the others are dynamic. Dynamic factors are circumstances or conditions in a youths life that can potentially

PAGE 126

126 be changed, such as the youths friends or schoo l performance. Dynamic factors are used to guide the rehabilitative effort. Protective factors are circumstances or events in th e youths life that reduce the likelihood of the youth committing a crime, those positive thi ngs that help the youth overcome adversity. Domains: The PACT full assessment includes 12 major domains related to juvenile delinquency and continued criminal activity base d on the research lite rature: (1) Criminal History; (2) Gender; (3) School; (4) Use of Fr ee Time; (5) Employment; (6) Relationships; (7) Family/Current Living Arrangements; (8) Al cohol and Drugs; (9) Mental Health; (10) Attitudes/Behaviors; (11) Aggression; and (12) Skills. Pre-Screen vs. Full Assessment: The PACT implementation consists of a two-stage process. The first stage is a pre-screen a ssessment completed for all youth placed on probation. The pre-screen is a shortened version of the full assessment that quickly indicates whether a youth is of low-, moderate-, moderate-high-, or high-risk to re-offend. Each of these levels is presumed to have distinctly different recidivism rates. The second stage, a full assessment, is requ ired by DJJ business rules, for youth assessed as moderate-high or high-risk on the pre-screen. The full assessment identifies a youths risk and protective factor profile and then forwards only those factors that are dynamic to the YES Case Plan to guide rehabilitative efforts. The YES Case Plan Report displays the dynamic factors th at are influencing the youths anti-social behavior, and it provides the facts (e vidence) needed to develop a Case Plan to effectively address these prioritized criminogenic needs. In a very focused way, the PACT and the YES Case Plan help JPOs and judges define the youths unique beha vioral problems and create goals and actions to resolve those problems.

PAGE 127

127 Figure A-2. YES Case Plan The responsivity principle: The Assessments.com software simplifies and organizes this process and provides a critical reporting function for monito ring individual and systemic progress toward the desired positive outcomes. Howe ver, it is up to the case workers to use their professional judgment to choose which behavioral problems to work on based on responsivity issues; and to choose those actions which the you th and the parents are most ready, willing and able to take to help turn the youths life around. If implemented su ccessfully, the Case Plan will increase the youths protec tive factors and lower the risk factors, and will result in lower overall recidivism.

PAGE 128

128 The risk principle: It is important, of cour se, to target all this e ffort, not at the low-risk, but at higher-risk youth. The PA CT provides the facts (eviden ce) for judges to focus their resources on higher-risk youth a nd assign lower-risk youth to mini mumsupervision caseloads, bringing savings and more effectiveness to the entire juvenile justice system. Re-assessments. Either the pre-screen or the full assessments can be re-administered to measure changes in risk and protec tive factors as interim outcomes for court interventions. This comparative-over-time information can help a judge make good senten cing decisions. For example, a youth who at intake is not attending school, but who returns to school as a result of court-ordered sanctions and then performs well as measured by a PACT re-assessment can be expected, on average, to be asso ciated with reduced recidivism.

PAGE 129

129 APPENDIX B PACT PRE-SCREEN ASSESSMENT DOMAIN 1: Record of Referrals Referrals, rather than offenses, are used to assess the persistence of re-offending by the youth. Include only referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether successfully completed). 1. Age at first offense: The age at the time of the offense for which the youth was referred to juvenile court for the first time on a non-traffic misdemeanor or felony that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. O Over 16 O 16 O 15 O 13 to 14 O 12 and Under Felony and misdemeanor referrals: Items 2 and 3 are mutually exclusive and should add to the total number of referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. 2. Misdemeanor referrals: Total number of referrals for which the most serious offense was a non-traffic misdemeanor that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether successfully completed). O None or one O Two O Three or four O Five or more 3. Felony referrals: Total number of referrals for a felony offense that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether successfully completed). O None O One O Two O Three or more Against-person or weapon referrals: Items 4, 5, and 6 are mutually exclusive and should add to the total number of referrals that involve an against-person or weapon offense, including sex offenses that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether successfully completed). 4. Weapon referrals: Total referrals for which the most serious offense was a firearm/weapon charge or a weapon enhancement finding. O None O One or more 5. Against-person misdemeanor referrals: Total number of referrals for which the most serious offense was an against-person misdemeanor a misdemeanor involving threats, force, or physical harm to another person or sexual misconduct (assault, coercion, harassment, intimidation, etc.). O None O One O Two or more 6. Against-person felony referrals: Number of referrals involving force or physical harm to another person including sexual misconduct as defined by FDLE as violent felonies. O None O One or two O Three or more Sex offense referrals: Items 7 and 8 are mutually exclusive and should add to the total number of referrals that involve a sex offense or sexual misconduct that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court.

PAGE 130

130 7. Sexual misconduct misdemeanor referrals: Number of referrals for which the most serious offense was a sexual misconduct misdemeanor including obscene phone calls, indecent exposure, obscenity, pornography, or public indecency, or misdemeanors with sexual motivation. O None O One O Two or more 8. Felony sex offense referrals: Referrals for a felony sex offense or involving sexual motivation including carnal knowledge, child molestation, communication with minor for immoral purpose, incest, indecent exposure, indecent liberties, promoting pornography, rape, sexual misconduct, or voyeurism. O None O One O Two or more 9. Confinements in secure detention where youth was held for at least 48 hours: Number of times the youth was held for at least 48 hours physically confined in a detention facility. O None O One O Two O Three or more 10. Commitment orders where youth served at least one day confined under residential commitment: Total number of commitment orders and modification orders for which the youth served at least one day confined under residential commitment. A day served includes credit for time served. O None O One O Two or more 11. Escapes: Total number of attempted or actual escapes that resulted in adjudication. O None O One O Two or more 12. Pick Up Orders for failure-to-appear in court or absconding supervision: Total number of failuresto-appear in court or absconding supervision that resulted in a pick up order being issued. Exclude failure-to-appear warrants for non-criminal matters. O None O One O Two or more Domain 2 : Social History Current is defined as behaviors occurring within the last six months 1. Youths Gender O Male O Female 2a. Youth's current school enrollment status, regardless of attendance: If the youth is in home school as a result of being expelled or dropping out, check the expelled or dropped out box, otherwise check enrolled. O Graduated, GED O Enrolled full-time O Enrolled part-time O Suspended O Dropped out O Expelled 2b. Youth's conduct in the most recent term: Fighting or threatening students; threatening teachers/staff; overly disruptive behavior; drug/alcohol use; crimes, e.g., theft, vandalism; lying, cheating, dishonesty. O Recognition for good behavior O No problems with school conduct O Problems reported by teachers O Problem calls to parents O Calls to police 2c. Youth's attendance in the most recent term: Full-day absence means missing majority of classes. Partial-day absence means attending the majority of classes and missing the minority. Habitual truancy as defined in FS includes 15 unexcused absences in a 90-day period. O Good attendance with few absences O No unexcused absences O Some partial-day unexcused absences O Some full-day unexcused absences O Habitual truant

PAGE 131

131 2d. Youth's academic performance in the most recent school term: O Honor student (mostly As) O Above 3.0 (mostly As and Bs) O 2.0 to 3.0 (mostly Bs and Cs, no Fs) O 1.0 to 2.0 (mostly Cs and Ds, some Fs) O Below 1.0 (some Ds and mostly Fs) 3a. History of anti-social friends/companions: Anti-social peers are youths hostile to or disruptive of the legal social order; youths who violate the law and the rights of others and other delinquent youth. (Check all that apply.) Never had consistent friends or companions Had pro-social friends Had anti-social friends Been a gang member/associate 3b. Current friends/companions youth actually spends time with: (Check all that apply.) No consistent friends or companions Pro-social friends Anti-social friends Gang member/associate 4. History of court-ordered or DCF voluntary out-of-home and shelter care placements exceeding 30 days: Exclude DJJ residential commitments. O No out-of-home placements exceeding 30 days O 1 out-of-home placement O 2 out-of-home placements O 3 or more out-of-home placements 5. History of running away or getting kicked out of home: Include times the youth did not voluntarily return within 24 hours, and include incidents not reported by or to law enforcement O No history of running away/being kicked out O 1 instance of running away/kicked out O 2 to 3 instances of running away/kicked out O 4 to 5 instances of running away/kicked out O Over 5 instances of running away/kicked out 6a. History of jail/imprisonment of persons who were ever involved in the household for at least 3 months: (Check all that apply.) No jail/imprisonment history in family Mother/female caretaker Father/male caretaker Older sibling Younger sibling Other member 6b. History of jail/imprisonment of persons who are currently involved with the household: (Check all that apply. ) No jail/imprisonment history in family Mother/female caretaker Father/male caretaker Older sibling Younger sibling Other member

PAGE 132

132 6c. Problem history of parents who are currently involved with the household: (Check all that apply). No problem history of parents in household Parental alcohol problem history Parental drug problem history Parental physical health problem history Parental mental health problem history Parental employment problem history 7. Current parental authority and control: O Youth usually obeys and follows rules O Sometimes obeys or obeys some rules O Consistently disobeys, and/or is hostile 8a. Youths history of alcohol use: (Check all that apply.) No past alcohol use Past alcohol use Alcohol caused family conflict Alcohol disrupted education Alcohol caused health problems Alcohol interfered with keeping pro-social friends Alcohol contributed to criminal behavior Youth needed increasing amounts of alcohol to achieve same level of intoxication or high Youth experienced withdrawal problems 8b. Youths history of drug use: (Check all that apply.) No past drug use Past drug use Drugs caused family conflict Drugs disrupted education Drugs caused health problems Drugs interfered with keeping pro-social friends Drugs contributed to criminal behavior Youth needed increasing amounts of drugs to achieve same level of intoxication or high Youth experienced withdrawal problems

PAGE 133

133 8c. Youths Current alcohol use: (Check all that apply.) No current alcohol use Current alcohol use Alcohol causing family conflict Alcohol disrupting education Alcohol causing health problems Alcohol interfering with keeping pro-social friends Alcohol contributing to criminal behavior Youth needs increasing amounts of alcohol to achieve same level of intoxication or high Youth experiences withdrawal problems 8d. Youths current drug use: (Check all that apply.) No current drug use Current drug use Drugs causing family conflict Drugs disrupting education Drugs causing health problems Drugs interfering with keeping pro-social friends Drugs contributing to criminal behavior Youth needs increasing amounts of drugs to achieve same level of intoxication or high Youth experiences withdrawal problems For abuse and neglect, include any history that is suspected, whether or not reported or substantiated; exclude reports of abuse or neglect proven to be false. 9a. History of violence/physical abuse: Include suspected incidents of abuse if disclosed by youth, whether or not reported or substantiated, but exclude reports investigated but proven to be false. (Check all that apply.) Not a victim of violence/physical abuse Victim of violence/physical abuse at home Victim of violence/physical abuse in a foster/group home Victimized by family member Victimized by someone outside the family Attacked with a weapon 9b History of witnessing violence : ( Check all that apply) Include perpetrators and victims of violence as having witnessed violence. Has not witnessed violence Has witnessed violence at home Has witnessed violence in a foster/group home Has witnessed violence in the community Family member killed as result of violence

PAGE 134

134 9c History of sexual abuse/rape: Include suspected incidents of abuse if disclosed by youth, whether or not reported or substantiated, but exclude reports investigated but proven to be false. (Check all that apply.) Not a victim of sexual abuse/rape Sexually abused/raped by family member Sexually abused/raped by someone outside the family 10. History of being a victim of neglect: Include suspected incidents of neglect, whether or not reported or substantiated, but exclude reports investigated but proven to be false. O Not victim of neglect O Victim of neglect 11. History of mental health problems: Such as schizophrenia, bi-polar, mood, thought, personality, and adjustment disorders. Exclude substance abuse and special education since those issues are considered elsewhere. Confirm by a professional in the social service/healthcare field. O No history of mental health problem(s) O Past history of mental health problem(s) diagnosis (more than six months ago) O Diagnosed with mental health problem(s) O Only mental health medication prescribed. If yes, list _______________________ O Only mental health treatment prescribed O Mental health medication and treatment prescribed Domain 3: Mental Health 1. History of suicidal ideation: Include any previous thoughts, threats, plans and attempts even if youth indicates they were manipulative or there was no intent. (Check all that apply) Has never had serious thoughts about suicide Has had serious thoughts about suicide Has made a plan to commit suicide. If yes, describe _____________________ Has attempted to commit suicide. If yes, describe attempt(s) and date(s)________________________ Feels life is not worth living no hope for future. Knows someone well who has committed suicide. If yes, who, when and how __________________________ Engages in self-mutilating behavior____________________ 2. History of anger or irritability: O No history of anger/irritability O History of occasional feelings of anger/irritability O History of consistent feelings of anger/irritability O History of aggressive reactions to feelings of anger/irritability.

PAGE 135

135 3. History of depression or anxiety O No history of depression/anxiety O History of occasional feelings of depression/anxiety O History of consistent feelings of depression/anxiety O History of impairment in every day tasks due to depression/anxiety 4. History of Somatic Complaints: Bodily or physical discomforts associated with distress, such as stomachaches or headaches O No history of somatic complaints O History of one or two somatic complaints O History of three or four somatic complaints O History of 5 or more somatic complaints 5. History of thought disturbance O No unusual thoughts or beliefs O Presence of hallucinations (auditory or visual) O Presence of beliefs that the youth is controlled by others or others control the youth. 6. History of traumatic experience: Lifetime exposure to events such as rape, abuse or observed violence, including dreams or flashbacks O No presence of traumatic event O Presence of traumatic event O Flashbacks to traumatic event Domain 4: Attitude/Behavior Indicators 1. Attitude toward responsible law abiding behavior: O Abides by conventions/values O Believes conventions/values sometimes apply to him or her O Does not believe conventions/values apply to him or her O Resents or is hostile toward responsible behavior 2. Accepts responsibility for anti-social behavior: O Accepts responsibility for anti-social behavior O Minimizes, denies, justifies, excuses, or blames others O Accepts anti-social behavior as okay O Proud of anti-social behavior 3. Belief in yelling and verbal aggression to resolve a disagreement or conflict: O Believes verbal aggression is rarely appropriate O Believes verbal aggression is sometimes appropriate O Believes verbal aggression is often appropriate 4. Belief in fighting and physical aggression to resolve a disagreement or conflict: O Believes physical aggression is never appropriate O Believes physical aggression is rarely appropriate O Believes physical aggression is sometimes appropriate O Believes physical aggression is often appropriate

PAGE 136

136 5. Reports/evidence of violence not included in criminal history: (Check all that apply.) No reports/evidence of violence Violent outbursts, displays of temper, uncontrolled anger indicating potential for harm Deliberately inflicting physical pain Using/threatening with a weapon Fire starting Violent destruction of property Animal cruelty 6. Reports of problem with sexual aggression not included in criminal history: (Check all that apply.) No reports/evidence of sexual aggression Aggressive sex Sex for power Young sex partners Child sex Voyeurism Exposure

PAGE 137

137 APPENDIX C FULL ASSESSMENT DOMAIN 1: Record of Referrals Referrals, rather than offenses, are used to assess the persistence of re-offending by the youth. Include only referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regardless of whether successfully completed). 2. Age at first offense: The age at the time of the offense for which the youth was referred to juvenile court for the first time on a non-traffic mi sdemeanor or felony that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. O Over 16 O 16 O 15 O 13 to 14 O 12 and Under Felony and misdemeanor referrals: Items 2 and 3 are mutually exclusive and should add to the total number of referrals that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. 5. Misdemeanor referrals: Total number of referrals for which the most serious offense was a nontraffic misdemeanor that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court (regar dless of whether successfully completed). O None or one O Two O Three or four O Five or more 6. Felony referrals: Total number of referrals for a felony offense that resulted in diversion, adjudication withheld, adjudication, deferred prosec ution or referral to adult court (regardless of whether successfu lly completed). O None O One O Two O Three or more Against-person or weapon referrals: Items 4, 5, and 6 are mutually exclusiv e and should add to the total number of referrals that involve an against-person or weapon offense, including sex offenses that resulted in diversion, adjudication withheld, adjudication, deferred prosecuti on or referral to adult court (regardless of whether successfully completed). 7. Weapon referrals: Total referrals for which the most serious offense was a firearm/weapon charge or a weapon enhancement finding. O None O One or more 12. Against-person misdemeanor referrals: Total number of referrals for which the most serious offense was an against-person misdemeanor a mis demeanor involving threats, force, or physical harm to another person or sexual misconduct (ass ault, coercion, harassment, intimidation, etc.). O None O One O Two or more 13. Against-person felony referrals: Number of referrals involving force or physical harm to another person including sexual misconduct as defined by FDLE as violent felonies. O None O One or two O Three or more Sex offense referrals: Items 7 and 8 are mutually exclusive and should add to the total number of referrals that involve a sex offense or sexual misconduct that resulted in diversion, adjudication withheld, adjudication, deferred prosecution or referral to adult court. 14. Sexual misconduct misdemeanor referrals: Number of referrals for which the most serious offense was a sexual misconduct misdemeanor in cluding obscene phone calls, indecent exposure, obscenity, pornography, or public indecency, or misdemeanors with sexual motivation. O None O One O Two or more

PAGE 138

138 15. Felony sex offense referrals: Referrals for a felony sex offens e or involving sexual motivation including carnal knowledge, child molestation, communication with minor for immoral purpose, incest, indecent exposure, indecent liberties, pr omoting pornography, rape, sexual misconduct, or voyeurism. O None O One O Two or more 16. Confinements in secure detention where youth was held for at least 48 hours: Number of times the youth was held for at least 48 hours physically confined in a detention facility. O None O One O Two O Three or more 17. Commitment orders where youth served at least one day confined under residential commitment: Total number of commitment orders and modification orders for which the youth served at least one day confined under residential commitment. A day served includes credit for time served. O None O One O Two or more 18. Escapes: Total number of attempted or actual escapes that resulted in adjudication. O None O One O Two or more 19. Pick Up Orders for failure-to-appear in court or absconding supervision: Total number of failures-to-appear in court or absconding supervision that resulted in a pick up order being issued. Exclude failure-to-appear warrants for non-criminal matters. O None O One O Two or more DOMAIN 3A: School History 1. Youth is a special education student or has a formal diagnosis of a special education need: (Check all that apply.) No special education need Learning Mental retardation Behavioral ADHD/ADD 2. History of expulsions and out of school suspensions since the first grade: O No expel/suspend O 1 expel/suspend O 2 or 3 O 4 or 5 O 6 or 7 O More than 7 3. Age at first expulsion or suspension: O No expulsions O 5 to 9 years old O 10 to 13 years old O 14 to 15 years old O 16 to 18 years old 4. Youth has been enrolled in a community school during the last 6 months, regardless of attendance: O No, graduated/GED and not attending school, do not complete Domain 3B O No, dropped-out or expelled for more than six months, do not complete Domain 3B O Yes, must complete Domain 3B DOMAIN 3B: Current School Status O For Initial Assessments, current is the most recent term in last 6 months. O For Re-assessments and Final Assessments, current is the last 4 weeks in the most recent term. 1. Youths current school enrollment status, regardless of attendance: If the youth is in home school as a result of being expelled or dropping out, check expelled or dropped out; otherwise check enrolled, if in home school. O Graduated/GED O Enrolled full-time O Enrolled part-time O Suspended O Dropped out O Expelled 2. Type of school in which youth is enrolled: Name of School O Public academic O Vocational O Alternative O GED program O Private academic O Home school O College O Other____

PAGE 139

139 3. Youth believes there is value in getting an education: O Believes getting an education is of value O Somewhat believes education is of value O Does not believe education is of value 4. Youth believes school provides an encouraging environment for him or her: O Believes school is encouraging O Somewhat believes school is encouraging O Does not believe school is encouraging O Not close to any teachers, staff, or coaches 5. Teachers, staff, or coaches the youth likes or feels comfortable talking with: O Close to 1 O Close to 2 O Close to 3 O Close to 4 or more 6. Youths involvement in school activities during most recent term: School leadership; social service clubs; music, dance, drama, art; athletics; other extracurricular activities. O Involved in 2 or more activities O Involved in 1 activity O Interested but not involved in any activities O Not interested in school activities 7. Youths conduct in the most recent term: Fighting or threatening students; threatening teachers/staff; overly disruptive behavior; drug/alcohol use; crimes (e.g., theft, vandalism); lying, cheating, dishonesty. O Recognition for good behavior O No problems with school conduct O Problems reported by teachers O Problem calls to parents O Calls to police 8. Number of expulsions and suspensions in the most recent term: O No expel/suspend O 1 expel/suspend O 2 or 3 O More than 3 9. Youths attendance in the most recent term: Partial-day absence means attending majority of classes and missing minority. Full-day absence means missing majority of classes. Habitual truancy as defined in FS includes 15 unexcused absences in a 90 day period. O Good attendance; few excused absences O No unexcused absences O Some partial-day unexcused absences O Some full-day unexcused absences O Habitual truant 10. Youths academic performance in the most recent school term: O Honor student (mostly As) O Above 3.0 (mostly As and Bs) O 2.0 to 3.0 (mostly Bs and Cs, no Fs) O 1.0 to 2.0 (mostly Cs and Ds, some Fs) O Below 1.0 (some Ds and mostly Fs) 11. Interviewers assessment of likelihood the youth will stay in and graduate from high school or an equivalent vocational school: O Very likely to stay in school and graduate O Uncertain if youth will stay and graduate O Not very likely to stay and graduate

PAGE 140

140 DOMAIN 4A: Historic Use of Free Time 1. History of structured recreational activities within the past 5 years: Youth has participated in structured and supervised pro-social community activities, such as religious group/church, community group, cultural group, club, athletics, or other community activities. O Involved in 2 or more structured activities O Involved in 1 structured activity O Never involved in structured activities 2. History of unstructured pro-social recreational activities within the past 5 years: Youth has engaged in activities that positively occupy the youth's time, such as reading, hobbies, etc. O Involved in 2 or more pro-social unstructured activities O Involved in 1 pro-social unstructured activity O Never involved in pro-social unstructured activities DOMAIN 4B: Current Use of Free Time O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. 1. Current interest and involvement in structured recreational activities: Youth participates in structured and supervised pro-social community activities, such as religious group/church, community group, cultural group, club, athletics, or other community activity. O Currently involved in 2 or more structured activities O Currently involved in 1 structured activity O Currently interested but not involved O Currently not interested in any structured activities 2. Types of structured recreational activities in which youth currently participates: (Check all that apply. ) None Community/cultural group Hobby group or club Athletics Religious group/church Volunteer organization 3. Current interest and involvement in pro-social unstructured recreational activities: Youth engages in activities that positively occupy his or her time, such as reading, hobbies, etc. O Currently involved in 2 or more pro-social unstructured activities O Currently involved in 1 pro-social unstructured activity O Currently interested but not involved O Not interested in any pro-social unstructured activities DOMAIN 5A: Employment History 1. History of employment: O Too young for employment consideration O Never been employed O Has been employed 2. History of successful employment: O Never successfully employed O Has been successfully employed 3. History of problems while employed: O Never fired or quit because of problems O Fired or quit because of poor performance O Fired or quit because he or she could not get along with employer or coworkers 4. History of positive personal relationship(s) with past employer(s) or adult coworker(s): O Never had any positive relationships O Had 1 positive relationship O Had 2 or more positive relationships

PAGE 141

141 DOMAIN 5B: Current Employment O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. 1. Understanding of what is required to maintain a job: O Lacks knowledge of what it takes to maintain a job O Has knowledge of abilities to maintain a job O Has demonstrated ability to maintain a job 2. Current interest in employment: O Currently employed O Not employed but highly interested in employment O Not employed but somewhat interested O Not employed and not interested in employment O Too young for employment consideration 3. Current employment status: O Not currently employed O Employment is currently going well O Having problems with current employment 4. Current positive personal relationship(s) with employer(s) or adult coworker(s): O Not currently employed O Employed but no positive relationships O At least 1 positive relationship DOMAIN 6A: History of Relationships 1. History of positive adult non-family relationships not connected to school or employment: Adults, who are not teachers and not part of the youths family, who can provide support and model pro-social behavior, such as religious leader, club member, community person, etc. O No positive adult relationships O 1 positive adult relationship O 2 positive adult relationships O 3 or more positive adults relationships 2. History of anti-social friends/companions: Anti-social peers are youths hostile to or disruptive of the legal social order; youths who violate the law and the rights of others and other delinquent youth. (Check all that apply.) Never had consistent friends or companions Had pro-social friends Had anti-social friends Been a gang member/associate DOMAIN 6B: Current Relationships O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. 1. Current positive adult non-family relationships not connected to school or employment: Adults, who are not teachers and not part of the youths family, who can provide support and model pro-social behavior, such as religious leader, club member, community person, etc. O No positive adult relationships O 1 positive adult relationship O 2 positive adult relationships O 3 or more positive adults relationships 2. Current pro-social community ties: Youth feels there are people in his or her community who discourage him or her from getting into trouble or are willing to help the youth. O No pro-social community ties O Some pro-social community ties O Has strong pro-social community ties 3. Current friends/companions youth actually spends time with: ( Check all that apply.) No consistent friends or companions Pro-social friends Anti-social friends Gang member/associate

PAGE 142

142 4. Currently in a romantic, intimate, or sexual relationship: O Not romantically involved with anyone O Romantically involved with a pro-social person O Romantically involved with an anti-social person/criminal 5. Currently admires/emulates anti-social peers: O Does not admire, emulate anti-social peers O Somewhat admires, emulates anti-social peers O Admires, emulates anti-social peers 6. Current resistance to anti-social peer influence: O Does not associate with anti-social peers O Usually resists going along with anti-social peers O Rarely resists goes along with anti-social peers O Leads anti-social peers DOMAIN 7A: Family History 1. History of court-ordered or DCF voluntary out-of-home and shelter care placements exceeding 30 days: Exclude DJJ residential commitments. O No out-of-home placements exceeding 30 days O 1 out-of-home placement O 2 out-of-home placements O 3 or more out-of-home placements 2. History of running away or getting kicked out of home: Include times the youth did not voluntarily return within 24 hours, and include incidents not reported by or to law enforcement. O No history of running away or being kicked out O 1 instance of running away/kicked out O 2 to 3 instances of running away/kicked out O 4 to 5 instances of running away/kicked out O Over 5 instances of running away/kicked out 3. History of petitions filed: Include all petitions regardless of whether the petition was granted. (Check all that apply.) No petitions filed CINS/FINS Dependency 4. History of jail/imprisonment of persons who were ever involved in the household for at least 3 months: (Check all that apply.) No jail/imprisonment history in family Mother/female caretaker Father/male caretaker Older sibling Younger sibling Other member 5. Youth living under any adult supervision. Adult supervision must be someone who is responsible for the youths welfare, either legally or with parental consent. O No, living with peers without adult supervision, do not complete Domain 7B O No, living alone without adult supervision, do not complete Domain 7B O No, transient without adult supervision, do not complete Domain 7B O Yes, living under adult supervision, must complete Domain 7B DOMAIN 7B: Current Living Arrangements O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.

PAGE 143

143 1. All persons with whom youth is currently living: (Check all that apply.) Living alone Transient (street) Biological mother Biological father Non-biological mother Non-biological father Older sibling(s) Younger sibling(s) Grandparent(s) Other relative(s) Long-term parental partner(s) Short-term parental partner(s) Youths romantic partner Youths child Foster/group home Youths friends 2. Annual combined income of youth and family: O Under $15,000 O $15,000 to $34,999 O $35,000 to $49,999 O $50,000 and over 3. Jail/imprisonment history of persons who are currently involved with the household: (Check all that apply.) No jail/imprisonment history in family Mother/female caretaker Father/male caretaker Older sibling Younger sibling Other member 4. Problem history of parents who are currently involved with the household: (Check all that apply.) No problem history of parents in household Parental alcohol problem history Parental drug problem history Parental physical health problem history Parental mental health problem history Parental employment problem history 5. Problem history of siblings who are currently involved with the household: (Check all that apply.) No siblings currently in household No problem history of siblings in household Sibling alcohol problem history Sibling drug problem history Sibling physical health problem history Sibling mental health problem history Sibling employment problem history 6. Support network for family: Extended family and/or family friends who can provide additional support to the family. O No support network O Some support network O Strong support network 7. Family willingness to help support youth: O Consistently willing to support youth O Inconsistently willing to support youth O Little or no willingness to support youth O Hostile, berating, and/or belittling of youth 8. Family provides opportunities for youth to participate in family activities and decisions affecting the youth: O No opportunities for involvement providedO Some opportunities for involvement provided O Opportunities for involvement provided 9. Youth has run away or been kicked out of home : Include times youth did not voluntarily return within 24 hours, and include incidents not reported by or to law enforcement. O Has not run away/kicked out of home O Has run away/kicked out of home O Is currently kicked out of home or is a runaway

PAGE 144

144 10. Family member(s) youth feels close to or has good relationship with: (Check all that apply.) Does not feel close to any family member Feels close to mother/female caretaker Feels close to father/male caretaker Feels close to male sibling Feels close to female sibling Feels close to extended family 11. Level of conflict between parents, between youth and parents, among siblings: O Some conflict that is well managed O Verbal intimidation, yelling, heated arguments O Threats of physical abuse O Domestic violence: physical/sexual abuse 12. Parental supervision: Parents know whom youth is with, when youth will return, where youth is going, and what youth is doing. O Consistent good supervision O Sporadic supervision O Inadequate supervision 13. Parental authority and control: O Youth usually obeys and follows rules O Youth sometimes obeys or obeys some rules O Youth consistently disobeys and/or is hostile 14. Consistent appropriate punishment for bad behavior: Appropriate means clear communication, timely response, and response proportionate to conduct. O Consistently appropriate punishment O Consistently overly severe punishment O Consistently insufficient punishment O Inconsistent or erratic punishment 15. Consistent appropriate rewards for good behavior: Appropriate means clear communication, timely response, and response proportionate to conduct; rewards mean affection, praise, etc. O Consistently appropriate rewards O Consistently overly indulgent/overly protective O Consistently insufficient rewards O Inconsistent or erratic rewards 16. Parental characterization of youth's anti-social behavior: O Disapproves of youth's anti-social behavior O Minimizes, denies, justifies, excuses behavior, or blames others/circumstances O Accepts youth's anti-social behavior as okay O Proud of youth's anti-social behavior DOMAIN 8A: Alcohol and Drug History Disrupted functioning involves having a problem in any of these five life areas: education, family conflict, peer relationships, crime, or health, and usually indicates treat ment is warranted. Use that contributes to criminal behavior typically precipitates the commission of a crime; there is evidence or reason to believe the youths 1. History of alcohol use: (Check all that apply.) No past alcohol use Past alcohol use Alcohol caused family conflict Alcohol disrupted education Alcohol caused health problems Alcohol interfered with keeping pro-social friends Alcohol contributed to criminal behavior Youth needed increasing amounts of alcohol to achieve same level of intoxication or high Youth experienced withdrawal problems

PAGE 145

145 2. Youths history of drug use: (Check all that apply.) No past drug use Past drug use Drugs caused family conflict Drugs disrupted education Drugs caused health problems Drugs interfered with keeping pro-social friends Drugs contributed to criminal behavior Youth needed increasing amounts of drugs to achieve same level of intoxication or high Youth experienced withdrawal problems 3. Youths history of referrals for alcohol/drug assessment: O Never referred for drug/alcohol assessment O Diagnosed as no problem O Referred but never assessed O Diagnosed as abuse O Diagnosed as dependent/addicted 4. History of attending alcohol/drug education classes for an alcohol/drug problem: O Never attended drug/alcohol education classes O Voluntarily attended drug/alcohol education classes O Attended classes by parent, school, or other agency request O Attended classes at court direction 5. History of participating in alcohol/drug treatment program : O Never participated in treatment program O Participated once in treatment program O Participated several times in treatment programs 6. Youth currently using alcohol or drugs:O No, do not complete Domain 8B O Yes, must complete domain 8B DOMAIN 8B: Current Alcohol and D r u g s O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. 1. Youths current alcohol use: (Check all that apply.) No current alcohol use Current alcohol use Alcohol causing family conflict Alcohol disrupting education Alcohol causing health problems Alcohol interfering with keeping pro-social friends Alcohol contributing to criminal behavior Youth needs increasing amounts of alcohol to achieve same level of intoxication or high Youth experiences withdrawal problems 2. Youths current drug use: (Check all that apply.) No current drug use Current drug use Drugs causing family conflict Drugs disrupting education Drugs causing health problems Drugs interfering with keeping pro-social friends Drugs contributing to criminal behavior Youth needs increasing amounts of drugs to achieve same level of intoxication or high Youth experiences withdrawal problems

PAGE 146

146 DOMAIN 9A: Mental Health History History of suicidal ideation: Include any previous thoughts, threats, plans and attempts even if youth indicates they were manipulative or there was no intent. (Check all that apply.) Has never had serious thoughts about suicide Has had serious thoughts about suicide Has made a plan to commit suicide. If yes, describe ____________________________ Has attempted to commit suicide. If yes, describe attempts and dates _________________________________ Feels life is not worth livingno hope for future Knows someone well who has committed suicide, If yes, who, when and how _____________________________ Engages in self-mutilating behavior For abuse and neglect, include suspected incidents of abuse, including those disclosed by youth, whether or not reported or substantiated, but exclude reports of abuse/neglect investigated but proven to be false. History of violence/physical abuse: Include suspected incidents of abuse, whether or not substantiated, but exclude reports proven to be false. (Check all that apply.) Not a victim of violence/physical abuse Victim of violence/physical abuse at home Victim of violence/physical abuse in a foster/group home Victimized or physically abused by family member Victimized or physically abused by someone outside the family Attacked with a weapon History of witnessing violence: (Check all that apply.) Include perpetrators and victims of violence as having witnessed violence. Has not witnessed violence Has witnessed violence at home Has witnessed violence in a foster/group home Has witnessed violence in the community Family member killed as a result of violence History of sexual abuse/rape: Include suspected incidents of abuse if disclosed by youth, whether or not reported or substantiated, but exclude reports investigated but proven to be false. (Check all that apply.) Not a victim of sexual abuse/rape Sexually abused/raped by family member Sexually abused/raped by someone outside the family 3. Type of drugs currently used: (Check all that apply.) No current drug use Marijuana/hashish Inhalants (glue/gasoline) Cocaine (crack/rock) Cocaine (coke) Amphetamines (Meth/uppers/speed/ecstacy) Barbiturates (Tuinal/Seconal/downers) Tranquilizers/sedatives (Valium/Libnum/Dalmane/ Ketamine) Hallucinogens (LSD/acid/mushrooms/GHB) Phencyclidine (PCP/angel dust) Heroin Other opiates (Dilaudid/Demerol/Percodan/Codeine/ Oxycontin) 4. Current alcohol/drug treatment program participation: O Alcohol/drug treatment not warranted O Not currently attending needed alcohol/drug treatment program O Currently attending alcohol/drug treatment program O Successfully completed alcohol/drug treatment program

PAGE 147

147 History of being a victim of neglect: O Not a victim of neglect O Victim of neglect History of ADD/ADHD: Confirmed by a professional in the social service/healthcare field. O No history of ADD/ADHD O Diagnosed with ADD/ADHD O Only ADD/ADHD medication prescribed O Only ADD/ADHD treatment prescribed O ADD/ADHD medication and treatment prescribed History of mental health problems: Such as schizophrenia, bi-polar, mood, thought, personality, and adjustment disorders. Exclude conduct disorder, oppositional defiant disorder, substance abuse, and ADD/ADHD. Confirmed by a professional in the social service/healthcare field. O No history of mental health problem(s) O Past history of mental health problem(s) diagnosis (more than six months ago) O Diagnosed with mental health problem(s) O Only mental health medication prescribed. If yes, list ___________________ O Only mental health treatment prescribed O Mental health medication and treatment prescribed History of anger or irritability O No history of anger/irritability O History of occasional feelings of anger/irritability O History of consistent feelings of anger/irritability O History of aggressive reactions to feelings of anger/irritability 1. History of depression or anxiety O No history of depression/anxiety O History of occasional feelings of depression/anxiety O History of consistent feelings of depression/anxiety O History of impairment in everyday tasks due to depression/anxiety 2. History of somatic complaints: Bodily or physical discomforts associated with distress, such as stomachaches or headaches. O No history of somatic complaints O History of one or two somatic complaints O History of three or four somatic complaints O History of 5 or more somatic complaints 3. History of thought disturbance O No unusual thoughts or beliefs O Presence of hallucinations (auditory or visual) O Presence of beliefs that the youth is controlled by others 4. History of traumatic experience: Lifetime exposure to events such as rape, abuse or observed violence, including dreams or flashbacks. O No presence of traumatic event O History of traumatic event O History of flashbacks to traumatic event 5. Currently has health insurance: O No health insurance O Public insurance (Medicaid, KIDCARE) O Private insurance

PAGE 148

148 6. Current mental health problem status: O No current mental health problem(s), do not complete Domain 9B O Current mental health problem(s), must complete Domain 9B DOMAIN 9B: Current Mental Health O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. 1. Current suicidal ideation: O Does not have serious thoughts about suicide O Has serious thoughts about suicide O Has recently made a plan to commit suicide. If yes, describe _______________________ O Has recently attempted to commit suicide. If yes, describe attempts and dates ______________________ O Feels life is not worth livingno hope for future O Knows someone well who has committed suicide. If yes, who, when and how _______________________ O Engages in self-mutilating behavior 2. Currently diagnosed with ADD/ADHD: Confirmed by a professional in the social service/healthcare field. Type of medication: _________________________O No ADD/ADHD diagnosis O No ADD/ADHD medication currently prescribed O Currently taking ADD/ADHD medication O ADD/ADHD medication currently prescribed, but not taking 3. Mental health treatment currently prescribed excluding ADD/ADHD treatment: O No current mental health problem O No mental health treatment currently prescribed O Attending mental health treatment O Treatment currently prescribed, but not attending 4. Mental health medication currently prescribed excluding ADD/ADHD medication: Type of medication: _________________________O No current mental health problem O No mental health medication currently prescribed O Currently taking mental health medication O Mental health medication currently prescribed, but not taking 5. Mental health problems currently interfere in working with the youth: O No current mental health problem O Mental health problem(s) do not interfere in work with youth O Mental health problem(s) interfere in work with youth DOMAIN 10: Attitudes/Behaviors O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks.

PAGE 149

149 1. Primary emotion when committing crime(s): O Nervous, afraid, worried, ambivalent, uncertain, or indecisive O Hyper, excited, or stimulated O Unconcerned or indifferent O Confident or brags about not getting caught 2. Primary purpose for committing crime(s) within the last 6 months: O Anger / Revenge O Impulse O Sexual desire O Money or material gain, including drugs O Excitement, amusement, or fun O Peer status, acceptance, or attention 3. Optimism: Youth talks about future in positive way with plans or aspirations of a better life that could include employment, education, raising a family, travel, or other prosocial life goals. O High aspirations: sense of purpose, commitment to better life O Normal aspirations: some sense of purpose O Low aspirations: little sense of purpose or plans for better life O Believes nothing matters; he or she will be dead before long 4. Impulsive; acts before thinking: O Uses self-control; usua lly thinks before acting O Some self-control; some times thinks before acting O Impulsive; often acts before thinking O Highly Impulsive; usually acts before thinking 5. Belief in control over anti-social behavior: O Believes he or she can avoid/stop anti-social behavior O Somewhat believes anti-soci al behavior is controllable O Believes his or her anti-social behavior is out of his or her control 6. Empathy, remorse, sympathy, or feelings for the victim(s) of criminal behavior: O Has empathy for his or her victim(s) O Has some empathy for his or her victim(s) O Does not have empathy for his or her victim(s) 7. Respect for property of others: O Respects property of others O Respects personal property but not publicly accessible property: Its not hurting anybody. O Conditional respect for personal property: If they are stupid enough to leave it out, they deserve losing it. O No respect for property: If I want something, it should be mine. 8. Respect for authority figures: O Respects most authority figures O Does not respect authority figures, and may resent some O Resents most authority figures O Defies or is hostile toward most authority figures 9. Attitude toward responsible law abiding behavior: O Believes pro-social rules/conventions apply to him or her O Believes some pro-social rules/conventions sometimes apply to him or her O Does not believe pro-social rules/conventions apply to him or her O Resents or is defiant toward pro-social rules/conventions 10. Accepts responsibility for antisocial behavior: O Accepts responsibility fo r anti-social behavior O Minimizes, denies, justifies, excuses, or blames others O Accepts anti-social behavior as okay O Proud of anti-social behavior 11. Youths belief in successfully meeting conditions of court supervision: O Believes he or she will be successful O Unsure if he or sh e will be successful O Does not believe he or she will be successful

PAGE 150

150 DOMAIN 11: Aggression Items 1. through 4.: O For Initial Assessments, rate items based on behavior during the last 6 months. O For Re-assessments and Final Assessments, rate items based on behavior during the last 4 weeks. 1. Tolerance for frustration: O Rarely gets upset over small things or has temper tantrums O Sometimes gets upset over small things or has temper tantrums O Often gets upset over small things or has temper tantrums 2. Hostile interpretation of actions and intentions of others in a common nonconfrontational setting: O Primarily positive view of intentions of others O Primarily negative view of intentions of others O Primarily hostile view of intentions of others 3. Belief in yelling and verbal aggression to resolve a disagreement or conflict: O Believes verbal aggression is rarely appropriate O Believes verbal aggression is sometimes appropriate O Believes verbal aggression is often appropriate 4. Belief in fighting and physical aggression to resolve a disagreement or conflict: O Believes physical aggression is never appropriate O Believes physical aggression is rarely appropriate O Believes physical aggression is sometimes appropriate O Believes physical aggression is often appropriate Items 5. and 6.: O For Initial Assessments, include the entire history of report. O For Re-assessments and Final Assessments, include reports within the last 4 weeks. 5. Reports/evidence of violence not included in criminal history: (Check all that apply.) No reports/evidence of violence Violent outbursts, displays of temper, uncontrolled anger indicating potential for harm Deliberately inflicting physical pain Using/threatening with a weapon Fire starting Violent destruction of property Animal cruelty 6. Reports of problem with sexual aggression not included in criminal history: (Check all that apply.) No reports/evidence of sexual aggression Aggressive sex Sex for power Young sex partners Child sex Voyeurism Exposure DOMAIN 12: Skills O For Initial Assessments, current means behaviors during the last 6 months. O For Re-assessments and Final Assessments, current means behaviors during the last 4 weeks. O Use a general pattern of current behaviors and not a single incident.

PAGE 151

151 1. Consequential thinking: O Does not understand there are consequences to actions O Understands there are consequences to actions O Identifies consequences of actions O Acts to obtain desired consequencesgood consequential thinking 2. Goal setting: O Does not set goals O Sets unrealistic goals O Sets somewhat realistic goals O Sets realistic goals 3. Problem-solving: O Cannot identify problem behaviors O Identifies problem behaviors O Thinks of solutions for problem behaviors O Applies appropriate solutions to problem behaviors 4. Situational perception: Ability to analyze the situation, choose the best pro-social skill, and select the best time and place to use the prosocial skill. O Cannot analyze the situation for use of a pro-social skill O Can analyze but not choose the best pro-social skill O Can choose the best skill but cannot select the best time and place O Can select the best time and place to use the best pro-social skill 5. Dealing with others: Basic social skills include listening, starting a conversation, having a conversation, asking a question, saying thank you, introducing yourself, introducing other people, and giving a compliment Advanced social skills include asking for help, joining in, giving instructions, following instructions, apologizing, and convincing others. O Lacks basic social skills in dealing with others O Has basic social skills, lacks advanced skills in dealing with others O Sometimes uses advanced social skills in dealing with others O Often uses advanced social skills in dealing with others 6. Dealing with difficult situations: Incl. making a complaint, answering a complaint, dealing with embarrassment, dealing with being left out, standing up for a friend, responding to frustration, responding to failure, dealing with contradictory messages, dealing with accusation, getting ready for a difficult conversation, and dealing with group pressure. O Lacks skills in dealing with difficult situations O Rarely uses skills in dealing with difficult situations O Sometimes uses skills in dealing with difficult situations O Often uses skills in dealing with difficult situations 7. Dealing with feelings/emotions: Includes knowing his or her feelings, expressing feelings, understanding the feelings of others, dealing with someone elses anger, expressing affection, dealing with fear, and rewarding oneself. O Lacks skills in dealing with feelings/emotions O Rarely uses skills in dealing with feelings/emotions O Sometimes uses skills in dealing with feelings/emotions O Often uses skills in dealing with feelings/emotions 8. Monitoring of internal triggers, distorted thoughts, that can lead to trouble: O Cannot identify internal triggers O Identifies internal triggers O Actively monitors/controls internal triggers 9. Monitoring of external triggers, events or situations, that can lead to trouble: O Cannot identify external triggers O Identifies external triggers O Actively monitors/controls external triggers 10. Control of impulsive behaviors that get youth into trouble: Reframing, replacing antisocial thoughts with pro-social thoughts, diversion, relaxation, problem solving, negotiation, relapse prevention. O Never had a problem with impulsive behavior O Does not know techniques to control impulsive behavior O Knows techniques to control impulsive behavior O Uses techniques to control impulsive behavior

PAGE 152

152 11. Control of aggression: Includes asking permission, sharing thoughts, helping others, negotiating, using self control, standing up for ones rights, responding to teasing, avoiding trouble with others, and keeping out of fights O Never had a problem with aggression O Lacks alternatives to aggression O Rarely uses alternatives to aggression O Sometimes uses alternatives to aggression O Often uses alternatives to aggression

PAGE 153

153 REFERENCES Akers, Ronald L. (1998). Social Learning and Social Structure: A General Theory of Crime and Deviance. Boston: Northeastern University Press. Allen, Francis A. (1981). The Decline of the Rehabilitati ve Ideal: Penal Policy and Social Purpose. New Haven, CN: Yale University Press. Andrews, Don A. (2001). Principl es of effective correctional pr ograms. In Laurence L. Motiuk and Ralph C. Serin (eds.), Compendium 2000 on Effective Correctional Programming. Ottawa, Ont., Canada: Correctional Service Canada. Andrews, Don A. (2006). Enhancing adherenc e to risk-need-responsiv ity: Making quality a matter of policy. Criminology and Public Policy, 5:595-602. Andrews, Don A., & Bonta, James. (2003). The Psychology of Criminal Conduct 3rd edition. Cincinnati, OH: Anderson Publishing Co. Andrews, Don A., Bonta, James, & Hoge, Robe rt D. (1990a). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17:19-52. Andrews, Don A, Bonta, James, & Wormith, J. St ephen. (2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52:7-27. Andrews, Don A., Dowden, Cr aig, & Gendreau, Paul. (1999). Clinically Relevant and Psychologically Informed Approaches to Reduc ed Re-offending: A Meta-analytic Study of Human Service, Risk, Need, Responsivit y and Other Concerns in Justice Contexts. Carleton University, Ottawa, Canada. Andrews, Don A., & Kiessling, Jerry J. (1980). Program structure and effective correctional practices: A summary of the CaVIC resear ch. In R.R. Ross & P. Gendreau (eds.), Effective Correctional Treatment. Toronto, Canada: Butterworth. Andrews, Don A., Zinger, Ivan, Hoge, Robert D., Bonta, James, Gendreau, Paul, and Cullen, Francis T. (1990b). Does correctional treat ment work? A clinically-relevant and psychologically informed meta-analysis. Criminology, 28:369-404. Andrews, Don A., Zinger, Ivan, Bonta, James, Gendreau, Paul, and Cullen, Francis T. (1990c). Does correctional treatment work? A ps ychologically informed meta-analysis. Criminology, 28:369-404. Antonowicz, D.H., & Ross, R.R. (1994). Essent ial components of successful rehabilitation programs for offenders. International Journal of Offende r and Comparative Criminology, 38:97-104. Aos, Steve. (1998). Watching the Bottom Line: Cost-Effective Interventions for Reducing Crime in Washington. Olympia, WA: Washington State In stitute for Public Policy.

PAGE 154

154 Auerhahn, Kathleen. (1999). Selective incapac itation and the problem of prediction. Criminology, 37(4):703-735. Auerhahn, Kathleen. (2006). Conceptual and me thodological issues in the prediction of dangerous behavior. Criminology & Public Policy, 5(4):771-778. Baird, S.C., Heinz, R.C., & Bemus, B.J. (1979). Project Report #14: A Two-year Follow-up. Madison, WI: Department of Health and Soci al Services, Case Cl assification/ Staff Deployment Project, Bureau of Community Corrections. Barnoski, Robert. (2004). Outcome Evaluation of Washington States Research-Based Programs for Juvenile Offenders. Olympia, WA: Washington State Institute for Public Policy. Belkin, Jacob, Blumstein, Alfred, and Glass, Willi am. (1972). Recidivism as a feedback process: An analytical model and empirical validation. Journal of Criminal Justice, 1:7. Belknap, J. (1996). The Invisible Woman: Gender, Crime, and Justice. Cincinnati: Wadsworth. Benekos, Peter J. & Merlo, Alida V. (1995). Three strikes and youre out! The political sentencing game. Federal Probation, 59:3-9. Blackmore, John & Welsh, Jane. (1983). Selective incapacitation: Sentencing according to risk. Crime & Delinquency, 29:504-529. Blumstein, Alfred, Cohen, Jacqueline and Nagin, Daniel, (eds.), (1978). Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. Washington, D.C.: National Academy of Sciences. Blumstein, Alfred, Cohen, Jacqueline, Rot h, Jeffrey A., and Visher, Christy. (1986). Criminal Careers and Career Criminals, Vol.1. Washington, D.C.: National Academy Press. Blumstein, Alfred, Cohen, Jacqueline, Das, Somnath, and Moitra, Soumyo D. (1988). Specialization and seriousness dur ing adult criminal careers. Journal of Quantitative Criminology, 4:303-345. Broidy, Lisa, & Agnew, Robert. (1997). Gender and crime: A General Strain Theory perspective. Journal of Research in Crime and Delinquency, 34:275-306. Bonta, James. (1995). The responsivity principle and offender rehabilitation. Forum on Corrections Research, 7:34-37. Bonta, James, Law, Moira, & Hanson, Karl. ( 1998). The prediction of criminal and violent recidivism among mentally disorder ed offenders: A meta-analysis. Psychological Bulletin, 123:123-142.

PAGE 155

155 Bonta, James, Wallace-Capretta, Suzanne, & Rooney, Jennifer. (1998). Restorative Justice: An Evaluation of the Restora tive Resolutions Project (User Report 1998-05). Ottawa, Ontario: Solicitor General Canada. Bonta, James, Wallace-Capretta, Suzanne, & Rooney, Jennifer. (2000). A quasi-experimental evaluation of an intensive reha bilitation superv ision program. Criminal Justice and Behavior, 27:312-329. Brown, Shelley L. (2003). The dynamic predicti on of criminal recidi vism: A three-wave prospective study 1995-2002. Ph.D. dissertation, Queens University, Kingston, Ontario, Canada. Brown, Brenner, & Campbell, Robin. (2005). Smoothing the Path from Prison to Home: A Roundtable Discussion on the Lesso ns of Project Greenlight. New York: Vera Institute of Justice. Available at http://www.vera.org. Bureau of Justice Statistics. (1992). Nationa l Corrections Reporting Program 1988. Washington, D.C.: U.S. Department of Justice. Burrell, William D. (2000). Reinventing proba tion: Organizational culture and change. Community Corrections Report, 7:49-64. Canela-Cacho, Jose A., Blumstein, Alfred, and C ohen, Jacqueline. (1997). Relationship between the offending frequency (A) of imprisoned and free offenders. Criminology, 35:133-175. Carey, Mark. (2002). Social lear ning, social capital and correc tional theories: Seeking an integrated model. Paper presented at Inte rnational Community Co rrections Association conference, November, 2002. Chaiken, Jan M., & Chaiken, Marcia R. (1982). Varieties of Criminal Behavior. Santa Monica, CA: Rand. Chaiken, Marcia R., & Chaiken, Jan M. (1984). Offender types and public policy. Crime& Delinquency, 30:195-226. Chesney-Lind, M. (1989). Girls crime and a woman s place: Toward a feminist model of female delinquency. Crime & Delinquency, 35:5-29. Chesney-Lind, M. (2001). What about the gi rls? Delinquency programming as if gender mattered. Corrections Today, 63:38. Chesney-Lind, M., & Rodriguez, N. (1983). Women under lock and key. Prison Journal, 63:4765. Chesney-Lind, M., & Sheldon, R.G. (1992). Girls, Delinquency and Juvenile Justice. Pacific Grove, CA: Brooks Cole.

PAGE 156

156 Clark, John, Austin, James, and Henry, D. Alan (1997). Three Strikes and Youre Out: A Review of State Legislation. Washington, D.C.: National Institute of Justice. Available at http://www.ncjrs.org. Clear, Todd R., and Barry, Donald M. (1983). So me conceptual issues in incapacitating offenders. Crime and Delinquency, 29:529-545. Cohen, Jacqueline. (1983). Incapacitation as a strategy for crime cont rol: Possibilities and pitfalls. In Michael Tonry & Norval Morris (eds.), Crime and Justice: An Annual Review of Research, vol.2. Chicago: University of Chicago Press. Cohen, Jacqueline. (1984). Incapacitation. University of Illinois Law Review, 2:253-290. Cohen, Stanley. (1985). Visions of Social Control: Crim e, Punishment and Classification. Oxford: Polity Press. Copas, John. (1983). Some statistical questions in the prediction of dangerous offending. In Hinton, John, W. (ed.), Dangerousness: Problems of Assessment and Prediction. London: Allen & Unwin. Corbett, Ronald P., Beto, Dan R., Coen, Brian, DiIu lio, John J., Faulkner, J. Richard, Fitzgerald, Bernard L., Gregg, Irwin, Helber, Norman, Hinzman, Gerald R., Malvestuto, Robert, Paparozzi, Mario, Perry, John, Pozzi, Ro cco, & Rhine, Edward E. (1999). Broken Windows Probation: The Next Step in Fighting Crime. Center for Civic Innovation, Manhattan Institute, New York. Covington, S.S. (2001). Creating gender-responsive programs: Th e next step for womens services. Corrections Today, 61:85-87. Covington, S.S. (2003). A womans j ourney home: Challenges for female offenders. In J. Travis and M. Waul (Eds.), Prisoners Once Removed: The Impact of Incarceration and Reentry on Children, Families, and Communities. (pp.67-103). Washington, D.C.: The Urban Institute. Cullen, Francis T. (2002). Rehabi litation and treatment programs. In James Q. Wilson and Joan Petersilia (eds.), Crime and Public Policy, 2nd Edition. San Francisco, CA: ICS Press. Cullen, Francis T., & Gendreau, Paul. (1989). The e ffectiveness of correc tional rehabilitation: Reconsidering the nothing works debate. In L. Goodstein and D. MacKenzie (eds.), American Prisons: Issues in Research and Policy. New York: Plenum. Cullen, Francis T., & Gilbert, Karen E. (1982). Reaffirming Rehabilitation. Cincinnati, OH: Anderson. Currie, Elliot. (1998). Crime and Punishment in America. New York, NY: Metropolitan Books.

PAGE 157

157 Daly, Kathleen. (1992). Womens pa thways to felony court: Feminist theories of lawbreaking and problems of representation. Southern Review of Law and Womens Studies, 2:11-52. Daly, Kathleen. (1994). Gender, Crime, and Punishment. New Haven, CT: Yale University Press. Daly, Kathleen., & Chesney-Lind, M. (1988). Feminism and criminology. Justice Quarterly, 5:497-538. Dawes, Robyn M., Faust, David, & Meehl, Paul E. (1993). Statistic al prediction versus clinical prediction: Improving what works. In G. Karen & C. Lewis (eds.), A Handbook for Data Analysis in the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. DeLisi, Matt. (2005). Career Criminals in Society. Thousand Oaks, CA: Sage Publications. Dershowitz, Alan M. (1973). Preventive confinement: A suggested framework for Constitutional analysis. Texas Law Review, 51:1277-1324. Dowden, C., & Andrews, Don A. (1999). What works for female offenders: A meta-analytic review. Crime& Delinquency, 45:438-452. Fabiano, Elizabeth A., Porporino, Frank J., & R obinson, David. (1991). Canadas cognitive skills program corrects offenders faulty thinking. Corrections Today, 53:102-108. Farrington, David P. (1986).Age and crim e. In M. Tonry & N. Morris (eds.), Crime and Justice: An Annual Review of Research vol. 7:189-250). Chicago, IL: University of Chicago Press. Farrington, David P. (1987). Predic ting individual crime rates. In Don Gottfredson and Michael Tonry, (eds.), Prediction and Classification: Criminal Decision Making Chicago: University of Chicago Press. Farrington, David P., Snyder, Howard N., and Fi nnegan, Terrence A. (1988). Specialization in juvenile court careers. Criminology, 26:461-485. Feeley, Malcolm M. & Simon, Jonathan. ( 1992). The new penology: Notes on the Emerging strategy of corrections and its implications. Criminology, 30:449-474. Flanagan, Timothy J. (1996). Reform or punish: Am ericans views of the correctional system. In Timothy J. Flanagan and Dennis R. Longmire (eds.), Americans View Crime and Justice: A National Public Opinion Survey. Thousand Oaks, CA: Sage. Foley, M. (1991). Risk assessment: The developmen t of an instrument to predict failure on probation for juveniles. Unpublished docto ral dissertation, Fordham University, New York.

PAGE 158

158 Funk, Stephanie J. (1999). Risk assessment fo r juveniles on probation: A focus on gender. Criminal Justice and Behavior, 26:44-68. Garland, David & Young, Peter. (eds.). (1983). The Power to Punish: Contemporary Penalty and Social Analysis. London: Heinemann. Gendreau, Paul. (1996). The principles of effectiv e interventions with offenders. In Alan T. Harland (ed.), Choosing Correctional Options Th at Work: Defining the Demand and Evaluating the Supply. Thousand Oaks, CA: Sage. Gendreau, Paul, Andrews, Don A., Goggin, Claire & Chanteloupe, F. (1992). The development of clinical and policy guidelines for the predic tion of criminal behavior in criminal justice settings. Unpublished manuscript availabl e from the Department of Psychology, University of New Brunswi ck, St. John, New Brunswick. Gendreau, Paul, Cullen, Francis T., Bonta, James. (1994). The Effects of Prison Sentences on Recidivism. Ottawa, Ontario: Solic itor General Canada. Gendreau, Paul, Little, T., & Goggin, C. (1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34:575-607. Gendreau, Paul, & Ross, Robert R. (1979). Effec tive correctional treatment: Bibliotherapy for cynics. Crime & Delinquency, 25:463-489. Gendreau, Paul, & Ross, Robert R. (1981). Corre ctional potency: Treatm ent and deterrence on trial. In R. Roesch & R.R. Corrado (eds.), Evaluation and Criminal Justice Policy. Beverly Hills, CA: Sage. Gendreau, Paul, & Ross, Robert R. (1987). Revivi cation of rehabilitation: Evidence from the 1980s. Justice Quarterly, 4:349-408. Glueck, Sheldon, & Glueck, Eleanor. (1950). Unraveling Juvenile Delinquency. Cambridge, MA: Harvard University Press. Goggin, Claire, & Gendreau, Paul. (2006). The imp lementation and maintenance of quality services in offender rehabilitation programs. In Clive R. Hollin and Emma J. Palmer (eds.), Offending Behavior Programs: Developm ent, Application, and Controversies. Chichester, U.K.: Wiley. In Press. Goldfarb, Ronald, & Singer, Linda. (1973). After Conviction. New York: Simon and Schuster. Gordon, Diana, R. (1991). The Power to Punish: Contemporary Penality and Social Analysis. London: Heinemann. Gordon, T. (1970). Parent Effecti veness Training. New York, NY: Wyden.

PAGE 159

159 Gottfredson, Don M. (1967). Assessment and pred iction methods in crime and delinquency. In Task Force Report: Juvenile Delinquency and Youth Crime. Washington, D.C.: U.S. Department of Justice. 171-187. Gottfredson, Don M., & Snyder, Howard N. (2005). The Mathematics of Risk Classification: Changing Data into Valid Inst ruments for Juvenile Courts. Office of Juvenile Justice and Delinquency Prevention, National Center for Juvenile Justi ce, U.S. Department of Justice. Gottfredson, Don M., & Tonry, Mich ael. (1987). Prediction and clas sification: Criminal justice decision making. In Crime and Justice: A Review of Research, vol.9. Chicago: University of Chicago Press. Gottfredson, Michael R. (1979). Pa role guidelines and the reducti on of sentencing disparity: A preliminary study. Journal of Research in Crime & Delinquency, 16(1):39-54. Gottfredson, Michael R., & Hirschi, Travis. (1986). The true value of lambda would appear to be zero: An essay on career criminals, crimin al careers, selective incapacitation, cohort studies, and related topics. Criminology, 24:213-234. Gottfredson, Michael R., & Hirschi, Travis. (1990). A General Theory of Crime. Stanford, CA: Stanford University Press. Gottfredson, Stephen D., & Gottfredson, Don M. (1979). Screening for risk: A comparison of methods. Washington D.C.: U.S. Department of Justice. Gottfredson, Stephen D., & Gottfredson, Don M. (1985a). Selective incapacitation? Annals of the American Academy of Political and Social Sciences, 478:135-149. Gottfredson, Stephen D., & Gottfredson, Don M. (1985b). Screening for risk among Parolees: Policy, practice, and method. In David P. Farrington and Roger Tarling (eds.), Prediction in Criminology. Albany, NY: SUNY Press. Greenberg, David. (1975). The incapacitative effect of imprisonment: Some estimates. Law & Society Review, 9:541-580. Greenwood, Peter W. (1998). Investing in prison s or prevention: The state policy makers dilemma (Special Issue: NCCD 90th Anniversary). Crime and Delinquency, 44:136-142. Greenwood, Peter W., & Abrahamse, Allan. (1982). Selective Incapacitation. Santa Monica, CA: Rand. Greenwood, Peter W., & Turner, Susan. (1987). Selective Incapacitation Revisited: Why the High-Rate Offenders are Hard to Predict Santa Monica, CA: Rand.

PAGE 160

160 Grove, William M., Zald, David H., Lebow, Boyd S., Snitz, Beth E., & Nelson, Chad. (2000). Clinical versus mechanical prediction: A meta-analysis. Psychology Assessment, 12:1930. Halleck, Seymour L., & Witte, Ann D. (1977). Is rehabilitation dead? Crime & Delinquency, 23(4):372-383. Hanson, R. Karl, & Bussiere, Monique T. (1998). Predicting relapse: A me ta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66:348-362. Hanson, R. Karl, & Harris, A.J.R. (2000). Where should we intervene? Dynamic predictors of sexual offense recidivism. Criminal Justice and Behavior, 27:6-35. Hanson, R. Karl, & Morton-Bourgon, K. (2004). Predictors of Sexual R ecidivism: An Updated Meta-Analysis (User report 2004-02). Public Sa fety and Emergency Preparedness, Ottawa, Canada. Harland, Alan T. (1996). Choosing Correctional Options th at Work: Defining the Demand and Evaluating the Supply. Thousand Oaks, CA: Sage. Hoge, Robert D. (2001). The Juvenile Offender: Theory, Research, and Application. Norwell, MA: Kluwer. Holtfreter, Kristy, & Morash, Merry. (2003). The needs of women offenders: Implications for correctional programming. Women and Criminal Justice, 14:137-160. Hubbard, Dana J. (2002). Cognitive-behavioral treatment: An analysis of gender and other responsivity characteristics and their effect s on success in offender rehabilitation. Ph.D dissertation, University of Ci ncinnati, Cincinnati, Ohio. Hubbard, Dana J., & Pratt, Travis C. (2002). A meta-analysis of the pr edictors of delinquency among girls. Journal of Offender Rehabilitation, 34:1-13. Hughes, John. (1971). Crime: Its causes, control, and effect on society. In 1971 Report of the New York State Joint Legislative Committee. State of New York Legislation Document Vol. 26. Izzo, Rhena L. & Ross, Robert R. (1990). Meta-a nalysis of rehabilitation programs or juvenile delinquents: A brief report. Criminal Justice and Behavior, 17:134-142. Kempf-Leonard, Kimbery, (1987). Specialization and the criminal career. Criminology, 25:399420. Lanza-Kaduce, Lonn, Gartin, P., Dundes, L., Blount, W.R., Boulinois-Manning, S., Kuhns III, J.B., Holten, N.G., Mahan, S., Taylor, D.W., Terry III, W.C., Frank, H., and Caouette, B. (1990). Prison utilization study: Risk assessmen t techniques and Floridas inmates. Vol.

PAGE 161

161 I: Application to Males and Females. Division of Economic and Demographic Research of the Joint Legislative Management Committe e, Florida Legislature, Tallahassee, FL. Latessa, Edward, Cullen, Francis, & Gendreau, Paul. (2002). Beyond correctional quakery: Professionalism and the possibili ty of professional treatment. Federal Probation, 66:4349. Law, M. (2004). A longitudinal follow-up of fe derally sentenced women in the community: Assessing the predictive valid ity of the dynamic characteristics of the Community Intervention Scale. Ph.D. dissertation, Carleton University, Ottawa, Canada. Lewis, Neal A. (1994). President Foresees Sa fer U.S.: Others See Crime Bill as Symbolic. The New York Times, August 27:6. Link, Bruce G., Andrews, Howard, & Cullen, Francis T. (1992). The violent and illegal behavior of mental patients reconsidered. American Sociological Review, 57:275-292. Lipsey, Mark W. (1989). The efficacy of interven tion for juvenile delinquency: Results from 400 studies. Paper presented at the 41st annual meeting of the American Society of Criminology, Reno, NV. Lipsey, Mark W. (1992). Juvenile delinquency treatment: A meta-analytic inquiry into the variability of effects. In T.D. Cook et al. (eds.), Meta-Analysis for Exploration: A Casebook. New York, NY: Russell Sage Foundation. Lipsey, Mark W., & Derzon, James H. (1998). Predic tors of violent or se rious delinquency in adolescence and early childhood: A synthesis of longitudinal re search. In R. Loeber & D. Farrington (eds.), Serious and Violent Juvenile Offenders: Risk Factors and Successful Interventions. Thousand Oaks, CA: Sage. Lipsey, Mark W., & Wilson, David B. (1993). Th e efficacy of psychological, educational, and behavioral treatment. American Psychologist, 48:1181-1209. Lipton, Douglas, Martinson, Robert and Wilks, Judith. (1975). The Effectiveness of Correctional Treatment: A Survey of Trea tment Evaluation Studies. New York: Praeger. Lowenkamp, Christoper T., Latessa, Edward J ., & Smith, Paula. (2006). Does correctional program quality really matter? The impact of adhering to the principles of effective intervention. Criminology and Public Policy, 5:575-594. Mannheim, Hermann, & Wilkins, Leslie T. (1955). Prediction Methods in Relation to Borstal Training. London: Her Majestys Stationery Office. Martinson, Robert. (1974). What works? Qu estions and answers about prison reform. The Public Interest, 35:22-54.

PAGE 162

162 McGuire, James. (2002). Evidence-based pr ogramming today. Paper presented at the International Community Corrections Asso ciation conference, Boston, MA, November, 2002. Mednick, Sarnoff, & Christiansen, Karl O. (1977). Biosocial Bases of Criminal Behavior. New York: Gardner. Meehl, Paul E., & Grove, William M. (1996). Comp arative efficiency of informal (subjective, impressionistic) and formal (mechanical, algorithmic) prediction procedures: The clinical-statistical controversy. Psychology, Public Policy, and Law, 2:293-323. Messinger, Sheldon. (1969). Strategi es of control. Ph.D. dissertation, Department of Sociology, University of California at Los Angeles. Messinger, Sheldon, & Ber ecochea, John. (1990). Dont Stay Too Long but do Come Back Soon. Proceedings, Conference on Growth and Its In fluence on Correctional Policy, Center for the Study of Law and Society, Univers ity of California at Berkley. Miller, William R., & Rollnick, Steven. (2002). Motivational Interviewing: Preparing People for Change. New York, NY: Guilford Press. Mitford, Jessica. (1973). Kind and Usual Punishment. New York: Alfred A. Knopf. Moffitt, T.E. (1993). Adolescentlimited and life-course persiste nt antisocial behavior: A developmental taxonomy. Psychological Review, 100(4), 674-701. Monahan, John. (1981). Predicting Violent Behavior: An Assessment of Clinical Techniques. Beverly Hills, CA: Sage. Moore, Mark H., Estrich, Susan R., McG illis, Daniel, and Spelman, William. (1984). Dangerous Offenders: The Elusive Target of Justice. Cambridge, MA: Harvard University Press. Mossman, Douglas. (1994). Assessing predictions of violence: Being accurate about accuracy. Journal of Consulting and Clinical Psychology, 62:783-792. Nagin, Danniel S. (1998). Deterrence a nd incapacitation. In Michael Tonry (ed.), The Oxford Handbook of Crime and Punishment. New York: Oxford University Press. Nagin, Danniel S. & Land, Kenneth C. ( 1993). Age, criminal career, and population heterogeneity: Specification and esti mation of a nonparametric mixed poisson model. Criminology, 31:327. National Institute of Corrections. (2004). Implementing Evidence-Based Practice in Community Corrections: The Principles of Effective Intervention. Washington, D.C.: National Institute of Corrections.

PAGE 163

163 ODonnell, C.R., Lydgate, T., & Fo, W.S.O ( 1971). The buddy system: Review and follow-up. Child Behavior Therapy, 1:161-169. Palmer, Ted. (1975). Martinson revisited. Journal of Research in Crime and Delinquency, 12:133-152. Palmer, Ted. (1983). The effectiveness issue today: An overview. Federal Probation, 46:3-10. Palmer, Ted. (1992). The Re-Emergence of Co rrectional Intervention. Newbury Park, CA: Sage. Petersilia, Joan. (1980). Criminal career research: A review of recent evidence. In Norval Morris and Michael Tonry (eds.), Crime and Justice: A Review of Research, vol.2. Chicago: University of Chicago Press. Piquero, Alex R., Farrington, David P., and Bl umstein, Alfred. (2003). The criminal career paradigm. In Michael Tonry (ed.), Crime and Justice: A Review of Research, vol.30. Chicago: University of Chicago Press. Pratt, Travis C., Maahs, J., and Stehr. (1998) The symbolic ownership of the corrections problem: A framework for understanding the development of corrections policy in the United States. Prison Journal, 78:451-464. Quinsey, Vernon L., Coleman, Garth, Jones, Brian, & Altrows, Irwin F. (1997). Proximal antecedents of eloping and re-offending among s upervised mentally disordered offenders. Journal of Interpersonal Violence, 12:794-813. Reichman, Nancy. (1986). Managing crime risks: Toward an insurance-based model of social control. Research in Law, Deviance and Social Control, 8:151-172 Reisig, Michael D., Holtfreter, Kristy, and Mo rash, Merry. (2006). Asse ssing recidivism risk across female pathways to crime. Justice Quarterly, 23:384-405. Ritzer, George. (1993). The McDonaldization of Society. Newbury Park, CA: Pine Forge. Shannon, Lyle W. (1991). Changing Patterns of Delinquency and Crime: A Longitudinal Study in Racine. Boulder, CO: Westview Press. Shannon, Lyle W., McKim, Judith L, Curry, James P., and Haffner, Lawrence J. (1988). Criminal Career Contin uity: Its Social Context. New York: Human Science Press. Sherman, Lawrence W., Gottfredson, Denise C., M acKenzie, Doris L., Eck, John, Reuter, Peter, & Bushway, Shawn D. (1997). Preventing Crime: What Work s, What Doesnt, Whats Promising. Washington, DC: National Institute of Justice. Shichor, David. (1997). Three strikes as a public policy: The convergence of the new penology and the McDonaldization of punishment. Crime and Delinquency, 43:470-493.

PAGE 164

164 Shinar, Shlomo & Shinnar, Reuel. (1975). The e ffects of the criminal justice system on the control of crime: A quantitative approach. Law and Society Review, 9(4):581-610. Simon, Frances H. (1971). Prediction Methods in Criminology. London, Her Majestys Stationery Office. Simon, Leonore M.J. (1997). Do criminal offenders specialize in crime types? Applied & Preventive Psychology, 6:35-53. Simourd, L., & Andrews, Don A. (1994). Correlates of delinquency: A look at gender differences. Forum on Corrections Research, 6:26-31. Simpson, S. (1989). Feminist th eory, crime, and justice. Criminology, 27:605-631. Slansky, David A. (1995). Cocaine, race, and equal protection. Stanford Law Review, 47:12831322. Snyder, H.N. (2003). Juvenile Arrests 2000. Wash ington, D.C.: Office of Juvenile Justice and Delinquency Prevention. Retrieved from http://ojjdp.ncjrs.org/ojstatbb /asp/JAR_Display.asp?ID=qa2300031502 (September 19, 2003). Snyder, H.N., & Sickmund, M. (1999). Juvenile Offenders and Victims: 1999 National Report. Washington, D.C. Office of Juvenile Justice and Delinque ncy Prevention: 115. Spelman, William. (1994). Criminal Incapacitation. New York: Plenum. Stolzenberg, Lisa, & DAlessio, Stew art J. (1997). Three strikes a nd youre out: The impact of Californias new mandatory senten cing law on serious crime rates. Crime and Delinquency, 34:457-470. Stouthamer-Loeber, Magda, Loeber, Rolf, Wei, Evelyn, Farrington, David P., & Wikstrom, PerOlof, H. (2002). Risk and promotive effects in the explanation of persistent serious delinquency in boys. Journal of Consulting and Clinical Psychology, 70:111-124. Sullivan, Christopher J., McGlogin, Jean M., Prat t, Travis C., and Piquero, Alex R. (2006). Rethinking the norm of offender generality: Investigating specialization in the shortterm. Criminology, 44:199-234. Swanson, Jeffrey W., Borum, Randy, Swartz, Ma rvin S., & Monahan, John. (1996). Psychotic symptoms and the risk of viol ent behavior in the community. Criminal Behaviour and Mental Health, 6:309-329. Tittle, Charles R., Grasmick, Harold G. (1998). Criminal behavior and age: A test of three provocative hypotheses. The Journal of Crimi nal Law & Criminology, 88(1):309-342.

PAGE 165

165 Tonry, Michael. (1995). Malign Neglect: Race, Crime and Punishment in America. New York: Oxford University Press. Van Voorhis, Patricia. (1987). Co rrectional effectiveness: The high cost of ignoring success. Federal Probation, 51:56-62. Visher, Christy A. (1986). The Rand inmate survey: A reanalysis. In Alfred Blumstein, Jacqueline Cohen, Jeffrey A. Rot h, and Christy A. Visher (eds.), Criminal Careers and Career Criminals, vol.2. Washington, D.C.: National Academy Press. von Hirsch, Andrew. (1972). Prediction of crimin al conduct and preventive confinement of convicted persons. Buffalo Law Review, 21:717-758. von Hirsch, Andrew. (1985). Past or Future Crimes: Deservedness and Dangerousness in the Sentencing of Criminals. New Brunswick, NJ: Rutgers University Press. von Hirsch, Andrew, & Gottfredson, Don M. (1984) Selective incapacitation: Some queries on research design and equity. Review of Law and Social Change, 12:11-51. Warner, Sam B. (1923). Factors determining pa role from the Massachusetts reformatory. Journal of Criminal Law & Criminology, 14:172-207. Weber, Max. (1968). Economy and Society. Totowa, NJ: Bedminster. Welsh, Wayne N. (2006). The need for a comprehensive approach to program planning, development, and evaluation. Criminology & Public Policy, 5:603-614. Welsh, Wayne N., & Harris, Philip W. (2004). Criminal Justice Policy and Planning, 2nd edition. Cincinnati, OH: LexisNexis Anderson Publishing Co. West, Donald J., and Fa rrington, David P. (1997). The Delinquent Way of Life. London: Heinenmann. Wiebush, R., Baird, C., Krisberg, B., & Onek, D. (1995). Risk assessment and classification for serious, violent, and chronic juveni le offenders. In J. Howell (Ed.), Guide for Implementing the Comprehensive Strategy for Serious, Violent, and Chronic Juvenile Offenders. Washington, D.C.: U.S. Department of Ju stice, Office of Juvenile Justice and Delinquency Prevention. Wilbanks, W.L. (1985). Predicting failure on paro le. In David P. Farrington and Roger Tarling (eds.), Prediction in Criminology. Albany, NY: SUNY Press. Wilkins, Leslie T. (1973). Crime and crim inal justice at the turn of the century. Annals of the American Academy of Political and Social Science, 408:13-29.

PAGE 166

166 Wilson, James Q. (1975). Thinking About Crime. New York: Basic Books. Wilson, James Q., and Herrnstein, Richard. (1985). Crime and human nature. New York: Simon and Schuster. Winokur, Kristin P. & Blankenship, Julia L. (200 7). Evidence-based risk assessment in the juvenile justice system: Floridas Positive Achievement Change Tool (PACT). PowerPoint presentation, The Justice Research Center. Wright, James D. & Rossi, Peter H. (1986). Armed and Considered Dangerous: A Survey of Felons and Their Firearms. New York: Aldine de Gruyter. Wolfgang, Marvin, Figlio, Robert M., and Sellin, Thorsten. (1972). Delinquency in a Birth Cohort. Chicago: Chicago University Press. Zamble, Edward, & Quinsey, Vernon L. (1997). The Criminal Recidivism Process. New York: Cambridge University Press. Zimring, Franklin E., & Hawkins, Gordon. ( 1988). The new mathmatics of imprisonment. Crime & Delinquency, 34:425-436. Zimring, Franklin E., & Hawkins, Gordon. (1995). Incapacitation: Penal confinement and the restraint of crime. New York: Oxford University Press.

PAGE 167

167 BIOGRAPHICAL SKETCH Michael Baglivio is a doctoral candidate in the Department of Criminology, Law, & Society at the University of Fl orida. He has a Bachelor of Sc ience in psychology, a Master of Health Sciences in rehabilita tion counseling, and a Master of Arts in sociology, all from the University of Florida. Michael is currentl y a co-Associate Project Director on a field demonstration grant sponsored by the Office of Juvenile Justice and Delinquency Prevention involving the effectivene ss of various treatments of juven ile offending youth with maltreatment histories. His research interests include criminol ogical theory, risk asse ssment, and life-course criminology.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101206_AAAABG INGEST_TIME 2010-12-06T07:48:57Z PACKAGE UFE0021569_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 47211 DFID F20101206_AAAIRB ORIGIN DEPOSITOR PATH baglivio_m_Page_014.pro GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
b3023727684004207333da4cba3f0cff
SHA-1
5a3599649f338c58360dcd93fce55971b1ccea02
1053954 F20101206_AAAIQN baglivio_m_Page_167.tif
46795017a34f0c8d99200401f5d58f7c
d88311504be5331dc19c27dbf0531287d419435a
F20101206_AAAIPY baglivio_m_Page_152.tif
235f46761154f24ae46dc0412db04326
cf3cc7d8da1fb941b79afe84425296d3e556fec7
38141 F20101206_AAAIRC baglivio_m_Page_015.pro
d23707cc5f6dfe46950a197cff4ba994
f7241cf913ac65d923e15f04c43c72ae76c6c7fe
7930 F20101206_AAAIQO baglivio_m_Page_001.pro
e2b698673e9b00d596d46de564b187cd
07f8dd4a45089d89fabbf47d8648c630e551fd75
F20101206_AAAIPZ baglivio_m_Page_153.tif
608267e5bea00c85b04a53945ea976f0
cabf82d92ba8569d852fd4c8d9e1e7773f42214c
53219 F20101206_AAAIRD baglivio_m_Page_016.pro
be5ea8dd04fc13d1425e7974804d5447
a4cdfbe9e54f2b513ed69f1422527b2769806db5
967 F20101206_AAAIQP baglivio_m_Page_002.pro
b857b8a754542848c95f217c701ff310
35599019de5a445d631197972bbca4d6b956c67d
51293 F20101206_AAAIRE baglivio_m_Page_017.pro
75a0b52edaa364fd08db6a7cefde862e
38c2bf3b5a460f74f10f67e2a4b611c5557853ef
1225 F20101206_AAAIQQ baglivio_m_Page_003.pro
16caedd9cdcad27cfb20e5b13506f041
7eb814cbaa8625f5d8c34500d10c210fb7d41c7f
56351 F20101206_AAAIRF baglivio_m_Page_018.pro
18bcd42f84059d2caa6dedf96fb2a4c4
bfd98f7f88107bc86c417636e2c0ba88e677855c
7614 F20101206_AAAIQR baglivio_m_Page_004.pro
ea7b6c57a393c2f319ab382459dbe81a
6aa4daa90264714e4672eb8e56b21bc69221504f
53005 F20101206_AAAIRG baglivio_m_Page_019.pro
988972fec8b508626b025680617b849c
fa2f96388e5d42b28518b5cc7e41d54fb2f60732
79316 F20101206_AAAIQS baglivio_m_Page_005.pro
8d51dd9b0ddd93d6dff616a45da3c5f1
6d6e75e77fa06c1d56246cdea31ca150a73f9f41
52408 F20101206_AAAIRH baglivio_m_Page_020.pro
305d5065c5fefa848ab814c5a4172807
15a5dc2ebeaaa5004e788713029747b0e98a37ca
55083 F20101206_AAAIRI baglivio_m_Page_021.pro
056160291b762b6d52deb9a8147ee76b
4df727d7441aeddbf106e34b39fc24f010b953a9
6710 F20101206_AAAIQT baglivio_m_Page_006.pro
1e8bd01190aaf6b515ec045e86d7ca07
a25fc7388dd2755a1ed09382a510325d14b7e2f3
30714 F20101206_AAAIRJ baglivio_m_Page_022.pro
5ab93d9eaac31afd9224a5f83bb70458
8e8f9d284c6ee8981159143b311cc26b70935f4b
69868 F20101206_AAAIQU baglivio_m_Page_007.pro
9ef2c2ffc28fc354a2d72497ee43e7a7
76d83c90170e37ae4e211963da49fa57c542ba92
49767 F20101206_AAAIRK baglivio_m_Page_023.pro
8af6fde2705aeb3a6ae9422cef9b46b2
0b6dd7a07ac6851035e76ac85e4b9bc5150c4c93
31798 F20101206_AAAIQV baglivio_m_Page_008.pro
f89d85cbfb1375c37f929015b6da9dca
f2340bc281c0f0426c408b33c9263ff8fe594c39
55817 F20101206_AAAIRL baglivio_m_Page_024.pro
d4f04f3d744609d615d0318b0d403fbe
4df1e8fe460ddd9b07f094b4aa1eb79e2cc4e1d8
8204 F20101206_AAAIQW baglivio_m_Page_009.pro
a4de117522ffe09ae7a35e734b70bf4a
d838d0429d7b2624f91ae267af6dc172397045dc
56679 F20101206_AAAIRM baglivio_m_Page_025.pro
34a2703c9e6e8c68f1f399d82c762590
faf625cbff1658d0d9d30ebf792447bdf6e7d834
33272 F20101206_AAAIQX baglivio_m_Page_010.pro
82acaa8f35fa6c3413361d5175207e85
4f07bd6440cc5b8abcf1dff0b8bbd7fc404b2954
55078 F20101206_AAAISA baglivio_m_Page_039.pro
b60d0024016267e131578cef0df50170
96bf1a7eb2910b2d949d55b5ad90cef2a3b46b87
67470 F20101206_AAAIRN baglivio_m_Page_026.pro
600b2b8cfac30ccfbe321a99262a0842
e39d2fb7764cc60afeb979d13fcec8a75e2c30a7
54788 F20101206_AAAIQY baglivio_m_Page_011.pro
b1d001f5808192b4e82a67b4252d80c2
0c94daf047f6d2603a6947e15039eec5b292b429
53381 F20101206_AAAISB baglivio_m_Page_040.pro
74c5d989ff59b14b1ec157e6ee3d2cf9
53ca5d0ebbcf686e159a8350f98634bcff7e0e49
57492 F20101206_AAAIRO baglivio_m_Page_027.pro
82f49ae3b42c01ee2bb1b3ba1fda48f2
6a922dfa6c297fe05e26158f6998946468c6bae5
51269 F20101206_AAAIQZ baglivio_m_Page_012.pro
127ffdf58d247e8c49ee5213bb40f2df
20a943aadc692610944c7739b5627b86d745756f
54050 F20101206_AAAISC baglivio_m_Page_041.pro
f6bb9292da87bd66888796bafd48fd44
778ac11afefb273fe8b9ac3cfe8561f91f41e4e7
56234 F20101206_AAAIRP baglivio_m_Page_028.pro
1225800917b96be7ab009024071919b9
b4c2b524a0549d4075632173881a755c404105d0
71862 F20101206_AAAISD baglivio_m_Page_042.pro
c90e7a7d967c2e4328635df54c0a90e0
22bc8793120e59ad9e15de341a5804d2203373aa
56139 F20101206_AAAIRQ baglivio_m_Page_029.pro
68ba7ec62af92f2abb5b276198134e9b
61eb0c112a4d57727689cd21a3d020b8a9d024bb
53537 F20101206_AAAISE baglivio_m_Page_043.pro
0d1e894c27c4e02840cf92c7c768ed9a
c6a308e4b0d81140c7281500529db803cd4cb7ea
53882 F20101206_AAAIRR baglivio_m_Page_030.pro
e7f4863f7eed9517625bb86a45f96780
f1c06c12cd353e781718fba911dbdf4b40a81f79
55888 F20101206_AAAISF baglivio_m_Page_044.pro
fcebf9c28b9f5b24dba14f9fa6492bbb
2f774548c209fed3c1d8864b4e5813c7b84a5b89
73978 F20101206_AAAIRS baglivio_m_Page_031.pro
3c73e9cc11b82d5762b0425e2ee2a3f6
0a33760e9fc0c594eacd0dfdde039542432c645e
54274 F20101206_AAAISG baglivio_m_Page_045.pro
0f02bafde113b24e766e3abb85802d32
22a2c8435652b2444b161caf0550fbbcf4c95b50
53600 F20101206_AAAIRT baglivio_m_Page_032.pro
ca843e6c33c81b34ef161d17f4f69ea7
e1ee07342d4e5ee53d8a524dfa25ccbd954d0a85
54252 F20101206_AAAISH baglivio_m_Page_046.pro
bf355c7f3d4e9a49fd4117de56802988
0e3d5154c2294a1a33b9afd50b4f9f93e765d9f5
54365 F20101206_AAAISI baglivio_m_Page_047.pro
672e1e8f380695792fed276f256a0eca
98d194a93dc0d469bd2e093debcfaa88e8ae8b08
55353 F20101206_AAAISJ baglivio_m_Page_048.pro
4c9d600af34dfb488f599ff5c936c183
0d93e19de352dedfa54b182771a2638cbb762b71
55511 F20101206_AAAIRU baglivio_m_Page_033.pro
d4cb98f4aaa5df6a427c9bfd4b663fe7
349906b60ff8f466df54620c6d6b07fecbd9db18
52943 F20101206_AAAISK baglivio_m_Page_049.pro
61da7791fb4b1d0293fd769e71967c2d
36e8698f59239592a6f57070156570b1dfa7bd00
54523 F20101206_AAAIRV baglivio_m_Page_034.pro
6f1d1c6822d1694e30152870ffde7490
0c342b042685aeca5d5a6b679401b53c43b32a36
54633 F20101206_AAAISL baglivio_m_Page_050.pro
9c930b370a6c777b389940f78921f378
822a49b6c3d7db4c54bea5e9c0e0997787b42230
54755 F20101206_AAAIRW baglivio_m_Page_035.pro
cf51dd4e6843a353906277f088af8a13
679b7cb93422277afce4aea5c97e3b07ea838db0
55924 F20101206_AAAITA baglivio_m_Page_065.pro
0afe45a6e3311ad892d152288d742be5
b19b6968f0400cd6e6725cb71e8d895577eeb14c
54580 F20101206_AAAISM baglivio_m_Page_051.pro
e672fd6a8e64407d1a778f9dcf046103
4abfa47d09c9a1e43cd2e75094a1a12b44680f63
54493 F20101206_AAAIRX baglivio_m_Page_036.pro
cb9ce721a371dec778fc46fc1a2aedae
1e06124cca128bcc2bb1e17f370f1eb3641fa65f
55613 F20101206_AAAITB baglivio_m_Page_066.pro
420de38683994f54fc59f0c67269ceeb
deae49e6e919ee5cb9d90a2f558282524f4f2949
55478 F20101206_AAAISN baglivio_m_Page_052.pro
d0dce766e9f4a96b4aa273534764f31e
c2b9b3614f9862c6ca9d38de134a3e3ca18bc634
52456 F20101206_AAAIRY baglivio_m_Page_037.pro
93392c2bf79ca2195694ffcc0b784d3d
ddc86977ecaf3536dd62b9d1134c9a7eca08c73a
49907 F20101206_AAAITC baglivio_m_Page_067.pro
0dd6006e2b8498689749d055fe6f54eb
09abfbe8eb20aa3766ab54d9f91a7cc65cf4d008
53453 F20101206_AAAISO baglivio_m_Page_053.pro
35c58d13f415adf4bf1b11867dc55893
7cfa46b4fcaeb987f90adab99654d8d2f8fe5046
49082 F20101206_AAAIRZ baglivio_m_Page_038.pro
0b02bfbd3f59863a87db518f5fd2abb2
939c49092f71554dc2caac6c17453df4791082cc
48234 F20101206_AAAITD baglivio_m_Page_069.pro
772eef7fcd3f2cd41b735fa8f5adc4b5
b555a00d00e59a0bae7f07b813085eeb2c8e165a
53122 F20101206_AAAISP baglivio_m_Page_054.pro
05c6a58d5641ee4211113711da12c9e7
55981049930fa72beccff5940b7f76fcb9622eec
51217 F20101206_AAAITE baglivio_m_Page_070.pro
fdd44b7d16fd6c7fb1ea59cf201471cf
4e49675d4ee96b19e685b43f34c13d0f0bfe90c9
56597 F20101206_AAAISQ baglivio_m_Page_055.pro
07c336463e72cb074671b548ab27bb8d
a56dea4ba1738f7ca049fcfa4a470708d3d4befb
52157 F20101206_AAAITF baglivio_m_Page_071.pro
b983c8999a246d824a8ee48313f672ef
4620bed141d73d2e3eac7c173941ee668db414bc
F20101206_AAAISR baglivio_m_Page_056.pro
fcd382ef322840e5b95c543a12e1931f
c60aca7e15c3dc6bcf7656a082cc572ed64c8c92
49881 F20101206_AAAITG baglivio_m_Page_072.pro
764468fad8e5b8d690ccaa910a5d51ef
781c9a5a873ac54172c07b1c0a0e43f6a2016d0c
54185 F20101206_AAAISS baglivio_m_Page_057.pro
e2edb641dae58d81c1d5ab1f7768a69d
4abf6b0b752197cea48fd44cf6b3b7df1b1aba93
54628 F20101206_AAAITH baglivio_m_Page_073.pro
35cadc2aeed476d7099b32329ecabdfc
a2344f913f18a98ccd7d55bafa012221f252c1ac
54492 F20101206_AAAIST baglivio_m_Page_058.pro
467da8aa1336fdd82dea773cc2ddf1c8
1a0dfad02866c86f104d362b3628207e6a84eb2a
56331 F20101206_AAAITI baglivio_m_Page_074.pro
2ea30d075eb2b8bfdae6347bf49a01c7
0c9791a04573416bb2651546d130f9b4147bb24b
56244 F20101206_AAAISU baglivio_m_Page_059.pro
fe95ffd4aa20b051f3f2b04ebcf34756
f29e01e53b6558a9aebc0f7786d9b16963544347
53223 F20101206_AAAITJ baglivio_m_Page_075.pro
3fd882cd2272546e7e0dc29ed6f72677
fbfc4135fd55ebb239e51f32697fae18f625230c
57589 F20101206_AAAITK baglivio_m_Page_076.pro
291f90baeb9e2feefffa0c2c679b81f5
6f674b7983d4015667508c376d6e980df23a812a
54184 F20101206_AAAISV baglivio_m_Page_060.pro
fae4fcc640349ed84e7c6f9e9c1df9fd
eda4a580b292f434168e92179a208b904a37b3f7
55175 F20101206_AAAITL baglivio_m_Page_077.pro
97cfb5d603993ed785f36b799c2127b4
c84ea6440661bc3453d25747c8e437c5d1209a89
54060 F20101206_AAAISW baglivio_m_Page_061.pro
03fec8ee1ab6b95003e2f989662ebdd7
ae483c22f341ec0ee526a0dd730ba100342499f2
56777 F20101206_AAAITM baglivio_m_Page_078.pro
42a507fb5cc097452108e563390e9441
6dfb5b370dfc8e5bd2587744f10300c53046c1cb
54052 F20101206_AAAISX baglivio_m_Page_062.pro
6ff827cca1e596a5848982017cffb17b
fbc5b5054d01318e49050b0ffa3b03c632a6b53b
53682 F20101206_AAAIUA baglivio_m_Page_093.pro
086e2b65cf2fde7be221c940f118c01b
7a3f2d88d22b97beb0cba357d8baf2a562e289a2
53557 F20101206_AAAITN baglivio_m_Page_079.pro
e40a5bf2bd2ccd2ee291cef8b3a7ee07
b010b47dff4c59d7bc0cda31da136e2b714fd21e
55919 F20101206_AAAISY baglivio_m_Page_063.pro
e067a9a802b1e0f5848f918d281acb5c
bac556c9916f45f37e357e5d8edefa0b19c86687
54926 F20101206_AAAIUB baglivio_m_Page_094.pro
42668d57cc36dc42c271a0e92d748b55
b2e85fa4a4ef45aa00d916827fc24d6e6d293f0a
50290 F20101206_AAAITO baglivio_m_Page_080.pro
410922ffad5af726f937a54e543c5102
02d251eaf3d9cdb52d9b3556965663fca97aa323
56449 F20101206_AAAISZ baglivio_m_Page_064.pro
3a5751d55ae2e5ab9efe1d3fedf3d308
d6bc365f8dbe5cd317ff11416524d122c81d4e95
53772 F20101206_AAAIUC baglivio_m_Page_095.pro
75f6983340f7f6a475fc3dead0c11509
d0f826b18e7c817e2d826bd76360607c121ab930
56196 F20101206_AAAITP baglivio_m_Page_081.pro
d2e1a6c9f3168082362ac3d33b82db09
cf5bbdb825f8aab2c0664c23cd0c0d05d407be2d
51646 F20101206_AAAIUD baglivio_m_Page_096.pro
da7cae06b76c581a2958f3b8a0a29f66
906f9c5d341f513987222f6392d46ded00f8960a
59081 F20101206_AAAITQ baglivio_m_Page_083.pro
4cbb6bee1ce4314abf498d8a0664abd1
000d3cceb225756271fe28abe6c781890da2ac20
47468 F20101206_AAAIUE baglivio_m_Page_097.pro
ca7b4d0d3aa276383a54a9d839ab12dc
1dd58b5c6dbe158f78dc82b6e10cf8eb509b8499
54689 F20101206_AAAITR baglivio_m_Page_084.pro
20dd968de2711f1557e4131f4a11eb05
e4c638075ab24d87246cc5708fe8477527a1b94d
2268 F20101206_AAAJAA baglivio_m_Page_084.txt
9d9b42918fb9e55e6e1e81f29e1bc364
e72055a3f1058587b2f10097d1e9a21b92aeecf6
49598 F20101206_AAAIUF baglivio_m_Page_099.pro
078b6ca1e2e3654107d93effb23212a1
e2f2b374304dfdddcca9fec2dcf1f644436411ed
51031 F20101206_AAAITS baglivio_m_Page_085.pro
d8ea2d557a9053a8108fbc9300a38294
d3e0dbce917414fa2ac35579a6b0df2a296e2fea
2225 F20101206_AAAJAB baglivio_m_Page_085.txt
42ba2dc72251a90b96fe558fda7f50bf
4fe44a08273dd48703c3bf3767fad0d7d6fcd3e3
55452 F20101206_AAAIUG baglivio_m_Page_100.pro
e1851649a24c5930a48dcdc5fef97b35
8f72e61a4cf74c34feaf51f9e041085f257065cb
54590 F20101206_AAAITT baglivio_m_Page_086.pro
26b710ebadacdee2657ef31fd84d85f6
f81a65131cc587aabbecccd7b7d0ec7fd496e6f6
61023 F20101206_AAAIUH baglivio_m_Page_101.pro
5b19ac332ffc27214d225048319e69c6
c375190c12668d25cf960ba7246688da7b652b44
53368 F20101206_AAAITU baglivio_m_Page_087.pro
9507719ab46135549dd3574ff6bfbba7
3bc9eb8d903e37a70cc2c71b6e2880a9fbfa7a30
2218 F20101206_AAAJAC baglivio_m_Page_086.txt
f3f1cab4cd7eeb4f417acc0b5230b25c
57339a4e44773e239ca8a38b50c5efdeee401891
50671 F20101206_AAAIUI baglivio_m_Page_102.pro
ca27abad5bcb1c310198e00b6e5b9c6a
11d21db632ca1f879ec8d79e51514def7d4d0b6d
58354 F20101206_AAAITV baglivio_m_Page_088.pro
58674f7697fa428ccff2ae2ef9cbdead
f97ded5703814c7671cf3a09dcd28ab8c2c0b781
2104 F20101206_AAAJAD baglivio_m_Page_087.txt
b45b389c5346bdfe266edae1bb4cf4cd
a49de5bee888de9f70d6ef7515b0cc31fd849797
55497 F20101206_AAAIUJ baglivio_m_Page_103.pro
ad96d6e04e514df901d7b85abb1df2e8
ca28ba5a87d36695d9e5fa7febff7804c12f43e1
2490 F20101206_AAAJAE baglivio_m_Page_088.txt
669cf315b5c7cfbee4ba71c125a34dfe
e7dd15be15539167e50b2ff2a5d6fdd863edbcf7
57556 F20101206_AAAIUK baglivio_m_Page_104.pro
b3e93c2aab5ab24656a75afd68e7951c
4272b5ecc818447d06ce0c1472d930ef5a99952a
50523 F20101206_AAAITW baglivio_m_Page_089.pro
b5d933e1f4976dfff28926a60d9733d4
266cea0bb563cd6fb161016b6d9690b042d8e88c
2132 F20101206_AAAJAF baglivio_m_Page_089.txt
70ee590ebe6c23af918f9d1356ea0318
97e62f2851f77d62dcc54f93a2a4da59c5222da6
57081 F20101206_AAAIUL baglivio_m_Page_105.pro
419531ec474cde5777c08b5f865942bc
072c5d26e866de96b32b7299783b4cb2c578c6ad
55367 F20101206_AAAITX baglivio_m_Page_090.pro
320422b457a62d14c4396b8243ff29d1
d2dc477ed76e34ee5c43e20de0fea94e3f154efa
2179 F20101206_AAAJAG baglivio_m_Page_090.txt
af3f6500ea35bebe0e904a88175433a8
a789ae20f22d09b93331d0043cca1e6c797f1599
53018 F20101206_AAAIVA baglivio_m_Page_120.pro
970c9df9bcc198e1a040d10a25e5c940
8182735085532c436b58278c6e56dcf37d0913b0
53265 F20101206_AAAIUM baglivio_m_Page_106.pro
fe55cd2b84bf4bdae11e0ed9b2232f9f
9ffece9c49741553b0c9c27ccf54ad71726e6fee
57538 F20101206_AAAITY baglivio_m_Page_091.pro
2a153577f738bdce99ec8f96850a5781
cfddf216488db35b05d0c34f92f49d977dd904a2
2602 F20101206_AAAJAH baglivio_m_Page_091.txt
0e1c70a66716931f72c230fb737b9d2d
ab538529a7282000b58b4a604a521231e2dcb635
45722 F20101206_AAAIVB baglivio_m_Page_121.pro
b174b6e9d6620f5f421ab8e010ed9487
59998e19b52117e89ceeba8657211be63346e006
55319 F20101206_AAAIUN baglivio_m_Page_107.pro
5e060aa5ffc98f51b1410e70e1f8c03e
e6256725ff45fb7c433787bba9ec14fc222589e6
57776 F20101206_AAAITZ baglivio_m_Page_092.pro
ea2e2a391d8174d215f9d9d40487337d
71e5fde0568a7b36b819338946a92e9ea377c07c
2435 F20101206_AAAJAI baglivio_m_Page_092.txt
019984148149387867ae53ff223f27ea
7a0707d58cb58bbdcbc250fae605039cb603a9bd
53282 F20101206_AAAIVC baglivio_m_Page_122.pro
e622a7666e142cb85d07b0f06e6f4b10
46cc1a9d7d6643449763067336e2cda92b64ece0
59039 F20101206_AAAIUO baglivio_m_Page_108.pro
fec7371c10c14f7e0332815ff9899a4b
a5193ac1d4fc1811a5cbf5f3962e84bd99460311
2313 F20101206_AAAJAJ baglivio_m_Page_093.txt
1c42ddca18bd840af217140ee04ec051
24f3cb1d28c63d11ad4e33c3f60a6921ae36a969
43458 F20101206_AAAIVD baglivio_m_Page_123.pro
4e9b4ac74f7ce9505963d359035d2b8d
679f0dcd3c723d0e9e958e45416007e086d61057
61155 F20101206_AAAIUP baglivio_m_Page_109.pro
2a802b1026dcfb820a1a882823dc2ac7
b83dd6e8a903dfb0c1bc7bb7f356ccd07f2a64c4
2239 F20101206_AAAJAK baglivio_m_Page_094.txt
75d04570edc301a545a20b943a3074a6
f1b7c824dbc6addaf75f0c62a3ce30668c0cf70e
33351 F20101206_AAAIVE baglivio_m_Page_124.pro
bab8fdb4d40073c1d6841dad17c244de
d0f4e8fc7aa161113fefa4f8dab9353092736ae3
56379 F20101206_AAAIUQ baglivio_m_Page_110.pro
0fa618cf7a79c415d1f79f5064499d47
922a51da2ea7ec128e9165dcb6a1b2c2b2a31d74
2201 F20101206_AAAJBA baglivio_m_Page_110.txt
198a84e57813a26125caca1f69a81efa
4fcde6f5f36ba7769c2f250b421f308e94303437
2240 F20101206_AAAJAL baglivio_m_Page_095.txt
c28a454fc48aadf2057291c40a580c6c
88e8c8c9ec9d7ebe02e365cef3e5ec0e3f5db2dd
56093 F20101206_AAAIVF baglivio_m_Page_125.pro
4357f167f87d0a98a40426673bb14701
7360f5769ff08536bc18519078f8f702dc39e6cc
52863 F20101206_AAAIUR baglivio_m_Page_111.pro
4adae79f4beb89698ce5158cd3d5b26e
89cdd95555e0f129bf1598d0cdf72dfeee60771f
2245 F20101206_AAAJBB baglivio_m_Page_111.txt
1c058b726092c95e562eb5beefcc9a6a
b567168d35c88c931c3a1515a8b8fce0920dc508
2287 F20101206_AAAJAM baglivio_m_Page_096.txt
d3c3efb0d9d9c416d2c9eadc7e7b8839
85830041ac8800ac24717a38d0fc6630ba31fa4e
50024 F20101206_AAAIVG baglivio_m_Page_126.pro
4c7d3644823899e0e02feaa0069318c1
c209815a78e8ffd71015c884fc6d4cb1cb14535d
52542 F20101206_AAAIUS baglivio_m_Page_112.pro
fb7116746c86bf6d74d5d5615d4b6bd9
5f03bec2c4a124576c97890d2703205bf5d4686d
2068 F20101206_AAAJBC baglivio_m_Page_112.txt
e4a7d46024aef761da727d0ddd0043d8
c0d32dd077f26e8848a02b2f2df836e244d7c417
2063 F20101206_AAAJAN baglivio_m_Page_097.txt
0598b47079c75c8fe53b9396a4d48ea4
a991029e907ab3943e2c1b906378bffe33594f17
40391 F20101206_AAAIVH baglivio_m_Page_127.pro
f1ee1b509656dc61c354abe980fe281f
544220e1a0df88d6b6aa12dc7644ef13b1662616
53830 F20101206_AAAIUT baglivio_m_Page_113.pro
e2a1d0334613dabee33075ceea87b6cb
bdaffa0b18df9c20bb3576c01f246ea771afc08d
2125 F20101206_AAAJAO baglivio_m_Page_098.txt
129842b5b017d71b889248d510cbdc59
9ebdf2aa3cd8afd1c761bc97215c8105d2971eaf
23357 F20101206_AAAIVI baglivio_m_Page_128.pro
dfb1b8862e58e207d6e60bcb5a8edbf3
f2a5548289a510b427adcd8ad3785cbf498fe43e
52231 F20101206_AAAIUU baglivio_m_Page_114.pro
4d67e25587a58185686fca526a749725
c228c785a06fad99ca81575bcffdb629c16828c7
2190 F20101206_AAAJBD baglivio_m_Page_113.txt
0d5f9db3020cd7436d24783141442513
1e04ab5196f742bed31c7fb374067e0c3feb671e
1962 F20101206_AAAJAP baglivio_m_Page_099.txt
bfc5971ad18b9bbbc115e9d86b04e0a2
9a274a55264d817a8227f96218c93419d799f5bf
75830 F20101206_AAAIVJ baglivio_m_Page_129.pro
0df33541166e99d579fb28a0d1cf7d08
9dfa4c2e1cfe607b59ea8a9b998e58650695c836
55591 F20101206_AAAIUV baglivio_m_Page_115.pro
242d72322c3dcd35ca3bfea46c370b19
048ad7daef0356ed7c2f2acd6e49af660f6343a0
2059 F20101206_AAAJBE baglivio_m_Page_114.txt
21d341fda52c96a01610b2194fb106c6
3c3ad8b040fd1a2051ffa1d2015f0fedc14b2543
F20101206_AAAJAQ baglivio_m_Page_100.txt
1b9f9149d8b5f9c26eb6b9025d0b3825
b01c88c412f84467eb0ed8417e615670f52c85bd
83472 F20101206_AAAIVK baglivio_m_Page_130.pro
1077b9f1303d94e2905c1ed9657557ea
fa7b3eb502dd61194daab73f515a845f49bad40d
56391 F20101206_AAAIUW baglivio_m_Page_116.pro
37bf65993aefc7f88ba1dd642fe8f410
3eed2395e0e0390b597f2089e64bd98e86d915f0
2185 F20101206_AAAJBF baglivio_m_Page_115.txt
1eecefcbcb503380ea40249774a84da4
0f2dd1666583cd6a0bb17a4508873a8ae28ae4ab
2646 F20101206_AAAJAR baglivio_m_Page_101.txt
f1b0b72f290aeb5f15c30c55a59b1b6c
596f231ecc56e1e0beee46229772b113913e3c6b
60056 F20101206_AAAIVL baglivio_m_Page_131.pro
d3ac26a997088299aaf71fe7700b4dfa
18965c190e307b24a3c6958f2156484a7d3d3f9a
2212 F20101206_AAAJBG baglivio_m_Page_116.txt
e7ae698caee91ba5ec0449c96421a18c
c7dae4e91c756903ee64537497416a1ddd73ad04
77180 F20101206_AAAIWA baglivio_m_Page_146.pro
cdc7a644a6f371d78eb17358f4f4031d
948eb67030dfdf31ba28bfd105d443e40c0fc0e3
2158 F20101206_AAAJAS baglivio_m_Page_102.txt
61e170e1a0c28335d2b4640c33ff97da
5adb139f35577e37ef6194a9b8a37f574bc1821b
39046 F20101206_AAAIVM baglivio_m_Page_132.pro
9d45880c79eb68a602a88f418365beed
3d1cd9341d9a7224ade08c76866fa9f7a90ff56c
54021 F20101206_AAAIUX baglivio_m_Page_117.pro
583d4f42cba1027d8ad3ecb9fca69e51
b0022d6ddcec81b6d9db6fccc59c4aac6918d59f
2128 F20101206_AAAJBH baglivio_m_Page_117.txt
ac21f704d4a56fd28dfb06c9d314eddf
b10bfe842c58de43cbe6a36b004061bde5a21e0c
65905 F20101206_AAAIWB baglivio_m_Page_147.pro
aa419fa74f8bc37231c38754db4735a8
3bdda9e01a056730b71cad9d612c736948140561
2376 F20101206_AAAJAT baglivio_m_Page_103.txt
01a3377541d070782228fe9953d2bd3d
5c62b050be56a00ece29a00d569094780367e597
53360 F20101206_AAAIVN baglivio_m_Page_133.pro
e832939dc4275698fd83c9eb78377471
46efde2cd0cbd1d907d7814d533ea046d0a16403
56108 F20101206_AAAIUY baglivio_m_Page_118.pro
fca06877c63746acc0da6751d2748360
7e24b792978be11734474fb48644afda650ea189
F20101206_AAAJBI baglivio_m_Page_118.txt
04f7e0e44e823a616d11a5f9c2e7164b
dec1d34a70f7246c4c23e255995b08e283a81f8d
58594 F20101206_AAAIWC baglivio_m_Page_148.pro
b9ecd606203dd5cca7a5478530bad7ad
a5dcb3d03c1329727272770c1d8320799a33bec5
2388 F20101206_AAAJAU baglivio_m_Page_104.txt
73fbd0794cc0752eb620f8cea2ad63aa
9e516b888aeaa22b44c51bd54a5f87f335ac7076
58629 F20101206_AAAIVO baglivio_m_Page_134.pro
9ee8d45bf5a06d34c26c7b3fdd047a57
7831e21cf7fd84f314c20956aa03560f956bd5b0
31217 F20101206_AAAIUZ baglivio_m_Page_119.pro
83480993c7141c349321f27cefab2e4e
1ba58eb74eb2a30fbacf5b868895b310c3bfe6f4
1244 F20101206_AAAJBJ baglivio_m_Page_119.txt
b5864c61f9f76c99c679fb963eac041b
51bf9f7441fa544a45372024db6b512ac9b825ac
84005 F20101206_AAAIWD baglivio_m_Page_149.pro
11a2ea64b397e389afbf91ed55317a3a
912cf512ad0767a3aaddaf823309dab247251d88
2496 F20101206_AAAJAV baglivio_m_Page_105.txt
e090ad7f1ca5f88727b7e8021bfac9e9
643bbc2d852fbe4e3f50354e88c079f2d273525c
60455 F20101206_AAAIVP baglivio_m_Page_135.pro
f6a36f0cf0d7c41327c68ae6c676cd90
82dc77985a9e1cb711b65b2173296df761c97ac5
2170 F20101206_AAAJBK baglivio_m_Page_120.txt
097aab46b1c5bea8f8eff19a7f6d141e
0bbfea14412918f1c6b7a850edfb948710641819
61896 F20101206_AAAIWE baglivio_m_Page_150.pro
c6b392453f4c3ed3ec9fb5aadddd8aea
bbb302d80a23a63042b6d7036d3f0c85f0617ad0
2093 F20101206_AAAJAW baglivio_m_Page_106.txt
90f1a82d36d388c2cc366065c48f6960
cbe170044bfd2c772d4713efcd192e9e9ef963cb
17594 F20101206_AAAIVQ baglivio_m_Page_136.pro
7687a8c92ddb8d04c879bae2f524d040
88419150a584c72a30e7f7448c54c0725165c12f
1212 F20101206_AAAJCA baglivio_m_Page_136.txt
a5e591557ca80a486c9fa158dd227ffb
d7c323cc44fefc451451ef7d3815d02fe133418c
1814 F20101206_AAAJBL baglivio_m_Page_121.txt
a37d4a74a8de68119c7143e12412b095
b2457b0a79b21e406c80d1a75628d6c62dfdfa39
99552 F20101206_AAAIWF baglivio_m_Page_151.pro
735b1b61cf81a4b2fbfc4c75ca111bd4
2281b60e6f9065f9ff8bde151f501e65a7e0d66a
2372 F20101206_AAAJAX baglivio_m_Page_107.txt
9834887886d1896d406bb9338e1b1758
cf4dfa1c9a6e6ad92722c1c275a2ce8767b5d922
85447 F20101206_AAAIVR baglivio_m_Page_137.pro
6438998f4facfc282967a4c6f0144b18
ca94d3d0de24c605474450514353fb781315e42e
4454 F20101206_AAAJCB baglivio_m_Page_137.txt
bcdb43e32e53d2e1d0793f0f1dbe7667
cd098795552ee9b83f75ce84f07a3bd32e95d20a
2169 F20101206_AAAJBM baglivio_m_Page_122.txt
9defea43a37ad79b2372ad06bd777b89
b3b675c3f468e6ae1d953c3bf33cf11dee34bb78
12451 F20101206_AAAIWG baglivio_m_Page_152.pro
35c49f5c86df8478b4fa785556a49def
57324c3bf67631b4c9fc52aaef763e6af43086b1
2826 F20101206_AAAJAY baglivio_m_Page_108.txt
b8e26ee8c56aa4a4bef277da0ffc4f03
707ff525e8976de7287d08944a9adeb34bb2e5e2
80778 F20101206_AAAIVS baglivio_m_Page_138.pro
246a0f9a1e4d99db5380067f575a8951
a6cef9e9a6eae6e1bf9855e9fed2f02946881574
4964 F20101206_AAAJCC baglivio_m_Page_138.txt
e8106d95500699eba0d80686b1a2e78f
e748771244df7ef51fc0727939f99a46cfccfd3f
1732 F20101206_AAAJBN baglivio_m_Page_123.txt
cb716969028e974994f726d95b7c3db3
20fcd7207347ac554f0844a8f72e9ccf46a724ac
64705 F20101206_AAAIWH baglivio_m_Page_153.pro
357c7a5a4982e7ea971f1ca6e012ca68
7329c46751fd5f6e700ff1c170ecfb7b389c2e1a
2753 F20101206_AAAJAZ baglivio_m_Page_109.txt
89f5dfb1906ed08c42f5abb0922dc810
bef3413b3c57235ccd8ccc7fa25f528b981140d0
68617 F20101206_AAAIVT baglivio_m_Page_139.pro
e53a1556afab0df3defdfafd0efeba9e
4fc03252f511f1e7648eb35851c719f30390b4b0
4253 F20101206_AAAJCD baglivio_m_Page_139.txt
5117c6a1a67dba48655ee39dd1d75691
51454969134cfa68cd47ac86502ed002ad8586c4
1907 F20101206_AAAJBO baglivio_m_Page_124.txt
a0c9fd44bfa529206c6757f64de10145
2f06d9dcd7ddf8c1e6ee5f3a73fad4663810512c
59209 F20101206_AAAIWI baglivio_m_Page_154.pro
efadae14c6028ebb5536a54fe9dbc096
14bfdcd433ad9e6c868f81e2f2096f107d51a42d
76862 F20101206_AAAIVU baglivio_m_Page_140.pro
69c60ee45695c0a8683330b5e7bc1893
85435c45e58e0ab7b3f481aa576a3de2c0adf12f
2241 F20101206_AAAJBP baglivio_m_Page_125.txt
364d48b4fb275709a03fbe82c34c536c
8b559e60c97e0b81273c691ccd6b461416e4430a
56575 F20101206_AAAIWJ baglivio_m_Page_155.pro
8e3736cc7ce93890e01296a2e33ede75
6f391923509892225c5bbac195bc556b3d8e55f4
79936 F20101206_AAAIVV baglivio_m_Page_141.pro
f3f4edcc00b0c7402fa005187bd3be91
3b0c3c4f77108662a27ed0a7141fdd70a4efc3cb
3751 F20101206_AAAJCE baglivio_m_Page_140.txt
ca00f76c0c322e9b5a92df6a4a7a39ef
05f0c3799dcce2b560ff3b76acc1997abb0287dd
1989 F20101206_AAAJBQ baglivio_m_Page_126.txt
6e3a10a7af4e3807a1bc7823e541dc66
4cb38e211046dfb5730e0b27b54c04e3f71c6dd2
62688 F20101206_AAAIWK baglivio_m_Page_156.pro
e10c1765f6bf3c3ac31ca497cab815f0
196814c896efe7f58a20d239b6e95f12c2d88ed6
69702 F20101206_AAAIVW baglivio_m_Page_142.pro
7fce1d1791f6a8497a1128d4dff37d8a
8243ad69da515e01e0a8d99dd1525fa9876e6d01
3683 F20101206_AAAJCF baglivio_m_Page_141.txt
f8b3363c1063f61e1f727d7b9c82183b
76f0a4d0e1b4e81dfc844c55a5b0463c0e8b49d1
1735 F20101206_AAAJBR baglivio_m_Page_127.txt
fdb95cba89b8db01a57c9e06e9701b85
44457bd8492b473fbd13c025e10bbd0983e1cdf2
60360 F20101206_AAAIWL baglivio_m_Page_157.pro
b40ecd36693efd4fd93422bea661503c
95e267a921e2af7288a4372d158f079e6a7c20fb
72827 F20101206_AAAIVX baglivio_m_Page_143.pro
b6dc34f2cc706db831a95d5515a54669
b71edb9df55aa821dd6cf4449c8b524bd6fa5b4f
3913 F20101206_AAAJCG baglivio_m_Page_142.txt
f80be703b5435b430f92aaffca9154ae
4a6d8067b31e61df07615e72958766c964fa2e4a
942 F20101206_AAAJBS baglivio_m_Page_128.txt
aaa0a670dfa86c1aed0769ce483a8fdd
661a72fcc501912290c9b09a714ecaed272f2457
59519 F20101206_AAAIWM baglivio_m_Page_158.pro
7c83b00733500f5816ca1aa24cdcf7b2
3deede76d57745d6f2a4dade646ed83cec586622
3245 F20101206_AAAIXA baglivio_m_Page_005.txt
439ff90d4b022d75ec1cd84c7b593f5c
3ee9dfc099e1ce310ff4382ed20bccb2f058202c
4492 F20101206_AAAJCH baglivio_m_Page_143.txt
9ac4f463aa4222a359751a9c407fc683
2ac8532a4f64dc55764cd33a40c54c3af97f75dc
4685 F20101206_AAAJBT baglivio_m_Page_129.txt
a5cef522543046cc02ab6d5971ab9404
f29052fafe34fcb998808105608b2de868c94a4c
60323 F20101206_AAAIWN baglivio_m_Page_159.pro
c7fedf10a07afac6afdbe64d4e355192
ddece0dc3475c43756071a158f1b23b7d0662509
76630 F20101206_AAAIVY baglivio_m_Page_144.pro
bee3b228dc2c27a8e299be4491743108
cfedf5b1fe9f99005e3cf8fe6379d523745b530c
396 F20101206_AAAIXB baglivio_m_Page_006.txt
3d20b174c5e3f1798b416202f752dce3
dba749d45ba828e439b0161a00a8d78066731010
4505 F20101206_AAAJCI baglivio_m_Page_144.txt
708e163176acf099b150d4e5ea187c45
66b17153363a76baa37f51fe4ddbade8a0d01d30
4793 F20101206_AAAJBU baglivio_m_Page_130.txt
cc1822e608e28394a7f9f8cd7e7b8f50
a643794208ed0ef42457aad98514f21f3ce01c45
61919 F20101206_AAAIWO baglivio_m_Page_160.pro
e4863e16ad781b957de2994b02522bc9
a08538daed63c33759c98a61982772652e980a5e
64769 F20101206_AAAIVZ baglivio_m_Page_145.pro
48c133c4d700d712b227dace7dac0621
0ea37271d097673897b6a037475ec4c6196cb1de
2757 F20101206_AAAIXC baglivio_m_Page_007.txt
94bf615cf50de5169f3f7083be9c7cbd
4261b8c2c38e9c18bd262c926c6473310515c36f
3668 F20101206_AAAJCJ baglivio_m_Page_145.txt
7d6b24a00ced35809b3955e9860e48bf
1a59cfa0f2c79ee56053552a55aa9c44fde6f3a5
4576 F20101206_AAAJBV baglivio_m_Page_131.txt
e92d2616202bd2168dc61a3d611d7a4f
3789bfe8a92cfdf65435e3f3f231514bc15e4886
64263 F20101206_AAAIWP baglivio_m_Page_161.pro
1cecb7c27d5bb2be1777e233094d1bdf
20c8c6dd7926ca55ec3333a571c7c4def59803a5
1261 F20101206_AAAIXD baglivio_m_Page_008.txt
f9ad985219fdf1ddb776922245f9b578
6c62c96259380ae62a8fc169e2fe8515fded1853
4226 F20101206_AAAJCK baglivio_m_Page_146.txt
c1a1f179c8ea84eb41c1375587d91c7d
98b3025e87096d9b18457a598de5d9ea9a9f841c
4864 F20101206_AAAJBW baglivio_m_Page_132.txt
151e99e9861b3259fa8f01b251064061
e9be70324d85709d183e3c8cf1836e3be82cec44
60496 F20101206_AAAIWQ baglivio_m_Page_162.pro
fecb65a581862aa609253198a859375a
7a3404975402504d443a2c94915ef43ebe25d552
378 F20101206_AAAIXE baglivio_m_Page_009.txt
70c19006ff13118dc2b74d436781b41f
384fb91b0bc848266a8cf813524ea69e2768bae4
2469 F20101206_AAAJDA baglivio_m_Page_162.txt
3b8f26ceaf17898b8179cd3a3d31012e
72f498b0e62601e36f41776cbba96be04141d21c
4077 F20101206_AAAJCL baglivio_m_Page_147.txt
93606b300abee6463ed811f7dc2c077c
39922e8a5361fc120a09bc6dab30d10e06122a1a
3182 F20101206_AAAJBX baglivio_m_Page_133.txt
b68cf90c9343fa937f4202a73713ee28
055099cb371e20556461fd7d115482e6b7822a26
1539 F20101206_AAAIXF baglivio_m_Page_010.txt
15802d9e48d4c37cbd349b2098bd92fa
c52316898984dfb8b14058d607bbfcd3815802ed
62916 F20101206_AAAIWR baglivio_m_Page_163.pro
966f09925deca2fdd632735c2c53c32b
c94f09e782bf8a4efcf8b7849f4cc4e994739055
2550 F20101206_AAAJDB baglivio_m_Page_163.txt
20a6e6611ff7d9f7508ecf88871ade0b
d02b317dbf87d2ca49514166701f01b991653283
3815 F20101206_AAAJCM baglivio_m_Page_148.txt
861909147079f099abe34f0e2309b645
776ba2c9f1656f63270a0b0fb5c347e358dcef33
2951 F20101206_AAAJBY baglivio_m_Page_134.txt
2448cfafaa12e7485c44eb76a7d56874
79318a1471a5c40404da63e842481087fc9f1f74
2233 F20101206_AAAIXG baglivio_m_Page_011.txt
1de9166489f74e4046402c38b4f9b9d8
bdaafae41727897a6828090b49a1e5f054f58bb1
58904 F20101206_AAAIWS baglivio_m_Page_164.pro
6979efd25d07634af103cec730afb8f3
1949890f54cf278a5ce90cf3d41e2a1b56f51f1d
2408 F20101206_AAAJDC baglivio_m_Page_164.txt
367250d515b07f8bb4d0e08945bfe159
7816d481238a82ebdd63af0753d0694539ea803a
4148 F20101206_AAAJCN baglivio_m_Page_149.txt
5b055961244d08c0ea93a68d51976932
c14cf45948a5c99142acf0c420669ec995c24d79
4094 F20101206_AAAJBZ baglivio_m_Page_135.txt
62f35cf7edcd18db0a20d5ae362484b6
6073d507a0e7f9e32b6d9466fb48cad2c75da2e2
89392 F20101206_AAAIAA baglivio_m_Page_062.jpg
f4fc0253a4aba1481e99d1c446a0c04b
2a58c6c953eda9b6efb02943a19acaacb7a735f7
2023 F20101206_AAAIXH baglivio_m_Page_012.txt
2a49879e6a48ceda057a90e038bb8f19
46cf96e124a4863f793d1479925afedfe46ae57d
59070 F20101206_AAAIWT baglivio_m_Page_165.pro
f988075997f8ffe6c9e0ddd80b136755
6e6e08c41ac5f6c37ac4034e0848f2265402dcb7
2411 F20101206_AAAJDD baglivio_m_Page_165.txt
7d80e77f0f885e5969bb84be777b0c52
87e102a824ee8b7248855b90ef229c3bffb0edd6
3351 F20101206_AAAJCO baglivio_m_Page_150.txt
5e2f86570117f6c5b03f7050475d1868
07e294757952559f7cc1f57e3230614901657ab7
90713 F20101206_AAAIAB baglivio_m_Page_063.jpg
040dbbf167ef0f7495380f89b0b02823
9068a07fce6371c9a1444522cfdab072988929e8
2130 F20101206_AAAIXI baglivio_m_Page_013.txt
d848c42498cfab8ba1f2be0bd43f29a5
bed15f38268c5b18fe5f8697a9709f98678d8d98
27878 F20101206_AAAIWU baglivio_m_Page_166.pro
5e0026b819fcc998fbfa6d5bc4e6ce52
ce216b5ca0c3ad71caae1d6b2823ddb358e478dc
1150 F20101206_AAAJDE baglivio_m_Page_166.txt
1decdb2f1a735b1ada2af0e3a9c1289d
397a348164dfe54d1528f35fe0b5184556647eb9
4696 F20101206_AAAJCP baglivio_m_Page_151.txt
8baea570d22f40e8d2ba88f81fc13d9f
91652319ae8920ab5dc9d1a8a3de2132fac66b4f
93353 F20101206_AAAIAC baglivio_m_Page_064.jpg
71ecd7c664f5e9417105f40e3ae24f0a
2cc61758194ad26b61de38c9e22b0310cc056001
1960 F20101206_AAAIXJ baglivio_m_Page_014.txt
1aeebb2037acbae6a7fb648c594ea750
bf382361597fa2370d4aa46a3b2cec282191fc18
17933 F20101206_AAAIWV baglivio_m_Page_167.pro
647177137785a3ac41eb04ff5309f3f4
7b5cf63b0bdede53ad351ad019da41740c029b76
478 F20101206_AAAJCQ baglivio_m_Page_152.txt
e1a638408c8c15682a781d59c306b62a
535de6b2d0e986475735e7d465ee5bd79c4bdb4b
1592 F20101206_AAAIXK baglivio_m_Page_015.txt
55587a79c42da1cc3843d6d1a3fc4857
b89d326a787be3eb9b01f3a1996d663ff6ccf635
485 F20101206_AAAIWW baglivio_m_Page_001.txt
69784ee1a2563613d7bb972d581b411a
d5bd7eea8984f1d96b726ea54643f5cf2151842b
753 F20101206_AAAJDF baglivio_m_Page_167.txt
f349ee7ee4e95d46f98bf802bcae3eac
3e30f618cd88465fe68fc55c2229b0a0555b5090
2635 F20101206_AAAJCR baglivio_m_Page_153.txt
37d19b441c4c0ea1a08558ddf514b7d5
16e987e8dd44122eef250547032964a9196ad80c
2096 F20101206_AAAIXL baglivio_m_Page_016.txt
cd9de8a7da4495c840ea2d665f4377af
a679992ae53ff95edfcd30b8fef6b6115e762448
93 F20101206_AAAIWX baglivio_m_Page_002.txt
f9dc4793d2089a40be5b8109181ff3de
d2f77cc44c27ce3a884d3709274d8fcc269a42d6
91515 F20101206_AAAIAD baglivio_m_Page_065.jpg
9b059b38809c6f06fd72a0d80b07d998
53a678536de04f28acb072079937f1da75314711
1586 F20101206_AAAJDG baglivio_m_Page_001thm.jpg
926212c1f5eb4880d9eb8c2df6b96000
886f66b8ff290be9b23ff58d9da59ab502fa7034
2418 F20101206_AAAJCS baglivio_m_Page_154.txt
da0a3adfc5cef9e7de3237a023954f5f
6f864d09839fde630650e808486b56d020246daf
2831 F20101206_AAAIYA baglivio_m_Page_031.txt
a1adfdf2236191109c71d8bf6393ed73
67d768390090b1f960e77cb0747ac3199a0df41b
2027 F20101206_AAAIXM baglivio_m_Page_017.txt
656a13c362c86f50bf7b1bd18976cc1e
9fc6b51c7d29e6b5cbf93524d38131526c347093
101 F20101206_AAAIWY baglivio_m_Page_003.txt
305098b6a240e705bd7352cda4904a81
1a8e944d0099f7a21109f7e4128b0b84bb2af7ee
93059 F20101206_AAAIAE baglivio_m_Page_066.jpg
0a5568748743b5aee70744390136de90
eed4ecb5cb71d4ca1aefe23dbaef7bc02e0ddfba
481141 F20101206_AAAJDH baglivio_m.pdf
c25b8aaadb394d8b35e261b02fadb2ae
224fe2bfc327c246548bc4ac6a9a975944ccc74a
2330 F20101206_AAAJCT baglivio_m_Page_155.txt
fa9e3cb33724719af16701c5672bbe54
8019e64fff19b30233c59172ea80c6a559819bf4
2109 F20101206_AAAIYB baglivio_m_Page_032.txt
921394655dd8335de983e53f98cd338b
6b9f90fadb8ecdbf5b3275e8ef6630f788563ee0
F20101206_AAAIXN baglivio_m_Page_018.txt
9221db6282d1e63daea0337b9918b0b0
b44f4b3d22c0d2128bbd817f2a5a9de246ada222
82329 F20101206_AAAIAF baglivio_m_Page_067.jpg
99a4ffd7c905cefba8954731c39a859d
cbdb70af9b3ae3c8b9b94a5c49bad76103ca4e4f
27939 F20101206_AAAJDI baglivio_m_Page_162.QC.jpg
87e586098f834a986f8294274f37e23d
116458ed870b471bc8f8515172d27114839c3e4a
2564 F20101206_AAAJCU baglivio_m_Page_156.txt
335decd8784b77bb2fba6b10c7070794
2fd610d96e9bb0f48dbf7c6a4e53362e79065f53
2184 F20101206_AAAIYC baglivio_m_Page_033.txt
da9b619fc70e40557f819d41136e6990
698b1146cc9128a7312ebf749ff162407c1280a4
2089 F20101206_AAAIXO baglivio_m_Page_019.txt
d08518531c17c200d358f73693ac92fa
da0d2bb8a7681986f62f06bdbcb945622f304339
343 F20101206_AAAIWZ baglivio_m_Page_004.txt
d86400da709d99af7004b9541023a9b8
9405ffb5ec106f434013324b091a35acb533bd6b
41621 F20101206_AAAIAG baglivio_m_Page_068.jpg
c0ff60e1c076117691183343347375e1
ea11a4a07635074a29e535d42446617e48e5fd6c
27095 F20101206_AAAJDJ baglivio_m_Page_054.QC.jpg
318ff2402296cfcf16fa2ed4625cc5af
a9699d35e9d490dfed0ceb3521af4322974eeead
2461 F20101206_AAAJCV baglivio_m_Page_157.txt
f547311262a52f84237060b0aded1010
5c3053dbd2b0fdfded90c64e7595d3c25f1edbe6
2140 F20101206_AAAIYD baglivio_m_Page_034.txt
9888c63db373de2b7177569cf098de45
60c14fdcd21c62814a4da55266a2470960304167
2069 F20101206_AAAIXP baglivio_m_Page_020.txt
fe1342e04d070b0b1413f834b60ccd4f
2d8d7726c389a7d29d49297721f2d44a9d75ee5b
80000 F20101206_AAAIAH baglivio_m_Page_069.jpg
22d6658f169234f4ae1aab806b378e7f
438f22e387c0a785ee8cf2e160545dcf8dcf2adf
29081 F20101206_AAAJDK baglivio_m_Page_118.QC.jpg
e6a1cd0500e4ed0de5d7e21b6fb9ee57
36b32d1973f5ffbe7c50180f439ce5aad3c877a3
2439 F20101206_AAAJCW baglivio_m_Page_158.txt
f5a515dd28f2dcd5950c14745208584e
c0e9296619f811bafda098f981a631760d3a71e6
2148 F20101206_AAAIYE baglivio_m_Page_035.txt
e0c6c33954b932640f7fd942f3fb0f5e
18793d2a59717bf81c4a827069a599b32a6390da
2206 F20101206_AAAIXQ baglivio_m_Page_021.txt
d186abd7ed556b5033d38e4689097357
ab429f3d0b496d72055af20e77585f0c8c46e981
84192 F20101206_AAAIAI baglivio_m_Page_070.jpg
cb80b405d5de026ac0b2ec9241df20be
87f6efc13c4f314f94b860fcfb89500809225c1f
6806 F20101206_AAAJEA baglivio_m_Page_044thm.jpg
7c73473673323620573c350a40e15454
1a426ff21bab1730abe3b34d67ee91b1a60079a7
26679 F20101206_AAAJDL baglivio_m_Page_012.QC.jpg
1bfdfde88ae8b21674ec91fe975ff15d
84c3e6b6ef514a90a3fa3b223800cd01805e3159
2474 F20101206_AAAJCX baglivio_m_Page_159.txt
44bc65d5eeae05d2c0b836b1ee848248
5891d2950224d92cfc7a9831d0c026ce13440f79
2177 F20101206_AAAIYF baglivio_m_Page_036.txt
a5cee89a592b6c0afe3a6b0f3bcd83ee
12b7d2b562b351031e7f843939e1e44ec491c137
1337 F20101206_AAAIXR baglivio_m_Page_022.txt
d2b32c7d6af5f68dddc77badd3fe53f2
72575027997310d9d2b9c8bba91c9950acba0d58
86681 F20101206_AAAIAJ baglivio_m_Page_071.jpg
8e7904a6a120efa2ac2265000e89bc94
d23c5ea3fd97c8db1908e93df79345d89cd691c6
6503 F20101206_AAAJEB baglivio_m_Page_075thm.jpg
bdd4d8cd63d7d6c8f56a99c8c9e6f91a
29f90e8ef4221172419a918227877f47093ab24e
23535 F20101206_AAAJDM baglivio_m_Page_072.QC.jpg
9b218ab9152185c44a89d1c1e401a43a
baeab5439d7c97ddc52d3dfc53c9d3ffa7ea6aea
2527 F20101206_AAAJCY baglivio_m_Page_160.txt
e44d3bc3ca43cac584b14eb3101bd47d
1cfcfab598a8d8420c2a13a095103a3949bd136a
2065 F20101206_AAAIYG baglivio_m_Page_037.txt
bdfabce1fbb4e347f819ba310abd1915
bcbccab871ea782adf31420b0b1e69f2d8f2f15c
F20101206_AAAIXS baglivio_m_Page_023.txt
54c1bec63d823cab96ab1bea35d41a2e
4cb184ca30809e61cfc68268e6352d8e2562935d
78574 F20101206_AAAIAK baglivio_m_Page_072.jpg
79ef362b366d9c28d2bfa4ee00f4c30a
449913cd1fe7533b0520c156a54318f4856012d0
29137 F20101206_AAAJEC baglivio_m_Page_047.QC.jpg
64cf595795c84636cc324d52538b5c6c
9a161722d878f0142187b595c37f03dcf871b886
6935 F20101206_AAAJDN baglivio_m_Page_026thm.jpg
cb6021393a75c16d6eb92ee218aa8c05
1e071274a61c88bf77b899c3555812c3797628d6
2639 F20101206_AAAJCZ baglivio_m_Page_161.txt
4365c3b0430600094a009d26c1510901
2f71706041c184f72b231d8a1c720e03187d7027
1973 F20101206_AAAIYH baglivio_m_Page_038.txt
09d51eff080f757281076f407fa351b6
6f3916d4a2e907f56536a9fb4b964f3ccffa62ab
2197 F20101206_AAAIXT baglivio_m_Page_024.txt
fa76f0be486b45b7c373e13116d9dc62
10ad4e3c0476948d81c05ff312a9b9235bd564f2
78799 F20101206_AAAIBA baglivio_m_Page_089.jpg
c8d64fb3546f8adfe43ae5c403e7077c
e360e7c7eb684fc80b6801bf38ba5923de7a4b0d
88713 F20101206_AAAIAL baglivio_m_Page_073.jpg
ff3aaf4f241a341c169fb456aaca8154
d4f20721a865744a261f996e03cdae41227aa376
6642 F20101206_AAAJED baglivio_m_Page_065thm.jpg
49b25c225019cedd65dd5a5e25974c6f
ca7132d24731047f6d672faf94cdc45b258dbbea
1112 F20101206_AAAJDO baglivio_m_Page_009thm.jpg
1bf3d7952fcaa1a92b1651aae343ec19
d6fda123e8bfd075b5f6e90621abe6a7afd1f29e
F20101206_AAAIYI baglivio_m_Page_039.txt
b118a99da83d35cc7d4fe64855cb0bd2
62ba18654993040476ff14882ff47535ce0b386e
2223 F20101206_AAAIXU baglivio_m_Page_025.txt
abb5e67c15b7ff4f452b6d1a8ad4f3b3
2fa5cd0af8949f71d59710f587ad91a744776e60
90523 F20101206_AAAIBB baglivio_m_Page_090.jpg
f18b370969eaf9b833ed3a9b158a5cc9
284828ead9e7fb5151e7e453f707b044221e42c7
92147 F20101206_AAAIAM baglivio_m_Page_074.jpg
681f7f26a4067b9eba61f50e462a8d14
602f57e4fe15c3e31be83f6db4f6803b6d1566c6
4681 F20101206_AAAJEE baglivio_m_Page_015thm.jpg
3e957d7b253ca09061d98a690c871288
ba7d75c5251728f978d2ef9d0b79798dbbae6bca
30573 F20101206_AAAJDP baglivio_m_Page_149.QC.jpg
5182a56ad8d220467b7fe8afeb4f8a8b
88d8a0afa7790f1cefbf44fa6ea81484a6f7ab7f
2102 F20101206_AAAIYJ baglivio_m_Page_040.txt
60fb0e56568821bd935d5ea5709f66f6
909d7f5c1a2857fe131aa4baf9384638174230dc
2603 F20101206_AAAIXV baglivio_m_Page_026.txt
dc99abc95c16600616315b9a7b2629d8
44ea53dc6d2fef049dd3cb4c39b89f8bcc26c906
90386 F20101206_AAAIBC baglivio_m_Page_091.jpg
6b060f68412a596f9fca8229fd5747ec
745a372646ec6dddf40d27ede5df0b0879496eec
87449 F20101206_AAAIAN baglivio_m_Page_075.jpg
cf1ea505a83c7d7b2584f028a3d5e4e8
9faa357d20af44f4fd4825372747a9e99fa3c005
27077 F20101206_AAAJEF baglivio_m_Page_074.QC.jpg
86b79a9e563515d63e0ef5cca8d493f5
786616635e8c89c0de01f63c866664df9a017321
976 F20101206_AAAJDQ baglivio_m_Page_006thm.jpg
56a4608a74e5a56ad089b55925e9fd50
642d42210790ca32e111ee36ca4b99d9ed89846a
2165 F20101206_AAAIYK baglivio_m_Page_041.txt
a2505751d276caf9aae9fa1d818f5e15
177aaa79f06ea994f6ecfc178586cb5c1f28d0ea
2283 F20101206_AAAIXW baglivio_m_Page_027.txt
3c77a475f96ad9d89e385421d98484a4
000aaff1f0db24871439bc28dd0debe94ec58d04
85372 F20101206_AAAIBD baglivio_m_Page_093.jpg
e5f7d59681098caa64e3f16468ed6a9c
7444ad6e7d4fa73590b63ac59064210ccdea4212
93718 F20101206_AAAIAO baglivio_m_Page_076.jpg
e3b724cefb6925207267c8b6b5ebe66b
d4dd55c22a1b300104fcb7499bc05d034bebaf9d
3394 F20101206_AAAJDR baglivio_m_Page_068thm.jpg
35f4f8d0fd04a7fee2541f424b68dab5
13d2f527911de8ba1b55183c7062381f82946f9b
2756 F20101206_AAAIYL baglivio_m_Page_042.txt
696ff0688d0817f1dc0b4bf29ee70b52
a3da62518f51d292e76f1374d429015f9c9b0700
2214 F20101206_AAAIXX baglivio_m_Page_028.txt
4a916af6089033e211d73b3877d76dea
03012b7ea7f5daac528afaa765ecec493661112b
90095 F20101206_AAAIAP baglivio_m_Page_077.jpg
cf796893646abc29be75fee693dde5ae
eee756284f53c95aa79ef80b88f2747fb2551ffc
6832 F20101206_AAAJEG baglivio_m_Page_155thm.jpg
190f70afca1f1d38ee49596b70c214b7
4be3f6699b434bd5fb4cb3c5596ba1a8bb7bac77
6958 F20101206_AAAJDS baglivio_m_Page_159thm.jpg
231fca7c32e11de66cd0e19f5992d2bb
9ff081898d507b0843dcb349bf5be9032d08aa7b
2107 F20101206_AAAIYM baglivio_m_Page_043.txt
063ce62624f72aba607b7cc02c03c8da
9f39d7e9ba77637b4073941fde15a1d95fce57d7
2203 F20101206_AAAIXY baglivio_m_Page_029.txt
69e69507899de4534424da7e3efdb63d
3b9fcdbe4333f3f315c9a76048bf33e41848a8e2
88610 F20101206_AAAIBE baglivio_m_Page_094.jpg
32586f43228eab08da2e2b8278a56b08
dd5b13a924152f9298be7dab96a4a0b72002813a
93365 F20101206_AAAIAQ baglivio_m_Page_078.jpg
971a3b687935405b97c1233b7dbd7aa7
c6d981b9ed5799042212e447d7a6caea7ca3c447
F20101206_AAAIZA baglivio_m_Page_057.txt
5d4c9e9c76b03dd85fd29626036dad99
d203eeba3e3f113a062d6f421d3a3ded3d39ddd3
27517 F20101206_AAAJEH baglivio_m_Page_050.QC.jpg
136f1a881b089ad5a2f7d6cdfd38be9a
168b0192beffb562d7673d574c827d3d517a308c
6554 F20101206_AAAJDT baglivio_m_Page_011thm.jpg
5e810f4614fec1e993b6c8fdc91f209e
9a1a86a567f6504442a797dcfd8b77de01cf302c
2194 F20101206_AAAIYN baglivio_m_Page_044.txt
da48b4508b196d44349192416bafdc22
0b35ea47dfaf053d1b2778594a79f7da529eba71
2116 F20101206_AAAIXZ baglivio_m_Page_030.txt
e1b914d7553baa34541f3b3c5382fe0b
93c45e0f1d1513acf4666f5f9ecf8deb838d5dc6
85741 F20101206_AAAIBF baglivio_m_Page_095.jpg
825956782dcf2ccec8f1beabf575aa13
aef5ca67afee6184a04fa12a5e6e4c9565e6bc38
87645 F20101206_AAAIAR baglivio_m_Page_079.jpg
e252a6aa0f77a0fc6dc48a3d86328a90
9f4b87228381ac6ca0011ba42d9a2300b496edc3
2139 F20101206_AAAIZB baglivio_m_Page_058.txt
3a05598429e94cdf1914fd207c4500b2
7ef31cb17a50189640c0d8556f187999f2c7d422
29596 F20101206_AAAJEI baglivio_m_Page_163.QC.jpg
fa13c17f315f1ebadbaf1e091071fd06
9e726c6624a62d18e9d25590bd35e40759c152c0
6555 F20101206_AAAJDU baglivio_m_Page_120thm.jpg
6ec73d13d733864aa16c059d4167b1cc
f4a68f9bf638f79b89bc6c5c96525d0fb3e61fa7
2135 F20101206_AAAIYO baglivio_m_Page_045.txt
96253c3f3c5de63d6c68bdbd3536c94d
d37d629054cb90d8b417a82958f454087426ebb5
80488 F20101206_AAAIBG baglivio_m_Page_096.jpg
330c06cf06458239e3a5d59b686c6e27
c84c9f8cb76a9f1b152e80deeb123849488a45d7
83905 F20101206_AAAIAS baglivio_m_Page_080.jpg
331d599e9f6dbfbf997f3fffdea1ba5a
e7e14f9bf08894a6237fd2473e5f402d3a078242
2204 F20101206_AAAIZC baglivio_m_Page_059.txt
8577c669a373f1af24509c9f60efb15f
1ed0952b6684e29e790bc67b2aa6c4726c90d273
28506 F20101206_AAAJEJ baglivio_m_Page_090.QC.jpg
cfb060f62988192c285153b3af8312c2
37deea9bd902a9cd1f7ad788506c99aaff8f91d7
10145 F20101206_AAAJDV baglivio_m_Page_167.QC.jpg
adf83f5c18a6c2af7b0d6d4a02daab0b
f86da9d418b9667398e2629493712e7c56bf1a02
2164 F20101206_AAAIYP baglivio_m_Page_046.txt
bf2182f6b6c30c2c9fbf4c9fdc10c7f8
f2ccc77502c69f929acd4d6500de27677c2e8a23
74598 F20101206_AAAIBH baglivio_m_Page_097.jpg
c252d4b6d69eca5f02289638d5519465
0e4042be78d60437a8b725aca6d9091d46a8ffc5
91078 F20101206_AAAIAT baglivio_m_Page_081.jpg
5f2341445a221693a20b9cb4e806c67b
4ae27de6422555928f43223a182d3ff92d580d95
2127 F20101206_AAAIZD baglivio_m_Page_060.txt
32ba56772f6f54503fe71f567a99e16c
3f382303abc4e74209a1d514c52c414c3beb0d84
6842 F20101206_AAAJEK baglivio_m_Page_160thm.jpg
5dd5d8641e9d321cee3ea6f46e7948de
fef37717f87acc3a693fc0e0e307a9f2af4ee297
28928 F20101206_AAAJDW baglivio_m_Page_116.QC.jpg
5c2435db6949b4296d4eab49ce266960
3fc9f82f097976b9042e0afeb42efa797b39839a
F20101206_AAAIYQ baglivio_m_Page_047.txt
712c40fa84c05e6a67cca1478bf533d1
3fc0197aac7fbece878cd57a22a92f92e0081a6a
87878 F20101206_AAAIBI baglivio_m_Page_098.jpg
09498bb9409de506273b5995984d70df
856be7ae77369bafc14200170054f2cac449d924
89111 F20101206_AAAIAU baglivio_m_Page_082.jpg
71a1e4b1d2444d3f35db544ac98d3958
d5eddf0c0f5f9990fef796cb51fe74d0df4d07e9
F20101206_AAAIZE baglivio_m_Page_061.txt
390d71d0b36a6ea858cf6fc4637682fd
90e7086295955baa7f5a37a56e69368302cbf80b
6531 F20101206_AAAJFA baglivio_m_Page_106thm.jpg
c793d82ecfb495d96519313bf53a4bde
af3b4d6eb4adb301f90ff971d5c7b04b4e7793de
6587 F20101206_AAAJEL baglivio_m_Page_112thm.jpg
ddfa5ba5d2a6d28f70cfb152b5d7a861
701489d707799497451ea06929bb94c91e63277a
7594 F20101206_AAAJDX baglivio_m_Page_136.QC.jpg
753e84d88f4f662f94ac790867e979d8
c49ced58e5b81c0f8fe763aaeda773ecbe1558b5
2172 F20101206_AAAIYR baglivio_m_Page_048.txt
eeba0569e39c5a717024451851d04adf
92fb63759ce9c1d9d6a625cb4b9a43cd6bb5dfe1
81544 F20101206_AAAIBJ baglivio_m_Page_099.jpg
683ef1e8c2fe3c28b408d33c9e541900
f05ef0600f49aa572d417dbeb24efb96a9e86b6a
89491 F20101206_AAAIAV baglivio_m_Page_084.jpg
3222b5c7a86746735abeff3be77b87a7
48e5a8be4f14a6fa10cb3184868023e84221b7e9
F20101206_AAAIZF baglivio_m_Page_062.txt
e05c401e3582af7cf5c7fd18a2945224
dd6bc3d343f65f59950d4abbcce270b3c920e9c5
6438 F20101206_AAAJFB baglivio_m_Page_113thm.jpg
f1258151063dc9243693ec1be78c5b02
59998cae9cdfa9eaffc02aebe6db2b2476f45aba
6347 F20101206_AAAJEM baglivio_m_Page_105thm.jpg
6eb97e13ebc537bbdbaf980b54f13448
e6764d8c2818653d276aced295014956234a85ab
30039 F20101206_AAAJDY baglivio_m_Page_141.QC.jpg
7a246ec0896e2e91dc466925745bb932
13a1547af603bc3ac8e87c7a5c25ee3678d552c0
2091 F20101206_AAAIYS baglivio_m_Page_049.txt
7030584a5e77ac231b8b72f93e523546
c4ba6712ac316f7b72b7711919ed2539cbadf15b
92399 F20101206_AAAIBK baglivio_m_Page_100.jpg
08d6579fb191c20ee617077048999005
c4bc9d74186843c8ab38ba491474373366189bc0
81849 F20101206_AAAIAW baglivio_m_Page_085.jpg
9f445e8cde099ce460d03531e26c1e98
75c5b22e1827fd127fe26d1c394e79f07fea15fc
2199 F20101206_AAAIZG baglivio_m_Page_063.txt
16386c019da34867ef20f75bbc94997f
27f2ebfaecc079ba4c0e98067900502449e94d6c
6452 F20101206_AAAJFC baglivio_m_Page_037thm.jpg
cf3db7dacc83ce668d08d86c4ea23a3b
76b7238c0d130e6f743e5c52e1c32f88f142c276
6676 F20101206_AAAJEN baglivio_m_Page_087thm.jpg
65c64fa201eaad9081a29fb3fb37300f
d03ad84dcc4a0f6950e9d72ecc46ca86d1a00f41
4673 F20101206_AAAJDZ baglivio_m_Page_133thm.jpg
ff16588aafadb360f35d102a7b75d236
4e47b7c9357dc49ea7f73e5af312d85eeb8cadaa
2155 F20101206_AAAIYT baglivio_m_Page_050.txt
6814dc61ce029db680b58bfef2c11bb4
7bc0a5a193d35b3870bb366689ec115149371ac2
91701 F20101206_AAAICA baglivio_m_Page_116.jpg
569141847bef4ba348a83e0923cc17a0
a5cad006380345f22e4eac276f62b9283ff446ba
93656 F20101206_AAAIBL baglivio_m_Page_101.jpg
e824bac755f7d68f66e65f4db2572f57
ce956813be7224227e18dfdd415b7d626ce05c77
89416 F20101206_AAAIAX baglivio_m_Page_086.jpg
14e52cc31d29ae5dab3656bc88e71aa9
9c9e89536351138997d80fb2256a4eef1020eaa2
2256 F20101206_AAAIZH baglivio_m_Page_064.txt
2fd5028b3f2b35930f11fc17fe51f1ff
e2774f44e6ed3e87c6bec50b2f218c1054581b7d
6596 F20101206_AAAJFD baglivio_m_Page_162thm.jpg
4de46de286c494d4a75a64c394bb6a05
9fa6645321ac1374a750bb8f7df0e7f58ad2f869
6874 F20101206_AAAJEO baglivio_m_Page_034thm.jpg
68b266fd1a2791b581e7818a750419d2
9b47daed29d8c44d2374a2e44f119af00b703d68
2144 F20101206_AAAIYU baglivio_m_Page_051.txt
98567866364d1e3f04de491a5ab80724
1d8c4e051858565eb8505f2579ab29d64d60936e
89365 F20101206_AAAICB baglivio_m_Page_117.jpg
b3fff4b3981fa95fe266c0d856fdfd22
29728eb490fde23513f441ea0d980b5428c712f1
79239 F20101206_AAAIBM baglivio_m_Page_102.jpg
b556f8e73e5bbb49b3ace9deefc1e4b7
482250a21bf626e69831f5c7102fbf1997ef3f87
87574 F20101206_AAAIAY baglivio_m_Page_087.jpg
354301abd9d202ebb2a4a16987d83b69
c234090c38b0efaabc867314323fa9fec8df5a94
F20101206_AAAIZI baglivio_m_Page_066.txt
bc3ed40258b8eec05401b1762b608fd0
03da116c8543c7a59fd724e28a6762d318659659
5388 F20101206_AAAJFE baglivio_m_Page_134thm.jpg
5981de8a8142ef17669ab90596cfb7cc
92dfef484c0da863ab936ddaf32854b6bf012247
27342 F20101206_AAAJEP baglivio_m_Page_073.QC.jpg
f4118c37609402f2ea2c21ed36884dc1
94860bf441563bd48c73e3c2e317ddcc3c8b2f0d
F20101206_AAAIYV baglivio_m_Page_052.txt
5d27fe641d6a15415899f6a122cc1853
c443738894575f94847b4f8b2777d24bbb2f7308
91956 F20101206_AAAICC baglivio_m_Page_118.jpg
0482ee1100a2ae0b8a275a94df68b2f8
5cef34c181b8b1726429f80dba020b6cccf763ab
87584 F20101206_AAAIBN baglivio_m_Page_103.jpg
bee5c58788423495d11dc97067ea6c1e
b6a36818ba752df4c0e2e956b52aa1c53feb34c1
91292 F20101206_AAAIAZ baglivio_m_Page_088.jpg
f8ebd10bdbbcf439317f11fd7e54079a
ba3a8246af74a62273f99ae4b8895693c716bf49
1986 F20101206_AAAIZJ baglivio_m_Page_067.txt
bc5327b720e9eb330c17903b6901f6dc
f2b61b9b0573e6d66ea86ca1b6ec2b317d1e1893
6570 F20101206_AAAJFF baglivio_m_Page_084thm.jpg
d4d0e8fe51b19ed740b569384b9693c9
6785852b4cd7a181d4c7c38a4bf0beeddd168c92
4465 F20101206_AAAJEQ baglivio_m_Page_005thm.jpg
73748a4b7dcc5d263b7b1528d1a4661c
612e864ea75806bd2f20ad874fe24e30f9b39f80
F20101206_AAAIYW baglivio_m_Page_053.txt
567c868e4b39ddd8deedb2b311936ef1
3f5f5484af8470c7404040836656b18ad8a7f594
54193 F20101206_AAAICD baglivio_m_Page_119.jpg
248762050f8249b13593d30b666275f8
c90fb263b85f79974bc22b86f8ce2d145db92200
90455 F20101206_AAAIBO baglivio_m_Page_104.jpg
4bf40a4a608b43b65d64668dcc064bf4
b01d9c359688de0a949b8f28922247afe0c1e043
1074 F20101206_AAAIZK baglivio_m_Page_068.txt
7b5af1c760f7800e2103504147cb389b
7af48f6f5e8814576bb3ef9cc375ea85bb26a8f9
251523 F20101206_AAAJFG UFE0021569_00001.xml FULL
0ec5eadb35ab7d8a511d69384939f68f
592e7fe17566f594c6007bc2f3b0bcc8d80cc756
6335 F20101206_AAAJER baglivio_m_Page_124thm.jpg
c0c9570a8f7b4b439cd312dd7b00c0f3
8b24e59d1ef919bfd294d2553010b2d9dbdd86f4
F20101206_AAAIYX baglivio_m_Page_054.txt
d074db2cd95fa4b58d7a8f97abab062d
ba834505d92d541ba895c07775955a6fa798d936
87976 F20101206_AAAICE baglivio_m_Page_120.jpg
b40af6b748faffeca0225415764ee674
25c94abdc3c55a8cbc797a5a5eed5092eebdb082
88270 F20101206_AAAIBP baglivio_m_Page_105.jpg
7228a0b2100af8ccd4f3f7ee5b374e55
523f956ec3df2a425c1e4a15bbe95c109459416c
1977 F20101206_AAAIZL baglivio_m_Page_069.txt
4b99abbe55199e0f2c67029f4ba01c8d
ce153f4b980d9b3546344bf958a4059241dfd22f
7156 F20101206_AAAJES baglivio_m_Page_163thm.jpg
42babce657e554491cb5162c31a076ed
9e44a543e035a0bdf6063570b3af18b1eb6177dd
2219 F20101206_AAAIYY baglivio_m_Page_055.txt
85104e974c80d1ab822b416c3e908edb
2934a27c56665f3de45111cce4db23ec1e131e47
88374 F20101206_AAAIBQ baglivio_m_Page_106.jpg
b51ed8656fcc89c1fa7fb2926da1003d
270e447f8ccef91fb73ab8ef0d5cbc9b6b64b5f8
2016 F20101206_AAAIZM baglivio_m_Page_070.txt
5584a16d68eeb9641b6a3b0c55d5a12a
360d563ca985c9d8298652f96f6d2f0f3d51fca6
6811 F20101206_AAAJFH baglivio_m_Page_001.QC.jpg
3f69632f6720804bd85d2ff3e6955e4f
715ba65148535cf8b50d5e66fd63efa93c9631de
26395 F20101206_AAAJET baglivio_m_Page_070.QC.jpg
2672daaad7e83766ce8da7364fbece2d
7eafd0cad2405bc808ad325c8f8e8faa03166c4f
F20101206_AAAIYZ baglivio_m_Page_056.txt
f4d7333834ed925e8da013c16fe1f95c
ca74e231169812015cd318500ea5bf52f89fd505
75739 F20101206_AAAICF baglivio_m_Page_121.jpg
2d8605a6229d0c5a5983da389022e454
bc1deccd6f604f4348e8c1245fcf2f31d5664796
85950 F20101206_AAAIBR baglivio_m_Page_107.jpg
740c4498823c10b7651fa8009ae5dd6a
511f5008f75cda71b6cd583828f4042b40abd64b
2083 F20101206_AAAIZN baglivio_m_Page_071.txt
f2451be06195df2a1f97b34b11a58f88
747318348136a9b68124d5ef9c6659fd2c1a07c3
476 F20101206_AAAJFI baglivio_m_Page_002thm.jpg
f53cef3140ee6b0486219b8e917229c6
0ed7b1be908205e4aae4fcca81363e5042c4b1c2
1305 F20101206_AAAJEU baglivio_m_Page_004thm.jpg
d2a9716ddf1fce10c4d8c24f675a339d
52895a446ba4cae08916c995cdce4273169a282d
87821 F20101206_AAAICG baglivio_m_Page_122.jpg
bf28cc52504844f69e02d0359c76b3d3
82499600a99b380dd2ee5931b2bc9f27388b7e0f
91948 F20101206_AAAHXA baglivio_m_Page_092.jpg
924524accd72e5fa0c42725fe06bc488
ed51235dcc6af0a24ae4e6acd03d807c7e78a7b4
84254 F20101206_AAAIBS baglivio_m_Page_108.jpg
458363505ae7f2af5cfad5248ec2ca51
51edc5d42d3bcb6f643b8db36abf6768323e7eb7
F20101206_AAAIZO baglivio_m_Page_072.txt
ed5ec9f9b09f27c8a9fcd560714e7e2d
4179aaf355983cbf1f798684d25f7efc3eacf746
F20101206_AAAJFJ baglivio_m_Page_002.QC.jpg
436a7eb154d93ea962ea053901f5dca1
ce66d52f0d300dcb67d9c5b0ce1fc1364896651d
28958 F20101206_AAAJEV baglivio_m_Page_044.QC.jpg
ee84c476a72b3535e09cd23923aed342
03465df2d76b0c2e954362f5525fde6cf5db5d79
73750 F20101206_AAAICH baglivio_m_Page_123.jpg
62eb525ff93e5a6b51d4df8f01fcf25b
ae064125c953a080e84573b95c760bbbd13dfb85
54007 F20101206_AAAHXB baglivio_m_Page_098.pro
4ab75763af96930e9172735983e65c0e
032a833dafd868bc520b7fc5eb1aa70cd6e4f2d6
92366 F20101206_AAAIBT baglivio_m_Page_109.jpg
02f21f080e05b7be3c8171c4e69fdf19
306ecf2c6c878b28d9f4f7719edf65b78ea2b715
2150 F20101206_AAAIZP baglivio_m_Page_073.txt
3c20c576b9fa9c9f8ca1dc4fd959e2c8
3f70d9732fd26f246de7b3a490c05c8ed1183687
513 F20101206_AAAJFK baglivio_m_Page_003thm.jpg
5195bfa1a9420a5df5bb7b76419cfe3d
87a6715f4b79a6400a36d306d61e97aaacb965c8
5630 F20101206_AAAJEW baglivio_m_Page_072thm.jpg
20a1138bdecfe150ffe820506bca7150
afb7251471f049dad92c6a43309b6646178d2198
75907 F20101206_AAAICI baglivio_m_Page_124.jpg
b5d70b7464a30042cf778db7fbe75159
61c67fc168c0fbe26e606052f4ba548e72598083
1003280 F20101206_AAAHXC baglivio_m_Page_134.jp2
39001027d0d3e84f37016c308a31ff52
757728389872e4e380e7e3877721722be497c377
86415 F20101206_AAAIBU baglivio_m_Page_110.jpg
338dd1178683f29f5ea622b8f3895594
563bf9bc01ce92b080e8248ba8b0c691253b4b2f
2321 F20101206_AAAIZQ baglivio_m_Page_074.txt
e81c14e5c61e36b241d3a9839aa2ff7a
a00b9239083495078b609597da92bfb103f77da3
5832 F20101206_AAAJGA baglivio_m_Page_014thm.jpg
201871d347f90abbf3ffb2050ed4fe48
ef067fd6818adae62690ae11d4ae40b4e5ea4885
1207 F20101206_AAAJFL baglivio_m_Page_003.QC.jpg
7d90063ba98d07e7d863af67745ee997
57445bcdb4c67ef695707ee058a31d07f6488c58
6506 F20101206_AAAJEX baglivio_m_Page_053thm.jpg
aa4639ec5681261bb406a341a33910c0
fa5766af8d87c309a2044dad041be4ad958ed849
89968 F20101206_AAAICJ baglivio_m_Page_125.jpg
e5217f27562e6d215617de9c7f5c883e
aa5574c4528445aec9d54600e049c8c95031fc78
23226 F20101206_AAAHXD baglivio_m_Page_097.QC.jpg
f5020c34efc5c5ca95c62da2e2c2adea
7df9b358f6cf64fbbe72c7c1166e5c0fe3fd3cd7
85831 F20101206_AAAIBV baglivio_m_Page_111.jpg
644efe8276c3557303fc05691a3db61b
df1be72bd5a7ad1759add443534f25b70386a4bc
2099 F20101206_AAAIZR baglivio_m_Page_075.txt
ca9b4535489290d6b99722478ad8dc5b
38be603db5a35cbb3558be56df527a57ad6b6d4e
23322 F20101206_AAAJGB baglivio_m_Page_014.QC.jpg
bd3ae3a630dec5a33febb20a45f8c421
6e2a33cdbf50f6f1644ec7e9ef5e576bf845eb16
4665 F20101206_AAAJFM baglivio_m_Page_004.QC.jpg
62d965200608d717ef3f6ee33733938d
dd75c76593e570165429a03b147f68cbb55e9f34
25690 F20101206_AAAJEY baglivio_m_Page_111.QC.jpg
5d7ec0a5403c2553fc5bb1e16b0f89c9
0f01472f0fedbf46f4ec6f7095f9906b8aa0aefb
83211 F20101206_AAAICK baglivio_m_Page_126.jpg
27afd1ea0bee2b322d419224e7c673e2
e874822cde89e0aad629115eecb40e7b880a331c
F20101206_AAAHXE baglivio_m_Page_151.tif
fd4b087cab299740cfb19c22d0a2c745
8be101f621d701981189bcb33478808f73f9cf22
85580 F20101206_AAAIBW baglivio_m_Page_112.jpg
a2fb3a8c46a9cada49c7df29997ac5f3
a007c2e16060a1eb5cdf9343106f6d8d4dedd449
2252 F20101206_AAAIZS baglivio_m_Page_076.txt
d5851a4dca2a352bea1fb773e8ff3111
f69a32dde478ae5a380dcb6048f67ab99490aa48
19285 F20101206_AAAJGC baglivio_m_Page_015.QC.jpg
17f7ed5c452a649e510a8a7820b02484
006167c4e67a6802a4d30af3cb016253ba4db568
17837 F20101206_AAAJFN baglivio_m_Page_005.QC.jpg
f3a1eeb5e5f78032e0b4725bc04009df
958239896ada191030d0e2d336b584d070258a9a
6423 F20101206_AAAJEZ baglivio_m_Page_032thm.jpg
aa4bb20c35fec35536c49a471db18fcf
06b42e7f726b0b564408a0e7a955a31804dc48a3
79397 F20101206_AAAICL baglivio_m_Page_127.jpg
04b953f441b82d988fb6388628843e10
32f2f5299d03fd7fb6f18f23123aa96f71b1ab13
22923 F20101206_AAAHXF baglivio_m_Page_145.QC.jpg
2e41b54b061989bf267f0961bac59139
3ad99d344b74a16afcc5f1b8973acbf2d2112bab
89652 F20101206_AAAIBX baglivio_m_Page_113.jpg
79c7fcb042a22225a2dad7c79306cbf1
868782c7356666439ff71735c23a607aafe5330a
2166 F20101206_AAAIZT baglivio_m_Page_077.txt
390a4c2bde1b6d4489af465df2fef04e
0db03d7f29707e689f91470ccf3ba6b7b8e45256
92418 F20101206_AAAIDA baglivio_m_Page_143.jpg
e36a87acce48c4e166fe4cd017717100
657f40ca85ce020e3d473d69367c9f2276095de2
6371 F20101206_AAAJGD baglivio_m_Page_016thm.jpg
945c79a4eb8f25ed4219ef215e12f322
11b29d8b062acb7d6a7d769fbc7a72cca50f4cb8
2808 F20101206_AAAJFO baglivio_m_Page_006.QC.jpg
f3539c8db951895f46a79d8ddaa5d364
ccbd8814da4875ea21ca449492fd184024264d94
41737 F20101206_AAAICM baglivio_m_Page_128.jpg
b18adea8b2b9d9486beba86f4cc5f485
50933883d1de39def0fb295549902fb13acefd0e
28777 F20101206_AAAHXG baglivio_m_Page_160.QC.jpg
00137135301206b89936662adef8f706
a84fa897633caaf57054a2d15ace3026759b10a2
86195 F20101206_AAAIBY baglivio_m_Page_114.jpg
247001a00714a844b5245e189d4df7e5
b1b577dfbb9014528c0efee946622ff411bba782
2226 F20101206_AAAIZU baglivio_m_Page_078.txt
d8ba9b730c39f13eaee7fa8a973887dd
621cdecd62702363c4a450305e48f989ecf6e8a4
99802 F20101206_AAAIDB baglivio_m_Page_144.jpg
37df12a007edd341e678821c377cc728
2f5424542e7ae9980ea20df58b5fec63cceeb57f
26000 F20101206_AAAJGE baglivio_m_Page_016.QC.jpg
76ef537d05693187d757def9f23e800b
f50f68d700a868dc3d7d823ef19698be950fa672
5749 F20101206_AAAJFP baglivio_m_Page_007thm.jpg
230df55df8cc67212ac54487a22e8ee8
211456c9e88eee5ffae9e6c7c0f42557989d28eb
101556 F20101206_AAAICN baglivio_m_Page_129.jpg
32caaf262a4213970dd65885b4f48876
462a7d5277066820c6a130159e719cd263e71566
25702 F20101206_AAAHXH baglivio_m_Page_068.pro
855e7d03032e6ab1978ee8302e5ad530
86c97574d21276392f4bc86fa9a107b492d30063
90020 F20101206_AAAIBZ baglivio_m_Page_115.jpg
3e5447894d6fbe8e59a948175993462f
1a3be3165f230d41a2e098bc0b2228b3e79161fb
2146 F20101206_AAAIZV baglivio_m_Page_079.txt
18f26ca28731769f62f0de4d914e82c8
0ace3a4bbcad4b0b4fb9fd2fc4c0df26e3eb5f2f
86383 F20101206_AAAIDC baglivio_m_Page_145.jpg
4aede6c427f9b92b2cd8ac071f62cb59
e8ee05080debd926528105ea3faed290b9ba3431
5799 F20101206_AAAJGF baglivio_m_Page_017thm.jpg
68492fd7702bd0387d68e7f2e672652e
14b0bcfeac14d8278e43e4e49d093788bb675b84
25174 F20101206_AAAJFQ baglivio_m_Page_007.QC.jpg
4acea0ebe973bc5081e8ea89f481aaf8
c1326e003b1057f523f3aebeb9928c9b069fe5fa
103526 F20101206_AAAICO baglivio_m_Page_130.jpg
39090e85404369658b96dbbf13e37b72
7b37c87610a6f2d17f0204d16a5daa2ca0c30ea2
138872 F20101206_AAAHXI baglivio_m_Page_031.jp2
51a39db6aa184bc6529e1e8b478422a5
d0a222712735638acfa3fb953fea66ead035f300
87937 F20101206_AAAHWU baglivio_m_Page_046.jpg
47c6de6289f2a9ccb664f19d43d1b09b
4cdf9bbaba923eee894f6115efb94ac85d49cbd4
1981 F20101206_AAAIZW baglivio_m_Page_080.txt
5a9fd393a9337147f91bc2079f367023
183af7e7397147157a831453ce4d19d2202fadfd
99624 F20101206_AAAIDD baglivio_m_Page_146.jpg
f9008c28ffbb34599e899371907945d4
53a334021d3d6c7f519fc4e01c3197a5c248f3dc
25421 F20101206_AAAJGG baglivio_m_Page_017.QC.jpg
82f9e55b220476d98f1e9115cd03fa59
344ea22e0b56a6e004b5470c3b4c6c0d2074e9b1
3211 F20101206_AAAJFR baglivio_m_Page_008thm.jpg
1e7271b15f9159b384eee03bf7a7a0af
8ebe68325f2f85bbf81fc6c807357e8675ff5a04
76424 F20101206_AAAICP baglivio_m_Page_131.jpg
a4f4f3e5d66469cc6b0f22fe70706f0c
bbb39b5ca1a25717176cd332255ff11dd17fc264
113277 F20101206_AAAHXJ baglivio_m_Page_081.jp2
70b21cdad72000b22bf7dc326de24d5f
341782e548d1eed1d836728cff4711b557e431c1
6241 F20101206_AAAHWV baglivio_m_Page_036thm.jpg
79f1cc545b3a700b51144e12befef8c8
31a6e856ae6a01bb80f9d637aa831e0c4232aa9b
F20101206_AAAIZX baglivio_m_Page_081.txt
68301694e66f3b5e222be6b4580422b3
ef6a0f1122ee535fb44a0c0d59235abae7b81285
88512 F20101206_AAAIDE baglivio_m_Page_147.jpg
79a02165a3316c46cf2ff5e6be3f9b6c
6a8753cf58e976509dc9ee4469224e9057bde4af
6434 F20101206_AAAJGH baglivio_m_Page_018thm.jpg
b7fec9dcde0ae185cff95c261d8616c4
7165d4670542df5990e40236f5c1940d71017be4
13441 F20101206_AAAJFS baglivio_m_Page_008.QC.jpg
bf0fc3c2285913ac9eae2c1877f71fea
edec83cafb13246b6576590cf4b4355aeebcbf94
55132 F20101206_AAAICQ baglivio_m_Page_132.jpg
2d87e4116447b47f28105e14200f85dc
8127d900289c191df419fabef2ce145f4fea356e
6751 F20101206_AAAHXK baglivio_m_Page_042thm.jpg
abee5f0a8d5666c896c332bff1b1f57b
2d9e3bf4defe6e94023e1a57b39183267b2ab90d
56517 F20101206_AAAHWW baglivio_m_Page_082.pro
15e274f4549fd554bedce9abb3c19890
47c1ed29bce49afd90a8e33348c900f3e85430da
2410 F20101206_AAAIZY baglivio_m_Page_082.txt
2ce89e9f5d20edd2628a721122b17720
0df0124679769b8c2113f5f551850d8add93e07a
85072 F20101206_AAAIDF baglivio_m_Page_148.jpg
b591507c1d7bc26850aa9877ab3c4a3f
64426e064547f0623c75d418866b5f265eb7d99b
3341 F20101206_AAAJFT baglivio_m_Page_009.QC.jpg
ba41a75109ab25d5eb1124e8d70e135e
d4af2eb0c1ee938b1db0eee07ee832119edf1d2d
69435 F20101206_AAAICR baglivio_m_Page_133.jpg
0d518a9c3a13009538bbfdef21caf655
ae24165736172c8c6ad82fbb2f0962551462d031
121988 F20101206_AAAHXL baglivio_m_Page_059.jp2
0625bde3a74054072db06ad611aa057d
6d0e773dc799973b1b3740c87f923af5cd505cab
F20101206_AAAHWX baglivio_m_Page_065.txt
787bda9195613f4b25755879f80d8cbc
af29ba44ba0dbc8421dcfb2fa97fd76147143c0b
2633 F20101206_AAAIZZ baglivio_m_Page_083.txt
8230c1a083d551d016457fcf432ee619
55fd4805736dbd54771a162021bec4e9a977b89a
27632 F20101206_AAAJGI baglivio_m_Page_018.QC.jpg
9b0b47bbc720c2851b0ccf1ca769c4da
4f18093eba672f846658979390455bdde8d1e3c1
4480 F20101206_AAAJFU baglivio_m_Page_010thm.jpg
dbbe32e3f778daa07217740a2ac96870
8e77cd05ab2c787d9bc7fd416699f4c95b091ed1
10996 F20101206_AAAHYA baglivio_m_Page_009.jpg
42abc069a6af4e19845bc3f85ab840ca
61872576734cb7e5047f2b7b5faa32f6ae091766
78000 F20101206_AAAICS baglivio_m_Page_134.jpg
4a7f7d7b4ac142f351ceb3e8c542c32e
f3eaa961ff49f00e987e6cdfba611816f20a9b71
F20101206_AAAHXM baglivio_m_Page_028.tif
f8b1c7338e8c9d9f8da5de2b479d6440
64473d3716f039a703aa33e4eadd0daab4f8ef41
118497 F20101206_AAAHWY baglivio_m_Page_065.jp2
f78b63c06dcae607d4a71098a71ca8c0
f905ddc7ed9b34230c4c34f8c46db36b72f47a2f
113505 F20101206_AAAIDG baglivio_m_Page_149.jpg
f3b788f6df966cfb0e7289f97f0c8680
376768dbc5a6dd80414bf487ebc9be565e67d616
6128 F20101206_AAAJGJ baglivio_m_Page_019thm.jpg
d3cc4a75b6db4b47798589fc8736de27
656b31fa9b8bec1362e8837b6d5d55079b1c2537
17967 F20101206_AAAJFV baglivio_m_Page_010.QC.jpg
64b471c610be0749741faf495169956d
03f061488ccc13c308334f63358367bc37763e43
60757 F20101206_AAAHYB baglivio_m_Page_010.jpg
cbc80818bd895b2db190de409749c136
25fec10a62a46c1c55bb80194807313c2b92fea6
82592 F20101206_AAAICT baglivio_m_Page_135.jpg
288ca5fdda6e9cea600d2d9e4011e9ba
b998ccef64867e1be0284be2c16eebafca3b3537
95150 F20101206_AAAHXN baglivio_m_Page_142.jpg
bda5c16f0f0c0911862b5b6cf1956bd3
3f970647f596a79aa878d384ebc49d3a85e7854f
90227 F20101206_AAAHWZ baglivio_m_Page_083.jpg
1e5cef1a8846734fc0d4bb6367939d3c
a294468bbb30b2612570eb0e46d121c272479667
89144 F20101206_AAAIDH baglivio_m_Page_150.jpg
3b26a1a084e336a554fa9a06ca77c78a
fab6b1a71df0f122825e39d39459fcb437919ff9
25948 F20101206_AAAJGK baglivio_m_Page_019.QC.jpg
e4e983fa16785aa1cb5f0cd51be3664e
0d351e5f9447719938ebd022389b9251b1c464a1
27884 F20101206_AAAJFW baglivio_m_Page_011.QC.jpg
6c0e534f35b30a040ebbbfd6e8cf3abb
9075b1f7c50514b92f04304c0b39c9e70ac216af
90374 F20101206_AAAHYC baglivio_m_Page_011.jpg
d31383f5942cababa74efacd902587ba
6ac06191b0df52a8d85d290cb4989832fee92ade
25642 F20101206_AAAICU baglivio_m_Page_136.jpg
cab3ea694f0f26d286dce5d30cde49e7
7a1be6ccea5691d98afba4de63bd966ec1068d8a
27973 F20101206_AAAHXO baglivio_m_Page_046.QC.jpg
f42b4b0214774136a963f357bc8e0938
b267ea30864702225f757caaad780e00946d2a27
119398 F20101206_AAAIDI baglivio_m_Page_151.jpg
29bc0d1486586a45ae5007b4f9854998
5209ec9c921d26d172aef83ed6ce9f9a078df19e
6853 F20101206_AAAJHA baglivio_m_Page_028thm.jpg
ca90de8ff6fa7830dfe9437bca81582b
0e3b760a61c860120df50024d8cc2ff72682202b
6146 F20101206_AAAJGL baglivio_m_Page_020thm.jpg
e3d31982312b0c64564f99dee95c0ffa
b4858776b52fa9031018c1f74a6fa3d8568b3680
6274 F20101206_AAAJFX baglivio_m_Page_012thm.jpg
a3cc313ab97b6b13a23705946b6d3735
ebcdd51201f2e8a89c4bb8385d8cc14d54a3165a
100916 F20101206_AAAICV baglivio_m_Page_137.jpg
edf45334201f0d29faa422291b0ad2f4
4ac5bfc089a59876ba55f972bd5a7fb397947708
194056 F20101206_AAAHXP UFE0021569_00001.mets
bec8798cd02e0bf7570bf16f52e2d817
8b7ff79a3d06f51dcce2c4b03e14e90a2d1d4dbe
18655 F20101206_AAAIDJ baglivio_m_Page_152.jpg
ff8863e93cb60af4496dd1dd4435abf9
d8120d17f6f3c67b866252a1dee60052679ec5f7
86125 F20101206_AAAHYD baglivio_m_Page_012.jpg
6005ddd422d3ef5101f8e52711d54f07
7b63792708ca53fe736614d9ce279aa0f8371f9a
29177 F20101206_AAAJHB baglivio_m_Page_028.QC.jpg
ebc950c2e88bbb1aab54fb688da067e3
313411e4d9b0938fde488e8cd10f72ac1c934a3c
25718 F20101206_AAAJGM baglivio_m_Page_020.QC.jpg
f2a603c23e4e69455bd95b92edefe937
39e6c9bf84bb7daa40c0113dcf61f9e83072e0f5
6305 F20101206_AAAJFY baglivio_m_Page_013thm.jpg
b1ae9c600f948f4cd2d9d71ec56f800c
b0ab4a59d604079306ba4dd6806c80ff213b54c6
101303 F20101206_AAAICW baglivio_m_Page_138.jpg
f86a0cae7168ac9fc9185380e3ce9893
190757873e7579aa542e7af286711c90ab339698
104878 F20101206_AAAIDK baglivio_m_Page_153.jpg
b1202611215d6abdc62ffd29a2a2a3ff
330ace1315500e763d82bd710ead970cc2815565
86402 F20101206_AAAHYE baglivio_m_Page_013.jpg
9e3949180fac30a10ec22e56b5654f26
985fa4f7d31c0166ed26d543284c4a1aff1d2645
6729 F20101206_AAAJHC baglivio_m_Page_029thm.jpg
748682aeb4e7513b17ce1d3f96992319
3b46123f96ba73f4b5c6e5f180592522710f2f61
6411 F20101206_AAAJGN baglivio_m_Page_021thm.jpg
ea8126caeb08ef559d1dab442bc86d08
da7b7ef6b8919a1163dfb0260862ea3207223a05
26605 F20101206_AAAJFZ baglivio_m_Page_013.QC.jpg
f0187b2eb8d8ab2ee725c9184020647c
4d52e125ce3e09c195498154851ea52e1d3d8c75
86691 F20101206_AAAICX baglivio_m_Page_139.jpg
0a3861e05cfef8f4712386c0d09de699
8a2f62b18b060c88d774007bf5f3b9a4f0b33cf3
5573 F20101206_AAAIEA baglivio_m_Page_002.jp2
72ce9cb4f48b2d4629130d7441f84a07
f486434e7062dcff5fa748748527627441af3e2f
96532 F20101206_AAAIDL baglivio_m_Page_154.jpg
b4f02973c5528481295e78b18d19622f
723619094071cd7c94d1f542b319a9a29205c8f8
77067 F20101206_AAAHYF baglivio_m_Page_014.jpg
862fd7ff443b236784c429e51ad7336a
46d7195faffd003ea198f5151a7655dee5348f68
27958 F20101206_AAAJHD baglivio_m_Page_029.QC.jpg
93f93b82dd64ac819cf05f41f2cf4b71
03976d9a7c7a19ebf951a58853c83afccf714e0c
27580 F20101206_AAAJGO baglivio_m_Page_021.QC.jpg
c027c6a141d5e147af5f96412b27301f
c55119b26e1755dc16c807069bee512c52777f88
107673 F20101206_AAAICY baglivio_m_Page_140.jpg
df7c5532940007871748746528a1dd39
54d32003bf08ea78024a0739e2c732578fefae82
22723 F20101206_AAAHXS baglivio_m_Page_001.jpg
6feb267054dd6c36f4f8d900ef87c3ce
7033cd955013348fe09b8841449abb8b54da355a
5948 F20101206_AAAIEB baglivio_m_Page_003.jp2
8ad1e10c8d303bd9160b385948c48839
1bd5f4616ec198d10a2b340116dc984bd27a8096
93085 F20101206_AAAIDM baglivio_m_Page_155.jpg
8e999b93a19c2caf4951fb756bd1c5d8
65ca7240b201d3c2e9c4f6cac441c88cdf634440
63436 F20101206_AAAHYG baglivio_m_Page_015.jpg
bb7b076247afdd8298767dad0a0045aa
e8635da7a831ff885bf4c8d969fc7dd18e60dc39
6564 F20101206_AAAJHE baglivio_m_Page_030thm.jpg
110c0cc231ced94ef22e4c96511b91f7
0c4f8b16776c8a1d6013aca438cdbb7c8581c58c
3930 F20101206_AAAJGP baglivio_m_Page_022thm.jpg
44afdeaee11fe6df368d556f6fc9c64e
96bfad8c39406399211addbcd0d671a2c5b94ec4
110230 F20101206_AAAICZ baglivio_m_Page_141.jpg
52dd737703f9cc14b8f168dc97ba2184
ae5477ef1f42a15c26fe59760df9bb43e7c709a7
3848 F20101206_AAAHXT baglivio_m_Page_002.jpg
91412a6b4a94dfd4c52c2e466dce5eb4
6d1ce40c3dc4452095b524fd139b0ebcd26ea81f
19868 F20101206_AAAIEC baglivio_m_Page_004.jp2
09aa4f6d337a6aace68e0edec736a6f5
01823869f22b84fb248f5cebd5b6d38132d996b7
102296 F20101206_AAAIDN baglivio_m_Page_156.jpg
28b01e9ea15fcc47deffa2206dfa4d4c
cca4ac976af9560e00135d68f8ef2d54765ef351
84950 F20101206_AAAHYH baglivio_m_Page_016.jpg
0097d11b66571e87992290a915b4cad3
353e44bcc041ba51456709e5550a73c58b46a3be
27192 F20101206_AAAJHF baglivio_m_Page_030.QC.jpg
61d29d3afbbae55029e567913b8c4af4
83ab533703df48caf36f4e69d377d0ef71e83678
16210 F20101206_AAAJGQ baglivio_m_Page_022.QC.jpg
535853da26211e46024e2d94e06daa62
3e6c4cfe169b48098fa5896c7af8b16a3ad6ee3b
4391 F20101206_AAAHXU baglivio_m_Page_003.jpg
e50c1d3f75f16358a0bc73fff8424e87
40bc4a3d5d724e781ac3de4a786954929abc4349
1051985 F20101206_AAAIED baglivio_m_Page_005.jp2
44cef0d662c445f00a61f7274ebc4b8f
c28ed5efa132e0c62f3bcecd6b5221be337dc1d5
99027 F20101206_AAAIDO baglivio_m_Page_157.jpg
a0c9349758dd5cba6ca738a75b5dc093
dcfefdb9e73f89547bc6121a89436cfcf45acd69
81750 F20101206_AAAHYI baglivio_m_Page_017.jpg
24b7cd294336f3e5712d7d478763c743
cf20de61315fbee501a49c6975e0502cdb2be5de
6825 F20101206_AAAJHG baglivio_m_Page_031thm.jpg
3910c529fb1055baf544b61b52057108
425898f409c7cc6567100cbb6debe876bcaf5ae1
6133 F20101206_AAAJGR baglivio_m_Page_023thm.jpg
42408961b399fbb2f3c4f1673446bafd
ef3a553db898392a835b331e603ca9a41072a306
16245 F20101206_AAAHXV baglivio_m_Page_004.jpg
4dbba7fa9cafc955c7f109fc1dc35c35
32509aa6fd73f8e6d2f3deeb09e763ba8f2da2c1
154639 F20101206_AAAIEE baglivio_m_Page_006.jp2
e4da4243b048a368926f4a81881acaf5
58a2fa14fcf77226073de262c7f412546f62e50d
97445 F20101206_AAAIDP baglivio_m_Page_158.jpg
c1a3dc73972f24c40bff8c01345f0879
90d25f504ca6c147311fbae5731bc6266332e70e
89097 F20101206_AAAHYJ baglivio_m_Page_018.jpg
b5b576f403df888bb8d3163d180f2ccb
72ccff1120bdba6df74504cf2e0939a118206b79
29410 F20101206_AAAJHH baglivio_m_Page_031.QC.jpg
c1365912b07d2ed9718dd7dde5333326
4b7d5a266f4bb137c7a103159efe17ef29eeb644
25532 F20101206_AAAJGS baglivio_m_Page_023.QC.jpg
834c37b99bea79dcdb7f606aa15d0442
1d1fc4197adf64a9269f9c405ffb659c73d4dd00
77139 F20101206_AAAHXW baglivio_m_Page_005.jpg
5b1bc6e1a5bb631e5b3c73b432c5c24e
7a424e33f84192ad7f2e3088e0d1117e104a84cb
1051967 F20101206_AAAIEF baglivio_m_Page_007.jp2
df21e2fcf92e96e5d9af7c9a0aec8523
76b2c62d0650774e5fd277855729156971f2b73e
101863 F20101206_AAAIDQ baglivio_m_Page_159.jpg
f250e642e7020f40cf90445a100f7ff1
f5a135cf2fbcf7e09c5985e1d5da479bb15b6908
84901 F20101206_AAAHYK baglivio_m_Page_019.jpg
bceb3548f075e02caeabd40577f5df44
1bba7eaa51fa6a14f963dc96f9c58ea28d76e80a
27513 F20101206_AAAJHI baglivio_m_Page_032.QC.jpg
67f783805e29d8f82945d5bbc8ea164c
db8a123f035c638f2967ee824336fea2e6531848
6756 F20101206_AAAJGT baglivio_m_Page_024thm.jpg
68007fa16bf6188ec968cd6403ceb942
263dccde0f1014b0bc4a6c0d3aa738f3cbaa17ea
9845 F20101206_AAAHXX baglivio_m_Page_006.jpg
f5256dbf3ff4338dda9ab63cdbd87203
d6c4cf88ce01aac9dd61cdb2126bd6617fd940f7
1051984 F20101206_AAAIEG baglivio_m_Page_008.jp2
619dbbc636bf1b8a77bd340da45dadca
0db99ffd4d05dacbb1a9d13d0b89859b0dd90299
101603 F20101206_AAAIDR baglivio_m_Page_160.jpg
c628a70ed619cc793b578196b78ccb7d
28d39fe0046a8b2017727ba195f8de46e5f6c877
82884 F20101206_AAAHYL baglivio_m_Page_020.jpg
487ff8dd5726bc087016e48b1a4e4748
223b00e4371ab1b4dd427de05b0fe65b57b844be
28547 F20101206_AAAJGU baglivio_m_Page_024.QC.jpg
e02b71d1b0a1ca94959401f9fb3b4474
8bd645ce495ddd988143ab37260c9b11dd3c8fe9
90150 F20101206_AAAHXY baglivio_m_Page_007.jpg
9880d6d54379acb54c51e5bfb16192ea
5f2efdf0006136c5416ef92f5ae769a71e68d4e1
88274 F20101206_AAAHZA baglivio_m_Page_035.jpg
489df4c15df2867e7eabd111b194366c
1a7322aab49ece319ddc3e172306fa5d0dec6839
104167 F20101206_AAAIDS baglivio_m_Page_161.jpg
e9e3ac2208f34c14ce87c28958075416
38f8539a178fbbffe606f0de9327b9588e8d019f
89462 F20101206_AAAHYM baglivio_m_Page_021.jpg
fbed9d80977c509ada5960ba9b40219f
79d5e6ce4832ec872181b4a4d49184411ba9bd5b
F20101206_AAAJHJ baglivio_m_Page_033thm.jpg
4babc47e7c16a71c1266082ca1e2abb1
b1c3c466283ac60fe0b10f5a109ea06ba97ebc1c
6896 F20101206_AAAJGV baglivio_m_Page_025thm.jpg
bb579c3157a613e0953c6e2c1fa31a81
52f8ab1d60b811dabbd8348ad331b8b1745ea9be
47402 F20101206_AAAHXZ baglivio_m_Page_008.jpg
f2cc3d7a2c6f312b2f25b7450d436a1f
840a0fa060143d0c0bf853dc6036a50aaaab9ec3
172774 F20101206_AAAIEH baglivio_m_Page_009.jp2
948b5be837865831435efb225dbce614
6ac0cb9797aee2cea880deb012771c56d2b45832
84927 F20101206_AAAHZB baglivio_m_Page_036.jpg
e06cb211939b97bbc1496c5b240119d7
33cc1f9511877393541851c67f9b049d806ace53
99243 F20101206_AAAIDT baglivio_m_Page_162.jpg
1f46ee21adf1863abaf44bf1cb5691cd
41c83445d3545984cb51c73663aa95f3d4edc8e6
53049 F20101206_AAAHYN baglivio_m_Page_022.jpg
f57cd7b654193040e29687b102063a7f
68eaa2866e840c5e08a06a4751254d92f458f413
28880 F20101206_AAAJHK baglivio_m_Page_033.QC.jpg
4c5298282f56af8804672e5433a1a4a8
bdf978371711368931f272d4ba7cd732857e21f8
28843 F20101206_AAAJGW baglivio_m_Page_025.QC.jpg
2e20e662501e0c07e7f6a82a0dc2aa50
923dc97cb5658ad59860e7daa8014b8209e95138
76154 F20101206_AAAIEI baglivio_m_Page_010.jp2
e6601626bea02b0a2ec342a29176507f
02854fab941dddcae1694b127d59f9aaba32bebe
85691 F20101206_AAAHZC baglivio_m_Page_037.jpg
a3ee4f80ba9477b6bcb6bba3e76757f3
67c115994d0d83aee3da46336147dfa296872c8e
106007 F20101206_AAAIDU baglivio_m_Page_163.jpg
e31d0235e3c5cdc1fecd31ab76da6495
9d9c824cffc9c013ec09568e5774d0a5a41db033
83308 F20101206_AAAHYO baglivio_m_Page_023.jpg
9f5aa3854144fc83e7f8185b831c736e
2e51afb793866c1fdeb1165d7d410f9b7e5db2ac
27031 F20101206_AAAJIA baglivio_m_Page_043.QC.jpg
8eae8629d074cf844137f64230d8c8ac
56af1e378060ad44e6cd0659aa70cd7e5b25d2ca
28162 F20101206_AAAJHL baglivio_m_Page_034.QC.jpg
668bd3b2ee19ec4057f9d9b7b37bb937
b3ee0c933dd8571714b496c090c392148f00d962
30201 F20101206_AAAJGX baglivio_m_Page_026.QC.jpg
8d3fb75ed4d51fbfc32e788ebce80140
81414876948cc406f8ca5cbc131d7538d1e6ac4e
1051978 F20101206_AAAIEJ baglivio_m_Page_011.jp2
1c662af87ee24edf069dc8a20ef856ee
5563f73f957e5112475f67e16243fa0542dcddfb
80705 F20101206_AAAHZD baglivio_m_Page_038.jpg
75e8162efc2ae1b38eb083b6de21338c
a7b5fb0ebb84c45c1f54508f74dfc0d2e21a0030
104102 F20101206_AAAIDV baglivio_m_Page_164.jpg
2e89c578f4d8e158365ea424982e8e80
78a17483d3e48c5f3544bae01c79999f73fa1eab
92103 F20101206_AAAHYP baglivio_m_Page_024.jpg
b672525896e6bbf517b6e4967b966014
9de319bbaa9d075cbaa46551741cf709c661daee
6408 F20101206_AAAJIB baglivio_m_Page_045thm.jpg
03f6bd6f6220a2426026361e3f2d654a
93e326b0288af3f7943baea1f29c64106fc18157
6667 F20101206_AAAJHM baglivio_m_Page_035thm.jpg
b8f503b36d64eebc9d54a32a087dbcdf
8e3d9e1a5e2d90b1a714ed836ba79dc775fc2713
6733 F20101206_AAAJGY baglivio_m_Page_027thm.jpg
72a9f67bb14d711dc8a52e70d38f9f91
8bfb4ebc8dadc4d0ef5abdfab317d8547e0821b2
1051913 F20101206_AAAIEK baglivio_m_Page_012.jp2
52ef2e6df8edd28f66e4ed91570f7ee1
cff699d3566ef14daf52836697e8943ae6f408bf
91054 F20101206_AAAHZE baglivio_m_Page_039.jpg
7511eebbdc6952f9c9841e31e71c7cc7
ff858ced91f9a64db77b788d939b47fe0cdf36ce
97209 F20101206_AAAIDW baglivio_m_Page_165.jpg
21b850d9bf5fbc5419b542388e4a85fa
4d7c597e34bb4072389010b31408354f85f169a0
93207 F20101206_AAAHYQ baglivio_m_Page_025.jpg
7b6c9ad7e4c5f542bb275aef20a337f7
fd4e0e0b3b26dbcf65d0e082a6f8ae004515acf9
27432 F20101206_AAAJIC baglivio_m_Page_045.QC.jpg
8ed5b739ce9e62352e18e41c3aa2cc56
d6988cf50dca1973aa3a4c28b0ef932a6961e106
27927 F20101206_AAAJHN baglivio_m_Page_035.QC.jpg
c2374c5547f1f9096913673dd69a0ffb
a533ddc4adfe22d9b158a6faf1a2035a76283172
29037 F20101206_AAAJGZ baglivio_m_Page_027.QC.jpg
92a15fe4f3302749f005756cc086dc6e
36e5c44b79d583ff93665d1b46bad5422243ee84
123167 F20101206_AAAIFA baglivio_m_Page_028.jp2
8e63fc4e387c049d06f62d14dd20c0db
581c0c2a264a7c174de8ff95ece2a38d924f9568
1051970 F20101206_AAAIEL baglivio_m_Page_013.jp2
5e2996a3ab16b12cacd60960b3c5b301
81d60aac62b962f72ee5a74632bfc2debafc7db6
87522 F20101206_AAAHZF baglivio_m_Page_040.jpg
29780fad4a694051a045c7eefd8642aa
212c0f81059bcbb189f38ae9f8f13cd197dd3819
50938 F20101206_AAAIDX baglivio_m_Page_166.jpg
157dc0b99e3d0fb587ef9a5c06d7e255
74e7505d42e6eba6bc197c9efb653f4f3c77e0a6
99927 F20101206_AAAHYR baglivio_m_Page_026.jpg
bb5e9a7e94343211ab199880fd417389
8c36517ef1fdb04204694f8136762599a9672f91
6215 F20101206_AAAJID baglivio_m_Page_046thm.jpg
1ead8596afd61b9dd764b3da00fe53b6
a935e56037960e480ba57b2af1b42ca14d17eef2
25776 F20101206_AAAJHO baglivio_m_Page_036.QC.jpg
ebc854a3a983c8b419f83ce4f225dae5
b949c9e5ba262f555a4543ae5c3bbb9f5d4fa24f
1015211 F20101206_AAAIEM baglivio_m_Page_014.jp2
491b6dda71f631452b368f73417ebb97
2ddfcfcf2bee3e41b52680614aeec47acf2498b4
90117 F20101206_AAAHZG baglivio_m_Page_041.jpg
08f4a5749aaffcf82650714fb3f0e67b
dd948dcd79489c05039c6441078d3729b3a1f9c4
33528 F20101206_AAAIDY baglivio_m_Page_167.jpg
ac55f8757db1b3e9436080af0a8e82a3
5d2a0b17e6c39fd323fded349473968850146561
92738 F20101206_AAAHYS baglivio_m_Page_027.jpg
e71deee03f305e23105d18cebbb34791
e098a7acdb04393c93b985d28b54d53c90596c5f
120324 F20101206_AAAIFB baglivio_m_Page_029.jp2
a6969e010adddf1e0af6d25c25f8c2fd
b368e94071c91dc949f135b9a5ff570a644a6f07
6534 F20101206_AAAJIE baglivio_m_Page_047thm.jpg
f36533230e8a50c450eb1f246ba63938
ec1c95aec5d7a2ad6a65267b75e02e5c3cae4205
26512 F20101206_AAAJHP baglivio_m_Page_037.QC.jpg
985656c3fdf5d49978fffe273bdd3329
0c271b762d42a8a6d9b9eef6992aa1b55a46da76
820432 F20101206_AAAIEN baglivio_m_Page_015.jp2
3dd3d22e69dd2b85cdc4f6e1d33bcc30
0bd3541837f4bf9d4b507cefa1bca1aeff822f5f
101363 F20101206_AAAHZH baglivio_m_Page_042.jpg
cc598d044deea710002125ba6aefd1c9
f8554472d53b243e6e6fc70343736d7008e5f089
24203 F20101206_AAAIDZ baglivio_m_Page_001.jp2
d26f380111d25988cc5f834c1ec98843
e8f58aa4c96859bc2e976fec73b3640e2429456d
93384 F20101206_AAAHYT baglivio_m_Page_028.jpg
636fcaef28088bca753f4c57c41501b5
a83bb93c021ea60fbeb2f7b7b0ff451e8e175bd4
116966 F20101206_AAAIFC baglivio_m_Page_030.jp2
a198b43d909dd9a32b570a46cb1af01b
15e764f0733dfe7404b5e675848cebf88c25fcdc
6701 F20101206_AAAJIF baglivio_m_Page_048thm.jpg
54e73225d3e17081599cef6d009aa437
64fce8e627b8d27ba591a501601111d0634b17c5
6281 F20101206_AAAJHQ baglivio_m_Page_038thm.jpg
a18f4dc65d9fad1954b99c63976e205a
ffa9b9e34bd84c0a98e07e890af18fb5d39b19e6
1051966 F20101206_AAAIEO baglivio_m_Page_016.jp2
f72d78112cad49a0ed56deec5fc1cb9e
a347bef32738634f9e0f08cf0fc4176afd3237b5
88203 F20101206_AAAHZI baglivio_m_Page_043.jpg
a017d9fa97f442742b6bf85160ebb40f
cae7658762dea9f5a9d76324c73e317396e04b98
91362 F20101206_AAAHYU baglivio_m_Page_029.jpg
a584c326084db85fdfb4ab375b7d588e
a9b2cbee79137e3c8f83630a6a39664d0dab22ec
114834 F20101206_AAAIFD baglivio_m_Page_032.jp2
add71c1dbf9b5f527009408f575cd430
ed8f724f5693c2f7f023f34a79bef9dd85de578c
28197 F20101206_AAAJIG baglivio_m_Page_048.QC.jpg
6a094bd34f7d9474ac2556816845214d
a558b39e4a546c5e2c2fcd9d65f8ea121a5dd671
26166 F20101206_AAAJHR baglivio_m_Page_038.QC.jpg
ca04db336d5062aae5185dd45a3b28c1
51efee86b752ba82d80017fb0ee5f5820535c22d
1051971 F20101206_AAAIEP baglivio_m_Page_017.jp2
8d0f894dd46223db1afc8d5c014f90da
9845e82cdd3d1b946181ef9c89350544b09f6235
92980 F20101206_AAAHZJ baglivio_m_Page_044.jpg
e4849303d868eedf191f9592dcf4e609
3ab601824bc183fa0f1e69aa5f332844d7ab9e09
88054 F20101206_AAAHYV baglivio_m_Page_030.jpg
a016ae85a539f1c73f048b05487b8e72
3d32eddfc1e12ae6ab9d095ae8488ae5406ebb11
120272 F20101206_AAAIFE baglivio_m_Page_033.jp2
d505d8e5328fcd56cfa7e23bd6c67c5c
cc36132c88e319befe60e49ba02a03009d8dccd0
6189 F20101206_AAAJIH baglivio_m_Page_049thm.jpg
e8dd454f929876e0ca1b42ef3985804b
52f3708c61213c508329a142d766a82788872507
6752 F20101206_AAAJHS baglivio_m_Page_039thm.jpg
02ebcfddc4fe0344ea3f22d3afad993b
72f78c8182c22a5c7d50ac9c6133fffab0ca41cc
1051975 F20101206_AAAIEQ baglivio_m_Page_018.jp2
eef816dfde62475e8cf63290b298718a
bea2aaa81c83263795fa073e55944ed4bb5b0781
88235 F20101206_AAAHZK baglivio_m_Page_045.jpg
b139316a435b213fc058a243bffeeec3
ae64540bfed368f0669ec4a94a16975d6b461f41
104806 F20101206_AAAHYW baglivio_m_Page_031.jpg
0c73c091ef9ac5c2d7268e706eb19c46
54bfbe2e4789387d25b24d3a6e41b85e26353bfa
118224 F20101206_AAAIFF baglivio_m_Page_034.jp2
15fb380f89461fec001d0285a23ac0fb
1a2bca6ab5255ab4a826a1650f461ae92be4b466
26628 F20101206_AAAJII baglivio_m_Page_049.QC.jpg
28034a9a8a075304006c99559bfafeeb
758dc237b4cd325c9a0fd6e8144e3f513f6197dc
28592 F20101206_AAAJHT baglivio_m_Page_039.QC.jpg
512a1be942054cc785e6ec0087bb0cb9
6ca470d1f2321422b1871dceae73c95fe242228f
1051974 F20101206_AAAIER baglivio_m_Page_019.jp2
010171fe6a56d11c2a7df8940a3c8ab9
bb31dd9b6cd46b1dae9de93d78ab41cfe08e33a9
90763 F20101206_AAAHZL baglivio_m_Page_047.jpg
ac8bf2a4548279961cd8bc8a752580b2
ef346271083dac4bca1d8eafbad8d86888c19909
88042 F20101206_AAAHYX baglivio_m_Page_032.jpg
6635fb85a61bcbc1a4fea7315bd6d924
b3d1085e60fc76aa49c719a1dc31c9eebfe4a8e1
114752 F20101206_AAAIFG baglivio_m_Page_035.jp2
84e9ab6a94d52493239bca07c4950b95
1e513e12bd21cbc763958f2aa90582fe89086b81
6522 F20101206_AAAJIJ baglivio_m_Page_050thm.jpg
72638ba21da364f50e1cdad4f65f8d42
f7d678713e2423096c876649d5c773cee6d48698
6584 F20101206_AAAJHU baglivio_m_Page_040thm.jpg
93edba15e9c80b5695deb0bcdeb598b9
c3d3b66b6eaf060823d376bb4e893345416ceeed
1051979 F20101206_AAAIES baglivio_m_Page_020.jp2
738a9cad40a6ae712ed7eb8dbbc1b607
e54f432732074837e725c734e9a027fb7e797fe4
89353 F20101206_AAAHZM baglivio_m_Page_048.jpg
ea008144bc972f58998d2c97b9167910
93878851e24a621383818b2261b416da92a1b410
91338 F20101206_AAAHYY baglivio_m_Page_033.jpg
7b77968425518679d2758c2947648b51
18ce118ff7ff47dca871f981dc2464feb4a4b9fe
111572 F20101206_AAAIFH baglivio_m_Page_036.jp2
c703340958a036546ea6fd3922b229ac
d95b121b3ea3c20f5343156a2eb41f7aeae0a4ad
27510 F20101206_AAAJHV baglivio_m_Page_040.QC.jpg
c30d829bdffc9a720db6ba269a28f34e
d134c877d29ffc09a3cb3f2a0cdc77d5917fd8b0
1051953 F20101206_AAAIET baglivio_m_Page_021.jp2
9e53a143d9c9657901d055a3993b39f4
87464573ae824c72f416d4a3c0abf60fc2df11f0
85206 F20101206_AAAHZN baglivio_m_Page_049.jpg
9b2a7f7ab8fdb704ac815e570ac2a5db
8475f0328138eba4bff4cd40cb467d1540167ee7
F20101206_AAAHYZ baglivio_m_Page_034.jpg
4d5bf162059ffc17ae2afc995ebf2b0e
b50d4eaa80bee2f18ada6af563ce8c5ef4e340ac
F20101206_AAAJIK baglivio_m_Page_051thm.jpg
a7d983e190ffa34c131737e6fe085952
147deeccf6a531b640cdab8137c4b0fe7197c1cb
6369 F20101206_AAAJHW baglivio_m_Page_041thm.jpg
052165f8b84c38ea449028dbde7b104e
6c7a33ef01c02c04e73345e6471e62f8bdcceaf9
65516 F20101206_AAAIEU baglivio_m_Page_022.jp2
4c307c4cbc3a6807290336c06e07ff85
a6250ac48f8bb45777edb3fe0572c79140eea51e
90128 F20101206_AAAHZO baglivio_m_Page_050.jpg
9131a34e57c87e1ad45716867d819fd3
6c7807d72b5430e21fa2100615242c3114118bbb
111970 F20101206_AAAIFI baglivio_m_Page_037.jp2
4285587476574ee1013f4041a67ae74f
8c95d5bbc84b4fb8a6e66e96fae459c07ccb436b
F20101206_AAAJJA baglivio_m_Page_060thm.jpg
738fa5e567cb39d682706cfe795bf4f5
261ea751548cdacc657968252262d962b9a8f78b
27762 F20101206_AAAJIL baglivio_m_Page_051.QC.jpg
f77745b465912c9e4b62d2dc6aabbde8
42bfc8d4f1a4b7f484b0e660efe1a89ac7480b28
28273 F20101206_AAAJHX baglivio_m_Page_041.QC.jpg
6afe4bfa1582d42f5106a1faad0c238b
10a6f3117b29e872512f279d8b00b03d0feb3e8d
109200 F20101206_AAAIEV baglivio_m_Page_023.jp2
215c3d3202a0cacab654b99bd63bf326
e829e7575acc8c74acb3012fa235120631ddd465
90689 F20101206_AAAHZP baglivio_m_Page_051.jpg
0f1195cad779ac73241aab254b154b88
ef64640fab085595f5b96fb762fa505d69218244
108275 F20101206_AAAIFJ baglivio_m_Page_038.jp2
dc3f7265eaba1a35bfb23d35e51d2127
d8ce7ec2674db0fc49dd0b4bac499bfbfca693c5
26938 F20101206_AAAJJB baglivio_m_Page_060.QC.jpg
45d69386db45cdf970b23ffa4132a57e
c1175a70111f76f74efbd0f4285d92b7b16f0a26
6766 F20101206_AAAJIM baglivio_m_Page_052thm.jpg
bdfa91c4d528a1bee0d314e709692567
853fe8913d74355b92b6c59b89666f51a91be551
29323 F20101206_AAAJHY baglivio_m_Page_042.QC.jpg
9483d2cc756fda1a1389270a7b76dd56
2dd03b38cb243631f9cef78174c3063838a7eeca
120791 F20101206_AAAIEW baglivio_m_Page_024.jp2
12f611a6414f44dbaad6df1c70386fb5
6300ddccb50a311c2db65a8732e9d6120db35a61
F20101206_AAAHZQ baglivio_m_Page_052.jpg
88179ee1c59c0e97d4f6c48a8f81d998
baa908b330bec51e4812fb1f9dc0bbc46176a505
121041 F20101206_AAAIFK baglivio_m_Page_039.jp2
4735debeca539f89c36c7233a20af625
b6eb01d2b9ea3758870355750910d2809e8431cb
6771 F20101206_AAAJJC baglivio_m_Page_061thm.jpg
4548b2543e2348c0eddffaa12a50755d
3fa6c8cefde57ee81cc8e6d9a7fa356d79a2b0b4
28539 F20101206_AAAJIN baglivio_m_Page_052.QC.jpg
228c9d6d676b170adb8e449241271aa4
291303541887f3fee19c065bbf976970e9a818fb
6591 F20101206_AAAJHZ baglivio_m_Page_043thm.jpg
7a6383b4db2ff31acbf1639c125ea826
c4077f3bafc40ff5c448c492b3afbf36a32f6d6b
122141 F20101206_AAAIEX baglivio_m_Page_025.jp2
34b8723be89f140b37532182e52b8448
915f8afd45e47c3ca96d8ef89f90390474056373
88692 F20101206_AAAHZR baglivio_m_Page_053.jpg
8e93bc2ad4a452076710b167c1f2b381
b67a8645dac8bfc2b407e21b328b4a3827d67ad1
121357 F20101206_AAAIGA baglivio_m_Page_055.jp2
27f4d31b570c53ed4a28d018bb1b2cd5
58a10930c52126d094fe368f3d0f0a650fb11c1e
115392 F20101206_AAAIFL baglivio_m_Page_040.jp2
7467433c10458b9a6cb47f85e93721c5
38c103c11a2d8352657071c69b498a569b442539
27935 F20101206_AAAJJD baglivio_m_Page_061.QC.jpg
bc53d9ba10d27c08f9d9f16ec6128776
5933c6dec3d64ef8125b90ae60668c2f2b4c5daf
27527 F20101206_AAAJIO baglivio_m_Page_053.QC.jpg
b5a83dd5b2a148ee7187b2e98643cba3
e6a7c860a46c011c30cc3ec37695b7db25da7b15
131380 F20101206_AAAIEY baglivio_m_Page_026.jp2
c49e2ef28b222844022f5e7a170edfb8
673e315425e9c9d0afd5cc3d419fd90f7de661b8
88116 F20101206_AAAHZS baglivio_m_Page_054.jpg
40ded26edbe352a2a0c57918225ad994
341939bc91526d313b45693d70fc9c0547a4faad
117753 F20101206_AAAIGB baglivio_m_Page_056.jp2
b0acd03aa814e135ca20dc1d9aadca3a
e84c19ef9b8757fad97c05776d1be2ffb6350ab8
116936 F20101206_AAAIFM baglivio_m_Page_041.jp2
8d217d3f6daf578b35c7f87c4ca7c22e
f5895d19d171454777564b952bac83214e9cd32e
6519 F20101206_AAAJJE baglivio_m_Page_062thm.jpg
4f7116c6954f7243f0d0624e09a072f8
b800f763d6bde5ffc9635a898fb16f1d1859ca23
F20101206_AAAJIP baglivio_m_Page_054thm.jpg
3e6d2d3aef7ba352e38887a838236e1d
a25916b884889d1d82499d8069210170389b4052
123406 F20101206_AAAIEZ baglivio_m_Page_027.jp2
47539bf035a696019e6a6d9963905a47
25f1813b069949a543451e1308d9cff6af9a781d
91533 F20101206_AAAHZT baglivio_m_Page_055.jpg
44b885c32128715cd481ccedd855d34a
c745029f018e034cb65efad04f6981ca60b754f9
117909 F20101206_AAAIGC baglivio_m_Page_057.jp2
d5cb0b7816868a4f84cb3a75c7fdbdd6
8ecc85b82b7911b5afde18ca3d28a142e3caac0d
134713 F20101206_AAAIFN baglivio_m_Page_042.jp2
815f29063a847558859c667a7ece3257
91152b6b3dabf6b927e0463e8590f9f581819103
27631 F20101206_AAAJJF baglivio_m_Page_062.QC.jpg
3fa05e567bb498f9022f780cf558b40e
84175a7bce437ca9eac3fb5bb99c92b991538991
6831 F20101206_AAAJIQ baglivio_m_Page_055thm.jpg
7ba962df81ab5cca1c666a75f10a9444
65f11065e22fb6505f64e9f0c902fefa42927460
89017 F20101206_AAAHZU baglivio_m_Page_056.jpg
4efcbac5424ae1428012b2d279138dbc
6e81af5bcdd82cb49fec747dd3006a292a211066
118263 F20101206_AAAIGD baglivio_m_Page_058.jp2
f85769732178ad77c3b9276295a76a56
2629167256f3a9fa40369e7c23dc5fdc0713a95e
116305 F20101206_AAAIFO baglivio_m_Page_043.jp2
0a24e2515a038fba12a92d6c50b4a11e
23b2b244e634d726ec124401e1706f92fd7eca3b
6623 F20101206_AAAJJG baglivio_m_Page_063thm.jpg
017519c18e706819abe9a068b1a62ea8
f6bd76d79e7fcf770e9f3c96b017ee51b610889c
28313 F20101206_AAAJIR baglivio_m_Page_055.QC.jpg
f1486828f01521001dd51a429b33899d
3ace67ac6e2fc24d7ab1bb6fb81aa95d46342437
89305 F20101206_AAAHZV baglivio_m_Page_057.jpg
4d574ef50701e36b50c0ad7dc9b9c02f
50774bb2561ec4b668de74e7f6ffad4b6c2beabf
117183 F20101206_AAAIGE baglivio_m_Page_060.jp2
00711e10689403bef06b57e9192b099e
252850c07167df69a2aeaf9f9161e088dc1518dd
121684 F20101206_AAAIFP baglivio_m_Page_044.jp2
d156cebafbccd38b3869408b67e45c03
bd7045ed9271792702a00fd3c1480552fa5d3815
27996 F20101206_AAAJJH baglivio_m_Page_063.QC.jpg
696d1f4935ad4835cd572a7ef55eb2d5
4c12eabefae31f50902da325c52232519612eebb
6588 F20101206_AAAJIS baglivio_m_Page_056thm.jpg
11685d04c1e25d7c56270b9f5b66aa93
55b2de00b9645f326b21c7702a8df4f9ad7d5a42
90708 F20101206_AAAHZW baglivio_m_Page_058.jpg
208e623a0bcc3ff8055f5905641bd76f
9cb14aa1a9a73f343af60e5d81d48fd2afb7ceac
119816 F20101206_AAAIGF baglivio_m_Page_061.jp2
d0d561bc1b183012a262f939f824b1bf
fbf781078fe40e0286ca758c4f60cdbc239761ec
114364 F20101206_AAAIFQ baglivio_m_Page_045.jp2
615b9476a753b91e866471f84811c6d8
c445e4fc28f4517af5219aad362656fd5438ef45
6574 F20101206_AAAJJI baglivio_m_Page_064thm.jpg
608b72c3c06853388f39fb6c98f6cf2e
8df5c56d16718da09b2cb7bb00544c1ef98fc714
27951 F20101206_AAAJIT baglivio_m_Page_056.QC.jpg
fb3a4837680692a2a506cc54184cfcd9
43b2c667c597f6f947f1636c5c1c361afae56d0d
92792 F20101206_AAAHZX baglivio_m_Page_059.jpg
b05f55141c1ca6bffa80c5fb31552faa
5dec67659c028348e69b6ba21cbf53136c61d55c
117165 F20101206_AAAIGG baglivio_m_Page_062.jp2
30e3b5dc2274e197d690d256c4074ab0
fc0bf7f271107bdff817b44f178bbac48ef481ac
115638 F20101206_AAAIFR baglivio_m_Page_046.jp2
4b17a835e4c8e28a8046ac640d0fd878
2b196cb12d92551b0591169e46845d611e8c1d10
29049 F20101206_AAAJJJ baglivio_m_Page_064.QC.jpg
8801071f7c847a41514623afa9497327
92b906618356594b1d559e78d002901e533ec5ab
6695 F20101206_AAAJIU baglivio_m_Page_057thm.jpg
8d95f0357efe1c6412f45dd8418e76aa
96a761062a4756ca7f6f2648a26afab3cc138fc8
88311 F20101206_AAAHZY baglivio_m_Page_060.jpg
5c0d3d4e33107652ac6c4b25103aee26
1c963b57d81194d8fc55410db215d474e26f4e0e
119353 F20101206_AAAIGH baglivio_m_Page_063.jp2
71f21b315aa4ffa51e26c818ca22f8f4
7503eebcab5d6ed4ce7022d479ec584b9042c47c
119419 F20101206_AAAIFS baglivio_m_Page_047.jp2
49e3cfd277d810b72c3fe602b5c8fbee
3fb8a11063c51d01067ac22afd0caef187b8cb82
28376 F20101206_AAAJJK baglivio_m_Page_065.QC.jpg
4e11963cd474dca0695ecab553ca4281
623e1072bd4b34f7f59060a67b6449fd304bd4c0
27635 F20101206_AAAJIV baglivio_m_Page_057.QC.jpg
3ffe96e5e71419c19bbf1e00366efc99
4614ec3d0fca1499976e07beedaa67c88a90ef0e
89570 F20101206_AAAHZZ baglivio_m_Page_061.jpg
6f16f6cddf71d4e96c0d164263148872
51613fa772565ff841119aa584a143059be0b832
119540 F20101206_AAAIGI baglivio_m_Page_064.jp2
204b8e2757cf44c8c507dc0da3923b7a
8399677cf9abb1b2a554ed51761a92d221712aed
117987 F20101206_AAAIFT baglivio_m_Page_048.jp2
e20e343da63d725fc490d477ff366d44
68fa40f3cda51051292285ba49d64be20b7442e0
6716 F20101206_AAAJIW baglivio_m_Page_058thm.jpg
a54db8670eadf649ae5c4cea23d4a1c5
f967ca40c6b82d81edbf0a462953acf67175b7b4
112817 F20101206_AAAIFU baglivio_m_Page_049.jp2
729bae21560f6752ac20fefd121cf639
1e80a29ebeb64abe03da7bc4e3023d3418131d08
6781 F20101206_AAAJKA baglivio_m_Page_077thm.jpg
85ca23a94f3f18cf3ddec5706a7c909c
a773e95df2344bdd4f2758fe219c71a7e4a04700
6870 F20101206_AAAJJL baglivio_m_Page_066thm.jpg
e56560493f5f61a3e585e106a5a6c11d
e7f8adc971560e1ba0a49299f965eade7c6dd2f4
27919 F20101206_AAAJIX baglivio_m_Page_058.QC.jpg
f90f56d3a9690badc331812581977f3c
0b87f4c21cfa3d68f56713adedbe21c4cc60c5c1
121605 F20101206_AAAIGJ baglivio_m_Page_066.jp2
6c4927ad4a4a6f3c524cc45a57107166
aab212f997dcf8928b77d4809be95e66c96b5c47
116392 F20101206_AAAIFV baglivio_m_Page_050.jp2
e14d4b5b288cd4a4de9ba960d9cbb28d
05a091838fe52141551ba1b8f6b48b2a49986834
F20101206_AAAJKB baglivio_m_Page_077.QC.jpg
c8e1859ec6180eaec983cc3bc4e39ca3
137bde7482fc3bbfab574b5e50f03e2ebd6371fc
28715 F20101206_AAAJJM baglivio_m_Page_066.QC.jpg
f3fd7966dceb6fec3f54c5f692aa2a43
7eac0384965864506fd59c57650f5f956dff426d
6734 F20101206_AAAJIY baglivio_m_Page_059thm.jpg
2346babef6d3ff9335f06c8ab17f3142
cb9a3bd3abc56e603ee858ba51cd23d5ded3432f
107854 F20101206_AAAIGK baglivio_m_Page_067.jp2
6cc2c9f79efe04f09f7734f04f9e5cb8
8ab378633e52b8676a2970913f44845a180ae8c1
115752 F20101206_AAAIFW baglivio_m_Page_051.jp2
743f214c54753ee51f62825cfba074c6
3f3cf961935d2d8099692edfeb9e73a303dc6fd6
6978 F20101206_AAAJKC baglivio_m_Page_078thm.jpg
778e10cee0a8298eeb33fddc01e4c7cb
050e79dd2b98b8d8a86f896341a612ec13cd135d
6057 F20101206_AAAJJN baglivio_m_Page_067thm.jpg
15d5f089f8bf762da14a91f9051af210
3c867a6eeee1d1ffa33771df90339527098047e7
29403 F20101206_AAAJIZ baglivio_m_Page_059.QC.jpg
c9e0ab31808900e1c5de57e0f6daebbe
2f10bb41f0f6e69c395ad0f8a3a7abf72f966d72
113514 F20101206_AAAIHA baglivio_m_Page_084.jp2
63a942ee533617a5fdaa429658c114cb
e08584ce372e1e80bab2e583b6b0bff3f2b3e65c
56128 F20101206_AAAIGL baglivio_m_Page_068.jp2
73a167175be85092866a6eb8c73b54b4
1d09c0f0fab0ee2d1c8038d5f802e38055454116
119862 F20101206_AAAIFX baglivio_m_Page_052.jp2
ea6b7cdc59f46c8c7473505885584df0
6c74cbf47e1d4c183e9133595cfc59c4336e96f4
29047 F20101206_AAAJKD baglivio_m_Page_078.QC.jpg
955bcca79759d08c4bdd2e87d17d39a2
9c1fee7845a5b38f832728dde79670554561fc78
25585 F20101206_AAAJJO baglivio_m_Page_067.QC.jpg
983dcb789e0f9823b65c475d69b4129d
b8a18a70abbef86df376ee17895936bf732b6178
104367 F20101206_AAAIHB baglivio_m_Page_085.jp2
3cccaf4e191feab9e7d225dc3ffc1e90
19efec3ce2edfa975f859b55d27cb1babe0fa23b
102959 F20101206_AAAIGM baglivio_m_Page_069.jp2
5f0667fdd46610dbead2b1b9e8d4685e
e39379d9039f1d51e34f8fe9bc409a47ad1eb98c
116654 F20101206_AAAIFY baglivio_m_Page_053.jp2
de4d2ec0dbbe171026b1eaa9b4b1455e
b6c25e07d8911694976f2587cb4151cf7a883b7c
6630 F20101206_AAAJKE baglivio_m_Page_079thm.jpg
92dfe58723e38fb21ba3787e975f2109
99bdb9ae21317b7cc963f1310c01c74bc1cf0c4c
12575 F20101206_AAAJJP baglivio_m_Page_068.QC.jpg
88df6279dda297c6d0449acff0fe2d09
1cb78927e3daa8526d341635f148110ee2e637ef
113309 F20101206_AAAIHC baglivio_m_Page_086.jp2
bd5f4678f30e1f5583ce9e1c1a47e2b5
fe62ba0a26ecfc087a07e17266a19113241794fc
107927 F20101206_AAAIGN baglivio_m_Page_070.jp2
4e261e977291fe499c42c63c3cd4002b
52fb123c3a258e23935e2b07a9508d537d7a5814
116296 F20101206_AAAIFZ baglivio_m_Page_054.jp2
7bfebc17094cb092501f23c65554afb7
914aa484e9959bacc681e696312d0eeedecad43a
27620 F20101206_AAAJKF baglivio_m_Page_079.QC.jpg
207fff781cb2299cb809eb629ba272b9
fc4c4daea93edc342780b09f7e5a5a8418bc0b39
5989 F20101206_AAAJJQ baglivio_m_Page_069thm.jpg
ce0b11b0009dbbb2a901b4d2920d94c5
52be930bf26d10c3c5b8a68daa36c8489f32c621
113379 F20101206_AAAIHD baglivio_m_Page_087.jp2
80aae6e6474b8050fe3e177f2f6438f7
42f4dbbe60467bf9880693f36b04b8df2ea2ba83
111913 F20101206_AAAIGO baglivio_m_Page_071.jp2
b4ecdf27b766cde14b63a6c2cb82591e
42edb9ed11c01b5e6d6124949f435c1710e83e7c
F20101206_AAAJKG baglivio_m_Page_080thm.jpg
d0ca533c6f2aea5d2fc5cbf2246b9eba
6c4782cb31bfd0f33139ed016e16d90482f4f6d7
24630 F20101206_AAAJJR baglivio_m_Page_069.QC.jpg
03602a0a6200c01f6cf5b57ce8de7506
9630f10a2570dc0ec7ab30cfb290a0ca8973dffb
116791 F20101206_AAAIHE baglivio_m_Page_088.jp2
530b3c8461f99dc687f63696174afd11
50c879d28d14d6d1b44d3181a7d711bc485fabd8
99052 F20101206_AAAIGP baglivio_m_Page_072.jp2
461752980cf728c22e6387b9173dbc44
38d93e14f37ceb19855e7b0180c7a4cd63e568ce
25820 F20101206_AAAJKH baglivio_m_Page_080.QC.jpg
5ec2a574bb03482fd3f30ef6f91ae1a7
2a75fded5f90534d28815cd8b73a5092a43c564f
6397 F20101206_AAAJJS baglivio_m_Page_070thm.jpg
e43d8397e114bcf8515c0fb84e0df63f
eed7e23896cccef9234373b4a22e52361147271b
102103 F20101206_AAAIHF baglivio_m_Page_089.jp2
8e418c1ba016dfdb9f50f96d05d1a359
6b6eea6b719297ed1ab59c0d1c514016ac9f027f
115021 F20101206_AAAIGQ baglivio_m_Page_073.jp2
84b66688641f85ed73df8c12782a0843
d9f0e3f6ed3701eb46956e5e9a106b9ed624d91d
6019 F20101206_AAAJKI baglivio_m_Page_081thm.jpg
aa7dd0a64633d8115ef79ad8e83d08f5
137275a95f79ba647d1996de9799b32d64705c78
6358 F20101206_AAAJJT baglivio_m_Page_071thm.jpg
593edc26442cc51898818521693b6865
a881d9bfd9d1b19054a4e011617e421bcaa30671
117992 F20101206_AAAIHG baglivio_m_Page_090.jp2
3b71197bddcbf45fb03622cdd7339aae
d4e66b684517bb86ef8abc8f14873bb5262cd85c
114916 F20101206_AAAIGR baglivio_m_Page_074.jp2
3e577d2bde4a7cfed81f9d390a5a53c5
0ae123245e23b27f3d1edaa86954571a1417fef2
25141 F20101206_AAAJKJ baglivio_m_Page_081.QC.jpg
0f8c824a2a2ba5046843008d68a0e9ac
00b3e367cf2ee6aa75c917c749d5165c8ad5a308
27075 F20101206_AAAJJU baglivio_m_Page_071.QC.jpg
26986194d4f6af7bd023a7fcceb9903d
364201a6d553da3997c8b8c1847b37cb109f0864
110767 F20101206_AAAIHH baglivio_m_Page_091.jp2
97900e2b78988997baad5b49ae09c3ce
354e5564dbe239d0ca98de580e7847d1c00afc32
112932 F20101206_AAAIGS baglivio_m_Page_075.jp2
6b0b7239434f477b974253e5580b432c
25777c1973ebc1b9c6b535d3edc577de519a822c
6453 F20101206_AAAJKK baglivio_m_Page_082thm.jpg
49462a26bc122941e7eb4dfb0a335256
8292447e9e261520026379451d48a8e58fb7b80a
6546 F20101206_AAAJJV baglivio_m_Page_073thm.jpg
744775950e83fc068da946b7eacaff02
4e59dc31cc8cddee8c672928d36696e21f345980
116856 F20101206_AAAIHI baglivio_m_Page_092.jp2
69e8172242f603743fba1d8e7fe4c96a
1d6ae6b46807bf36187d7a6809782b13e0699c16
121242 F20101206_AAAIGT baglivio_m_Page_076.jp2
f9f451f5d7703bb08e5b4ec2109c435a
409aa7d39a21a6f273eb016ff5601ab372043a34
27537 F20101206_AAAJKL baglivio_m_Page_082.QC.jpg
c2062587ad4a490f9d3ce2262132565e
bfef6cd9856a9184788639dbc097e8d4df841488
6336 F20101206_AAAJJW baglivio_m_Page_074thm.jpg
7886116571dcf0a6fd36df06d1cbcf4c
4f2f366b177e06bb88ae14951bf2be87209af66a
106051 F20101206_AAAIHJ baglivio_m_Page_093.jp2
4610d548df5139189d17f95951ba1fba
9851133fab3cfb6d1ab88236c6db0a3903163dd8
118099 F20101206_AAAIGU baglivio_m_Page_077.jp2
9210f37a386573784be4128d8e1eeebe
d320b26a08fc5da1f8840a757e03250ddf6266f0
24932 F20101206_AAAJLA baglivio_m_Page_091.QC.jpg
1115cf5e6e24fc4e6a9e906cfc4ceeed
bbba916f372ae76204a13c93d9c50233e385e168
27484 F20101206_AAAJJX baglivio_m_Page_075.QC.jpg
1da19565e615e6039dee94444d321043
2c54ecc8441a182815af26fda00677895a0ff916
120390 F20101206_AAAIGV baglivio_m_Page_078.jp2
2ca8ca810350501cf513d2116b6e0c64
298872292f0a2d813c683f9c05ce86f8ccd565e2
6521 F20101206_AAAJLB baglivio_m_Page_092thm.jpg
6c420518995d9de53fcf8be9c2090c2a
bcdff4888d27ad6405883c9d451ca5ff590bfbb8
6391 F20101206_AAAJKM baglivio_m_Page_083thm.jpg
9e42d875272b22575893e002e65d56ec
cf6f00ef02393388102f5c1cdb3f47f4c1030265
6905 F20101206_AAAJJY baglivio_m_Page_076thm.jpg
b7184e77413c9b301777a178b065fa68
c941b7808ce9e81cf27fae73ff1834862f744630
114474 F20101206_AAAIHK baglivio_m_Page_094.jp2
7cdd22c4bef7a1ab14b810abb9515061
e597af3104dae86652552f567710d682022cca96
113318 F20101206_AAAIGW baglivio_m_Page_079.jp2
46024086adfcef8ddc2dce639d5f2f8a
12cc4f9f1b2ea8e8c8c0556164f4ae524c1ba177
28525 F20101206_AAAJLC baglivio_m_Page_092.QC.jpg
16470726b70db21afdad0bba6c58fa62
b2ab57c4805fb92bc76c330181f31fe45ee13419
26209 F20101206_AAAJKN baglivio_m_Page_083.QC.jpg
cac0d0294417be0712245c172b70766e
8a9fcfe549507fed85a37800ea3a505c5f081178
29392 F20101206_AAAJJZ baglivio_m_Page_076.QC.jpg
806f4648781a79fec61c71afda1bd0cf
afc8a05276183d420b5d475d11d860577a09b765
111319 F20101206_AAAIHL baglivio_m_Page_095.jp2
4b4d466b0864e86c18623c360a88b34c
63a45d5be98c38960b6f69e0d936409177f86bca
108311 F20101206_AAAIGX baglivio_m_Page_080.jp2
8b2fe8dcff85d129094354f924c8504c
2247000619185c26147afb81f8cb0ff1c4f91eef
111523 F20101206_AAAIIA baglivio_m_Page_110.jp2
382ac2e26550dc0843d0cd5349164992
d8e8cfa330559c5dcbd1556550a33dbdf9e747ac
5930 F20101206_AAAJLD baglivio_m_Page_093thm.jpg
40b848420c2699ac1bdc8781cc9e2321
61af296edb4ba54e723ddf15500f75376991857b
27536 F20101206_AAAJKO baglivio_m_Page_084.QC.jpg
a3d6fc15a2f6e5f1c1195807ec75b0db
543aaa8c8d50bf53c8bf4e673fb3806a594cde86
101793 F20101206_AAAIHM baglivio_m_Page_096.jp2
f3f309c4b0f787a699abee414c81c652
be1c7547357c9e0c2cb8a810900fcb15df579f5f
115178 F20101206_AAAIGY baglivio_m_Page_082.jp2
48c284a3b7ca9b1ccdc9797bd1c97482
fec9aadd4a9b80c5b657ebdb07c9c9e0af7c7203
108028 F20101206_AAAIIB baglivio_m_Page_111.jp2
2c300aee8880cd58066660193e9de75c
641267058c601e7be953fa58ddf0cf014ca24036
24811 F20101206_AAAJLE baglivio_m_Page_093.QC.jpg
555d73dc2a4b21f3243b279a95e43001
9231eb235620a0e2becd0eec0e3384a5f575179b
6250 F20101206_AAAJKP baglivio_m_Page_085thm.jpg
966109ded6f9c0feed580e85ae24577c
2dd381a07644fa08b323880007263c9ed424d7ed
96986 F20101206_AAAIHN baglivio_m_Page_097.jp2
ad328cba4871fe97221b6646c905b234
20766ae66a4c2342b2cfaf180709073ae62cf183
112441 F20101206_AAAIGZ baglivio_m_Page_083.jp2
0e650ed45945d397a77cd16890f3be23
890f628a7dfbd3d53a23162acb2b06bc07302b73
110828 F20101206_AAAIIC baglivio_m_Page_112.jp2
0b9685c13e7c1501cdfa4622545f710e
4adc3209daea09ee1322d9f1e584d6373024a213
6664 F20101206_AAAJLF baglivio_m_Page_094thm.jpg
eef389be74f97873dc9f3d2b58924990
185d743a39166950cc9b3818dd2b94eeaf344fe5
25552 F20101206_AAAJKQ baglivio_m_Page_085.QC.jpg
3391c67bf90244b64db4d02c064d4bf5
508a4935e87aee366f394d0d0d5007e76640844c
112538 F20101206_AAAIHO baglivio_m_Page_098.jp2
bb8bbf9ac4293ccc9e4b54998c1b4e05
54ccf516cef1d82b667bb4f2a2c7bd386d1080a6
114110 F20101206_AAAIID baglivio_m_Page_113.jp2
872392f672d2088172df6696199bf56d
53543cf669000fb72214cb090a543c484bc84a53
27328 F20101206_AAAJLG baglivio_m_Page_094.QC.jpg
450c82e7b9c03a4ba8f9a409f0e9de9b
7b6e117dcf85351f637a793d9f716b4d1f6040de
6719 F20101206_AAAJKR baglivio_m_Page_086thm.jpg
63090a55ec976d9c4509dcb66965074d
8efb7ebbc209ff72cc93ea9d72b13b176e42f83f
106036 F20101206_AAAIHP baglivio_m_Page_099.jp2
a4c9496b3622010696a28cc643d34575
295c401e558c048f900f8780570aeccd20b179ee
111011 F20101206_AAAIIE baglivio_m_Page_114.jp2
62904fad759a51e97b84053dc4b2f71c
fbf1e14f30bdb6d5a0f46033d6e94fa38ddc8e26
6469 F20101206_AAAJLH baglivio_m_Page_095thm.jpg
43dd46a8b5f707fcf61ee2dd8d28bd50
7c3a50ffd9e42ec52a49425d685a13bfc444b121
27873 F20101206_AAAJKS baglivio_m_Page_086.QC.jpg
77f1cb910edba6a5c8e50696494e47a9
c2d8285a8a27860352630c3f3bb790fdc5ffe549
117167 F20101206_AAAIHQ baglivio_m_Page_100.jp2
1e7d70292106e656bc27cd6e7dfcc084
e6e1814c4b2bef951efd3245050a91172dae9e56
116940 F20101206_AAAIIF baglivio_m_Page_115.jp2
de454be3bb1da113e7531b7bc6321c80
b879b50c3fa75ea5cffd2d8f6ba1425707aa7371
27252 F20101206_AAAJLI baglivio_m_Page_095.QC.jpg
ffdd31f101af24f0c19c17782c2f3cb3
55101e6b5c22685b4ae75b0df6a4497eee4241aa
27706 F20101206_AAAJKT baglivio_m_Page_087.QC.jpg
268fa0b4d0bbbba0595617d758638231
31cbbcffbfb5f1f6789ca4ae439a91729e5f7509
119157 F20101206_AAAIHR baglivio_m_Page_101.jp2
088a9936da9ecdce9d9f850f44cf988b
c1ec4bec898eadda477827f962d0025db905ace7
119322 F20101206_AAAIIG baglivio_m_Page_116.jp2
0fc044e81ac2f4a1421087a774efe942
166d3ff065b9f671e3957f79b36f464ea22f44e4
5924 F20101206_AAAJLJ baglivio_m_Page_096thm.jpg
389fc6d1b12a747aab769c5e199bcae1
a4ad28438282ad631ed9e1fccc6d4704c94e9ef6
6638 F20101206_AAAJKU baglivio_m_Page_088thm.jpg
f388a58c66ce3027bbd666ef4d1ee9ab
17fdf6b7c792292c3587657eb72ffdaaf0bb2e8a
102784 F20101206_AAAIHS baglivio_m_Page_102.jp2
665667b2508cf7ac21d8175136f787c1
9264ab91883d6382bd7939037d1719fff1648892
115428 F20101206_AAAIIH baglivio_m_Page_117.jp2
adfdf38de416160b59d2f4230e8c95bf
b4519168d2be6513c968d4a22ee78e087ed9b08a
24489 F20101206_AAAJLK baglivio_m_Page_096.QC.jpg
bd7fe8cf11a144eba7594dd0293cd60b
3b22c4b31e39ccaa6488273af9eb44c921111923
28064 F20101206_AAAJKV baglivio_m_Page_088.QC.jpg
819fe7903b5c7673215202665783f70c
61cd25aab35d615cde4f56dbafc6b0b404bd2762
110603 F20101206_AAAIHT baglivio_m_Page_103.jp2
8839e95389c57e34b1e2d554515be309
ac320f8ac661daad8cc19587560da1d3c4635eca
120245 F20101206_AAAIII baglivio_m_Page_118.jp2
f8a1364357e19e6de669576bc947a474
b45c5a2e2ec276b2d2bd31edaa4ae801cffddae2
6051 F20101206_AAAJLL baglivio_m_Page_097thm.jpg
eae9c18fab0a9e55a4e50f299f6ca1c0
6830ee8f3b4f76ecae071fcdb355a9c2a8290002
6299 F20101206_AAAJKW baglivio_m_Page_089thm.jpg
20024975185edde19158015759e29a42
7d5d45144ed0094082973738f4ccca79a87a8360
116885 F20101206_AAAIHU baglivio_m_Page_104.jp2
0ea35e4e6a3bc8622b1801b21a67fad6
7a8b356551fbc0b82063e8592f7fd386a0ebca4d
69686 F20101206_AAAIIJ baglivio_m_Page_119.jp2
50e3a0ea017716e48666fb794efed1f7
e39fc74fbc0c5f2ecc683137fb30059b4593fdae
27000 F20101206_AAAJMA baglivio_m_Page_105.QC.jpg
5a0653fe4980cb2c7e4b2b3e28079273
b797b1095771269b0d3808d9d2a3714c3de13ac0
6545 F20101206_AAAJLM baglivio_m_Page_098thm.jpg
c48be0d9777117abff52cb57c8c067ca
292e4337c92117db95daae6df40f7dc8dca12b62
24484 F20101206_AAAJKX baglivio_m_Page_089.QC.jpg
12d606950cdd94cea9ff3a3c1b03448d
b182e7af2643fa725795dc3004095ca97c5656e2
114426 F20101206_AAAIHV baglivio_m_Page_105.jp2
0fbc37b9bcc42aac27e5ce7cdc000f2a
f28be5aac226ce8da9f283abbd0f157815c8c5f1
113287 F20101206_AAAIIK baglivio_m_Page_120.jp2
0729086169a7deaf80235403f27194fe
5335fc61c967c7fde8710871af1af63003b9f5a4
27215 F20101206_AAAJMB baglivio_m_Page_106.QC.jpg
2e0d9dc98b56932557c121a3ad848948
4c17556da34071e6026551a863625f94eba6c886
6366 F20101206_AAAJKY baglivio_m_Page_090thm.jpg
e14068bf97fc49aa6f7064dc0ed2d8d6
6e0f28e557cca639b9a76873fd227a6eca67ebd9
114904 F20101206_AAAIHW baglivio_m_Page_106.jp2
1daf429a2d71bc5e1c71c88658039399
ad797bd75f7b31b086c413c1d048888acd71f908
6549 F20101206_AAAJMC baglivio_m_Page_107thm.jpg
97721b48bab90c8198166eeb7e43312a
cd38e95442d0be567ba383e609f3afd4490b3faf
27005 F20101206_AAAJLN baglivio_m_Page_098.QC.jpg
5bb0c243c3a4cbfe1b8c36ad43f88af1
47cb9c20802eb577830dcfed4ee9c7b9f73d1530
6008 F20101206_AAAJKZ baglivio_m_Page_091thm.jpg
c544f4bf9d778699742587a10ea94b7d
ada58e8a405d7fc98944cb9298d4dafb0d7bbe4b
111045 F20101206_AAAIHX baglivio_m_Page_107.jp2
f9f22580250f1e12215e097cc29d6f6c
ae0c08a36862ec45ea913dfe18c698a527653723
1051963 F20101206_AAAIJA baglivio_m_Page_137.jp2
0b4e31942b1d6a7f3da29229adcf8190
f55242c280c447cc5c9d4951e8872d5e56f6dbb4
97928 F20101206_AAAIIL baglivio_m_Page_121.jp2
c4658ffaaa5d8859a4929f4b70131c2e
645c1fcf0214ab4ef90756e6a7bdc7be9686e1e2
26268 F20101206_AAAJMD baglivio_m_Page_107.QC.jpg
41945f8e3597fe7efd87598d19977583
2bccf725201614a67d7eb6e7f98c79102341c8c6
F20101206_AAAJLO baglivio_m_Page_099thm.jpg
9b2c4ad5aa575dc386657ec7e5c30bcc
387cc92b71482f00432799d6d14d3432d9fd83e4
105072 F20101206_AAAIHY baglivio_m_Page_108.jp2
b270123506d398c16880fdbc0a944390
fc7f47b41973b56bc0ea6dcbc8435e3c90f509a2
F20101206_AAAIJB baglivio_m_Page_138.jp2
f574f2a4db04c77ea850361e0c9f6ff8
a641928037835695efd1471961b19f9807e13be6
114212 F20101206_AAAIIM baglivio_m_Page_122.jp2
8d274322506f440dc04dac1c823e1add
f4935404d155c19daf6ddbef3dede18b801e0177
5815 F20101206_AAAJME baglivio_m_Page_108thm.jpg
3c6d0ed9c530cfa7b55e112cce8baf61
61f1867799b1e51fbbbba04551232ebaf269cea0
25661 F20101206_AAAJLP baglivio_m_Page_099.QC.jpg
e098206980ce98df9e8aeb8f82a25fbc
72e2d75e5b9b5484d844d371a26d22cff462f3b0
1051986 F20101206_AAAIJC baglivio_m_Page_139.jp2
5fa67dffa79e31d386a0fbc96edad6dd
fdf77a62eba3a4b6d160c692e62268b39ba2c627
95973 F20101206_AAAIIN baglivio_m_Page_123.jp2
5ddf40dce3f8f0d0697c6b30d9caafeb
4838062c9ae3ab2cd9924f866f8d267fccfef0cd
117570 F20101206_AAAIHZ baglivio_m_Page_109.jp2
9e6ddda1666f6a01c254b83ea7d007b4
8e0132e55017b3c4470387eefdfbb9300147dadb
24767 F20101206_AAAJMF baglivio_m_Page_108.QC.jpg
ccf47dfe68903ec1c0cff1d622a4bb4d
80b7cc7672325eeb9487d2f7a33efd333e429f43
6647 F20101206_AAAJLQ baglivio_m_Page_100thm.jpg
c272a4bde605248a3f4adecc9b06eeac
2db202a9a01105c238070dbd884b0c5bfbcd5840
133519 F20101206_AAAIJD baglivio_m_Page_140.jp2
908a859398b9fe3122c9d8c41e4a4789
55f9240e97816b87bef3b2dab6371c91c92e47bf
839596 F20101206_AAAIIO baglivio_m_Page_124.jp2
7c7c11b4122de2552ac993ce0bc74f15
8a78f9c48e423dd046c12f85396669e04c153b60
6432 F20101206_AAAJMG baglivio_m_Page_109thm.jpg
5e43e3491c4e4e26a5cfac5f2c6be8b4
c14dc6b6515ce71e001fe138d165c40bab8f166c
28509 F20101206_AAAJLR baglivio_m_Page_100.QC.jpg
c27949123ee21676a9de31271c799cd7
fc45a43cf4fb1227c921487fd08fbbaa43477607
F20101206_AAAIJE baglivio_m_Page_141.jp2
f34fcffff5be43c294281de3c46682cf
f9120095ce4e819976d5d2a54a28eeda44ff5b3e
F20101206_AAAIIP baglivio_m_Page_125.jp2
9fdde61c3d9ee13ec5fe0460bd7cb1d0
612bbe0d01bd221f685b41f494f33ecd4be3ac00
27616 F20101206_AAAJMH baglivio_m_Page_109.QC.jpg
5d813dcacfb0aca1ad8c838cb6e8dd24
5d1abc4cb1122df04c2c7b8c9e5dcf5912ac16ae
F20101206_AAAJLS baglivio_m_Page_101thm.jpg
c7235f00b5fde7f00580da8fd01dcd47
647c268afab44f5a8dbb928bbb4fda112f712f54
F20101206_AAAIJF baglivio_m_Page_142.jp2
9c1e59f669c4cdc78615c2fa1222834c
49735571fe815486410ed67ec29c54a032abab51
108176 F20101206_AAAIIQ baglivio_m_Page_126.jp2
61ccbf4e1bbef5bce7936fb5d1fb2355
9c60f58c27229ed3e1c8f51baaa08bb857080fce
6342 F20101206_AAAJMI baglivio_m_Page_110thm.jpg
f113a47a57e6614967515bb175ec7b94
83b483d2e0273cf3324cb727b9929215afa16b1e
27795 F20101206_AAAJLT baglivio_m_Page_101.QC.jpg
d7b57e68f32fc5a1ad30fe4e1c4cbe28
8492be55c671bff4082aed3f280649e83d263edb
F20101206_AAAIJG baglivio_m_Page_143.jp2
d03e31240856501f676e6a6edc4e6abd
49c60f6eb82856925e7c73353afa655539299fac
1050239 F20101206_AAAIIR baglivio_m_Page_127.jp2
d3e81253b52ba8db4f26b3aa175008e0
cbf1cfa172695aca3eceb12868536fbeaeb06fd8
26401 F20101206_AAAJMJ baglivio_m_Page_110.QC.jpg
97274f60c829669a265fa4fdd55086e8
71762bb58769147a97edf9514d4679023242f7d9
6237 F20101206_AAAJLU baglivio_m_Page_102thm.jpg
8d24d89ba49ba7deb544fc2cccbd4296
c1cc44bbcd5329aab7d0899cf7fb1f1c5a73f134
1051910 F20101206_AAAIJH baglivio_m_Page_144.jp2
b0881cf181de22b1d17ca3aa0a46dfef
81213b9a91e9f9b8564a7958233a0b6096dc21da
52709 F20101206_AAAIIS baglivio_m_Page_128.jp2
ee4670974d9f186aa5f352edc2fa5232
e5b71ffa0bc56668e4accb37e26944ac2a34ef7d
6332 F20101206_AAAJMK baglivio_m_Page_111thm.jpg
123b2095248ca6370422ebf514779813
ff81b27b9a6d702f6cd5fb342a18ea9b19eb7564
25063 F20101206_AAAJLV baglivio_m_Page_102.QC.jpg
00c1393d64411f27887497ac5ab6e1ac
fb2279e52ca1bdd7ad4b46ac237280ab3466cae7
1051950 F20101206_AAAIJI baglivio_m_Page_145.jp2
f621366ea6577a9e807c47df896a0b7f
2f673192ed212d285c9e1fcadcc15a2ee585155e
152630 F20101206_AAAIIT baglivio_m_Page_129.jp2
246848e378e08aa5e97e125249dc7c34
2e029e6ffa40c05ebd16fe8094fc9944b1fd2e4e
27122 F20101206_AAAJML baglivio_m_Page_112.QC.jpg
0548bf1f5b788eaf492d0965cf3d2ae4
ac8ada1b7a04cfacc3c7832a717c043a13ccbbf8
6528 F20101206_AAAJLW baglivio_m_Page_103thm.jpg
1b7d96ffd8b33a8537354cd8c995127a
1f80ee7872c6863fa3a65b20718c24ab4e8e9fb3
F20101206_AAAIJJ baglivio_m_Page_146.jp2
81a23b76412749403d61ef5fa1f79950
cd6f90f0111e9b5a5b77acbb2ca9e627992f84ea
141661 F20101206_AAAIIU baglivio_m_Page_130.jp2
5b3f1a6bc37b1e35ce6c3f58eb3e6425
0d5d89d3ac63765c8dc2455f73c428b67c850ec3
6486 F20101206_AAAJNA baglivio_m_Page_122thm.jpg
b670e1c11e465e8cc040872c7362cd10
e8a59d00e43ac52732c6c24c28ae7ed1b84e6742
27581 F20101206_AAAJMM baglivio_m_Page_113.QC.jpg
bc9895d6468a79c777ee0603707d4405
e0fdf3b6f125bef84c8a533d454dfccdcecd5624
27034 F20101206_AAAJLX baglivio_m_Page_103.QC.jpg
56dcc409ad568774b48b9313e6390cc6
eec6034fbc2745a37ff575ea684ece72797f62a2
1051952 F20101206_AAAIJK baglivio_m_Page_147.jp2
0fa6ded2b071ae3261c87cdb2a49bbbb
409e4fcc62ec6d6d0434c64b4d6b86ae0b2e8cc9
94346 F20101206_AAAIIV baglivio_m_Page_131.jp2
1e99ee7fc24a73902dc34c43bfad1118
358816ee927ffac7b99781e4c3d51aa0a981e3ba
F20101206_AAAJNB baglivio_m_Page_122.QC.jpg
c87bd48e31a4ffc502de37aa4691a0cb
42c6fdc9fdfdb491ae183d475719eee2cfe9328f
6444 F20101206_AAAJMN baglivio_m_Page_114thm.jpg
5aa853d4b97d7c5ff2dab9d8c83a8d91
8dddd89f30a5ff0c7a80e03ab5de3888a3a6007c
6586 F20101206_AAAJLY baglivio_m_Page_104thm.jpg
33ccd42c9a80cd9e911d63e636aa225e
db9bd700acffc7cd38f4db361e99c9ee97516c61
102873 F20101206_AAAIJL baglivio_m_Page_148.jp2
940e54b2620875947e0e889d9f6ea1c3
eb041e8092a3679d3e1f9ac9da230593873cd434
65615 F20101206_AAAIIW baglivio_m_Page_132.jp2
f4ec66b2e277db023ac12b2057eed455
a8cdf6b6bca7c08585f50c9a3c8100b50847ae52
5508 F20101206_AAAJNC baglivio_m_Page_123thm.jpg
a26b4be38ab073a5fa7d62d9115725cd
3d3714c18e9825daa286df7ac20bd5ade65bdf48
28192 F20101206_AAAJLZ baglivio_m_Page_104.QC.jpg
68ed359971e1b3c9f502e0e15261874c
cecd47a70ad1bf8c9fdb1150b4801f97744d21d0
87270 F20101206_AAAIIX baglivio_m_Page_133.jp2
e27a4e737f3321318f535255196ccf3c
8f42d6cf975cb7a659c6d875c22e2a2084828769
137372 F20101206_AAAIKA baglivio_m_Page_163.jp2
c475fe88016a88065112a1da66f618cc
31afd793897c596bb5d5636b35b5c9a545091391
23621 F20101206_AAAJND baglivio_m_Page_123.QC.jpg
8f3b0a90da3726e567dc3fa8092a4598
26f950b86cdd710fef22ed30e4747da726404ab4
26665 F20101206_AAAJMO baglivio_m_Page_114.QC.jpg
0454a4352787178da1ac1a3d008963f7
51a6020d63e2916e5ddbe51010627b8b43dad9d4
1051982 F20101206_AAAIJM baglivio_m_Page_149.jp2
fbe333d213d54bd653e4f773f9ce6dea
5dc686a23fad0d10d0fa39c953f6c4373999a79c
101246 F20101206_AAAIIY baglivio_m_Page_135.jp2
9cbcac6185ab635552b07fdb32d1b57b
f2635f18986739766caec6fd831da983a44f9780
1051932 F20101206_AAAIKB baglivio_m_Page_164.jp2
a104591fd32be11d9df84d201ad8d413
757fcc7e8d89854411e4a22e3f2024c0295dcb9e
24217 F20101206_AAAJNE baglivio_m_Page_124.QC.jpg
ffac9595ea95fbfe2fb74c129384e2c7
2c06f7d6308de07d3e1d599066ebae1883d38f5c
6626 F20101206_AAAJMP baglivio_m_Page_115thm.jpg
5f9ed19b76e8fcb15269a3e8522f1ea4
6457ccb55256961b40862905ef873a6940b831fe
1051935 F20101206_AAAIJN baglivio_m_Page_150.jp2
c8db9791aa70fdd0b621ce599f8c6a6e
8a43a72af605f905f9e793a882395c7ee0e26b76
32202 F20101206_AAAIIZ baglivio_m_Page_136.jp2
0646297a76b73fcc103806ddc99d914d
1534b9e3ba71691468b5a3324ab278c98d5cc7d6
130538 F20101206_AAAIKC baglivio_m_Page_165.jp2
b7137b5e5960cf930edf1760fdecc1ba
23a05a83e4e509b82b290c10292631f682dd7f99
6565 F20101206_AAAJNF baglivio_m_Page_125thm.jpg
1fdfa22bb663741b8e57eece0d36b911
550ffc438f93358cec8a866f9257e4cf2599792b
27674 F20101206_AAAJMQ baglivio_m_Page_115.QC.jpg
bc90299f571c1fc6e66f6212ec18eccf
3a6c029746483a3234c9140db344954596bd90ce
162140 F20101206_AAAIJO baglivio_m_Page_151.jp2
339da6fb6ca1bf3732eee81832d406ce
3ff6bacbd6aa4d60de6d5440cd3787eb02483cc5
66223 F20101206_AAAIKD baglivio_m_Page_166.jp2
c870bc0fb45f345dc739c3ea46c5ad68
d225d3c96918de3926cc6c128abe6c4f898f56e1
27468 F20101206_AAAJNG baglivio_m_Page_125.QC.jpg
a1e7bc1e58e14c7155927621e6667184
778942a6099eb9849c190dee5df165464decf7da
6797 F20101206_AAAJMR baglivio_m_Page_116thm.jpg
eb3083dca299c0a9037ec3104c49c00b
f16624aced51dcb11f88e042344028fa75f622ed
26407 F20101206_AAAIJP baglivio_m_Page_152.jp2
eef89ab8feab4a32dbb6dce1864780f9
456fbe912eef20f2496eda2c3a1113e09226bfaf
42586 F20101206_AAAIKE baglivio_m_Page_167.jp2
7b3221eef37698f4534f831de1a474ab
9aecd0d96b91705574ee43009dd267f2b4891039
6353 F20101206_AAAJNH baglivio_m_Page_126thm.jpg
93f40740fca62bfbfe5eda99e6bbb384
ba1c0899c7c6eb43299ab15de8d3bb7c747d4fb4
6558 F20101206_AAAJMS baglivio_m_Page_117thm.jpg
1e2acd99c0d8b1b53ffb2832d528c0ff
579fedbacd1d86c5680d28354ccc9e813866fcc1
144092 F20101206_AAAIJQ baglivio_m_Page_153.jp2
7a6c7e38d438e5bfa8db1f009bb54312
ae5216b12902479ed4531bd6ab2b748a605efcfc
F20101206_AAAIKF baglivio_m_Page_001.tif
2c22ddfe1376cabb138681ecaf4463cd
664da74c52b14c08f97dfcecfefc1c181d36153a
26049 F20101206_AAAJNI baglivio_m_Page_126.QC.jpg
d07b8d43b1294d12422a847cb2431a92
715d7906aa66106775f703f096e2885ecd34249e
27818 F20101206_AAAJMT baglivio_m_Page_117.QC.jpg
062371f9c657bbdf335032c468294532
3d4492942a7e9793feaae2ef55f4d308791b1a5a
131413 F20101206_AAAIJR baglivio_m_Page_154.jp2
e92996f4ed6178c36fd280657a3ab7c0
8e41d65300d0e581d74f4a357322f86ed7ca991b
F20101206_AAAIKG baglivio_m_Page_002.tif
34659843d6fcb4c613ae3234fd4303ba
24d1cb5049080bfac0ada01d932b8497b371ffe9
6577 F20101206_AAAJNJ baglivio_m_Page_127thm.jpg
99d9a24b5123bd58a9bc04cef538b99a
0a239dad15b1be270c90ee4c74bc122f0ddf3c2f
6895 F20101206_AAAJMU baglivio_m_Page_118thm.jpg
7c7fa196e1bf63618721e51968910b21
ba5a4192eec54fd02ad773f4b924adccd97209f9
126692 F20101206_AAAIJS baglivio_m_Page_155.jp2
7a4f3f80d7c92faba974c3b7bce3e6dc
c996f569b638e56598cea34c3b5e2de35100fbfb
F20101206_AAAIKH baglivio_m_Page_003.tif
2be99f51df2bfe8cc6eed383df53a973
6de9aa0790ed687a7032b80d1da93b48594c8f05
24057 F20101206_AAAJNK baglivio_m_Page_127.QC.jpg
0dde2b45da4749b08f668980bdbb6239
fd9167b35e50175344774f1bf47b6dc13ec73db0
4172 F20101206_AAAJMV baglivio_m_Page_119thm.jpg
50f8bd25433470921d61b1afd579f953
87ca22e1dbc55b8f9bdcf97ca6c45f1254d4920d
135133 F20101206_AAAIJT baglivio_m_Page_156.jp2
b22711e71a044381e85b04dd3e6b9a89
1acaae3184331e737718862a89c7bfbf5bc864c5
F20101206_AAAIKI baglivio_m_Page_004.tif
50dee0db377e7fdffaa7b36ff5f82b76
8172a4eb597dd360617984c32ae94e09c0476fa9
3279 F20101206_AAAJNL baglivio_m_Page_128thm.jpg
dee81c37ddf26634aa4296196950c4f7
4061c930213b51933470fa9441b2d3e8d3d4fc2a
16827 F20101206_AAAJMW baglivio_m_Page_119.QC.jpg
d0b6b425277a6ca8a6c83594ee16fa68
ff139eda7aa05075225e2c425ade3b1c764a3593
132254 F20101206_AAAIJU baglivio_m_Page_157.jp2
c57a42998c9760b3969fde3c6e1f5657
f255c50fe16fc2b9e5b046308878906ea97121cc
25271604 F20101206_AAAIKJ baglivio_m_Page_005.tif
7146a04ef3e618703d1c0be31e4f7104
8dc8e36de184030c6e2c9351afa01323995bf59b
5838 F20101206_AAAJOA baglivio_m_Page_137thm.jpg
f84fb6b8b2e1a3b61e96993cfa1b0fbf
6431dea9a6dab6d1d899d50cad9641d8bd56ced4
13435 F20101206_AAAJNM baglivio_m_Page_128.QC.jpg
5333449fdd251b93caa153c9f0f0bb1d
168dad66db44ee601228eb4cde7b3eb0b0de0fd0
27457 F20101206_AAAJMX baglivio_m_Page_120.QC.jpg
c759cfa01bac4c2aa42343ce618c808e
2aab96936ee1fc7092ba020861d175ab1b4f643b
131477 F20101206_AAAIJV baglivio_m_Page_158.jp2
90af04c75faff55b67864eaedc0a827f
234fa8b62f41fa9b9817407bbf94287bcd704196
F20101206_AAAIKK baglivio_m_Page_006.tif
238e99ce5c2f3114e0407a9a0662e72a
bce139dba54471a8bdfc180b78a6b0cd2cb6b18f
26193 F20101206_AAAJOB baglivio_m_Page_137.QC.jpg
bf1f757a1aa12adb4dd86e4a609f7d01
ea3463b24eb33d136db6686a0e012da43e6ffecb
6799 F20101206_AAAJNN baglivio_m_Page_129thm.jpg
f5bc790cf885c4129b5a1e37e36acef6
ab1e74046024ec53b75219077294dc353e0a9857
5485 F20101206_AAAJMY baglivio_m_Page_121thm.jpg
592639c64e1342acd7e85da10639a5d7
98649b5c2a3bdb2a764e7e3cd338b0487a80acd4
133340 F20101206_AAAIJW baglivio_m_Page_159.jp2
15b4e1654f55a7140d553e27040b39bd
955d3546ea37d180dbe08fcf69de6328b7ca0c57
F20101206_AAAIKL baglivio_m_Page_007.tif
733ffa14d55f9dbffb2861d56760d339
d14e4eb72f1e549de304241337c34d1d776705f6
7136 F20101206_AAAJOC baglivio_m_Page_138thm.jpg
34a74736968ead7375473cb83f8128fc
41306f74de1ba68e9dedb8df0029f7ed674bc3d4
27078 F20101206_AAAJNO baglivio_m_Page_129.QC.jpg
9b088872debab1289cf9c5b7dd75b2c1
e26cb22717304cbe5e202d94e505cbfdf1604ba5
23810 F20101206_AAAJMZ baglivio_m_Page_121.QC.jpg
89f7ca9d91d0cec0c3fac0e2d48ead41
471e4922db8c8a81395a6be52e9b88bb589910ae
133712 F20101206_AAAIJX baglivio_m_Page_160.jp2
2315781a14bbafdb1f94f12f70fc3173
0445222ad4ea5650de82f07469f915e72301d592
F20101206_AAAILA baglivio_m_Page_022.tif
15c74e1c67f0c3948cdf91588c730c2b
4f8559961ce95ff52e6331137cf0e356a990452a
F20101206_AAAIKM baglivio_m_Page_008.tif
7e4459b2dd11c31cbf3fb0040c6782aa
83e083a4679fdcd90a4445c5ab329a9401a4b607
28199 F20101206_AAAJOD baglivio_m_Page_138.QC.jpg
a69d50527d3ee34e518f3b2311448591
5f83240f5113e777bd2cf03fcc1429d6eff40592
138799 F20101206_AAAIJY baglivio_m_Page_161.jp2
3265defa457b2ab5e55f812480d00e20
b10bf0742318adc4aab08e9e3871c0972511cb25
F20101206_AAAILB baglivio_m_Page_023.tif
c244314f08e74645a9839974be4a6990
c8ac214016e3c55ac1fe4a1f03a3ea60e4cd4ba5
6118 F20101206_AAAJOE baglivio_m_Page_139thm.jpg
150538405afc835e1ae0cedfa6f15453
840097b0eb8c25d9fd7551b4ec09088931ceba1c
6600 F20101206_AAAJNP baglivio_m_Page_130thm.jpg
343be505de193a2d6622c94cafd74204
2f656386e4005529fc5ba2249b5b2b3ff649294c
134975 F20101206_AAAIJZ baglivio_m_Page_162.jp2
143d66b9742ef0c82ceec8149a797027
7377e22c02397c932c2a22ce1e52451a51e3e59f
F20101206_AAAILC baglivio_m_Page_024.tif
4c33e07e4956a37939978255f7cfd90e
71616d912212108daa9d0a30aac792963405fb0e
F20101206_AAAIKN baglivio_m_Page_009.tif
7ec07092b3bfcff0ff5e0f8a02004f7c
5b4c5e9b20668d131d7e6411c0101300b850ae65
24362 F20101206_AAAJOF baglivio_m_Page_139.QC.jpg
49d8939c732df59467255fc7aa26540e
ca74fc9cf6ad3eef57f648e215f0db07e07d80f6
27502 F20101206_AAAJNQ baglivio_m_Page_130.QC.jpg
1914604ecd0fa177c85ac75cf59b472f
79d02b4070e627f91c025fca379ef04866c45726
F20101206_AAAILD baglivio_m_Page_025.tif
361e37564520306eb7bff4b1680d4f4d
77795e03b00a162d5eef30e5db4585a8bbea6a1e
F20101206_AAAIKO baglivio_m_Page_010.tif
a555759de6d6539b8f137d083e753815
32df79058254b9e49f705b15a2e235f0347b4ac3
7198 F20101206_AAAJOG baglivio_m_Page_140thm.jpg
5083d35106b935b0056d5f56cea3ac0a
9f697578e14b73c292d24e80bca25e6b3dfb6bb4
5313 F20101206_AAAJNR baglivio_m_Page_131thm.jpg
362c5bc2b0218bec1af052b661ed1759
c38a49f2b7be976cdce41b7b68009d62fee16c81
F20101206_AAAILE baglivio_m_Page_026.tif
1cf3c6eee3a799a8826c7e6fd4ea04d1
39902dff88aadb0fe536f07b88bacde60719d066
8423998 F20101206_AAAIKP baglivio_m_Page_011.tif
8cc276770c74c411e02382ff29158b23
ea43833de259ce453da20700b24cd4bc4bf2d8f4
29435 F20101206_AAAJOH baglivio_m_Page_140.QC.jpg
ff8fa897d2bf3b120691ace8a8efb137
e32a4f135c612efd504e6e38e860a51dc287cd78
20717 F20101206_AAAJNS baglivio_m_Page_131.QC.jpg
5a036aa07876d1a60d2b51d1e8ed0398
2e90a5c09ad10f21da45037a9710747982012dcb
F20101206_AAAILF baglivio_m_Page_027.tif
f03f34b0676241678e918cd47d960c8b
31ccaaf1715ddb9dd94dfd55ad052d38af91d976
F20101206_AAAIKQ baglivio_m_Page_012.tif
49997a38c2fcd08fcf0c882d70fffc40
91b8286873ef400c1f1a586f0be549829660f961
7208 F20101206_AAAJOI baglivio_m_Page_141thm.jpg
579e3a2d8cc7c19fa4c609c5ef1d4f1e
95482908053aac31e7cfa3238c243d455de73a30
3904 F20101206_AAAJNT baglivio_m_Page_132thm.jpg
14fda49a383c0d7d5b871907b6eebf7f
743fb33611199f1da2073971e1a6084edd24b89a
F20101206_AAAILG baglivio_m_Page_029.tif
37897805e55cbaa4629daa0b91277fdd
b0ead7d82413a578ff4b045314034ee71e3ec154
F20101206_AAAIKR baglivio_m_Page_013.tif
f2099af7c66e8da986b7894118ef971a
9b2c45f96d23a0f1039a3dd8e469343cd32e7cfb
F20101206_AAAJOJ baglivio_m_Page_142thm.jpg
cb935609700b89c42349c07d53130e5d
dca7d4f3ce1a00ab02002900215ea45c05293d66
15483 F20101206_AAAJNU baglivio_m_Page_132.QC.jpg
6790739ef9dcb291b31dc0d0a06df8e5
a70f5e8da6db1648c0cfc23b721c8236d943240d
F20101206_AAAILH baglivio_m_Page_030.tif
220a8006e62181d2497d1f8417ae84f5
ecdfc90ca2062e0e669135bc8491e54849271b7f
F20101206_AAAIKS baglivio_m_Page_014.tif
12105b0b14da72ebe498fc8e80d069c0
763679b9074e87cfec5bc61d49164ff370b1495d
25752 F20101206_AAAJOK baglivio_m_Page_142.QC.jpg
b16c75970f44f21897781ec8a9943d10
d2a4c884e8e3acce162e5a98f0ea61bb79c4df99
19091 F20101206_AAAJNV baglivio_m_Page_133.QC.jpg
6f680335df40bee1813eb346e5d6a49a
0eeac9ec939b1b97f11d7f8e991bd1c0ae822058
F20101206_AAAILI baglivio_m_Page_031.tif
b8dc5283f3e2cf6772a7a25260257b15
29b95e8bd6e65322913c67dd87021b0f89b0fd14
F20101206_AAAIKT baglivio_m_Page_015.tif
edc6458ccd0a90a3ee88621658516232
eb48918ade58168044a0ab5baec1320d16091f15
6385 F20101206_AAAJOL baglivio_m_Page_143thm.jpg
bfc15729c62e996088fb5467afcb1e5e
58f863595c64a3a4f2d55cb1ce589cf44fd66403
21680 F20101206_AAAJNW baglivio_m_Page_134.QC.jpg
67d7771eca4102179df512d6b38a79ee
18bbd352f110792ffb1111ba71ee8f15e10ff2f6
F20101206_AAAILJ baglivio_m_Page_032.tif
1cf87e14eaede4a5158bbaf39adddf8d
55b07098395c6ee2fd3b036d051c6550ae624ef4
F20101206_AAAIKU baglivio_m_Page_016.tif
b3b323065ff46937271b0b67f936af54
5da24167dc58b51283dfa7fa19362238a470a0bd
31301 F20101206_AAAJPA baglivio_m_Page_151.QC.jpg
0fc02b4586e528333d58212246eff80f
f43a847616206c3adefb98b6bf2ef2bbc2bea120
25492 F20101206_AAAJOM baglivio_m_Page_143.QC.jpg
7babcddfdfd33b206257ef3d9912d2ac
be37b2d5ed22f98c8f0bd561d1bb9f94ad80926d
6139 F20101206_AAAJNX baglivio_m_Page_135thm.jpg
97f73e54405b73ed7cf7024ebf0e9e83
99fcb0473c2634438e42ed1fe2ff5f5a821eea70
F20101206_AAAILK baglivio_m_Page_033.tif
6ce7256aa37623ca22869c1f10dfd031
50a0707c6be0406bb2560602e241423555c94871
F20101206_AAAIKV baglivio_m_Page_017.tif
ad09053b9a24d5aeac9f5eda3ffca3e9
addbc3ddf1be494467bec291c9f82f5182035036
1458 F20101206_AAAJPB baglivio_m_Page_152thm.jpg
6302573895e17b097a4c075512f47376
f6911d10bd94eb9090c91844ddae9ee067fc98b1
F20101206_AAAJON baglivio_m_Page_144thm.jpg
7a8dad82dedf9f0fc2322164b0d83b58
d0f19be36f8b64bcf30e175caf710735c3b39c24
23482 F20101206_AAAJNY baglivio_m_Page_135.QC.jpg
1a4606a6c9498191a86a76f216edab88
61dde9a67c89d598db7fd9799bb206cce6f349b7
F20101206_AAAILL baglivio_m_Page_034.tif
f7108f97de39cdd58ad30c5ded3bbb7d
12582a498a8408ef3b94a6f3358de9aeed390ecb
F20101206_AAAIKW baglivio_m_Page_018.tif
fbe25d4f7e0e95049b6c58be4adf7efb
fee83dee9f0017222a0a25dc5e2a4ed6b3a709c3
5023 F20101206_AAAJPC baglivio_m_Page_152.QC.jpg
ae5d8f80988058c88addf812a5c50ccb
a3ec2f2a3352e97817883162c1ee1b6fe7cf8f3d
26847 F20101206_AAAJOO baglivio_m_Page_144.QC.jpg
4918f74391c63033e7a54cfb1a51258b
c4c67f05fc99ded4ef86f0a43621955bf6aa511e
2015 F20101206_AAAJNZ baglivio_m_Page_136thm.jpg
1f0c41936ee1a5139a3797f8faeaeab7
1ee282b07b177a02ebcd13b0a3cda0cee25dfd1b
F20101206_AAAIMA baglivio_m_Page_049.tif
a466a925978b19adb34813ce1ef44dfa
707c5bae377c5982436bebab93adb23b9858e002
F20101206_AAAILM baglivio_m_Page_035.tif
b565f1ee641dacc44da261d91e70ac51
4caef14ef3b6779441d8282d29595e1014461043
F20101206_AAAIKX baglivio_m_Page_019.tif
2d78aaf529d642baa4d67d5fd3605aa2
248a288f41de4345a400fa32934c3aaeff5a0de3
7008 F20101206_AAAJPD baglivio_m_Page_153thm.jpg
12731a5a32b639e2630f795019dfab3f
5d7e6ec0ab25f2742fbc41adfa3c6cb437983d40
5625 F20101206_AAAJOP baglivio_m_Page_145thm.jpg
950d404f2cbcbabb217ef907581d974a
b8cf47cbecc7a2f5baf0c7ec4c19e1d06b7e8666
F20101206_AAAIMB baglivio_m_Page_050.tif
9e50d1973b3b35e11621b8f83d8d7156
d7f1e6b2a6a118853d25bbfb466b45b5cdd37cf1
F20101206_AAAILN baglivio_m_Page_036.tif
1866c85dde70e55b571fa70d2c6051ae
50886fcfc86b0d5410efcc5b878ea1089c6f77da
F20101206_AAAIKY baglivio_m_Page_020.tif
9927464784908b248b77c13e87acd427
213328e13d90a7f77451bdfe880dd93c474e583a
28857 F20101206_AAAJPE baglivio_m_Page_153.QC.jpg
aac49cf7df9faf4292e8a40f170eea91
7e786407a59b5e768df3e4e72d886ecab7fdea40
F20101206_AAAIMC baglivio_m_Page_051.tif
b36492f7f7594558cce0d600aaf02914
53b00220ffd188dd0595a0b8d9db73b08a2694ee
F20101206_AAAIKZ baglivio_m_Page_021.tif
2487997b4924ca99faaa34efa1d59c72
5c875404cf4178a37b191738cb0e1ff8992c1d41
6535 F20101206_AAAJPF baglivio_m_Page_154thm.jpg
a8540a790f4e4343206ce09b87e97511
4ad950f405603cf96e410184d426294b6eb38494
F20101206_AAAJOQ baglivio_m_Page_146thm.jpg
1e833fe1c4bbbf0066dcaeb698d7769e
aa2225820df66276198fbd26fc9b66c14c5a8bd5
F20101206_AAAIMD baglivio_m_Page_052.tif
5febc757bbbe0ef9c726a09730218fb6
30845e3bc7a9c133c6df85f166778c3804bf9db6
F20101206_AAAILO baglivio_m_Page_037.tif
27cc406d5f4d60c862d1b7b76d7d17ab
7b6989b0cff18b20af8135bf4e6d05f18c9244a2
27218 F20101206_AAAJPG baglivio_m_Page_154.QC.jpg
c6d869f0203dc7b0940dc29c80bfb7cf
9ad9329d4f9df03a96104d16aca4baf1cd1ac354
26596 F20101206_AAAJOR baglivio_m_Page_146.QC.jpg
b2023b9500dfc93d37e875e60b1ed089
e05f7f24a2a232202431b936666a34d2fe2a5790
F20101206_AAAIME baglivio_m_Page_053.tif
f5f5a526204020c2a53aa1ddac2c5272
819cdb4c25061f3f2b9adec124e9672aae212fc7
F20101206_AAAILP baglivio_m_Page_038.tif
15dee03a61ffd9bb1f7c2c21aa703151
d7b2eaa748cfc764ab51c5585e80f90d98e51d14
26709 F20101206_AAAJPH baglivio_m_Page_155.QC.jpg
bddc3afcddf3205833140a932619a46a
adeba6c27e89dff89415c39e65c7a539b3440dec
5963 F20101206_AAAJOS baglivio_m_Page_147thm.jpg
0b3900eb921040a5f975ef1459a8c759
8de86929e8f4052ad6b41696aa95908124a0d0b4
F20101206_AAAIMF baglivio_m_Page_054.tif
85cb40e1930eafa01bd88c617e486619
9848671fc699983ec00ec2243615de26b6b689be
F20101206_AAAILQ baglivio_m_Page_039.tif
2e8f1e2881c556d2b04f2f3c7b98de57
bca75505f57f1a89a04277b90e075e16243aeb44
6998 F20101206_AAAJPI baglivio_m_Page_156thm.jpg
9b1b020a718c5b6ad1bd5e823632ae56
e43c433532cdbeb5642d0cea0ffd0766b1995f26
23925 F20101206_AAAJOT baglivio_m_Page_147.QC.jpg
12bb62b964f08513c05598ee81249958
ba4f785020f751f3ce1343b51477fd6bd1f87d9c
F20101206_AAAIMG baglivio_m_Page_055.tif
1734d8cf0b692c6930f75f2dcb3b985e
b3e3584e8865087361cc9b2b72a29e98a385485f
F20101206_AAAILR baglivio_m_Page_040.tif
5f414003245987ca7f8b6596b6c7ac82
e091ad1f6585ea69db8c5bd4cb0aa38b2640808b
28405 F20101206_AAAJPJ baglivio_m_Page_156.QC.jpg
54523c8a0696a750f23bc0e031b97ff9
5fde3ea1e650f9d108d29ba6b9f5c21f01813ba3
6466 F20101206_AAAJOU baglivio_m_Page_148thm.jpg
a776c6c871891a44cc1ac624805480c0
9513f4360f5d0baea1b299e041590610d112a45d
F20101206_AAAIMH baglivio_m_Page_056.tif
473a5b8805c40a978f55f365e426926e
162966229241ba7343e70ac0dfe097f277999cf2
F20101206_AAAILS baglivio_m_Page_041.tif
afc861af3e06a261abe1082000fc9ea4
df4da5ee55466b38eff616ffd803f9fd211767ed
7048 F20101206_AAAJPK baglivio_m_Page_157thm.jpg
96aa609eaeb084363be843d6d3a0c033
71c0c5efdb1a861e73a972145a71f2531ca4175f
24162 F20101206_AAAJOV baglivio_m_Page_148.QC.jpg
25cc2e3cd2cb0f9fe913315c69c7cec9
26f2edc6fb3e8251d6988796292a984045cf8f6a
F20101206_AAAIMI baglivio_m_Page_057.tif
8801d61216d07ce3a1efc83c4bea48e4
eec922a07ed472a306392d0f44420f7a3e43621b
F20101206_AAAILT baglivio_m_Page_042.tif
53f91861b0fca1e569527175d27d45eb
65e69c37090652783acc07552cba84637799228c
27147 F20101206_AAAJPL baglivio_m_Page_157.QC.jpg
f6711c91db2e3f725c5bc4f4b7a0d3b4
fa034527ad77ce9076fd9b7dac5e48f5913ad094
7269 F20101206_AAAJOW baglivio_m_Page_149thm.jpg
659000624f3cdbfd514f3f90bd8b931c
ec6c72974c393fe326cdb672631e52055bcfad61
F20101206_AAAIMJ baglivio_m_Page_058.tif
f9452e6de6f832f9013ed889e1a0aa7f
4c860fe14828e07ad92d3ac43a7c2ed62317da55
F20101206_AAAILU baglivio_m_Page_043.tif
7b24f43c1861cfbf3e463ecfd626bdd4
a71639dea09460e5a4079b29907b09296ee2b189
6786 F20101206_AAAJPM baglivio_m_Page_158thm.jpg
a4fa380842ca0f3df9dc1942d1ce6157
dce425962b2fb2b036a33f01ad5ce6cc641e8ba4
5696 F20101206_AAAJOX baglivio_m_Page_150thm.jpg
a625a4c79306c4508f6675aeeb7086fd
b8465edf01415661a77fe9ddd4ca09a2c8b8eed4
F20101206_AAAIMK baglivio_m_Page_059.tif
dfc8aa5ce63aee9ad614cf32cb738cf1
0437a5f1b1fd9aa89d82122375e74266d7620cdc
F20101206_AAAILV baglivio_m_Page_044.tif
151071ca97d5befc65f28c596d56392b
a50084d17103976efde30f726f37e58475bd7bd3
27231 F20101206_AAAJPN baglivio_m_Page_158.QC.jpg
2a92698da693debf36bd13d3c75fa23b
2a8f147ff029bddce0091a4f890d5ae1fd65123a
23760 F20101206_AAAJOY baglivio_m_Page_150.QC.jpg
a7f37ddc842db91a5655879b3fc51d0e
38f95604410b00e5c6e1bee26410da0ef5f494df
F20101206_AAAIML baglivio_m_Page_060.tif
fd4743f89da91b02ba83e76dfa028b97
b31152d2b3ac283c5e139e7faf0c26fd17e68104
F20101206_AAAILW baglivio_m_Page_045.tif
1e230f09845103aaa4a65059d91235fa
925e458c332cbf237c2d12fd49d9cccacd3bff02
27854 F20101206_AAAJPO baglivio_m_Page_159.QC.jpg
52cda20c6f29f82cd7043b660ff2382a
f21c210ead78dec810caf6c92ea3eea430aad0b0
7410 F20101206_AAAJOZ baglivio_m_Page_151thm.jpg
83750c3b88ddcdeec5b87d3d8e87f08a
fab1aad8ec9c2159ae213f4f317864a3a5507c13
F20101206_AAAIMM baglivio_m_Page_061.tif
05b4be7f67a03e93834f4c4798ed62d9
7d4f43826ae28a00cfc0b0ef441e0b93903aeb00
F20101206_AAAILX baglivio_m_Page_046.tif
6246d14bdf7f805c31b301ed89dfbffa
b69f4eac5b0696d0c79a817186a9ebb92802121b
F20101206_AAAINA baglivio_m_Page_075.tif
97bae265022f37726eb07afed395fbd7
0917f8bf3bd16a2b4e25ce1851c4e1b8a999bee9
6839 F20101206_AAAJPP baglivio_m_Page_161thm.jpg
de26c32d98a90c00114daf3879600927
71ae7e9a691f953a0c1c88082175f7d771e4a8a0
F20101206_AAAIMN baglivio_m_Page_062.tif
2bfa4a5896bc572ecda6e9084c5efc85
884ab295c444d3bb8edba38807d3579b0eec9f05
F20101206_AAAILY baglivio_m_Page_047.tif
b8599e62fa490ee24c84e494e0a6dcda
acb9a4794385d8f589a42de6953bd1a4d1d0bc97
F20101206_AAAINB baglivio_m_Page_076.tif
9549c40b499d23cc02a881d9fcdba7c4
733899148863fa1adc4d69bed36db8c9bf6afeb5
28277 F20101206_AAAJPQ baglivio_m_Page_161.QC.jpg
0c9469f425494feb8e075a15f379b4e1
35cfa00635d6a523bdc2bacecabcba2f797f489c
F20101206_AAAIMO baglivio_m_Page_063.tif
94a4cdb0899341d6109e194a7ce0e7fe
5824fc9b35f93d6e0119d6f97a22e26dcefc1a58
F20101206_AAAILZ baglivio_m_Page_048.tif
c7e06aba6b6541118c462dd8f6040bb9
81cc9c182262eb6fccef1dc381723dc73d9da0a0
F20101206_AAAINC baglivio_m_Page_077.tif
9605d160db2e97e545289e5eb6039994
9baf624f73e2eef6e8435d313c649368b7e3720c
F20101206_AAAIND baglivio_m_Page_078.tif
1647b16f4da2da5c3269ae9d744860dd
071b39fd8a4bf43dea42f1da289f185aa27e0745
7065 F20101206_AAAJPR baglivio_m_Page_164thm.jpg
d39b157d289d8718d5191340de9e1b21
4beee9c297606fe77287dd285607a445c6f49b59
F20101206_AAAIMP baglivio_m_Page_064.tif
49a5eab833808b8b7ad29747d1e45c4d
0976543e49d748f98a2f9df7dd6553312c8d5888
F20101206_AAAINE baglivio_m_Page_079.tif
605dc5ba554499276ac1942ae85e2746
730750445a234e43477d201f35176f0e746539ef
28665 F20101206_AAAJPS baglivio_m_Page_164.QC.jpg
5ed877faac3f2b2180e13275928ffebf
bfe6f8b8a070ced1d24aad9b7a12176530367f7c
F20101206_AAAIMQ baglivio_m_Page_065.tif
95f346a568bcd67a872fddf7227d1bde
52d5a2b2cd33501c44bb0db48eca6805f194b687
F20101206_AAAINF baglivio_m_Page_080.tif
b7b06d02d99ef94bbfeddd683a6897c5
632713f10b3d4452474854f401233d6704a96bc8
6776 F20101206_AAAJPT baglivio_m_Page_165thm.jpg
d62dc581fd2394cc0ab1eb9de432d9c0
0f4b48b86271d59d9f2b8778a07400cdb0e99b77
F20101206_AAAIMR baglivio_m_Page_066.tif
e77e7c8f64cbe913e16720f70fb553ea
f9510ac088d2ff1a8d523546f904531627c4dc0c
F20101206_AAAING baglivio_m_Page_081.tif
8b4c99c79a8cbd2233c198c424db67e6
f98ece00e39084cca4ef944327ebf9cfbcaaba6b
27202 F20101206_AAAJPU baglivio_m_Page_165.QC.jpg
1f01b2c9ea4aae843140b4ea6a92a066
d7fb33847ff4a494fcbef103353fc1be2fda6687
F20101206_AAAIMS baglivio_m_Page_067.tif
f9fbab4840edbc97365fb78c8df4dd15
f1ca6fc86bd70f8e3bc1dc5225b5e48be53d1079
F20101206_AAAINH baglivio_m_Page_082.tif
baef412bad031026b39b36fb8a49fd45
0244d19062b72dc3187e2ea8d4ac5e688035aa53
3736 F20101206_AAAJPV baglivio_m_Page_166thm.jpg
ad590dd38ad66bd631c8ef49aaa4e606
7bd18c3970da02d7e0b7b24e9187a752808ec64e
F20101206_AAAIMT baglivio_m_Page_068.tif
1cc41581898d158b512d52e6b1a920cf
e91e8374a3ddbf4e679b807d3a5cb5ebb7a87772
F20101206_AAAINI baglivio_m_Page_083.tif
3f16678ff9f3ccf39b4eadf6192f6528
83d08fdbc05fb4790fffee50b3615440f635b103
14904 F20101206_AAAJPW baglivio_m_Page_166.QC.jpg
581353a268c8481c7645143bb5d1f8d8
decc35a0691b40bdb8f5b7184bc671b73d36228f
F20101206_AAAIMU baglivio_m_Page_069.tif
964404c2a229d352f6c3101f3d163e26
b35489a217a6697bea0c28fdf817d596c1044ddd
F20101206_AAAINJ baglivio_m_Page_084.tif
84d90f7e2ed32bf0ccd0accd38d10951
99b0ea2d2b51154d48eadec7a21ad8ca4bb3277e
2522 F20101206_AAAJPX baglivio_m_Page_167thm.jpg
c067fc6c94c3e12e513f7188167fe38e
ce83eae758dc353c4963d810a0080867da6d8409
F20101206_AAAIMV baglivio_m_Page_070.tif
fea5df1da600bf69ed03407e29cde7a3
3c65dc70c09b55811475d1ed92a28e40dbea2042
F20101206_AAAINK baglivio_m_Page_085.tif
5e47aa52e54550c474bb2a785902999b
7e28bd681b4e52df46a736cc32793c5f6f7273ae
F20101206_AAAINL baglivio_m_Page_086.tif
eb28b0e3957e9f6bc165548bc750cf7f
528e22202259009045442998af2079e8c4f98285
F20101206_AAAIMW baglivio_m_Page_071.tif
209a1f3b814b8d8084c681f2a8519e1b
1c2ad3b561193f80a451f929b31b0c29ceae54ee
F20101206_AAAIOA baglivio_m_Page_101.tif
b880c84b1863c17f2b74843749e368cf
65967be958e25460003166dcbb33607d604edfdb
F20101206_AAAINM baglivio_m_Page_087.tif
ab5c842895c5163ea1636c302683d371
2fcb1c0e1ef44a054a7a6cf6cfc405a70a2bbd6a
F20101206_AAAIMX baglivio_m_Page_072.tif
f07553430d5bb9cbdce9851fb8552757
2f72b3f2040677904eb17a5d503412bef2a7dd10
F20101206_AAAIOB baglivio_m_Page_102.tif
56f416bfe1f1b7e8b7b62648d8a07da4
c8c07b6bdbbd613626d3ce0f60a000b106282026
F20101206_AAAINN baglivio_m_Page_088.tif
f3f9fb5b3c4b9c3fd32fca6912627711
a78d668293449243f14175154a6b791c93168425
F20101206_AAAIMY baglivio_m_Page_073.tif
edddaf5b5af32c79e5534c0a96ca5003
f7d93e9a62ff206a7839eb6413bf19d3dfd13c13
F20101206_AAAIOC baglivio_m_Page_103.tif
5852aa2292600c051f504a029327cf6d
29df06b073d2d08ea24cb54d1cc1dc26cc03d5ee
F20101206_AAAINO baglivio_m_Page_089.tif
52af6fccb3a95886afef35ed80890fd9
2c8967300987ffd1dad4ff119a39e93de9a4811e
F20101206_AAAIMZ baglivio_m_Page_074.tif
3c23ce68bf42428ba2a9119ef5eb48d5
00cbd65a528f7792118e2ce47a848faaf8a628e6
F20101206_AAAIOD baglivio_m_Page_104.tif
7e38312901959f03193e77fb58e6ddbb
14f0035dfcdc919d7824e83607454e8058204574
F20101206_AAAINP baglivio_m_Page_090.tif
a8cd73f1059a2a29c1dcb9852e6a23bc
a12b690003c254d50363ab182b6c600e0a0511ae
F20101206_AAAIOE baglivio_m_Page_105.tif
e21d529b5b574b91754e3262f2b6e96e
3dafa2d36551b23283550be143c2bfa49a43e5d7
F20101206_AAAIOF baglivio_m_Page_106.tif
cbd078d84a2191aea4e45d6500866039
c8e3b8dc8d2e68753ed3fb4bf78d8e320d22a0bd
F20101206_AAAINQ baglivio_m_Page_091.tif
679ba402f27d2dc3353fa32959dfcf05
4c0c38d53354a63ec11251f8396630b36a5e6063
F20101206_AAAIOG baglivio_m_Page_107.tif
cd3700029797748ccc1a1ba8e09b6855
6aabbf0cc97569fcd37e608e78129893cac9db6d
F20101206_AAAINR baglivio_m_Page_092.tif
5d05118600864e27229e1fb91a7ebd99
3d2926a721d114b7e3796d52cce0fc5938184e9b
F20101206_AAAIOH baglivio_m_Page_108.tif
6734fc66ed91f779c7a6c269adea977a
36396736845684f4a55cddc1fa26bf681b35371c
F20101206_AAAINS baglivio_m_Page_093.tif
e88ee2109c56c49fe3a319ffb1ae948b
8469f2256cb011e762e7830c75f56ac11b5ea098
F20101206_AAAIOI baglivio_m_Page_109.tif
491e7948c9d53582d079cb72955897f6
5d8eb0c17c2444ae368f12975b34777abf3cab17
F20101206_AAAINT baglivio_m_Page_094.tif
d76f48b652530a87ce5a8fecc3e68c5d
ee48fdfb53f45440d0106c7f3d0a265dec4a6036
F20101206_AAAIOJ baglivio_m_Page_110.tif
49b471a8b9dfd460d50f9b3cc3b91f25
03b2e17693a9ddb35493482233a56ae8eb9d4c33
F20101206_AAAINU baglivio_m_Page_095.tif
dae7a1b529707f16ba0e8c4bed5dad62
866954a21a3435393c37afb4c20e8893391c4ce5
F20101206_AAAIOK baglivio_m_Page_111.tif
67fae1f10eb1125d85deaea9f766289e
46d30d9d440a51dc082dd126f1caaab496379ee9
F20101206_AAAINV baglivio_m_Page_096.tif
306544f2e41de518f71f4fefd491cb6d
26e827b4059d3d028878b5c87ac31d84dafebeea
F20101206_AAAIOL baglivio_m_Page_112.tif
58b48c86efe921282924f7ece017d19c
8bb57950b8e19bda6fe13c370cd4a88e1cdc331c
F20101206_AAAINW baglivio_m_Page_097.tif
6858b2543d160f3bf35caf08f9ffd9ac
77d2300ab9d2585775f2996bf6a92e306d25dba3
F20101206_AAAIOM baglivio_m_Page_113.tif
57391851eb4a4b48fc3832e4731cd2fc
6beb65f0538eeaf7ada82394bc92a56267808886
F20101206_AAAINX baglivio_m_Page_098.tif
50d0f8d742bb79adce9b92773e1018a5
a37d2ba3deea227dcb499eadbacf1e71144b9f59
F20101206_AAAIPA baglivio_m_Page_127.tif
81d7dfea1c9fe8006dd498ccbd3dbbb7
beb6853d7aa4efb46ae9ef6cc9b1c951d03ea4a5
F20101206_AAAION baglivio_m_Page_114.tif
bdad7e7aa9136615c90dce193f97db3f
fda69076e05591964161ce37c04cbe77bda26f41
F20101206_AAAINY baglivio_m_Page_099.tif
4cae0da7739c5a75018650ae03fecfee
91bb7285bab5c38993ed709b0bd3c89d899b10e5
F20101206_AAAIPB baglivio_m_Page_128.tif
420a12eaa7810fe7f314127ce3701221
866badde8f9d7122b0a51a726c633c6c0f19b9f5
F20101206_AAAIOO baglivio_m_Page_115.tif
f4862abbf92219d66a9288e45491779e
49049e5fe3a437831a17fca52a359e20c62c2fd0
F20101206_AAAINZ baglivio_m_Page_100.tif
c765c70291bc2abbbd3a15b671ffd701
6c08f172a050aa1501bf79c2763eaec76f3e0791
F20101206_AAAIPC baglivio_m_Page_129.tif
fa40662667273c8e6e1616cd79237d6e
67391741318734d2dda780207416aac4c30a4a51
F20101206_AAAIOP baglivio_m_Page_116.tif
ade3ebec2dafce4e39bb5cd8dd528f03
1be8457de1ffa6f678e61ea9802ed3779c6188be
F20101206_AAAIPD baglivio_m_Page_130.tif
782e7915b71ca5c654840ece744b951c
6118dffda414d157eca3082257bdcec51c5f6d54
F20101206_AAAIOQ baglivio_m_Page_117.tif
4ac26791a6e7145c90a36b95517a2b11
5976daf3801d63ca03d6d78d9e9e17de5568edfc
F20101206_AAAIPE baglivio_m_Page_131.tif
4c22d3a5a171af9da5c5a953b3d26ca8
a7ebb984f0c7f2a87955c45ade6500a11a415c04
F20101206_AAAIPF baglivio_m_Page_132.tif
e71f016960fc75bcf65fe96c9330abbe
4fb2af7b9fd6aecd8cc759a67444d29317132dfe
F20101206_AAAIOR baglivio_m_Page_118.tif
4d7e4b6e8c682b13a81e197b1091d1cc
ed07247fd74c6e5a36f2ec29564b79358f189567
F20101206_AAAIPG baglivio_m_Page_133.tif
4dd6bc734c1947b26a150b08345ebd07
f4f102120a985711034fec21caae71a96aa2e2dc
F20101206_AAAIOS baglivio_m_Page_119.tif
02555c2ca43cf05c789697813f6b7d36
45a4ffc64992de5e9680a18c430c432f8071e7f8
F20101206_AAAIPH baglivio_m_Page_134.tif
fad9b1e19a872128c826574ebec11f54
7fbdd9f12c795e9fe52ada10d74ec4565533e010
F20101206_AAAIOT baglivio_m_Page_120.tif
a8faefd943c8eba95d17123b8caa92ed
4dfb4e12ebc774d71aebd8282d06f5607d5bcb19
F20101206_AAAIPI baglivio_m_Page_135.tif
1373700d4a8920e2a21d6a474989ad23
eb12b26c5b79e7cc81b9785855e8b473fe4217fc
F20101206_AAAIOU baglivio_m_Page_121.tif
a1073c38d663f4a65e6273f529f3c540
a76479ebc2010a69f9b63200944bb8a7c1ff71e1
F20101206_AAAIPJ baglivio_m_Page_136.tif
cd43e77717d61a6097cc2cc6449d8550
08f08a37817e41e2f84477c4c132ac9c652faaf2
F20101206_AAAIOV baglivio_m_Page_122.tif
15db2984ccecd39ec8ffe0b0a0d24981
644a9b7536b476792b6f50d413392a502728992d
F20101206_AAAIPK baglivio_m_Page_137.tif
47ca74a012aff9e426807c9e1bfdbc41
45711508fd6c10228fbdbcc70e0db76eb0f16148
F20101206_AAAIOW baglivio_m_Page_123.tif
e0208860ef79f2cf9f1373560fe5b866
186ffb94b3acae0b1b2c86f8625c048a0e1ce10b
F20101206_AAAIPL baglivio_m_Page_138.tif
69595ce8ef6c16c1111edd8ea827200d
3f24a0ce2e115dad5bbd76e34b7a72b2261f60ea
F20101206_AAAIOX baglivio_m_Page_124.tif
9a419eb00f5ebb7123bf4e81d1f01cac
e44166d19cf60eacd7fd63821228371f4ff682d1
F20101206_AAAIQA baglivio_m_Page_154.tif
893a580daf7a1438a104f44c1f86272f
a47f0b22013a162380354e888e22452b2eaf95ff
F20101206_AAAIPM baglivio_m_Page_139.tif
7090b2ba2ab9cf89c6dbe8882cc6def2
9c7c3e33f5e1513a006001d7ef667992833db48e
F20101206_AAAIOY baglivio_m_Page_125.tif
1eac06124bcfa72c7a2c47d14258fdc5
78d9ecc0bdae806999ba6339a6fb7d571fadbc9a
F20101206_AAAIQB baglivio_m_Page_155.tif
1afb23c6e4bcba0df3a087d152587f0b
ce787515d4d9eed4282840de5d85a455fba3bde8
F20101206_AAAIPN baglivio_m_Page_140.tif
93ae343aaf1cb660dc8fdb4ac67cfe05
3f5163ebab44cb1743c7beef152deb7030ffea7b
F20101206_AAAIOZ baglivio_m_Page_126.tif
7e15af3f65e0cde1a14d514b5448c702
71def25649a76a7e4b0292cf544cc7ec05468636
F20101206_AAAIQC baglivio_m_Page_156.tif
cf40a4bf27e83a8738e92f5999301e45
8c7b0713b510c99752606ba3af533cea708b7284
F20101206_AAAIPO baglivio_m_Page_141.tif
528c4b0cc915aec4bf8072591ddc57b8
ebb762124e12953d3f7c485c735f8b63e9d07bc2
F20101206_AAAIQD baglivio_m_Page_157.tif
f1aac9a81da6d9d530685711a919829b
25945f7d027d40b51d77ca7b8cd740e86af7bfd5
F20101206_AAAIPP baglivio_m_Page_142.tif
056b9e788085f242c6c02ec10b3f1622
f117f08159cc0200d12896f534261ab39f762b40
F20101206_AAAIQE baglivio_m_Page_158.tif
13b3cc8b46fedbc8fa4cd2807e01f7b0
cdb3a5f49dcce8d2589ec6d46d4c2aab70aa7ca5
F20101206_AAAIPQ baglivio_m_Page_143.tif
b588fa80b8f98656845c80f0f1f07d60
ac9320f15419647fc2182e72d8ecc37d6bcf224a
F20101206_AAAIQF baglivio_m_Page_159.tif
7ca311002b8b2283177e0fe6506b3509
d8c415c6f14a61331834baffca26ab4157f48c43
F20101206_AAAIPR baglivio_m_Page_144.tif
cc77d59454f680525bff89fa38e4df49
245760805ae36c2f9f166a5bee2fe4842a37c9d9
F20101206_AAAIQG baglivio_m_Page_160.tif
dae8e1f00da6af891b4e2243088c576f
e6ba15e3f537504b554d1f0680fabfce0a53671f
F20101206_AAAIQH baglivio_m_Page_161.tif
908da03f62a7e02df24caae32d16a6c6
440c93737f7d9e67a70e0a50801b48291df46ccc
F20101206_AAAIPS baglivio_m_Page_145.tif
309bb11c4da21312128880d41b29a03c
5bc0cafb3a6b9372ba6b215154ce5700b2fe75b3
F20101206_AAAIQI baglivio_m_Page_162.tif
c4714860943aa30f06aff93eed50debe
90189698ce22206530e3afd6f77a7f997d7eed75
F20101206_AAAIPT baglivio_m_Page_146.tif
9fe4531621a12bb0e1d817b3bd2b26af
161d4009249c82c668cab77ad74008e425e068c4
F20101206_AAAIQJ baglivio_m_Page_163.tif
e732e552bf3fa47e1995950d56e90d35
b32366fb4c195542766000a2add639a528792d80
F20101206_AAAIPU baglivio_m_Page_147.tif
e4136d76448755664a40ee56b6c272a9
f2d0911a20f6675934e38b8204b200d601d03c9f
F20101206_AAAIQK baglivio_m_Page_164.tif
3c78db745cbd575c16719a2a0b3bb0fe
b862ac3567e85cc060178aed9b14800cf8e96cdd
F20101206_AAAIPV baglivio_m_Page_148.tif
0f6161526e2e7e86d278a6f571e2868b
fc151787769f04f675c029b1a3eda65e29749a57
F20101206_AAAIQL baglivio_m_Page_165.tif
1dd5d71452ad16cb924077e3813dba81
cf2201ad3d36da4c8aaa8b7901ffab001fb587d3
F20101206_AAAIPW baglivio_m_Page_149.tif
5030775259e5cbf67a066fb8afff2f5a
35503726f824c764c84b2e5a9d358a72d35b26f4
53188 F20101206_AAAIRA baglivio_m_Page_013.pro
52dcb0e76b446b38d6b4319d846d6fbf
a6ffb9da96014beec9a9343c782e59a66ebd0a4a
F20101206_AAAIQM baglivio_m_Page_166.tif
e4cf5dff543b015d2158857b128c3d9c
421291079d4b1aad3a98db8947c07a4ab2565801
F20101206_AAAIPX baglivio_m_Page_150.tif
54ac0a020b245c66e29f70034db4d48a
20c82f12618f4b6702bfe8f03777dce60992a28d